Pediatric Neuroradiology [1 ed.] 3540410775, 9783540410775 [PDF]

Zitiervorschau

Pediatric Neuroradiology Brain Head and Neck Spine

Preface

Paolo Tortori-Donati · Andrea Rossi In collaboration with Roberta Biancheri

Pediatric Neuroradiology

I

II

Preface

Two-volume set consists of: Pediatric Neuroradiology Brain Pediatric Neuroradiology Head and Neck Spine

Preface

Pediatric Neuroradiology Brain Paolo Tortori-Donati and

Andrea Rossi In collaboration with

Roberta Biancheri Foreword by

Charles Raybaud

With 1635 Figures in 4519 Separate Illustrations, 207 in Color and 202 Tables

123

III

IV

Preface

Paolo Tortori-Donati, MD Head, Department of Pediatric Neuroradiology G. Gaslini Children’s Research Hospital, Genoa, Italy

Andrea Rossi, MD Senior Staff Neuroradiologist Department of Pediatric Neuroradiology G. Gaslini Children’s Research Hospital, Genoa, Italy

Roberta Biancheri, MD, PhD Consultant Pediatric Neurologist Department of Pediatric Neuroradiology G. Gaslini Children’s Research Hospital, Genoa, Italy

Library of Congress Control Number: 2004118036

ISBN 3-540-41077-5 Springer Berlin Heidelberg New York This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitations, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer-Verlag. Violations are liable for prosecution under the German Copyright Law. Springer is part of Springer Science+Business Media http//www.springeronline.com © Springer-Verlag Berlin Heidelberg 2005 Printed in Germany The use of general descriptive names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Product liability: The publishers cannot guarantee the accuracy of any information about dosage and application contained in this book. In every case the user must check such information by consulting the relevant literature. Medical Editor: Dr. Ute Heilmann, Heidelberg Desk Editor: Wilma McHugh, Heidelberg Production Editor: Kurt Teichmann, Mauer Typesetting: Verlagsservice Teichmann, Mauer Cover-Design: Frido Steinen-Broo, eStudio Calmar, Spain Printed on acid-free paper – 21/3150xq – 5 4 3 2 1 0

Preface

Dedication

To my wife Lella, inseparable companion of my life, and to my son Raffaele for giving it a meaning.

V

Preface

Foreword

Pediatric neuroradiology has always been a distinct entity within neuroradiology, because pathologies are different, more diverse and more often related to development than adult diseases. This has become still truer since the introduction of modern imaging modalities, in particular magnetic resonance imaging (MRI). In the past two decades entire fields of pathology have been made accessible to the physician on a daily basis: demyelinating and dysmyelinating diseases, malformations of cortical development, perinatal disorders and, moreover, prenatal brain disorders. The evaluation of tumors and malformations is now considerably more detailed. These changes are leading the pediatric neuroradiologist to develop much closer links with other specialists in neuropathology, genetics, neurophysiology and neurochemistry, to name but a few, to better comprehend the diseases she/he is dealing with. At the same time, the availability of an imaging modality practically devoid of noxious effects makes it possible to accumulate large series of patients studied in vivo, while more historical material has consisted only of a few specimens derived from autopsy or surgery. Easier patient surveillance has improved our understanding of the natural history of many diseases. Striking examples are the discovery of the transient nature of lesions (including some tumors) in neurofibromatosis type 1 and the routine observation of the relocation of cognitive functions after cortical injuries as a mechanism of brain plasticity. All this has given the pediatric neuroradiologist a unique opportunity to approach the central nervous system from a fresh, genuine point of view. She/he can now contribute significantly to the advances in clinical pediatric neurosciences, so much so that imaging has become the crucial step in diagnostic assessment, bringing information not only about structure (including the water sectors through DWI), but also about metabolism and function. With so many diseases and so much new knowledge, in such a short time: the need for comprehensive syntheses is immense, and Paolo Tortori-Donati, with his colleagues Andrea Rossi and Roberta Biancheri, has responded to the challenge with this new book. The 57 contributors of the 45 chapters are predominantly from Italy, and especially from Genoa, but also from France, the United States, Switzerland, the United Kingdom and Saudi Arabia. They represent more than 25 hospitals or universities and constitute a substantial representation of world pediatric neuroradiology. Despite the great number of contributors, the final result demonstrates a profound unity. The credit is to be given to Paolo Tortori-Donati and his colleagues. The book can be divided into three sections: brain (the largest), skull and neck, and spine. Each section begins with embryology. The three chapters written by Martin Catala (Chaps. 1, 28 and 38) do not only address the classical morphologic changes (descriptive embryology); they make up an exhaustive, state-of-the-art review of modern data regarding brain development: comparative embryology, tissue induction processes, neurulation, segmentation, migration, tissue interactions, cellular guidance, and specialization,

VII

VIII

Foreword

all with their respective genetic control. These chapters are extremely rich and detailed, clearly written and easy to read. They help bridge the gap between disease, morphogenesis, and the general principles of the development. While most of the pathology rests upon MRI, chapters concluding each section deal extensively with sonography. Brain MR spectroscopy and diffusion, perfusion and functional MRI are also extensively dealt with, as these techniques have now become dedicated parts of the routine evaluation of patients. Each pathology chapter is an exhaustive review, each is lucidly didactic, beautifully illustrated, up to date, and written by eminent experts. It would take up too much space, and be tiresome for the reader, to analyze all of them. However, the chapter on metabolic disease, composed by Zoltan Patay (Chap. 13), deserves a mention as truly representative of the spirit and aims of the book. Being thoroughly documented and education-oriented, indeed it could be a book by itself. The neuroradiologists of my generation remember “the Taveras book” (J.M. Taveras and E.H. Wood: Diagnostic Neuroradiology) as being the bible in which they found most of their knowledge of and enthusiasm for neuroradiology. Since then, not all neuroradiology texts have attained that ideal, but a few have come close, having a similar impact on neuroradiologists in training, and they have set the standard of the subspecialty. I believe that the new Pediatric Neuroradiology elaborated by Paolo Tortori-Donati and his group will take its place among these major books. Toronto, January 2005

Charles Raybaud

Preface

Preface

This book sees the light of day after 4 years’ gestation. The idea behind it originated in the early 1990s when, as I moved to the G. Gaslini Children’s Hospital, I realized the immense potential that was to be found there, both in terms of general pathologies and, more specifically, regarding central nervous system diseases, with a strong impact on diagnostic imaging, especially magnetic resonance imaging. The G. Gaslini Children’s Research Hospital in Genoa, Italy, where I hold the position of Head of the Department of Pediatric Neuroradiology, is a scientific institution and hospital dedicated to infancy and childhood, equipped with 400 beds for admission of patients under 14 years of age. Adults are not treated, with the sole exception of the Obstetrics and Gynecology department. All pediatric specialties are represented in the hospital. With particular regard to our field of interest, departments of Neurosurgery (with built-in spinal unit), Pediatric Hematooncology, Neurooncology, Infantile Neuropsychiatry, Ophthalmology, ENT, a Muscular Diseases Unit, and two Orthopedics Divisions are to be found. Moreover, special neuropathological competences are available within the Pathological Anatomy Unit. Within this environment, Neuroradiology has existed as an independent, individual unit since 1994. Very recently, the structural reorganization of the hospital has led to the definition of larger, “mother” departments. Within this new setting, Neuroradiology is housed within the Department of Diagnostic Imaging and is also functionally related to the Department of Neurosciences. Despite the unit’s relatively recent foundation, the neuroradiological experience in our hospital dates back to the introduction of CT in 1979 and MRI in 1988, as well as angiography, initially conventional and then digitalized, with neurovascular interventional activity since 1993. Ultrasound diagnosis has been developed, since its initial introduction, by the Department of Radiology, chaired by my friend Paolo Tomà, who also is a contributor to this book. The book is deliberately encyclopedic and was conceived to address pediatric neuroradiology in a different way from other existing publications in the field. I set out to create a book that would be useful not only to those who are directly involved with diagnostic imaging, but also to clinicians who care for infants and children with neurological conditions during their clinical practice. While the book retains a main diagnostic focus, clinical aspects, genetics, and treatment of the various conditions are also addressed. Embryology and neuropathology are also especially highlighted, the former being essential to the understanding of malformations and the latter because neuroradiology is nothing but the expression of the biological behavior of tissues. The book is divided into two volumes, one dedicated to the brain and the other to the head, neck, and spine. It is a long journey that starts from embryology, explores prenatal ultrasound and MRI diagnosis, reaches the fields of conventional CT, MRI and angiographic diagnosis, peers into the present and future of advanced imaging

IX

X

Preface

modalities and interventional techniques and then arrives at the clinical, metabolic, and genetic diagnosis. All the cases presented in this book have received a confirmed diagnosis, either pathological or clinical-metabolic, or genetic. Only those rare cases where a definite diagnosis could not be obtained are indicated as “presumed”. Many cases are presented globally, starting from the in utero appearance and extending to the final post-operative results and pathological features. Of the 45 chapters, about one third, the most relevant ones in terms of pathology and comprising the bulk of the book, were written by my group (except for the chapter on metabolic diseases), and belong to the listed authors in equal measure. The remaining chapters were written by leading, world-renowned authorities in the various fields. We undertook an enormous editorial revision of all contributions in other to give the book a homogeneous structure. This entailed editing the text, especially with regard to redundancies, and adding our own case material where I considered it useful to enrich the various presentations. I also made an extensive search, with the help of many friends and colleagues, for outstanding cases that could be added where we did not have sufficiently good demonstrations of individual diseases. Therefore, cases presented at various congresses or published in the literature were added to this book thanks to the gracious courtesy of these colleagues, who are individually acknowledged in the captions and to whom I extend my gratitude. The various clinical sections of the book were revised by one of my collaborators, specialized in Child Neurology, while a novice to the subject read the text to evaluate whether it was understandable at a basic level. We were fanatical about the quality of the illustrations. We drew most of the line drawings ourselves, and pathologic illustrations and anatomic preparations were included. Because the work on this book spanned several years, some illustrations from chapters submitted earlier were replaced with new, more recent and better ones obtained on modern imaging equipment. Furthermore, for the same reasons we also undertook a sometimes painful but hopefully thorough review of the more recent literature, so that I am confident the whole book embodies up-to-date knowledge on pediatric neuroradiology and the related fields. While a careful review of the existing literature was the foundation of this book, I also deliberately wanted to bring forth our own experience, based on our single-center case series, which I believe to be among the largest available. This experience, which has allowed me and my group to teach pediatric neuroradiology at various national and international congresses and courses, has eased the difficult task of putting together a clear, didactic and understandable text. I hope that the readers will appreciate this effort; only they will be able to judge whether I and my collaborators have succeeded in our goal. This book is dedicated not only to those who will use it in their everyday clinical practice, but especially to the children and their families who have had the ill luck to provide us with the material for our work.

Preface

Acknowledgements I would like to thank all those involved with the production of this book, without whom this work would not have been completed. Professor Claude Manelfe needs no presentation, because everyone working in our field knows him. At the time when he was the President of the European Society of Neuroradiology, he believed in me and my project and introduced me to Springer. Prof. Lorenzo Moretta, Scientific Director of the G. Gaslini Children’s Research Hospital, assigned a grant to Roberta Biancheri, whose collaboration was crucial to the success of the project. The anesthesiologists of our hospital, and particularly Dr Giorgio Salomone, without whom we would be unable to obtain our results as most young children at our institution are imaged under general anesthesia. My technologists Piero Sorrentino (Chief), Eleonora Cioetto and Claudia Mancini, with whom I have been collaborating for many years and to whom I feel connected by bonds of affection, rather than merely work. My nurses Fausta Lucchetti, Patrizia Massabò, Armanda Piana, Wilma Ponta, Claudia Ricci, Michela Rollando and Candida Soro, who have lovingly taken care of these unlucky children. My brotherly friend and colleague Armando Cama, Chief Neurosurgeon of our hospital, his collaborators Gianluca Piatelli and Marcello Ravegnani, and the neurooncologists Maria Luisa Garrè and Claudia Milanaccio: how many fights, discussions, and laughs…. Without this everyday confrontation we would not have grown. Paolo Nozza, young and enthusiast neuropathologist, whose input was crucial especially for the chapters on neoplastic diseases. My good friend and colleague, Paolo Tomà, Chief of the Radiology Department of our hospital, for the friendship and esteem he always showed to me and for allowing neuroradiology to grow freely in our institution. All people at Springer in Heidelberg, Germany, and especially Dr Ute Heilmann, Ms Wilma McHugh, and Mr Kurt Teichmann, for the enduring kindness, patience, and trust that they bestowed in us, graciously accepting our too often controversial requests, believing that we would eventually succeed, and especially for their enthusiasm in pursuing a common goal. Dr Roberta Biancheri, Child Neurologist in love with Neuroradiology, outstanding contributor and “clinical soul” of this book, has been an invaluable support in our moments of discomfort and fear of not being able to finish this work and also has taught me and my group a lot regarding clinical correlations to neuroimaging. Last but not least, no words would be sufficient to thank Dr Andrea Rossi, a man of lofty human and professional talents, my first pupil and my major collaborator both in putting together this book and in everyday work activity, who shared and amplified my enthusiasm for our work. I am proud to say that Andrea is recognized both nationally and internationally as an authority in our field, and I believe he will hold high the torch of future development of pediatric neuroradiology. Genoa, January 2005

Paolo Tortori-Donati

XI

Contents

Contents

Brain 1 Embryology of the Brain Martin Catala . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

2 Myelination Cecilia Parazzini, Elena Bianchini, and Fabio Triulzi . . . . . . . . . . . . . . . . . .

21

3 Malformations of the Telencephalic Commissures. Callosal Agenesies and Related Disorders Charles Raybaud and Nadine Girard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

41

4 Brain Malformations Paolo Tortori-Donati, Andrea Rossi, and Roberta Biancheri . . . . . . . . . . .

71

5 Magnetic Resonance Imaging of the Brain in Preterm Infants Luca A. Ramenghi, Fabio Mosca, Serena Counsell, and Mary A. Rutherford . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 6 Neonatal Hypoxic-Ischemic Encephalopathy Fabio Triulzi, Cristina Baldoli, and Andrea Righini . . . . . . . . . . . . . . . . . . . 235 7 Cerebrovascular Disease in Infants and Children Robert A. Zimmerman and Larissa T. Bilaniuk . . . . . . . . . . . . . . . . . . . . . . . . . 257 8 Arteriovenous Malformations: Diagnosis and Endovascular Treatment Laecio Batista, Augustin Ozanne, Marcos Barbosa, Hortensia Alvarez, and Pierre Lasjaunias . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287 9 Capillary-Venous Malformations Paolo Tortori-Donati, Andrea Rossi, and Roberta Biancheri . . . . . . . . . . . 319 10 Brain Tumors Paolo Tortori-Donati, Andrea Rossi, Roberta Biancheri, Maria Luisa Garrè, and Armando Cama . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329 11 Hemolymphoproliferative Diseases and Treatment-Related Disorders Paolo Tortori-Donati, Andrea Rossi, and Roberta Biancheri . . . . . . . . . . . . 437 12 Infectious Diseases Paolo Tortori-Donati, Andrea Rossi, and Roberta Biancheri . . . . . . . . . . . 469

XIII

XIV

Contents

13 Metabolic Disorders Zoltán Patay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 543 14 Neurodegenerative Disorders Ludovico D'incerti, Laura Farina, and Paolo Tortori-Donati . . . . . . . . . . . 723 15 Acquired Inflammatory White Matter Diseases Massimo Gallucci, Massimo Caulo, and Paolo Tortori-Donati . . . . . . . . . . 741 16 Phakomatoses Paolo Tortori-Donati, Andrea Rossi, Roberta Biancheri, and Cosma F. Andreula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 763 17 The Rare Phakomatoses Simon Edelstein, Thomas P. Naidich, and T. Hans Newton . . . . . . . . . . . . . . . 819 18 Sellar and Suprasellar Disorders Paolo Tortori-Donati, Andrea Rossi, and Roberta Biancheri . . . . . . . . . . . 855 19 Accidental Head Trauma Paolo Tortori-Donati, Andrea Rossi, and Roberta Biancheri . . . . . . . . . . . 893 20 Nonaccidental Head Injury (Child Abuse) Bruno Bernardi and Christian Bartoi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 929 21 Hydrocephalus, Cysts, and Other Disorders of the Cerebrospinal Fluid Spaces Paolo Tortori-Donati, Andrea Rossi, and Roberta Biancheri . . . . . . . . . . . 951 22 Epilepsy Renzo Guerrini, Raffaelo Canapicchi, and Domenico Montanaro . . . . . . . 995 23 MR Spectroscopy Petra S. Hüppi and François Lazeyras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1049 24 Diffusion-Weighted, Perfusion-Weighted, and Functional MR Imaging Ernst Martin-Fiori and Thierry A. G. M. Huisman . . . . . . . . . . . . . . . . . . . . . . . 1073 25 Brain Sonography Paolo Tomà and Claudio Granata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1115 26 Prenatal Ultrasound: Brain Mario Lituania and Ubaldo Passamonti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1157 27 Fetal Magnetic Resonance Imaging of the Central Nervous System Nadine Girard and Thierry A. G. M. Huisman . . . . . . . . . . . . . . . . . . . . . . . . . . 1219

Contents

Head and Neck 28 Embryology of the Head and Neck Martin Catala . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1255 29 Skull Development and Abnormalities Robert A. Zimmerman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1271 30 The Craniostenoses Cesare Colosimo, Armando Tartaro, Armando Cama, and Paolo Tortori-Donati . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1289 31 The Orbit Paolo Tortori-Donati, Andrea Rossi, and Roberta Biancheri . . . . . . . . . . . 1317 32 Disorders of the Temporal Bone Mauricio Castillo and Suresh K. Mukherji . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1361 33 Sinonasal Diseases Bernadette Koch and Mauricio Castillo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1391 34 Imaging of the Neck Zoran Rumboldt, Mauricio Castillo, and Suresh K. Mukherji . . . . . . . . . . . 1419 35 Cervico-Facial Vascular Malformations Jeyaledchumy Mahadevan, Hortensia Alvarez, and Pierre Lasjaunias . . 1459 36 Face and Neck Sonography Paolo Tomà . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1479 37 Prenatal Ultrasound: Head and Neck Mario Lituania and Ubaldo Passamonti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1503

Spine 38 Embryology of the Spine and Spinal Cord Martin Catala . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1533 39 Congenital Malformations of the Spine and Spinal Cord Paolo Tortori-Donati, Andrea Rossi, Roberta Biancheri, and Armando Cama . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1551 40 Tumors of the Spine and Spinal Cord Paolo Tortori-Donati, Andrea Rossi, Roberta Biancheri, Maria Luisa Garrè, and Armando Cama . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1609 41 Infectious and Inflammatory Disorders of the Spine Mauricio Castillo and Paolo Tortori-Donati . . . . . . . . . . . . . . . . . . . . . . . . . . 1653

XV

XVI

Contents

42 Spinal Trauma Paolo Tortori-Donati, Andrea Rossi, Milena Calderone, and Carla Carollo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1683 43 Spine and Spinal Cord: Arteriovenous Shunts in Children Siddhartha Wuppalapati, Georges Rodesch, Hortensia Alvarez, and Pierre Lasjaunias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1705 44 Spine and Spinal Cord Sonography Paolo Tomà . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1715 45 Prenatal Ultrasound: Spine and Spinal Cord Mario Lituania and Ubaldo Passamonti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1725

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1737

Preface

List of Contributors

Hortensia Alvarez, MD Diagnostic and Therapeutic Vascular Neuroradiology Unit Bicêtre Hospital Le Kremlin-Bicêtre, France

Roberta Biancheri, MD, PhD Consultant Child Neurologist Department of Pediatric Neuroradiology G. Gaslini Children’s Research Hospital Genoa, Italy

Cosma F. Andreula, MD Director, Diagnostic and Interventional Neuroradiology Unit Anthea Hospital Bari, Italy

Elena Bianchini, MD Department of Radiology and Neuroradiology V. Buzzi Children’s Hospital Milan, Italy

Cristina Baldoli, MD Department of Neuroradiology San Raffaele Scientific Institute Milan, Italy Marcos Barbosa, MD Diagnostic and Therapeutic Vascular Neuroradiology Unit Bicêtre Hospital Le Kremlin-Bicêtre, France Christian Bartoi, MD Department of Radiology Oakwood Hospital Dearborn, MI, United States of America and Department of Radiology Wayne State University and Michigan Children’s Hospital Detroit, MI, United States of America Laecio Batista, MD Diagnostic and Therapeutic Vascular Neuroradiology Unit Bicêtre Hospital Le Kremlin-Bicêtre, France Bruno Bernardi, MD Department of Neuroradiology Bellaria Hospital Bologna, Italy and Associate Professor, Department of Radiology Wayne State University and Michigan Children’s Hospital Detroit, MI, United States of America

Larissa T. Bilaniuk, MD Professor of Radiology, University of Pennsylvania Staff Neuroradiologist Children’s Hospital of Philadelphia Philadelphia, PA, United States of America Milena Calderone, MD Department of Neuroradiology Civic Hospital Padua, Italy Armando Cama, MD Head, Department of Pediatric Neurosurgery G. Gaslini Children’s Research Hospital Genoa, Italy Raffaello Canapicchi, MD Unit of Neuroradiology Clinical Physiology Institute National Council of Research Pisa, Italy Carla Carollo, MD Head, Department of Neuroradiology Civic Hospital Padua, Italy Mauricio Castillo, MD Professor of Radiology Chief of Neuroradiology University of North Carolina School of Medicine Chapel Hill, NC, United States of America Martin Catala, MD, PhD Laboratory of Histology and Embryology Pitiè-Salpêtrière School of Medicine University of Paris 6 Paris, France

XVII

XVIII

List of Contributors

Massimo Caulo, MD Department of Radiology ITAB- University of Chieti Chieti, Italy

Renzo Guerrini, MD Chief, Division of Child Neurology and Psychiatry University of Pisa and IRCCS Stella Maris Foundation Pisa, Italy

Cesare Colosimo, MD Professor of Radiology Department of Radiology G. D’Annunzio University Foundation Chieti, Italy

PD Thierry A.G.M. Huisman, MD Radiologist-in-Chief and Chairman Department of Diagnostic Imaging Radiology and Neuroradiology University Children‘s Hospital Zurich Zurich, Switzerland

Serena Counsell, MSc Superintendent Radiographer Imaging Sciences Department, Robert Steiner MR Unit Hammersmith Hospital London, United Kingdom Ludovico D’Incerti, MD Department of Neuroradiology C. Besta National Neurological Institute Milan, Italy Simon Edelstein, MD Senior Fellow Diagnostic and Interventional Neuroradiology Head and Neck Radiology Advanced Clinical Neuroimaging The Mount Sinai School of Medicine New York, NY, United States of America Laura Farina, MD Department of Neuroradiology C. Besta National Neurological Institute Milan, Italy Massimo Gallucci, MD Professor of Radiology University of L’Aquila L’Aquila, Italy

Petra S. Hüppi, MD Director of Child Development Unit Department of Pediatrics University Children’s Hospital Geneva, Switzerland Bernadette Koch, MD Department of Radiology Cincinnati Children’s Hospital Cincinnati, OH, United States of America Pierre Lasjaunias, MD, PhD Head, Diagnostic and Therapeutic Vascular Neuroradiology Unit Bicêtre Hospital Le Kremlin-Bicêtre, France François Lazeyras, PhD Department of Radiology University Children’s Hospital Geneva, Switzerland Mario Lituania, MD Director of Center of Fetal and Perinatal Medicine Galliera Civic Hospital Genoa, Italy

Maria Luisa Garrè, MD Director of Neurooncology Unit Department of Pediatric Hematology and Oncology G. Gaslini Children’s Research Hospital Genoa, Italy

Jeyaledchumy Mahadevan, MD Diagnostic and Therapeutic Vascular Neuroradiology Unit Bicêtre Hospital Le Kremlin-Bicêtre, France

Nadine Girard, MD Associate Professor Department of Radiology Hospital Nord Université de la Méditerranée Marseille, France

Ernst Martin-Fiori, MD Head, Neuroradiology & Magnetic Resonance Department of Diagnostic Imaging University Children’s Hospital Zurich, Switzerland

Claudio Granata, MD Department of Pediatric Radiology G. Gaslini Children’s Research Hospital Genoa, Italy

Domenico Montanaro, MD Unit of Neuroradiology Clinical Physiology Institute National Council of Research Pisa, Italy

List of Contributors

Fabio Mosca, MD Consultant in Neonatal Medicine Head, Department of Neonatology Mangiagalli Clinical Hospital Milan, Italy Suresh K. Mukherji, MD Associate Professor of Radiology Chief of Neuroradiology University of Michigan School of Medicine Ann Arbor, MI, United States of America Thomas P. Naidich, MD Department of Radiology (Neuroradiology) The Mount Sinai School of Medicine New York, NY, United States of America T. Hans Newton, MD Professor of Radiology The Moffitt Hospital University of California at San Francisco San Francisco, CA, United States of America Augustin Ozanne, MD Diagnostic and Therapeutic Vascular Neuroradiology Unit Bicêtre Hospital Le Kremlin-Bicêtre, France Cecilia Parazzini, MD Department of Radiology and Neuroradiology V. Buzzi Children’s Hospital Milan, Italy Ubaldo Passamonti, MD Center of Fetal and Perinatal Medicine Galliera Civic Hospital Genoa, Italy Zoltán Patay, MD, PhD Consultant Neuroradiologist Department of Radiology King Faisal Specialist Hospital and Research Centre Riyadh, Kingdom of Saudi Arabia Luca A. Ramenghi, MD Consultant in Neonatal Medicine Department of Neonatology Mangiagalli Clinical Hospital Milan, Italy Charles Raybaud, MD Professor and Head, Neuroradiology CHU Timone, Université de la Méditerranée Marseille, France Andrea Righini, MD Department of Radiology and Neuroradiology V. Buzzi Children’s Hospital Milan, Italy

Georges Rodesch, MD Head, Department of Diagnostic and Therapeutic Neuroradiology Foch Hospital Suresnes, France Andrea Rossi, MD Senior Staff Neuroradiologist Department of Pediatric Neuroradiology G. Gaslini Children’s Research Hospital Genoa, Italy Zoran Rumboldt, MD Department of Radiology Medical University of South Carolina Charleston, SC, United States of America Mary A. Rutherford, MD, FRCR, FRCPCh PPP/Academy Senior Fellow in Perinatal Imaging Honorary Senior Lecturer and Consultant in Pediatric Radiology Imaging Sciences Department, Robert Steiner MR Unit Hammersmith Hospital London, United Kingdom Armando Tartaro, MD Department of Radiology G. D’Annunzio University Foundation Chieti, Italy Paolo Tomà, MD Head, Department of Pediatric Radiology G. Gaslini Children’s Research Hospital Genoa, Italy Paolo Tortori-Donati, MD Head, Department of Pediatric Neuroradiology G. Gaslini Children’s Research Hospital Genoa, Italy Fabio Triulzi, MD Head, Department of Radiology and Neuroradiology V. Buzzi Children’s Hospital Milan, Italy Siddhartha Wuppalapati, MD Diagnostic and Therapeutic Vascular Neuroradiology Unit Bicêtre Hospital Le Kremlin-Bicêtre, France Robert A. Zimmerman, MD Vice Radiologist-in-Chief, Department of Radiology Chief, Neuroradiology Division/MRI Children’s Hospital of Philadelphia Philadelphia, PA, United States of America

XIX

Embryology of the Brain

1

Embryology of the Brain Martin Catala

1.1 Introduction

CONTENTS 1.1

Introduction

1.2

The Two Neural Inductions: Head and Trunk Inductions 2 Amphibian Chimeras and Neural Induction Genetic Regulation of Head Induction 3 Hesx1 and Septo-Optic Dysplasia 3

1.2.1 1.2.2 1.2.3 1.3 1.3.1 1.3.2 1.3.2.1

1.3.2.2 1.3.2.3 1.3.2.4 1.3.3 1.3.4 1.4 1.4.1 1.4.2 1.4.2.1 1.4.2.2 1.4.3 1.4.4 1.4.4.1 1.4.4.2 1.4.4.3 1.4.5 1.4.6 1.4.7 1.4.8 1.5 1.5.1 1.5.2 1.5.3

1

2

Segmentation of the Neural Primordium into Neuromeres 3 The Genetic Control of Insect Segmentation 4 The Rhombomeres 4 The Motoneurons That Constitute One Branchial Nerve Are Located in Two or Three Adjacent Rhombomeres 4 Even-Numbered Rhombomeres Differentiate Earlier Than Odd-Numbered Rhombomeres 4 Rhombomeres Represent Relative Cell Lineage Restriction Compartments 5 The Genetic Control of Anteroposterior Identity of the Rhombomeres 6 Embryonic Origin of the Cerebellum 7 The Prosomeres, a Model of Rostral Segmentation 7 New Insights About Cerebral Cortex Development 8 Partitioning the Telencephalic Vesicle into Two Hemispheres 8 Asymmetric Division of Neuroepithelial Cells in the Ventricular Zone 9 Neuroepithelial Cells Divide at the Ventricular Zone of the Neural Tube 9 Asymmetric Mitosis and Neurogenesis 9 Formation of the Preplate 10 From the Ventricle to the Cortical Plate: The Migration of Neuroblasts 10 Radial Migration and Radial Glial Cells 10 Two Modes of Radial Migration 11 Radial Glial Cells Are Bipotential Progenitors 12 The Basal Telencephalon Participates in the Formation of Cortical Neurons 12 Subplate Neurons Control the Development of Thalamic Axons 13 The Ventricular Neuroblasts Are Not Fully Specified at the Time of Their Birth 13 Formation of Cerebral Gyration 13 Control of Formation of the Pyramidal Tract 14 Chemoattraction and Netrins 14 L1 Permits Crossing of the Midline 15 Epha4 and Ephrin B3 Prevent Pyramidal Axons from Recrossing the Midline at the Spinal Level 16 References

16

The construction of the brain during embryonic life is a fascinating event. Indeed, the brain is the most complex organ of the whole body, and this is particularly evident in human beings. The human brain contains a huge number of cells, and each neuron is able to connect a great number of other neurons, leading to a very complex network of circuits. The development of such a complex structure is likely to be highly regulated in order to give rise to reliable anatomical regions that can perform their normal tasks after birth. Furthermore, the capacity for growth of the human brain is fantastic during fetal life; this can be illustrated by comparing the size of the brain at the beginning and the end of gestation (Fig. 1.1). In terms of anatomical segmentation, it is classical to distinguish the brain from the spinal cord. The brain is then divided into the hemispheres, the cerebellum, and the brainstem. The hemispheres are segmented into the cerebral cortex and the diencephalon, whereas the brainstem is separated into mesencephalon, metencephalon, and myelencephalon. This

Fig. 1.1. Comparison of the size of two human fetal brains at 10.5 weeks of gestation (arrow) and at 38.5 weeks of gestation (arrowhead) shows the tremendous capacity of growth of the brain

1

2

M. Catala

anatomical classification has been extensively used in terms of embryological description. This explains why the segmentation of the primitive neural tube has been described according to the same classification with the different vesicles (telencephalon, diencephalon, mesencephalon, metencephalon, and myelencephalon) (Fig. 1.2). Each embryonic vesicle was considered as the primordium for the adult region. This fixed and committed embryology is now really obsolete. First, it has been demonstrated that there are two separate regions in the neural tube: one rostral, comprising the telencephalon, diencephalon, and mesencephalon, and the other caudal. Indeed, the genetic regulations of these two regions are different.

another one, creating a chimera. After healing, it is thus possible to study the subsequent development of the chimera and to notice the consequences of such an operation. This was the beginning of the great American school with Ross Granville Harrison and the famous German school with Hans Spemann. In 1918, Hans Spemann described the peculiar property of the blastoporal lip of the amphibian gastrula (an early marker of the dorsal side of the gastrula in amphibians). Indeed, the graft of an extra lip on the ventral side of the gastrula leads to a duplication of the nervous system. However, it was impossible at that time to decipher between the tissues deriving from the host and those deriving from the donor. To address this question, it was necessary to have a perfect cell marker that can differentiate host tissues from donor ones. This was possible using two different species of Triton (T. taeniatus and T. cristatus), which can be differentiated thanks to the intrinsic pigmentation of the cytoplasm of their cells. These chimeras, in which donor and host embryos belong to different species, are called interspecific chimeras. This technique allowed the discovery of neural induction. This discovery was due to the experiments of Hilde Mangold (1924) in Hans Spemann’s laboratory [1]. Using the technique of interspecific chimeras, it was thus possible to note that the graft gives rise to the notochord, a part of the endoderm, the floor plate of the neural tube, and the medial part of the somites. In contrast, the remainder of the secondary neural tube was formed by the host (Fig. 1.3). This result allowed Spemann to describe the phenomenon of so-called Sek. D. R. sek. Pron. R. sek. Uw.

Fig. 1.2. Rostral extremity of a chick embryo (16-somite stage) showing the different vesicles: telencephalon (T), diencephalon (D), mesencephalon (M), rhombencephalon (R) or myelencephalon. Note that the eye vesicles (E) are developing and that the paraxial mesoderm is segmented into somites (S)

Sek Med. Pr. Med.

Sek. Ch.

1.2 The Two Neural Inductions: Head and Trunk Inductions 1.2.1 Amphibian Chimeras and Neural Induction At the end of the nineteenth century, experimental embryology was mainly based on the amphibian model, since it is possible to operate on the embryo and even to transplant a small part of one embryo to

Fig. 1.3. Section through an amphibian embryo after the graft of the dorsal lip of the blastopore (after Hans Spemann and Hilde Mangold, 1924). The cytoplasm of host cells is pigmented, whereas that of donor cells is not. Donor cells give rise to the secondary notochord (Sek. Ch.), the medial part of the somites (R. sek. Uw.), and the floor plate of the neural tube. The rest of the neural tube (Sek. Med.) derives from the host. [1]

Embryology of the Brain

neural induction. Neural induction means that the graft acts on the surrounding tissue and forces it to differentiate into neural fate. It is possible to reproduce the experiments of Spemann and Mangold in amniotes embryos by grafting Hensen’s node in birds [2] or the node in mice [3]. The problem of different inductors responsible for head and trunk induction was hypothesized by Spemann himself [1]. He distinguished three inductors: the head inductor, located in the upper blastoporal lip of the early gastrula; the trunk inductor in the same place of the advanced gastrula; and the tail inductor in the same place of the completed gastrula. These observations were at the origin of the concept of regionalized inductions.

1.2.2 Genetic Regulation of Head Induction The problem of head induction was recently strengthened by experiments performed in different models. The group of Eddy De Robertis [4] found that the secreted protein Cerberus is able to induce a head without inducing a trunk in the amphibian embryo. This experiment demonstrates that the formation of the head is independent of the formation of the trunk, and that the molecules involved in such a problem are different. In the mouse, two knock-out experiments selectively remove head structures while trunk and tail develop almost normally. This suggests that the genetic control of the head organizer is different from that of the trunk and tail organizer in the mouse. Lim1 is a gene coding for a protein carrying a LIM class homeodomain motif and two cyteine-rich domains. This gene is homologous to the Xenopus gene Xlim1 that is involved in the blastoporal functions. Heterozygous mice Lim1+/– are normal, whereas homozygous Lim1–/– died at embryonic day 10 [5]. Homozygous embryos showed a very abnormal neural plate with no neural structures anterior to rhombomere 3. Otx2 is one of the mouse cognates of the drosophila gene orthodenticle. This gene codes for a homeodomain containing protein and is particularly expressed in the rostral brain during development. The knockout of this gene in the mouse leads to very interesting results. First, heterozygous mice are either normal [6] or present a phenotype that is reminiscent of the otocephalic syndrome in humans [7]. The discrepancies between these results can be explained by the different genetic background in which the knock-out was performed. In the case of normal heterozygous, the background was CD1/129, whereas when the hetero-

zygous were abnormal, the background was C57Bl/6. These results show that there are other genes that can modify the phenotype of a single gene knock-out. In homozygous embryos, the head never develops and the embryos were truncated at the level of the third rhombomere [6–8]. Since Otx2 is expressed rostrally until the limit between the mesencephalon and the metencephalon [9], the absence of rhombomeres 1 and 2 (in which Otx2 is not expressed) is a secondary consequence of the absence of more rostral structures, indicating that these two rhombomeres need an interaction to survive or develop. In conclusion, from these two knock-out experiments, it can be proposed that the genetic control which is mandatory for the correct development of the head is different from that which is needed for the development of the body. It is thus possible to observe embryos lacking a head. The genetic control for the development of the head is probably complex, but it comprises both Otx2 and Lim1.

1.2.3 Hesx1 and Septo-Optic Dysplasia These experimental evidences lead to a tremendous amount of genetic work to try to decipher all the genetic markers involved in head development. I do not intend to make a thorough review of this field, but just to illustrate some important results. The pairedlike gene Rpx (now called Hesx1) is expressed very early during development in the most anterior part of the neural plate [10]. The knock-out of the gene leads in the homozygous mouse to a nervous phenotype which is reminiscent of septo-optic dysplasia (absence of the septum pellucidum, of the corpus callosum, microphthalmia, hypoplasia of the olfactory bulbs) [11]. This result prompted the authors to test the human homologue, HESX1, in familial cases of septo-optic dysplasia, and indeed they found a missense mutation in this gene [11].

1.3 Segmentation of the Neural Primordium into Neuromeres In insects, the body of the larva develops as separated and segregated repetitive units called compartments. This type of development is responsible for the metamerization of the adult body, and is considered a very ancestral mode of development. However, in vertebrates, including mammals, such repetitive units

3

4

M. Catala

can be evidenced in the adult body: such is the case with the vertebrae and ribs. The search for a metamerization of the central nervous system is really an old story. The first description of repetitive units in the central nervous system can be credited to von Baer in 1828 [12]. These compartments were termed neuromeres by Orr in 1887 [13]. This theory was so popular that segmentation of the neural tube was presented in all textbooks of those times (Fig. 1.4).

Cm

ment no zygotic genes are expressed in drosophila. The first sets of zygotic genes to be expressed are gap genes. These genes are expressed in broad domains corresponding roughly to three future segments. Gap genes control the expression of pair-rule genes, which are expressed according to a zebra-like pattern (one segment expresses the gene, the adjacent does not express and the next one expresses). Pair-rule genes control the expression of segment polarity genes that are expressed only by a subregion of a segment, defining a polarity within each segment. Then, gap genes, pair-rule genes, and segment polarity genes interact to induce the expression of homeotic selector genes, which determine the developmental fate of each segment. These homeotic genes belong to the antennapedia and bithorax complexes, and are homologous to Hox genes of vertebrates.

1.3.2 The Rhombomeres I

I II III IV V

II III IV V

au

gv

The rhombencephalon is the neural tube region where segmentation is obvious with repetitive units, called rhombomeres. In higher vertebrates, seven rhombomeres can be perfectly identified. The most caudal rhombomere (or rhombomere 8) is less defined, and its caudal border is not as sharp as the frontiers of the others.

v v

Fig. 1.4. The rhombomeres (I, II, III, IV, V) in a section of a pig embryo as illustrated in the famous Prenant’s Eléments d‘Embryologie de l’homme et des vertébrés (1896). au otic vesicle, Cm midbrain, gv ganglion of the X cranial nerve, v X cranial nerve

1.3.1 The Genetic Control of Insect Segmentation The aim of this chapter is not to account for a thorough description of the genetic pathways involved in insect segmentation, but to describe the salient features of this issue that could be used for the understanding of vertebrate segmentation. The drosophila embryo is first patterned by maternal effect genes. The transcribed mRNA and the corresponding proteins accumulate and define the anteroposterior axis of the future larva. The genes Nanos, Caudal, and Bicoid belong to this family. These genes are only from maternal origin, since at this stage of develop-

1.3.2.1 The Motoneurons That Constitute One Branchial Nerve Are Located in Two or Three Adjacent Rhombomeres

The roots of the different cranial nerves can be traced by using a retrograde marker injected in the nerve pathway. This technique allows study of the nervous origin of each branchial nerve. It is interesting to note that adjacent rhombomeres participate in the formation of a single branchial nerve; motoneurons that form the motor part of one branchial nerve are located in adjacent rhombomeres. For example, in the chick embryo, the motoneurons of nerve V are located in rhombomeres 1, 2, and 3; those of nerve VII in rhombomeres 4 and 5; and those of nerve IX in rhombomeres 6 and 7 [14]. 1.3.2.2 Even-Numbered Rhombomeres Differentiate Earlier Than Odd-Numbered Rhombomeres

Neurons can be characterized by several markers. One of these is the intermediate filament, neurofila-

Embryology of the Brain

ment, that can be evidenced using different specific antibodies. In the chick embryo, the pattern of neurofilament expression is very specific in the developing hindbrain. Indeed, neurofilament-positive neurons are present earlier in even-numbered rhombomeres than in odd-numbered ones [15]. 1.3.2.3 Rhombomeres Represent Relative Cell Lineage Restriction Compartments Lrd, a Suitable Marker for Lineage Analyses

Lineage studies are based on experimental paradigms that allow labeling of one single cell and all its derivatives. The most-used technique consists of injecting lysinated rhodamine dextran (LRD) into the cytoplasm of one cell by iontophoresis. This marker is fluorescent and is transmitted to daughter cells during mitosis. Its molecular weight prevents the diffusion to neighbor cells through gap junctions, allowing its use for lineage studies. Because of dilution during successive mitosis, it is possible to observe daughter-cells until the eighth symmetrical division.

The Inter-Rhombomeric Boundary Is Established Through Interactions Between Even- and OddNumbered Rhombomeres

If an even-numbered rhombomere is grafted adjacent to an odd-numbered one, the cells of the grafted rhombomere will not mix with the cells of the host. In contrast, if a odd rhombomere is grafted adjacent to another odd rhombomere, the cells of the two rhombomeres intensively mix [18]. So, the boundaries between rhombomeres are established by interactions of one rhombomere with the adjacent ones. Furthermore, this experiment shows that even- and oddnumbered rhombomeres display distinct properties. To maintain these frontiers requires that the cells are both able to recognize each other within the same segment and to repulse each other between two adjacent segments. A repulsive molecular system that is involved in rhombencephalon segmentation is constituted by the Eph-ephrin system. Eph belong to a family of 14 members that are subdivided into two groups in vertebrates. Eph class A (A1–A8) are GPI (glycosylphosphatidylinositol)-anchored proteins. Eph class B (B1–B6) contain a transmembrane domain followed by a cytoplasmic tail with tyrosine hydroxylase activity. Eph class A bind to Ephrins A

Lineage Analyses of the Rhombomeres

The spreading of the progeny of one single cell has been extensively studied in the rhombomeres of the chick embryo. First, if LRD injection is performed before the 13-somite stage (the stage of appearance of the rhombomeric boundaries), the progeny of the injected cell is not always restricted to one rhombomere [16]. In contrast, if the injection is performed after this stage and if the embryo is observed at stages 18–19, the progeny of the injected cell is limited by the rhombomeric boundaries [16] (Fig. 1.5). This demonstrates that the boundaries between rhombomeres are established progressively during development. These results allow consideration of the rhombomeres as true cell-lineage restricted compartments, such as the developing compartments of insects. However, if the embryos are analyzed later during development (stages 18–25), the results appear a little bit different. Indeed, greater than 80% of the clones are located within the rhombomeric limits and greater than 5% of the clones expand across the rhombomeric boundaries [17] (Fig. 1.5). For the other clones, the results are equivocal. This shows that the rhombomeres are not strict compartments as defined in insects, but they can be considered as relative compartments in which cell mixing is very limited.

a

r1

r1

r7

r7 b

Fig. 1.5a,b. Appearance of the rhombencephalic clones observed after a single cell injection. a If the injection is performed before boundary formation, some clones spread across the inter-rhombomeric limits (dark gray). b If the injection is performed after the formation of the boundaries, the majority of clones does not cross the limits (light gray). A few clones appear to cross the inter-rhombomeric limits (not shown)

5

6

M. Catala

(except for EphA4, which can bind numerous Ephrins A but also Ephrins B2 and B3). Eph class B bind to Ephrins B. The Eph-ephrin system generates repulsive behavior between cells expressing these proteins. EphA4 is expressed by rhombomeres 3 and 5, whereas ephrin B2 by rhombomeres 2, 4, and 6 [19]. If EphA4 is forced to be expressed by a cell, it will adopt an odd position in the rhombencephalon. On the contrary, if the expression of ephrin B2 is forced, the cell will adopt an even position. This shows that the Eph-ephrin system is functionally involved in the formation of inter-rhombomeric frontiers. Adhesive molecules, such as cadherins, may play a role favoring cell adhesion within a rhombomere. However, functional evidences are lacking to support this model. 1.3.2.4 The Genetic Control of Anteroposterior Identity of the Rhombomeres Numerous Genes Are Expressed by the Rhombomeres

It is obviously out of the focus of this chapter to review all the genes known to be expressed in the rhombencephalon. It is convenient to decipher these genes into two classes (Fig. 1.6): 1) Some genes are expressed by a few rhombomeres. For example, Hox b1 is expressed by rhombomere 4 [20], Krox 20 is expressed by rhombomeres 3 and 5 [20], lunatic fringe is expressed by rhombomeres 3 and 5 in the mouse [21] but in rhombomeres 2, 4, and 6 in the zebrafish [22], kreisler in rhombomeres 5 and 6 [23]. 2) Other genes are expressed in a more broad domain, poorly limited at its caudal part but with a strict r1

r1

r1

Hoxb 1

Krox 20

Hoxa 1

r5

r7

a

r7

b

This pattern of expression suggests that these genes may play a role in specifying the anteroposterior identity of the rhombomeres. Some Rhombomeres Are Missing in Mutant Mice

I will now present some illustrative mouse mutants to show that some genes are involved in the control of the development of the rhombomeres. The Kreisler Mouse

Kreisler is an X-ray-induced recessive mutation in the mouse. The gene that is impaired by this mutation (mafB) is expressed in rhombomeres 5 and 6. A careful analysis of this mutant shows that it lacks rhombomeres 5 and 6, which are replaced by an enlarged rhombomere 4 [23]. This result indicates that the Kreisler protein is mandatory for the acquisition of the identity of rhombomeres 5 and 6; in its absence, cells will adopt a r4 phenotype. Knock-Out of the Gene Coding for the Zinc Finger Protein Krox20

The knock-out of the gene coding for Krox20 leads to homozygous that die either in the first two weeks after birth [24] or shortly after birth [25]. In the two constructs, the hindbrain is severely affected with a dramatic reduction or disappearance of both rhombomeres 3 and 5, corresponding to the rhombomeres where the gene is expressed. Knock-Out of the Hox A1 Gene

r3

r3

anterior limit corresponding to an inter-rhombomeric boundary. Such is the case for Hox genes of the paralogues 1, 2, 3, and 4 (except for b1). For example, Hox a1 is expressed by the neural tube until the limit between rhombomeres 3 and 4.

Mice homozygous for Hox a1 invalidation present a marked rhombencephalic phenotype characterized by severe reduction of rhombomere 4 and absence of rhombomere 5 [26]. In this case, the phenotype is observed at the most rostral domain of expression of the gene. This is a quite general rule for Hox genes invalidation. Knock-Out of the Hox B1 Gene

c

Fig. 1.6a–c. Expression pattern of some genes expressed in the rhombencephalon; cells expressing the gene are represented in gray. a Hoxb1 is expressed by the sole rhombomere 4. b Krox 20 is expressed by both rhombomeres 3 and 5. c Hoxa1 is expressed by the neural tube up to a rostral limit corresponding to the boundary between rhombomeres 3 and 4

Mice homozygous for Hox b1 invalidation present a partial modification of r4 into r2 [27]. This change affects the territory of expression of the gene. Since the change is only partial, the protein Hox b1 is not mandatory for the acquisition of the r4 identity. This feature suggests that other genes are involved in such a control.

Embryology of the Brain

Interactions Between Rhombomeres

The situation is far from being so simple. Indeed, rhombomeres may be subjected to interactions that force them to change their fate. If rhombomeres are transplanted to the level of r7/8 in the chick embryo, their behavior is different according to their initial position. Rhombomeres 2, 3, and 4 do not change their fate in this new position, whereas rhombomeres 5 and 6 adopt a r7/8 identity [28]. The signals that induce this transformation are from two sources: the paraxial mesoderm [29, 30] and the posterior neural epithelium [29]. These results indicate that rostral rhombomeres differ from caudal ones for their genetic regulation. In this concept, it is interesting to note that quail embryos deficient for vitamin A present a very abnormal hindbrain with an absence of its caudal part [31] which is respecified into anterior hindbrain [32]. Such is also the case for the mouse embryos lacking the Raldh2 gene (which codes for one of the key enzymes involved in retinoic acid synthesis) [33]. These mutant embryos lack a caudal hindbrain, which is replaced by an abnormal r3–r4 identity. Raldh2 is expressed weakly by somites 1–3, whereas its expression is stronger from somite 4 caudalwards [34]. Furthermore, Hox b5, b6, and b8 are sensitive to retinoic acid, since they present a RARE (retinoic acid-response element) in their regulative sequences [35], indicating that these genes are direct targets for retinoic acid signaling. The action of retinoic acid on the caudal hindbrain is mediated through a gradient of concentration: higher concentrations lead to a caudal phenotype, whereas absence of signal leads to r4 identity [36]. These results in both quail and mouse embryos show that retinoic acid signaling is mandatory to promote the induction of the posterior hindbrain. Furthermore, FGFs are able to induce rhombomeric anterior markers in the hindbrain [37]. The patterning of the hindbrain is due to two antagonistic influences: FGF, which anteriorizes, and retinoic acid, which posteriorizes.

1.3.3 Embryonic Origin of the Cerebellum The classic conception is that the cerebellum is a pure rhombencephalic derivative. However, using the quail-chick chimera technique, a fate map of the cerebellum was constructed in the avian embryo at the 10–12-somite stage [38–41], which showed that both the mesencephalic and the metencephalic (upper rhombencephalon) vesicles contribute to the forma-

tion of the cerebellum. Furthermore, the results help to establish the exact origin of the different cell types of the cerebellum. Purkinje cells derive from progenitors located in the ventricular layer [39]. This is also the case with neurons of the molecular layer and small cells surrounding Purkinje cells. The external granular layer only yields the granule cells [39]. However, this conclusion of a dual origin of the cerebellum was challenged a few years later. The gene Otx2 coding for a transcription factor can serve as a reliable marker of the mesencephalo-metencephalic border [9] (Fig. 1.7). This gene enables the demon-

10-somite stage 40-somite stage

Fig. 1.7. Expression pattern of Otx2 in the mesencephalon during development of the avian embryo. Cells expressing the gene are represented in gray. At 10-somite stage, the caudal limit of expression does not correspond to the constriction between the two vesicles (arrows), whereas the concordance is perfect at 40-somite stage. This shows that the constriction is not a reliable anatomical marker for appreciating the limit between the mesencephalon and the metencephalon. [9]

stration that the constriction between the so-called mesencephalic and metencephalic vesicles is in fact not a fixed point, but moves rostrally during development. Therefore, at the 10-somite stage, the so-called mesencephalic vesicle is in fact formed by the mesencephalic vesicle itself and by the rostral part of the metencephalic vesicle. Using this new landmark, it is possible to demonstrate that the cerebellum is a pure metencephalic derivative [9]. The caudal limit of the presumptive cerebellar territory corresponds to the rostral limit of Hoxa2 expression [42]. The cerebellum arises precisely from rhombomere 1.

1.3.4 The Prosomeres, a Model of Rostral Segmentation The groups of Rubenstein at UCSF (United States) and Puelles in Murcia (Spain) carefully examined the

7

8

M. Catala

expression pattern of numerous genes in the rostral part of the neural tube (prosencephalon) [43, 44]. From these aspects, it can be concluded that regionalization of the forebrain results from different mechanisms. First, the neural tube is separated into units according to the antero-posterior axis. This leads to the formation of separate domains, whose development is highly regulated. The second mechanism is the formation of longitudinal domains that are due to polarization of the neural plate according to the medio-lateral axis, and of the neural tube according to the ventro-dorsal axis. The convergence of these two types of regionalization could allow one to distinguish segments that the authors proposed to called prosomeres. However, the exact biological significance of this model remains to be strengthened in the future.

1.4 New Insights About Cerebral Cortex Development Writing a thorough paper on the tremendous amount of works published on cortical development would lead to the publication of a complete book! Here, I just want to give the reader some new insights about the results gained by experimental embryology. These results have sometimes changed our conceptions of cortical development.

1.4.1 Partitioning the Telencephalic Vesicle into Two Hemispheres The human telencephalon begins to be evidenced at Carnegie stage 14 (embryonic day 32, 4 weeks and 4 days of gestation) [45]. The interhemispheric fissure begins to form as early as at Carnegie stage 16 (embryonic day 37, 5 weeks and 2 days of gestation) [46]. This important landmark is obvious at Carnegie stage 17 (embryonic day 41, 5 weeks and 6 days of gestation) [47]. The falx cerebri differentiates during Carnegie stages 22–23 (8th week of gestation) [48]. The formation of two hemispheres from the single telencephalic vesicle is due to the induction by bone morphogenetic proteins (BMPs) of the midline roof plate [49]. The telencephalic midline roof plate is characterized by a low level of cell proliferation and a high level of apoptosis [50]. The limits between the cortical ventricular zone and the midline roof plate are called the cortical hems.

The defect of formation of two hemispheres leads to the development of holoprosencephaly. In humans, at least 12 genetic loci (HPE 1–12) have been linked with this pathological condition [50, 51]. Seven mutated human genes out of these 12 loci are associated with holoprosencephaly [51]. Three involved genes belong to the Sonic Hedgehog pathway, two genes to the nodal signaling pathway, and the two last genes are not yet related to the other molecular pathways. The first gene (corresponding to the HPE3 locus) whose mutation was linked with human holoprosencephaly is Sonic Hedgehog (SHH) [52], which codes for a secreted molecule acting as a morphogen. SHH is a diffusible molecule that acts on a membrane receptor, Patched-1. The consequence of the binding of SHH to Patched-1 is the inhibition of the membrane receptor. Since Patched-1 inhibits another membrane receptor, Smoothened, the effect of SHH will be to relieve the endogenous inhibition of Smoothened. The activation of Smoothened is mediated to the cytoplasm by the proteins Gli (Gli1, 2, and 3). These proteins will be translocated to the nucleus, where they will act as transcription factors. SHH is expressed in various tissues, including the prechordal mesoderm (the most rostral part of the axial mesoderm) and the ventral telencephalon [50]. Holoprosencephaly in SHH mutation is considered to be due to impairment of the prechordal mesoderm, and not to a defect of the ventral telencephalon [50]. Two other genes belonging to the SHH pathway have been associated with holoprosencephaly: PATCHED-1 [53], which codes for the receptor of SHH, and GLI2 [54], which codes for one of the cytoplasmic effectors of this pathway. Furthermore, Dispatched is required for long-range activity of Hedgehog in Drosophila. The mouse mutant line Dispatched A–/– displays holoprosencephaly [55]. The homologous human gene is located in chromosomal region 1q42, corresponding to the locus of HPE10. However, no human mutations of this gene have yet been reported in cases of holoprosencephaly. The second signaling pathway that has been linked with holoprosencephaly is the Nodal signaling. Nodal is a secreted molecule that belongs to the TGFβ superfamily. In absence of Nodal, the prechordal plate is not present and SHH is absent in the rostral part of the embryo [51]. Mutations of the TGIF gene are linked with the HPE locus situated in the chromosomal region 18p11.3 [56]. Another locus has been found to be associated with mutation of the TDGF1 gene (homologous to the Crypto gene) [57]. These two types of mutations lead to a loss of function of the proteins encoded by these genes. It is quite surprising to note that TGIF acts as negative regulator of SMAD2, an effector of Nodal signaling, whereas

Embryology of the Brain

TDGF1 is a coreceptor that is mandatory to mediate Nodal signaling. The SIX3 gene corresponds to the HPE2 locus (mapped on human chromosome 2p21) [58]. Six 3 (SIX3 for humans) is closely related to optix, a drosophila member of the sine oculis family of genes [58]. This gene is expressed by the anterior neural plate (which corresponds to the eye field domain). The exact role of this mutant in the generation of holoprosencephaly remains to be established [51]. The ZIC2 gene corresponds to the HPE locus in the chromosomal region 13q32 [59]. Zic2 (ZIC2 in humans) codes for a zinc finger gene homologous to the odd-paired gene of Drosophila [59]. The expression of this gene at the level of the forebrain is mainly dorsal [60]. Mutations of this gene lead to holoprosencephaly without any craniofacial involvement [50].

1.4.2 Asymmetric Division of Neuroepithelial Cells in the Ventricular Zone 1.4.2.1 Neuroepithelial Cells Divide at the Ventricular Zone of the Neural Tube

Fig. 1.8. Section through the neural tube of a chick embryo at the 6th day of incubation (Feulgen-Rossenbeck’s staining). The mitoses of neuroepithelial cells are concentrated at the ventricular side of the neural tube (arrows)

It has been well known for a long time that neuroepithelial cells divide at the ventricular border of the neural tube (Fig. 1.8). The first author to describe mitosis in the ventricular part of the neural tube was Wilhelm His in the middle of the nineteenth century. However, at that time, he considered that the neural tube was made of two types of cells: germinal cells, that are able to divide and give rise to neuroblasts, and epithelial cells (or spongioblasts), that are not. The latter were considered by His as the precursors of glial cells. Now, it is well established that neuroepithelial cells follow a cell cycle leading to a mitosis. This general movement of the nucleus from the periphery to the ventricular zone was described by Sauer [61] (Fig. 1.9) and is termed karyokinesis. 1.4.2.2 Asymmetric Mitosis and Neurogenesis

In a pioneer work, Martin [62] observed that the mitotic spindle of dividing neuroepithelial cells is orientated, and that this spindle is perpendicular to the surface of the neural tube only during the phase of neuroblast formation. The author concluded that the orientation of the mitotic spindle is linked with the differentiation program of neuroepithelial cells. Now, it is firmly established that two types of mitosis

Fig. 1.9. Diagram of the alar plate of a pig embryo from Sauer (1935). Nuclei follow a general movement from the basal part of the neural tube to its apical pole where they divide. [61]

9

10

M. Catala

can take place in the ventricular zone of the neural tube. Symmetrical divisions lead to the formation of two identical daughter cells, whereas asymmetrical divisions lead to the formation of a young migratory neuron and a proliferative daughter cell. The difference between these two types of division is explained by the orientation of the mitotic spindle [63]. Asymmetric mitoses have been particularly well studied in Drosophila. They play a major role in generating neuronal cells. Different proteins, such as Glial-cells missing, Prospero, Numb, Miranda, Partner of Numb, Inscuteable, and Notch, are asymmetrically distributed in the dividing cell, leading to an asymmetrical repartition of these proteins in the two daughter cells [64]. In the mammalian telencephalon, asymmetric mitoses take place [63, 65–67]. However, the situation is far from simple. Indeed, there are differences according to the studied species: Numb is located at the apical part of the ventricular zone in the mouse [65], whereas it is located at the basal part of the ventricular zone in the chick [68]. Furthermore, the role played by Numb changes during development: this protein prevents cell differentiation during the early phase of neurogenesis [69], whereas Numb promotes neuronal differentiation during the late phase of neurogenesis [67]. In any case, all these results are highly suggestive for a pivotal role of asymmetric mitoses in the control of neurogenesis in the mammalian cortex.

1.4.3 Formation of the Preplate

term preplate was indeed coined to underscore that this layer is laid down before the formation of the cortex. The cells that compose the preplate derive from ventricular progenitors [70]. The role of the preplate has been firmly established by studying the reeler mutation in the mouse. In this case, the preplate is normally formed, but cortical organization is highly abnormal with an inverted outside-inside gradient. The reeler mutation affects the reelin gene, coding for a large extracellular protein synthesized by both Cajal-Retzius cells and granule neurons of the cerebellum [71]. Reelin interacts with Disabled-1, a cytoplasmic protein expressed by neurons of the cortical plate [72]. The couple Reelin–Disabled-1 is thus involved in the control of the end of migration of neuroblasts in the cortical plate. It is interesting to note that a human autosomal recessive disease, characterized by lissencephaly and cerebellar hypoplasia, has been associated with mutations in the human gene coding for Reelin [73]. Thus, Reelin also plays a role in the formation of the cortex in humans.

1.4.4 From the Ventricle to the Cortical Plate: The Migration of Neuroblasts 1.4.4.1 Radial Migration and Radial Glial Cells

Neuroblasts, generated by asymmetrical mitosis, must migrate from the ventricular zone to the superficial layer of the brain. This migration proceeds in the so-called intermediate zone (Fig. 1.11). When

The first-ever evidence of the future mammalian isocortex is the formation of the primordial plexiform layer, also known as the preplate (Fig. 1.10). This layer is composed of a superficial plexus, the CajalRetzius neurons, and the neurons of the subplate. The

Fig. 1.10. Section of the rostral neural tube of a human embryo (45 days post fertilization) showing the ventricular zone (VZ) and the developing preplate (Pre)

Fig. 1.11. Section of the telencephalic wall of a human embryo (12 weeks of gestation) showing a laminar organization orientated from the ventricle to the meninges (Me): the ventricular zone (VZ), the intermediate zone (IZ), the subplate (SP), the cortical plate (CP), and the molecular layer (M)

Embryology of the Brain

neuroblasts arrive at the level of the preplate, they stop their migration and split the preplate into two layers: the future layer I, i.e., the most superficial layer of the cortex, and the subplate. The first neurons to be generated are the future neurons of the deepest layers (i.e., layer VI) and, subsequently, neurons migrate and form more and more superficial layers. This mode of formation of the cortical plate has been termed the inside-outside gradient. Histological examination of developing cortex has shown the presence of radially orientated cells, i.e., radial cells (Figs. 1.12, 13), whose processes extend Fig. 1.13. Radial processes can be evidenced using the monoclonal antibody directed against GFAP (Glial Fibrillary Glycoprotein) (arrows)

of retrovirus [78], and construction of a transgenic mouse in which the lacZ gene is located on the X chromosome [79]. The discovery of nonradial migration of cells from the ventricle to the cortex is an important result, and shows that interpretations of malformative diseases in humans should be very careful and revisited using more modern data. 1.4.4.2 Two Modes of Radial Migration

It is possible to directly visualize the migrating cells by time-lapse imaging applied on brain slices. Using this technique, it was found that radial migration proceeds according to two different modes [80] (Fig. 1.14): Fig. 1.12. Reproduction of a drawing from Wilhelm His showing the radial cells (arrows) and the mitosis located in the ventricular zone (arrowhead)

from the ventricular pole to the brain surface. Migrating neuroblasts are preferentially apposed to the glial processes. These histological results lead to the proposal that ventriculo-cortical migration of neurons is strictly radial. This concept was so popular that the concept of radial unit is largely used in neuropediatric textbooks to explain cortical malformations. However, using defective retroviruses as cell markers, Walsh and Cepko [74, 75] found that clones arising from a single ventricular progenitor are widely dispersed across the cortex. This result leads to the proposal that the ventriculo-cortical migration is not solely radial. This conclusion is now firmly established by different techniques: marking cells on living slices [76], staining cells of telencephalic explants [77], use

a

Soma translocation

b

Cell locomotion

Fig. 1.14a,b. The two modes of radial migration in the telencephalon. The neuroblasts are represented in light gray. a soma translocation and b cell locomotion

11

12

M. Catala

1) Soma translocation: in this case, the neuroblast contains a long process that reaches the pial surface. The length of this process progressively shortens and the cell soma adopts a more superficial location. This movement is continuous, and is in fact independent of radial glial cells. 2) Cell locomotion: the neuroblast contains a leading edge apposed to glial radial cells that does not reach the pial surface. The length of this process is constant during migration. Migration is, in this case, discontinuous with periods of active movement and of pause, and is dependent upon radial glial cells.

the intermediate zone, and of the cortical plate [90] (Fig. 1.16). Cells coming from the medial ganglionic eminence follow a migration pathway located underneath the pia mater. This could correspond to the so-called subpial granular layer (Fig. 1.17), described by Brun in 1965 [91]. At last, the cells of the caudal ganglionic eminence participate in the formation of cortical GABAergic interneurons populating the layer V at the caudal level of the cortex [86]. In humans, the situation has been assessed by the group of Rakic [92]. Cortical GABAergic interneu-

1.4.4.3 Radial Glial Cells Are Bipotential Progenitors

It is now firmly established that radial glial cells can generate cortical pyramidal neurons and astrocytes [81–84]. During mitosis, the radial process persists and is asymmetrically inherited by daughter cells [85]. If the postmitotic neuroblast inherits the process, it will adopt the radial movement called soma translocation to reach the cortical plate. In contrast, if the postmitotic neuroblast is devoid of radial process, it will be guided by radial cells to reach the cortical plate. This can explain the two modes of radial migration observed on time-lapse microscopy.

Fig. 1.15. Organization of the embryonic telencephalon. OB olfactory bulb, CGE caudal ganglionic eminence, MGE medial ganglionic eminence, LGE lateral ganglionic eminence

1.4.5 The Basal Telencephalon Participates in the Formation of Cortical Neurons The existence of nonradial migration shows that the origin of cortical neurons is more complex that was commonly thought. In classic textbooks, it was thought that only the dorsal telencephalon is able to produce cortical neurons, whereas the ventral telencephalon gives rise to the striatum and pallidum. The ventral telencephalon is subdivided into three ganglionic eminences during development, i.e., lateral, medial, and caudal (Fig. 1.15). These three ganglionic eminences do not express the same subset of genes, indicating that they represent three different genetic compartments [86]. Staining the lateral ganglionic eminence allows to observe marked cells migrating in the intermediate zone of the dorsal telencephalon and then populating the cortical plate [87–90] (Fig. 1.16). These migrating cells differentiate into GABAergic neurons. Using the same technique, it is also possible to show that the medial ganglionic eminence gives rise to neurons that migrate and differentiate into GABAergic interneurons of layer I, of the subplate, of

Fig. 1.16. Frontal section of the telencephalon showing both the dorsal (DT) and the ventral telencephalon (VT). The dorsal telencephalon corresponds to the pallium, whereas the ventral telencephalon is composed of both medial and lateral ganglionic eminences (MGE and LGE). Neuroblasts originating from the two ganglionic eminences migrate to populate the pallium (arrows). These neuroblasts will eventually differentiate into GABAergic interneurons

Embryology of the Brain

1.4.7 The Ventricular Neuroblasts Are Not Fully Specified at the Time of Their Birth

Fig. 1.17. Section of the superficial layers of the telencephalic wall in a human embryo (13 weeks of gestation) showing the molecular layer (M) overlying the cortical plate (CP). The socalled subpial granular layer (SGL) corresponds to the most superficial part of the molecular layer, characterized by a higher cell density. Me meninges

rons arise from two different embryonic origins: the ganglionic eminences account for 35% of these interneurons, whereas 65% derive from the ventricular zone. Different behavioral cell movements have been associated with nonradial migration [93]: tangential pathways within the intermediate zone, branching cells that change the direction of their migration, and ventricle-directed migration in which a cell of the intermediate zone migrates to the ventricular zone.

1.4.6 Subplate Neurons Control the Development of Thalamic Axons The subplate is fated to disappear after birth, but its role during fetal development is crucial. During development, the neurons of the subplate develop axons which grow to reach subcortical targets, such as the internal capsule or the thalamus [94–96]. Furthermore, axons arising from subcortical nuclei form functional synapses with subplate neurons [97]. If occipital subplate neurons are destructed early during development by injection of kainic acid, the occipital cortex develops normally, but the axons arising from the lateral geniculate nucleus fail to enter the cortex and form a subcortical network [98, 99]. If the destruction of the subplate is performed later, the axons enter the cortex but make connections with wrong layers, showing that subplate neurons are necessary to control the correct cortical projection of geniculate axons [100].

Theoretically, it is possible to imagine two models for the ventricular zone. The protomap model postulates that the neuroblasts acquire their specification at the time they are generated [101]. In contrast, the protocortex model suggests that neuroblasts in the ventricular zone are still plastic and can change their fate according to their environment [102]. One experiment which can solve this discrepancy consists of transplanting ventricular cells from a donor into the ventricle of a host that has a different age (heterochronic transplantation). If the protomap model is correct, transplanted cells will not change their fate. After heterochronic transplantations, neurons belong to two populations: either they are fixed and do not change their fate or they adapt to this new environment [103]. The most striking result is that their adaptive behavior is dependent upon the phase of the cell cycle [104]. If the cells are very close to mitosis, they are specified and cannot change their fate. On the contrary, if the cells are transplanted during the S phase, they can adapt to their new environment. This shows that ventricular cells are sensitive to external clues, which can be conveyed by cortical neurons through their projecting axons [105]. In conclusion, even if cortical plate neurons develop far away from their progenitors, there are interactions between the two layers which can adapt the fate of the newly-generated neuroblasts to the cortical environment.

1.4.8 Formation of Cerebral Gyration One of the salient features of the development of human cortex is the progressive appearance of sulci and gyri during gestation. A gyrus is a cortical structure limited by either two sulci or a sulcus and a fissure. Sulci are classified according to their time of appearance. Primary sulci (i.e., Sylvian, calcarine, parieto-occipital, callosal, central, and olfactory sulci) develop after the 14th week of gestation [106]. It is important to note that primary sulci are quite invariant in all members of the species, suggesting that their development is highly regulated at the genetic level. The Sylvian and callosal sulci first appear during the 14th week of gestation (Fig. 1.18). Olfactory, parietooccipital, and calcarine sulci appear in the 16th week of gestation. Lastly, the central sulcus develops from the 20th week of gestation onwards [106] (Fig. 1.19).

13

14

M. Catala

Secondary sulci (i.e., precentral, postcentral, lunate superior, frontal, and orbital sulci) arise later and show individual variations even in twins. The last to be formed are tertiary sulci, which present huge interindividual anatomical variations (Fig. 1.20).

Fig. 1.20. Lateral aspect of the right hemisphere of a human fetus at term. Note the complex development of the sulci and gyri

1.5 Control of Formation of the Pyramidal Tract

Fig. 1.18. Lateral aspect of the right hemisphere of a human fetus at 19 weeks of gestation. Note that the Sylvian fissure (arrows) is well developed

The pyramidal tract, as all fasciculi of the white matter and all white commissures, is composed of axons developing from neurons. These axons grow thanks to the development of their growth cone. The development of these cones is highly regulated, so that the tracts are packed and organized within the white matter. I will briefly review the three elementary mechanisms that are involved in the control of the development of the pyramidal tract. These elementary mechanisms are likely applicable also to other tracts or commissures.

1.5.1 Chemoattraction and Netrins

Fig. 1.19. Lateral aspect of the right hemisphere of a human fetus at 23 weeks of gestation. Note the developing central sulcus (arrows)

Some cerebral regions may play a chemoattractive role on growth cones derived from cortical neurons [107]. Such an effect is found for the basal pons [107], the ganglionic eminence [108], and the internal capsule and floor plate of the neural tube [109]. These results indicate that some neural regions are able to synthesize and secrete chemoattractant factors that direct the growth of axons (Fig. 1.21). What could be the chemical nature of these signals? From chick brains, it was possible to isolate two diffusible proteins, netrin 1 and 2 [110], which are homologous to UNC-6, a C. elegans protein. The latter is involved in the control of migration of commissural cells in this species. UNC-6 acts on two different receptors: one (unc-40) responsible for a chemoat-

Embryology of the Brain

Explant of the cortex

Growing axons

Floor plate of the neural tube Fig. 1.21. The axons that develop from the cortex converge to the floor plate, suggesting that chemoattractant proteins are secreted by the floor plate

tractant effect, and the other (unc-5) responsible for a chemorepellent effect. In vertebrates, the scenario is far from being so simple: netrins 1 and 2 act on two receptors [DCC (deleted in colorectal cancer) and neogenin] involved in chemoattraction, and two others (unc-5h2 and unc-5h3) involved in chemorepulsion. The role of netrins has been established in vitro. Transfected COS cells (a cell line derived for monkey kidney) secrete netrin 1 or 2 and play a chemoattractive role in growing axons from commissural neurons of the spinal cord [111]. Such a chemoattractive effect can be elicited for pyramidal axons [112]. Thus, netrins are excellent molecular candidates to account for the directional growth of pyramidal axons during development. To test in vivo the effect of netrin 1 on the development of the brain in mammalian embryos, the knock-out of this gene has been performed [113]. Homozygous mice die during the first postnatal days because they fail to eat. The study of the central nervous system in these mice shows that the white ventral commissure of the spinal cord is reduced, the corpus callosum is absent with formation of Probst’s bundles, the fimbria is malformed, the hippocampal commissure is practically missing, and the anterior white commissure is absent. In contrast, both the posterior white and the habenular commissures are normal. Furthermore, new commissures develop in aberrant anatomic positions (i.e., a commissure located at the junction between the mesencephalon and the metencephalon). These results show the pivotal role played by netrin 1 during the formation of some commissures of the central nervous system. It

is important to note that not all cerebral commissures are impaired by this knock-out, highlighting the fact that molecular mechanisms controlling the development of the different commissures are various, and differ according to their anatomical position. In the homozygous mice, the thalamo-cortical fasciculi are normal from the thalamus to the internal capsule, where the axons are disorganized leading to a dramatic reduction of the cortical projections [114]. The results concerning the effect of this knock-out on the pyramidal tract are surprising. Indeed, the pyramidal tract is normal from the cortex to the level of the medulla oblongata. However, pyramidal tracts fail to decussate, leading to a dramatic reduction of the total number of axons of the pyramidal tract in the spinal cord [115]. A similar phenotype is observed for mutations of the netrin receptors DCC and Unc5h3 [115]. In conclusion, netrin 1 plays a chemoattractive role in some growing axons, and its presence is mandatory for the correct decussation of the pyramidal tracts at the level of the medulla oblongata.

1.5.2 L1 Permits Crossing of the Midline Other molecular systems are involved in the control of the development of the decussation of the pyramidal tract at the level of the medulla oblongata. Such is the case for L1, an adhesion molecule belonging to the immunoglobulin superfamily. This molecule acts through both homo- and heterophilic interactions. Mutations of the human gene have been reported in association with the so-called CRASH syndrome (corpus callosum hypoplasia, mental retardation, adducted thumbs, spastic paraplegia, and hydrocephalus). This X-linked syndrome shows great interand intrafamilial variability. The pyramidal tract is frequently absent in patients with CRASH syndrome (Fig. 1.22). Such feature is not specific, but is highly evocative for this syndrome [116]. This indicates that L1 plays an important role during the formation of the pyramidal tract in humans. The exact role of L1 has been extensively studied in mice after knocking-out this gene [117, 118]. In these mice, the pyramidal tract appears normal from the level of the cortex to the medulla oblongata, where the axons fail to cross the midline. Caudally, the hypoplastic tract stops at the cervical spinal cord level. Furthermore, the corpus callosum can be hypoplastic. It is important to note that such a neuropathological feature is highly variable, and depends upon the genetic background of the mouse lines [119]. This suggests that other genes may have a modulatory

15

16

M. Catala

Fig. 1.22. Section through the medulla oblongata of a human fetus (24 weeks of gestation) with CRASH syndrome. The pyramidal tract is hypoplastic (arrows), whereas the olives are hypertrophic (arrowheads)

role in L1 functions. The white anterior, posterior, and spinal commissures are normal in the knock-out mouse. The exact mode of action of L1 on pyramidal tract formation is still unknown. Three hypotheses may account for the observed phenotype [117]: (1) L1 could interact with CD24, one of its known ligands, expressed in the midline of the medulla oblongata; (2) axons of the pyramidal tract could interact with other L1 positive axons, such as those of the medial lemniscus that are formed earlier during development; and (3) formation of pioneer axons is independent of L1, but the subsequent growth of other axons requires the presence of L1. In conclusion, the adhesion molecule L1 plays a major role in controlling the crossing of pyramidal axons in the region of the medulla oblongata. The required interactions in terms of molecules are still unknown.

1.5.3 EphA4 and Ephrin B3 Prevent Pyramidal Axons from Recrossing the Midline at the Spinal Level The role of the Eph-ephrin system in the regulation of the development of the pyramidal tract has been demonstrated by studying the knock-out of the gene coding for EphA4 in the mouse [120]. EphA4 is expressed by growing axons of the pyramidal tract. Homozygous mice EphA4 -/- display an abnormal motor behavior (hesitation to move, lack of coordination, and abnormal synchronous movements of the hind limbs). Neuropathological examination of these mice shows that the pyramidal tract is severely impaired, with a lot of axons stopping at the level of the medulla oblongata. Furthermore, axons reaching

the spinal cord cross the midline in the spinal cord, in contrast to normal axons of the pyramidal tract that never cross the spinal midline. It is interesting to note that the white anterior commissure is absent. It is possible to copy the spinal phenotype by performing a deletion of the tyrosine kinase domain of the cytoplasmic tail [121]. In this case, the white anterior commissure is normal. This shows that tyrosine kinase activity is necessary for correct development of the pyramidal tract, whereas EphA4 controls the development of the white anterior commissure by kinaseindependent functions. Ephrin B3 is one of the ligands of EphA4. Axons expressing EphA4 interact with cells expressing Ephrin B3. EphA4-positive axons cannot grow, and cross a region of cells expressing Ephrin B3. Such a region forms a sort of molecular barrier. It is interesting to study the consequences of removal of the gene coding for Ephrin B3. This has been performed in the mouse [122]. Homozygous mice display the same abnormal locomotor behavior as EphA4 -/-. Neuropathological examination discloses that the decussation of the pyramidal tract crosses the midline at a more rostral position and the spinal pathway of the tract is normal in the dorsal bundle; however, at the level of the gray matter, axons cross the midline and contact contralateral motoneurons. The white anterior commissure is normal, indicating that another ligand of EphA4 is involved in the control of the formation of the white anterior commissure.

References 1. Spemann H. Embryonic development and induction. Yale University Press, New Haven, 1938. 2. Waddington CH. Induction by the primitive streak and its derivatives in the chick. J Exp Biol 1933; 10:38–46. 3. Beddington RSP. Induction of a second neural axis by the mouse node. Development 1994; 120:613–620. 4. Bouwmeester T, Kim S-H, Sasai Y, Lu B, De Robertis E. Cerberus is a head-inducing secreted factor expressed in the anterior endoderm of Spemann’s organizer. Nature 1996; 382:595–601. 5. Shawlot W, Behringer RR. Requirement for Lim1 in headorganizer function. Nature 1995; 374:425–430. 6. Ang S-L, Jin O, Rhinn M, Daigle N, Stevenson L, Rossant J. A targeted mouse Otx2 mutation leads to severe defects in gastrulation and formation of axial mesoderm and to deletion of rostral brain. Development 1996; 122:243–252. 7. Matsuo I, Kuratani S, Kimura C, Takeda N, Aizawa S. Mouse Otx2 functions in the formation and patterning of rostral head. Genes Dev 1995; 9:2646–2658. 8. Acampora D, Mazan S, Lallemand Y, Avantaggiato V, Maury M, Simeone A, Brûlet P. Forebrain and midbrain regions are deleted in Otx2-/- mutants due to a defective anterior

Embryology of the Brain neuroectoderm specification during gastrulation. Development 1995; 121:3279–3290. 9. Millet S, Bloch-Gallego E, Simeone A, Alvarado-Mallart RM. The caudal limit of Otx2 gene expression as a marker of the hindbrain – midbrain boundary: a study using in situ hybridisation and chick/quail homotopic grafts. Development 1996; 122:3785–3797. 10. Hermesz E, Mackem S, Mahon KA. Rpx: a novel anteriorrestricted homeobox gene progressively activated in the prechordal plate, anterior neural plate and Rathke’s pouch of the mouse embryo. Development 1996; 122:41–52. 11. Dattani MT, Martinez-Barbera J-P, Thomas PQ, Brickman JM, Gupta R, Martensson I-L, Toresson H, Fox M, Wales JKH, Hindmarsh PC, Krauss S, Beddington RSP, Robinson ICAF. Mutations in the homeobox gene HESX1/Hesx1 associated with septo-optic dysplasia in human and mouse. Nature Genet 1998; 19:125–133. 12. Vaage S. The segmentation of the primitive neural tube in chick embryos (Gallus domesticus). A morphological, histochemical and autoradiographical investigation. Berlin: Springer Verlag, 1969. 13. Orr HA. Contribution to the embryology of the lizard. J Morphol 1887; 1:311–372. 14. Marín F, Puelles L. Morphological fate of rhombomeres in quail/chick chimeras: a segmental analysis of hindbrain nuclei. Eur J Neurosci 1995; 7:1714–1738. 15. Lumsden A, Keynes R. Segmental patterns of neuronal development in the chick hindbrain. Nature 1989; 337:424– 428. 16. Fraser S, Keynes R, Lumsden A. Segmentation in the chick embryo hindbrain is defined by cell lineage restrictions. Nature 1990; 344:431–435. 17. Birgbauer E, Fraser SE. Violation of cell lineage restriction compartments in the chick hindbrain. Development 1994; 120:1347–1356. 18. Guthrie S, Prince V, Lumsden A. Selective dispersal of avain rhombomere cells in orthotopic and heterotopic grafts. Development 1993; 118:527–538. 19. Cooke J, Moens CB. Boundary formation in the hindbrain: Eph only it were simple… Trends Neurosci 2002; 25:260– 267. 20. Wilkinson D, Bhatt S, Chavrier P, Bravo R, Charnay P. Segment-specific expression of a zinc-finger gene in the developing nervous system of the mouse. Nature 1989; 337:461– 464. 21. Johnston SH, Ruaskolb C, Wilson R, Prabhakaran B, Irvine KD, Vogt TF. A family of mammalian Fringe genes implicated in boundary determination and the Notch pathway. Development 1997; 124:2245–2254. 22. Prince VE, Holley SA, Bailly-Cuif L, Prabhajaran B, Oates AC, Ho RK, Vogt TF. Zebrafish lunatic fringe demarcates segmental boundaries. Mech Dev 2001; 105:175–180. 23. McKay IJ, Muchamore I, Krumlauf R, Maden M, Lumsden A, Lexis J. The kreisler mouse: a hindbrain segmentation mutant that lacks two rhombomeres. Development 1994; 120:2199–2211. 24. Schneider-Maunoury S, Topilko P, Seitanidou T, Levi G, Cohen-Tannoudji M, Pournin S, Babinet C, Charnay P. Disruption of Krox-20 results in alteration of rhombomeres 3 and 5 in the developing hindbrain. Cell 1993; 75:1199–1214. 25. Swiatek PJ, Gridley T. Perinatal lethality and defects in hindbrain development in mice homozygous for a targeted mutation of the zinc finger gene Krox20. Genes Dev 1993; 7:2071–2084.

26. Mark M, Lufkin T, Vonesch J-L, Ruberte E, Olivo J-C, Dollé P, Gorry P, Lumsden A, Chambon P. Two rhombomeres are altered in Hoxa-1 mutant mice. Development 1993; 119:319–338. 27. Studer M, Lumsden A, Ariza-Mc Naughton L, Bradley A, Krumlauf R. Altered segmental identity and abnormal migration of motor neurons in mice lacking Hoxb-1. Nature 1996; 384:630–634. 28. Grapin-Botton A, Bonnin M-A, Ariza-Mac Naughton L, Krumlauf R, Le Douarin NM. Plasticity of transposed rhombomeres: Hox gene induction is correlated with phenotypic modifications. Development 1995; 121:2707– 2721. 29. Grapin-Botton A, Bonnin M-A, Le Douarin NM. Hox gene induction in the neural tube depends on three parameters: competence, signal supply and paralogue group. Development 1997; 124:849–859. 30. Itasaki N, Sharpe J, Morrison A, Krumlauf R. Reprogramming Hox expression in the vertebrate hindbrain: influence of paraxial mesoderm and rhombomere transposition. Neuron 1996; 16:487–500. 31. Maden M, Gale E, Kostetskii I, Zile M. Vitamin A-deficient quail embryos have half a hindbrain and other neural defects. Curr Biol 1996; 6:417–426. 32. Gale E, Zile M, Maden M. Hindbrain respecification in the retinoid-deficient quail. Mech Dev 1999; 89:43–54. 33. Niederreither K, Vermot J, Schuhbaur B, Chambon P, Dollé P. Retinoic acid synthesis and hindbrain patterning in the mouse embryo. Development 2000; 127:75–85. 34. Swindell EC, Thaller C, Sockanathan S, Petkovich M, Jessell TM, Eichele G. Complementary domains of retinoic acid production and degradation in the early chick embryo. Dev Biol 1999; 216:282–296. 35. Oosterveen T, Niederreither K, Dollé P, Chambon P, Meijlink F, Deschamps J. Retinoids regulate the anterior expression boundaries of 5’ Hoxb genes in posterior hindbrain. EMBO J 2003; 22:262–269. 36. Dupe V, Lumsden A. Hindbrain patterning involves graded responses to retinoic acid signalling. Development 2001; 128:2199–2208. 37. Marín F, Charnay P. Hindbrain patterning: FGFs regulate Krox20 and mafB/kr expression in the otic/preotic region. Development 2000; 127:4925–4935. 38. Martinez S, Alvarado-Mallart RM. Rostral cerebellum originates from the caudal portion of the so-called “mesencephalic” vesicle: a study using chick/quail chimeras. Eur J Neurosci 1989; 1:549–560. 39. Hallonet MER, Teillet M-A, Le Douarin NM. A new approach to the development of the cerebellum provided by the quailchick marker system. Development 1990; 108:19–31. 40. Alvarez Otero R, Sotelo C, Alvarado-Mallart RM. Chick/ quail chimeras with partial cerebellar grafts: an analysis of the origin and migration of cerebellar cells. J Comp Neurol 1993; 333:597–615. 41. Hallonet MER, Le Douarin NM. Tracing neuroepithelial cells of the mesencephalic and metencephalic alar plates during cerebellar ontogeny in quail-chick chimaeras. Eur J Neurosci 1993; 5:1145–1155. 42. Wingate RJT, Hatten ME. The role of the rhombic lip in avian cerebellum development. Development 1999; 126:4395–4404. 43. Shimamura K, Hartigan DJ, Martinez S, Puelles L, Rubenstein JLR. Longitudinal organization of the anterior neural plate and neural tube. Development 1995; 121:3923–3933.

17

18

M. Catala 44. Rubenstein JLR, Shimamura K, Martinez S, Puelles L. Regionalization of the prosencephalic neural plate. Annu Rev Neurosci 1998 ; 21:445–477. 45. Müller F, O’Rahilly R. The first appearance of the future cerebral hemispheres in the human embryo at stage 14. Anat Embryol 1988; 177:495–511. 46. Müller F, O’Rahilly R. The development of the human brain, including the longitudinal zoning in the diencephalon at stage 15. Anat Embryol 1988; 179:55–71. 47. Müller F, O’Rahilly R. The human brain at stage 17, including the appearance of the future olfactory bulb and the first amygdaloid nuclei. Anat Embryol 1989; 180:353–369. 48. O’Rahilly R, Müller F. The embryonic human brain. An atlas of developmental stages. New York: Wiley-Liss, 1994. 49. Furuta Y, Piston DW, Hogan BL. Bone morphogenetic proteins (BMPs) as regulators of dorsal forebrain development. Development 1997; 124:2203–2212. 50. Monuki ES, Walsh CA. Mechanisms of cerebral cortical patterning in mice and humans. Nature Neurosci (Suppl) 2001; 4:1199–1206. 51. Hayhurst M, Mc Connell SK. Mouse models of holoprosencephaly. Curr Opin Neurol 2003; 16:135–141. 52. Roessler E, Belloni E, Gaudenz K, Jay P, Berta P, Scherer SW, Tsui LC, Muenke M. Mutations in the human Sonic Hedgehog gene cause holoprosencephaly. Nature Genet 1996; 14: 357–360. 53. Ming JE, Kaupas ME, Roessler E, Brunner HG, Golabi M, Tekin M, Stratton RF, Sujansky E, Bale SJ, Muenke M. Mutations in PATCHED-1, the receptor for SONIC HEDGEHOG, are associated with holoprosencephaly. Hum Genet 2002; 110:297–301. 54. Roessler E, Du YZ, Mullor JL, Casas E, Allen WP, GillessenKaesbach G, Roeder ER, Ming JE, Ruiz i Altaba A, Muenke M. Loss of function mutations in the human GLI2 gene are associated with pituitary anomalies and holoprosencephaly-like features. Proc Natl Acad Sci USA 2003; 100:13424– 13429. 55. Ma Y, Erkner A, Gong R, Yao S, Taipale J, Basler K, Beachy PA. Hedgehog-mediated patterning of the mammalian embryo requires transporter-like function of Dispatched. Cell 2002; 111:63–75. 56. Gripp KW, Wotton D, Edwards ML, Roessler E, Ades L, Meinecke P, Richieri-Costa A, Zackai EH, Massague J, Muenke M, Elledge SJ. Mutations in TGIF cause holoprosencephaly and link NODAL signalling to human neural axis determination. Nature Genet 2000; 25:205–208. 57. De La Cruz JM, Bamford RN, Burdine RD, Roessler E, Barkovich AJ, Donnai D, Schier AF, Muenke M. A loss of function mutation if the CFC domain of TDGF1 is associated with human forebrain defects. Hum Genet 2000; 110:422–428. 58. Wallis DE, Roessler E, Hehr U, Nanni L, Wiltshire T, Richieri-Costa A, Gillessen-Kaesbach G, Zackai EH, Rommens J, Muenke M. Mutations in the homeodomain of the human SIX3 gene cause holoprosencephaly. Nature Genet 1999; 22:196–198. 59. Brown SA, Warburton D, Brown LY, Yu C-h, Roeder ER, Stengel-Rutkowski S, Hennekam RCM, Muenke M. Holoprosencephaly due to mutations in ZIC2, a homologue of Drosophila odd-paired. Nature Genet 1998; 20:180–183. 60. Brown LY, Kottmann AH, Brown S. Immunolocalization of Zic2 expression in the developing mouse forebrain. Gene Expr Patterns 2003; 3:361–367. 61. Sauer FC. Mitosis in the neural tube. J Comp Neurol 1935; 62:377–405.

62. Martin AH. Significance of mitotic spindle fibre orientation in the neural tube. Nature 1967; 216:1133–1134. 63. Chenn A, Mc Connell SK. Cleavage orientation and the asymmetric inheritance of Notch1 immunoreactivity in mammalian neurogenesis. Cell 1995; 82: 631–641. 64. Fishell G, Kriegstein AR. Neurons from radial glia: the consequences of asymmetric inheritance. Curr Opin Neurobiol 2003; 13:34–41. 65. Zhong W, Jiang M-M, Weinmaster G, Jan LY, Jan YN. Differential expression of mammalian Numb, Numblike and Notch1 sugests distinct roles during mouse cortical neurogenesis. Development 1997; 124:1887–1897. 66. Chambers CB, Peng Y, Nguyen H, Gaiano N, Fishell G, Nye JS. Spatiotemporal selectivity of response to Notch1 signals in mammalian forebrain precursors. Development 2001; 128:689–702. 67. Shen Q, Zhong W, Jan YN, Temple S. Asymmetric Numb distribution is critical for asymmetric cell division of mouse cerebral cortical stem cells and neuroblasts. Development 2002; 129:4843–4853. 68. Wakamatsu Y, Maynard TM, Jones SU, Weston JA. NUMB localizes in the basal cortex of mitotic avian neuroepithelial cells and modulates neuronal differentiation by binding to NOTCH-1. Neuron 1999; 23:71–81. 69. Petersen PH, Zou K, Hwang JK, Jan YN, Zhong W. Progenitor cell maintenance requires numb and numblike during mouse neurogenesis. Nature 2002; 419:929–934. 70. Hevner RF, Neogi T, Englund C, Daza RAM, Fink A. CajalRetzius cells in the mouse: transcription factors, neurotransmitters, and birthdays suggest a pallial origin. Dev Brain Res 2003; 141:39–53. 71. Bar I, Lambert de Rouvroit C, Goffinet AM. The evolution of cortical development. An hypothesis based on the role of the Reelin signalling pathway. Trends Neurosci 2000; 23:633–638. 72. Rice DS, Sheldon M, D’Arcangelo G, Nakajima K, Goldowitz D, Curran T. Disabled-1 acts downstream of Reelin in a signalling pathway that controls laminar organization in the mammalian brain. Development 1998; 125:3719–3729. 73. Hong SE, Shugart YY, Huang DT, Al Shahwan S, Grant PE, Hourihane J O’B, Martin NDT, Walsh CA. Autosomal recessive lissencephaly with cerebellar hypoplasia is associated with human RELN mutations. Nature Genet 2000; 26:93– 96. 74. Walsh C, Cepko CL. Widespread dispersion of neuronal clones across functional regions of the cerebral cortex. Science 1992; 255:434–440. 75. Walsh C, Cepko CL. Clonal dispersion in proliferative layers of developing cerebral cortex. Nature 1993; 362:632–635. 76. O’Rourke NA, Dailey M, Smith SJ, Mc Connell SK. Diverse migratory pathways in the developing cerebral cortex. Science 1992; 258:299–302. 77. Fishell G, Mason CA, Hatten ME. Dispersion of neural progenitors within the germinal zones of the forebrain. Nature 1993; 362:636–638. 78. Reid CB, Liang I, Walsh C. Systematic widespread clonal organization in cerebral cortex. Neuron 1995; 15:299–310. 79. Tan SS, Breen S. Radial mosaicism and tangential cell dispersion both contribute to mouse neocortical development. Nature 1993; 362:638–640. 80. Nadarajah B, Brunstrom JE, Grutzendler J, Wong ROL, Pearlman AL. Two modes of radial migration in early development of the cerebral cortex. Nature Neurosci 2001; 4:143–150.

Embryology of the Brain 81. Malatesta P, Hartfuss E, Götz M. Isolation of radial glial cells by fluorescent-activated cell sorting reveals a neuronal lineage. Development 2000; 127:5253–5263. 82. Noctor SC, Flint AC, Weissman TA, Dammerman RS, Kriegstein AR. Neurons derived from radial glial cells establish radial units in neocortex. Nature 2001; 409:714–772. 83. Noctor SC, Flint AC, Weissman TA, Wong WS, Clinton BK, Kriegstein AR. Dividing presursor cells of the embryonic corticla ventricular zone have morphological and molecular characteristics of radial glia. J Neurosci 2002; 22:3161– 3173. 84. Tamamaki N, Nakamura K, Okamoto K, Kaneko T. Radial glia is a progenitor of nocortical neurons in the developing cerebral cortex. Neurosci Res 2001; 41:51–60. 85. Miyata T, Kawaguchi A, Okano H, Ogawa M. Asymmetric inheritance of radial glial fibers by cortical neurons. Neuron 2001; 31:727–741. 86. Nery S, Fishell G, Corbin JG. The caudal ganglionic eminence is a source of distinct cortical and sucortical cell populations. Nature Neurosci 2002; 5:1279–1287. 87. De Carlos JA, Lopez-Mascaraque I, Valverde F. Dynamics of cell migration from the lateral ganglionic eminence in the rat. J Neurosci 1996; 16:6146–6156. 88. Anderson SA, Eisenstat DD, Shi L, Rubenstein JLR. Interneuron migration from basal forebrain to neocortex: dependence on Dlx genes. Science 1997; 278:474–476. 89. Tamamaki N, Fujimori KE, Takauji R. Origin and route of tangentially migrating neurons in the developing neocortical intermediate zone. J Neurosci 1997; 17:8313–8323. 90. Lavdas AA, Grigoriou M, Pachnis V, Parnavelas JG. The medial ganglionic eminence gives rise to a population of early neurons in the developing cerebral cortex. J Neurosci 1999; 19:7881–7888. 91. Brun A. The subpial granular layer of the fetal cerebral cortex in man. Its ontogeny and significance in congenital cortical malformations. Acta Pathol Microbiol Scand Suppl 1965; 179:7–98. 92. Letinic K, Zoncu R, Rakic P. Origin of GABAergic neurons in the human neocortex. Nature 2002; 417:645–649. 93. Nadarajah B, Alifragis P, Wong ROL, Parnavelas JG (2003) Neuronal migration in the developing cerebral cortex: observations based on real-time imaging. Cereb Cortex 2003; 13:607–611. 94. Mc Connell SK, Ghosh A, Shatz CJ. Subplate neurons pioneer the first axon pathway from the cerebral cortex. Science 1989; 245:978–982. 95. De Carlos JA, O’Leary DDM. Growth and targeting of subplate axons and establishment of major cortical pathways. J Neurosci 1992; 12:1194–1211. 96. Mc Connell SK, Ghosh A, Shatz CJ. Subplate pioneers and formation of descending connections from cerebral cortex. J Neurosci 1994; 14:1892–1907. 97. Friauf E, Mc Connell SK, Shatz CJ.Functional synaptic circuits in the subplate during fetal and early postnatal development of cat visual cortex. J Neurosci 1990; 10:2601–2613. 98. Ghosh A, Antonini A, Mc Connell SK, Shatz CJ. Requirement for subplate neurons in the formation of thalamocortical connections. Nature 1990; 347:179–181. 99. Ghosh A, Shatz CJ. A role for subplate neurons in the patterning of connections from thalamus to neocortex. Development 1993; 117:1031–1047. 100. Ghosh A, Shatz CJ. Segregation of geniculocortical afferents during the critical period: a role for subplate neurons. J Neurosci 1994; 14:3862–3880.

101. Rakic P. Specification of cerebral cortical areas. Science 1988; 241:170–176. 102. O’Leary DDM. Do cortical areas emerge from a protocortex? Trends Neurosci 1989; 12:400–406. 103. McConnell SK. Fates of visual cortical neurons in the ferret after isochronic and heterochronic transplantation. J Neurosci 1988; 8:945–974. 104. McConnell SK, Kaznowski CE. Cell cycle dependence of laminar determination in the developing cerebral cortex. Science 1991; 254:282–285. 105. Desai AR, McConnell SK. Progressive restriction in fate potential by neural progenitors during cerebral cortical development. Development 2000; 127:2863–2872. 106. Chi JG, Dooling EC, Gilles FH. Gyral development of the human brain. Ann Neurol 1977; 1:86–93. 107. Heffner CD, Lumsden AG, O’Leary DD. Target control of collateral extension and directional axon growth in the mammalian brain. Science 1990; 247:217–220. 108. Métin C, Godement P. The ganglionic eminence may be an intermediate target for corticofugal and thalamocortical axons. J Neurosci 1996; 16:3219–3225. 109. Richards LJ, Koester SE, Tuttle R, O’Leary DD. Directed growth of early cortical axons is influenced by a chemoattractant released from an intermediate target. J Neurosci 1997; 17:2445–2458. 110. Serafini T, Kennedy TE, Galko MJ, Mirzayan C, Jessell TM, Tessier-Lavigne M. The netrins define a family of axon outgrowt-promoting protein homologous to C. elegans UNC-6. Cell 1994; 78:409–424. 111. Kennedy TE, Serafini T, De La Torre JR, Tessier-Lavigne M. Netrins are diffusible chemotropic factors for commissural axons in the embryonic spinal cord. Cell 1994; 78:425–435. 112. Metin C, Deleglise D, Serafini T, Kennedy TE, Tessier-Lavigne M. A role for netrin-1 in the guidance of cortical efferents. Development 1997; 124:5063–5074. 113. Serafini T, Colamarino SA, Leonardo ED, Wang H, Beddington R, Skarnes WC, Tessier-Lavigne M. Netrin-1 is required for commissural axon guidance in the developing nervous system. Cell 1996; 87:1001–1014. 114. Braisted JE, Catalano SM, Stimac R, Kennedy TE, Tessier-Lavigne M, Shatz CJ, O’Leary DD. Netrin-1 promotes thalamic axon growth and is required for proper development of the thalamocortical projections. J Neurosci 2000; 20:5792–5801. 115. Finger JH, Bronson RT, Harris B, Johnson K, Przyborski SA, Ackerman SL. The netrin 1 receptors Unc5h3 and Dcc are necessary at multiple choice points for the guidance of corticospinal tract axons. J Neurosci 2002; 22:10346–10356. 116. Chow CW, Halliday JL, Anderson RM, Danks DM, Fortune DW. Congenital absence of pyramids and its significance in genetic disease. Acta Neuropathol 1985; 65:313–317. 117. Cohen NR, Taylor JSH, Scott LB, Guillery RW, Soriano P, Furley AJW. Errors in corticospinal axon guidance in mice lacking cell adhesion molecule L1. Curr Biol 1997; 8:26–33. 118. Dahme M, Bartsch U, Martin R, Anliker B, Schachner M, Mantei N. Disruption of the mouse L1 gene leads to malformation of the nervous system. Nat Genet 1997; 17:346–349. 119. Demyanenko GP, Tsai AY, Maness PF. Abnormalities in neuronal process extension, hippocampal development, and the ventricular system of L1 knockout mice. J Neurosci 1999; 19:4907–4920. 120. Dottori M, Hartley L, Galea M, Paxinos G, Polizzotto M, Kilpatrick T, Bartlett PF, Murphy M, Köntgen F, Boyd AW. EphA4 (Sek1) receptor tyrosine kinase is required for the

19

20

M. Catala development of the corticospinal tract. Proc Natl Acad Sci USA 1998; 95:13248–13253. 121. Kullander K, Mather NK, Diella F, Dottori M, Boyd AW, Klein R. Kinase-dependent and kinase-independent functions of EphA4 receptors in major axon tract formation in vivo. Neuron 2001; 29:73–84.

122. Kullander K, Croll SD, Zimmer M, Pan L, Mc Clain J, Hughes V, Zabski S, De Chiara TM, Klein R, Yancopoulos GD, Gale NW. Ephrin-B3 is the midline barrier that prevents corticospinal tract axons from recrossing, allowing for unilateral motor control. Genes Dev 2001; 15: 877–888.

Myelination

2 Myelination Cecilia Parazzini, Elena Bianchini, and Fabio Triulzi

CONTENTS 2.1 2.2 2.3

Introduction and Structure of Myelin Technical Aspects 21 Myelination Progression 23 References

21

40

2.1 Introduction and Structure of Myelin Myelination is a very important process of brain maturation because it is essential for neural impulses transmission. It is a dynamic process that starts during fetal life and proceeds predominantly after birth, at least until the end of the third year, in a welldefined, predetermined manner [1–3]. Myelin is a cell membrane with a lamellar structure that wraps around the axon. It is composed of alternating lipids layers with large proteins. The lipidic component is higher than in other cellular membranes, and is formed mainly of cholesterol, glycolipids, and phospholipids. Cholesterol and glycolipids have functional groups interacting with the water and form the outer lipid layer, whereas hydrophobic phospholipids constitute the inner layer. Proteins mainly include myelin basic protein (MBP) and proteolipid protein (PLP). MBP is an intracytoplasmatic protein attached to the inner surface of the cell membrane. It is thought to stabilize the myelin spiral. PLP spans the lipid layer and interacts in the extracellular space with similar PLP chains from other myelin membranes, leading to close apposition of adjacent myelin spirals [4, 5]. As the formation of myelin by the oligodendrocytes and the progressive thickening and tightening of the myelin sheaths proceed, an increase in brain lipid concentration and a decrease in water content takes place [3, 6–9]. Other maturational changes include an increase in cellular and synaptic density and dendrite formation, involving mainly the gray matter and occurring simultaneously with myelination.

Such changes contribute to reduce the amount of free water in the brain [9–11]. This results in a modification of the magnetic resonance (MR) signal, characterized by shortening of both T1 and T2 relaxation times. It is thought that the deposition of lipid is the main factor accounting for the high signal on T1weighted images, whereas the decreased number of water molecules results in the diminished signal on T2-weighted images [4, 9, 12]. Lipid deposition is the main biochemical change in the first step of myelination, whereas subtle changes related to water density predominate in the later phase of the process. As a consequence, T1-weighted sequences detect contrast changes within the myelinating white matter earlier than T2-weighted sequences [13]. The white matter in the newborn is mainly unmyelinated; hence, it has lower signal intensity than gray matter on T1-weighted images, and higher signal intensity than gray matter on T2-weighted images. With advancing maturation, there is an inversion of signal intensity until the adult pattern is reached (white matter high and gray matter low signal intensity on T1; white matter low and gray matter high signal intensity on T2) (Fig. 2.1). On T1-weighted images, the adult appearance is reached at about 12–15 months of age, and minimal changes are seen after this time, whereas on T2weighted images maturation is complete at about 3 years [4, 14]. Maturation of the brain also includes dramatic volumetric growth. The increase in volume is mainly evident in the corpus callosum, the most important commissural system of the brain. The adult thickness is reached at about 9 months of age (Fig. 2.2).

2.2 Technical Aspects The process of myelination is better evaluated on a 1.5 T MR unit. Spin-echo or fast-spin-echo T2 sequences with long repetition time (TR:5500/6000) and echo time (TE:120/200), and spin-echo T1 (TR:500/600;

21

22

C. Parazzini, E. Bianchini, F. Triulzi

a

c

b

Fig. 2.1a–d. Newborn and 3 years of age on T1-weighted images (a, c). Newborn and 3 years of age on T2 W images (b, d). The inversion of signal intensity of gray and white matter at two different ages is evident

d

a

b

c

Fig. 2.2a–c. T1-weighted images. Increase in volume of the corpus callosum. a Newborn, b 6 months, c 9 months

Myelination

TE:10) or inversion-recovery T1 with short inversion time (TR:2000; TE:10; TI:750) should be used. Slice thickness should be 4–5 mm. At least axial and sagittal sections should be obtained. Dedicated coils with high signal-to-noise ratio, such as a knee coil for newborns, are preferred. A perfect immobility of the patient is important; hence, sedation for infants older than 3–4 months is requested. Recently, new MR techniques such as diffusionweighted images and the measurement of the apparent diffusion coefficient (ADC) have been applied to the study of normal brain development. The ADC of the normal neonatal brain is higher than that of the adult brain, and its value progressively decreases with age [15]. Moreover, ADC maps in newborns show a strong contrast between gray and white matter, with ADC white matter values being higher that those of gray matter. In the adult, ADC maps show essentially equal values for white and gray matter. The cause of such age-related ADC decrease is not understood, although it has been postulated that the rapid decrease observed between early gestation and term is due to the decrease in overall water content [16]. Relative anisotropy (RA) values also differ between adult and pediatric brain. RA values for white matter areas are relatively low in infants and increase with age in two different steps. The first increase takes place before the histological appearance of myelin, and has been attributed to changes in white matter structure which accompany the “premyelinating state.” The second, more sustained increase in RA is associated with the histological appearance of myelin and its maturation. Whereas RA values of cortical gray matter in adult brain are close to zero, they are transiently nonzero in premature brain for the earliest gestational age studied (26 weeks), and decreased nearly to zero by 32 weeks [17]. This change reflects microstructural modifications of the cerebral cortex. During the gestational ages for which anisotropy values are nonzero, cortical cytoarchitecture is dominated by the radial glial fibers spanning across the cortical strata and by the radially oriented apical dendrites of the pyramidal cells. With time, this architecture is disrupted by the addition of basal dendrites as well as of thalamocortical afferents, which tend to be oriented orthogonal to the apical dendrites.

2.3 Myelination Progression The myelination process begins very early during intrauterine life. Recent data on preterm newborns

suggest that the water content of the brain decreases between 24 weeks and term. Myelin is evident in numerous gray matter nuclei and white matter tracts of the brainstem and cerebellum around 28 weeks of gestational age (Fig. 2.3, 2.4). New myelin is not visualized anywhere between 28 and 36 weeks gestational age. At 36 weeks, myelin becomes evident in the typical regions of the term newborn (posterior limb of internal capsule, corona radiata, corticospinal tracts). After 37 weeks gestational age, the dorsal brainstem appears diffusely myelinated, probably as a combination of nuclei and white matter tracts; however, it is no longer possible to delineate specific structures [9, 11]. During the first postnatal year, myelin spreads throughout the brain according to a preordered scheme of chronological and topographic sequences. Myelination proceeds centrifugally, from inferior to superior, and from posterior to anterior. In the brainstem, myelination proceeds from the dorsal to the ventral areas. In the cerebral hemispheres, it proceeds from the central sulcus towards the pole and from the occipital and parietal lobes towards the frontal and temporal lobes. Sensory fibers myelinate before motor fibers, and projection pathways earlier than association pathways; sensitive, visual, and auditory tracts are already myelinated at birth. In detail, in the term newborn myelination is evident in the cerebellar flocculi, inferior and superior cerebellar peduncles, cerebellar vermis, dentate nuclei, dorsal brainstem (cranial nerves nuclei and sensitive tracts), decussation of the superior cerebellar peduncles, ventroposterolateral thalamic nuclei, globi pallidi, posterior putamen, posterior portion of the posterior limb of internal capsules, central corona radiata, pre- and postcentral gyri (Figs. 2.5–2.9). At 3 months of age myelination is evident in both T1- and T2-weighted images in cerebellar peduncles, deep cerebellar white matter, and optic radiation. The anterior limb of the internal capsule, splenium of the corpus callosum, and subcortical central white matter appear myelinated only on T1-weighted images. The different timing of the appearance of myelination in T1- and T2weighted sequences becomes evident at this age (Figs. 2.10, 2.11). Myelination proceeds at 5 months in the genu of the corpus callosum, subcortical parieto-occipital white matter, and central parietal lobes, and begins to appear in the frontal lobes on T1-weighted images (Fig. 2.12). At the same time, it is evident on T2-weighted images in the splenium of the corpus callosum and subcortical paracentral white matter (Fig. 2.13). In the following months, myelination proceeds on T1-weighted images involv-

23

24

C. Parazzini, E. Bianchini, F. Triulzi

a

b

c

d

e

f

g

h

i

Fig. 2.3a–i. Preterm newborn (30 weeks gestational age): On T1-weighted images, myelin is evident in numerous white and gray matter structures. a gracile and cuneate fasciculi, b inferior cerebellar peduncles (short arrow), medial longitudinal fasciculus (long arrow), c superior cerebellar peduncles, d medial lemnisci, e lateral lemnisci, f subthalamic nuclei, g ventroposterolateral thalamic nuclei (VPL)

Myelination

a

b

cc

d

e

f

Fig. 2.4a–f. Preterm newborn (30 weeks gestational age): On T2-weighted images, myelin is present in a dentate nuclei of the cerebellum (short arrow), cerebellar vermis (long arrow), inferior cerebellar peduncles, b lateral lemnisci (long arrow), medial lemnisci (short arrow), c inferior colliculi (long arrow), decussation of the superior cerebellar peduncles (short arrow), subthalamic nuclei (large arrow), e ventroposterolateral thalamic nuclei (arrow). No myelination is evident in the posterior limb of internal capsule

ing the ventral portion of brainstem, peripheral cerebellar white matter, and subcortical frontal white matter (Fig. 2.14); at about one year of age the pattern of myelination on T1-weighted images is very similar to that of adults. On T2-weighted images, the anterior limb of the internal capsules and genu of the corpus callosum appear myelinated at 9 months of age; at the same age myelination is evident in the parieto-occipital regions and extends in deep frontal areas. No significant difference between gray and white matter is evident in the cerebellum (Figs. 2.15, 2.16, 2.17). After 1 year of age, myelination proceeds in the temporo-frontal areas (Fig. 2.18), and during the second and third year of age the subcortical regions are involved. Myelination is complete on T2weighted images at about 3 years (Figs. 2.19, 2.20).

Initially, myelination increases at a high rate, but after the first year of life it proceeds more slowly [18]. The last areas to myelinate (so-called terminal zones) are commonly considered to be the peritrigonal regions, i.e., a triangular region posterior and superior to the trigones of the lateral ventricles characterized by persistent hyperintensity on T2-weighted images. However, recent data suggest that T2 high signal in these regions may in fact result from perivascular spaces (Fig. 2.21). On the contrary, the true terminal zones of myelination seem to be the subcortical areas of the frontal and temporal lobes, in accord with autopsy studies in which the subcortical associative fibers, connected with the highest intellectual functions, complete their myelination early in adulthood [12, 19–21] (Figs. 2.22, 2.23).

25

26

C. Parazzini, E. Bianchini, F. Triulzi

a

b

c

d

e

f

g

h

i

Fig. 2.5a–i. Term newborn: On T1-weighted images, myelinated areas are bright. b cerebellar flocculi (short arrow), inferior cerebellar peduncles (long arrow), cerebellar vermis (large arrow), c superior cerebellar peduncles (arrow), d decussation of superior cerebellar peduncles (arrow), e ventroposterolateral thalamic nuclei (long arrow), globi pallidi (short arrow), f posterior portion of posterior limb of internal capsules (arrow), g central corona radiata (arrow), h precentral (short arrow) and postcentral (long arrow) central gyri, i precentral gyrus (arrow)

Myelination

a

b

c

d

e

f

g

h

i

Fig. 2.6a–i. Term newborn: On T2-weighted images myelinated areas are dark. b cerebellar flocculi (short arrow), inferior cerebellar peduncles (long arrow), c superior cerebellar peduncles (short arrow), dentate nucleus (long arrow), cerebellar vermis (large arrow), d decussation of superior cerebellar peduncles (arrow), e ventroposterolateral thalamic nuclei (short arrow) and posterior putamen (long arrow), f posterior portion of posterior limb of internal capsules (arrow), g central corona radiata (arrow), h, i precentral (short arrow) and postcentral (long arrow) gyri

27

28

C. Parazzini, E. Bianchini, F. Triulzi Fig. 2.7a–h. Term newborn: T2-weighted images, sensitive pathway. a gracile and cuneate nuclei, b, c, d medial lemniscus through the brainstem, e ventroposterolateral thalamic nuclei, f posterior limb of the internal capsules, g thalamic radiation, h postcentral gyrus

a

b

c

d

e

f

g

h

Myelination

a

b

c

Fig. 2.8a–c. Term newborn: T2-weighted images, visual pathway. a optic tract, b lateral geniculate nucleus, c calcarine fissure

a

c

b

Fig. 2.9a–c. Term newborn: T2-weighted images, acoustic pathway. a lateral lemniscus (dorsal pons), b medial geniculate nuclei, c inferior colliculi (short arrow), temporal cortex (long arrow)

29

30

C. Parazzini, E. Bianchini, F. Triulzi

a

b

c

d

e

f

g

h

i

Fig. 2.10a–i. Three months. T1-weighted images. Myelination proceeds in a middle cerebellar peduncles (long arrow) and deep cerebellar white matter (short arrow), e anterior limb of internal capsule (short arrow) and optic radiations (long arrow), f anterior limb of internal capsules (short arrow) and splenium of corpus callosum (long arrow), g–i subcortical central white matter

Myelination

a

b

c

d

e

f

g

h

i

Fig. 2.11a–i. Three months: T2-weighted images show progression of myelination in a deep cerebellar white matter, b middle cerebellar peduncles, e Optic radiations

31

32

C. Parazzini, E. Bianchini, F. Triulzi

a

b

c

Fig. 2.12a–c. Five months: FLAIR T1 images. Myelination proceeds in a, b subcortical parieto-occipital white matter (long arrow), genu of the corpus callosum (short arrow) and c paracentral areas, both parietal and frontal

a

b

c

Fig. 2.13a–c. Five months: T2-weighted images. b Splenium of the corpus callosum is myelinated as well as c subcortical white matter of the paracentral area

Myelination

Fig. 2.14. Nine months: T1-weighted images. Myelination extends in ventral portion of brainstem, peripheral cerebellar white matter, and subcortical frontal region

33

34

C. Parazzini, E. Bianchini, F. Triulzi

a

b

c

d

e

f

g

h

i

Fig. 2.15a–i. Nine months: T2-weighted images. Myelination is evident also in d anterior limb of internal capsules (long arrow), splenium of the corpus callosum (short arrow), genu of the corpus callosum (large arrow) e splenium of the corpus callosum (short arrow), subcortical parieto-occipital white matter (long arrow), g deep frontal region

Myelination

Fig. 2.16. One year: T1-weighted images. White matter is almost completely myelinated. The pattern of myelination is very similar to that of the adult

35

36

C. Parazzini, E. Bianchini, F. Triulzi

a

b

c

d

e

f

g

h

i

Fig. 2.17a–i. One year: T2-weighted images. a subcortical cerebellar white matter is myelinated and myelination proceeds in f the frontal white matter

Myelination

a

b

c

d

e

f

Fig. 2.18a–f. Fifteen months. T2-weighted images. Myelination is evident peripherically in a temporal lobes and b–d frontal lobes

37

38

C. Parazzini, E. Bianchini, F. Triulzi

Fig. 2.19. Three years. T1-weighted images. Myelination is complete

Fig. 2.20a,b. T2-weighted images. Linear hyperintensities in the peritrigonal areas, showing a radial course spanning from the ventricle wall towards the periphery, can be referred to perivascular spaces

a

b

Myelination

Fig. 2.21. Three years. T2-weighted images. Myelination is complete. The subcortical fronto-temporal white matter is myelinated

Fig. 2.22a,b. Twenty months. T2weighted images. Hyperintensity is recognizable in both subcortical temporopolar (a) and temporolateral (b) areas

a

b

39

40

C. Parazzini, E. Bianchini, F. Triulzi

a

b

c

Fig. 2.23a–c. T2-weighted images. Myelination of subcortical white matter at 21 (a), 33 (b), and 40 (c) months of age. a Subcortical white matter is bright in prerolandic area and along the first and second convolutions. b myelination is present in the prerolandic area. In c, T2 subcortical hyperintensity is no longer evident and myelination is complete

References 1. Dietrich RB, Bradley WG, Zaragoza EJ 4th, Otto RJ, Taira RK, Wilson GH, Kangarloo H. MR evaluation of early myelination patterns in normal and developmentally delayed infants. AJNR Am J Neuroradiol 1988; 9:69–76. 2. Martin E, Kikinis R, Zuerrer M, Boesch C, Briner J, Kewitz G, Kaelin P. Developmental stages of human brain: an MR study. J Comput Assist Tomogr 1988; 12:917–922. 3. Staudt M, Schropp C, Staudt F, Obletter N, Bise K, Breit A. Myelination of the brain in MR: a staging system. Pediatr Radiol 1993; 23:169–176. 4. Barkovich AJ, Kjos BO, Jackson DE, Norman D. Normal maturation of the neonatal and infant brain: MR imaging at 1.5T. Radiology 1988; 166:173–180. 5. Barkovich AJ. Concepts of myelin and myelination in neuroradiology. AJNR Am J Neuroradiol 2000; 21:1099–1109. 6. Hayakawa K, Konishi Y, Kuriyama M, Konishi K, Matsuda T. Normal brain maturation in MRI. Eur J Radiol 1990; 12:208–215. 7. Korogi Y, Takahashi M, Sumi M, Hirai T, Sakamoto Y, Ikushima I, Miyayama H. MR signal intensity of the perirolandic cortex in the neonate and infant. Neuroradiology 1996; 38: 578–584. 8. McArdle CB, Richardson CJ, Nicholas DA, Mirfakhraee M, Hayden CK, Amparo EG. Developmental features of the neonatal brain: MR imaging. Gray-white matter differentiation and myelination. Radiology 1987; 162:223–229. 9. Counsell SJ, Maalouf EF, Fletcher AM, Duggan P, Battin M, Lewis HJ, Herlihy AH, Edwards AD, Bydder GM, Rutherford MA. MR imaging assessment of myelination in the very preterm brain. AJNR Am J Neuroradiol 2002; 23:872–881. 10. Barkovich AJ: MR of the normal neonatal brain: assessment of deep structures. AJNR Am J Neuroradiol 1998; 19:1397–1403. 11. Battin M, Rutherford MA. Magnetic resonance imaging of the brain in preterm infants: 24 weeks’ gestation to term. In: Rutherford MA (ed) MRI of the neonatal brain. Edinburgh: WB Saunders, 2002:25–49.

12. Curnes JT, Burger PC, Djang WT, Boyko OB. MR imaging of compact white matter pathways. AJNR Am J Neuroradiol 1988; 9:1061–1068. 13. van der Knaap MS, Valk J. Magnetic resonance of myelin, myelination, and myelin disorders, 2nd ed. Berlin: Springer, 1995:31–52. 14. Bird CR, Hedberg M, Drayer BP, Keller PJ, Flom RA, Hodak JA. MR assessment of myelination in infants and children: usefulness of marker sites. AJNR Am J Neuroradiol 1989; 10:731–740. 15. Neil J, Miller J, Mukherjee P, Huppi PS. Diffusion tensor imaging in normal and injured developing human brain – a technical review. NMR Biomed 2002; 15:543–552. 16. Neil JJ, Shiran SI, McKinstry RC, Schefft GL, Snyder AZ, Almli CR, Akbudak E, Aronovitz JA, Miller JP, Lee BC, Conturo TE. Normal brain in human newborns: Apparent diffusion coefficient and diffusion anisotropy measured by using diffusion tensor MR imaging. Radiology 1998; 209:57–66. 17. Neil JJ, Mc Kinstry RC, Shiran SI, Snyder AZ, Conturo TE. Timing of changes on diffusion tensor imaging following brain injury in full-term infants. Ann Neurol 1998; 44:551. 18. van der Knaap MS, Valk J: MR imaging of the various stages of normal myelination during the first year of life. Neuroradiology 1990; 31:459–470. 19. Kinney HC, Brody BA, Kloman AS, Gilles FH. Sequence of central nervous system myelination in human infancy. Patterns of myelination in autopsied infants. J Neuropathol Exp Neurol 1988; 47:217–234. 20. Baierl P, Forster C, Fendel H, Naegele M, Fink U, Kenn W. Magnetic resonance imaging of normal and pathological white matter maturation. Pediatr Radiol 1988; 18:183– 189. 21. Parazzini C, Baldoli C, Scotti G, Triulzi F. Terminal zoneS of myelination: MR evaluation of children aged 20–40 months. AJNR Am J Neuroradiol 2002; 23:1669–1673.

Malformations of the Telencephalic Commissures

3

Malformations of the Telencephalic Commissures Callosal Agenesis and Related Disorders Charles Raybaud and Nadine Girard

3.1 Introduction

CONTENTS 3.1

Introduction 41

3.2

The Anatomy and Morphogenesis of the Forebrain Commissures 42 Normal Radiological Anatomy 42 Anterior Commissure 42 Corpus Callosum 44 Hippocampal Commissure 45 Septum Pellucidum 45 Limbic Structures 46 Midline Cysts 46

3.2.1 3.2.1.1 3.2.1.2 3.2.1.3 3.2.1.4 3.2.1.5 3.2.1.6 3.3

The Development of the Telencephalic Commissures 47

3.3.1 3.3.2

Comparative Anatomy 47 Morphogenesis: The Lamina Reuniens Forms Two Commissural Sites 47 The Role of the Specialized Glia: Glial Tunnel, Glial Sling, Glial Wedge, and Glia of the Indusium Griseum 48 Summary: A Practical Model of Commissuration 49

3.3.3

3.3.4

3.4

Imaging the Commissural Structures

3.4.1 3.4.2 3.4.2.1 3.4.2.2

Technical Issues 49 The Common Form of Commissural Agenesis 50 Classical Complete Commissural Agenesis 50 Classical Partial Posterior Commissural Agenesis 53 Commissural Agenesis with Meningeal Dysplasia 54 Commissural Agenesis with Interhemispheric Meningeal Cystic Dysplasia 55 Commissural Agenesis with Interhemispheric Lipomas 59 Agenesis of a Single Commissure 59 Isolated Agenesis of the Anterior Commissure 59 Isolated Agenesis of the Hippocampal Commissure 59 Isolated Agenesis of the Corpus Callosum 62 Other Commissural Defects 63 Septo-Commissural Dysplasia 64 The Case of the Commissure in Lobar Holoprosencephaly 65

3.4.3 3.4.3.1 3.4.3.2 3.4.4 3.4.4.1 3.4.4.2 3.4.4.3 3.4.5 3.4.6 3.4.7

3.5

Conclusions 66 References

67

49

“Callosal agenesis” is a misnomer. Modern imaging with magnetic resonance (MRI) allows for a detailed anatomic study of the malformation, and demonstrates that in at least 90% of the cases, absence of the corpus callosum is only part of the disorder: typically, the hippocampal commissure is also missing or incomplete. Therefore, the term “commissural agenesis,” or “agenesis of the (forebrain) commissures” should rather be used [1]. In our experience, defects of the telencephalic commissures are the most common malformation of the central nervous system found by MRI in utero [2, 3]. This raises difficult problems of management, as the functional prognosis is uncertain. Many authors have long contended that absence of the corpus callosum per se (the hippocampal commissure was neglected) would not be responsible for any clinical disorder [4]. Only associated malformations would generate the clinical features [5, 6]. In those times, most patients were diagnosed by autopsy or by reluctantly applied radiographic methods, such as pneumoencephalography or angiography. This may have introduced a bias in the detection of affected patients. More innocuous modern imaging modalities are used more liberally in patients with neurological or psycho-intellectual deficits, and demonstrate quite a strong correlation between the pathology and the presence of clinical disorders; the latter, however, seem to be extremely diverse and unpredictable, while the name of “callosal agenesis” is univocal and lacks specificity. Today’s imaging provides exquisitely detailed depiction of morphologic abnormalities. One may presume that a more detailed classification of the abnormalities based on precise anatomical features might uncover more pertinent radiological-clinical correlates, and therefore would allow for more secure prognostic expectations. Although the interhemispheric commissures have seemingly been rediscovered in the 1980s thanks to the acquisition of midline sagittal cuts from MRI, their agenesis has been known for a long time.

41

42

C. Raybaud and N. Girard

Bianchi [7] was the first to describe it (see Bull [8]) in 1749. Reil [9] produced the first detailed autopsy report of a woman with complete agenesis. Forel and Onufrowicz (quoted by Déjerine [10]) already emphasized that agenesis of the hippocampal commissure was associated with the agenesis of the corpus callosum. In 1892, Sachs [11] recognized that the bundle of fibers that run parallel and medial to the lateral ventricle is made of the rerouted fibers of the corpus callosum, and therefore considered it an “heterotopia of the corpus callosum,” rather than an agenesis. This was nine years before M. Probst [12] gave this bundle the name of Balkenlängsbündel, which was translated as the “longitudinal bundle of Probst,” or for simplicity “Probst’s bundle.” Later, the diagnosis became possible in vivo by pneumoencephalography, and Davidoff and Dyke [13] gave the first detailed radiological description of the malformation in 1934. More material was produced by Feld [14] on a pediatric series with clinico-radiologic correlates, and, in the same book, by Gross and Hoff [15] who reported on a large series of 40 autopsy cases, noting that even the longitudinal “heterotopic” bundle could be missing. Then, in 1968, Loeser and Alvord [16] gave a magisterial description of the malformation. In their paper, amongst other observations, they stressed the fact that the laminae of white matter which close the lateral ventricles medially are truly the leaves of the septum pellucidum (kept apart from the midline in the absence of commissuration and of the resulting interhemispheric approximation). This ill-located septal leaf contains the longitudinal bundle of Probst dorsally, and the longitudinal fornix ventrally. They noted that in rare cases, the corpus callosum alone could be missing, with a patent hippocampal commissure, and reported also on the total absence of callosal fibers, without a longitudinal bundle. The same year, Rakic and Yakovlev [17] reviewed the embryological concepts and provided their own analysis of the development of the telencephalic commissures, which after 35 years is still the reference paper. As clinically recognized cases were becoming more common, updates were published by Bull [8] and especially by F.P. Probst [18], whose sum is full of still valuable data. More recent works have mostly addressed modern imaging [19–22] and have attempted to organize the concepts about this malformation [1, 23]. For a clear understanding of the abnormalities encountered in the patients affected with commissural disorders, the following points will be addressed. Our description is based on the assumption that different morphologic types of commissural agenesis may represent different diseases.

● Radiological

anatomy, morphogenesis, imaging of the telencephalic commissures and midline cysts; ● Common form of commissural agenesis, either complete or partial; ● Commissural agenesis associated with meningeal dysplasias (either multicystic or more rarely lipomatous); ● Isolated agenesis of a single commissure; ● Other varieties of commissural dysplasias; ● Related malformations.

3.2 The Anatomy and Morphogenesis of the Forebrain Commissures 3.2.1 Normal Radiological Anatomy The telencephalic commissures are cortico-cortical bundles of white matter extending from one hemisphere to the other, typically but not absolutely in a symmetrical fashion (i.e., corpus callosum) (Fig. 3.1). Association bundles are cortico-cortical bundles of white matter extending from one area of the cortex to another area within one hemisphere (i.e., superior longitudinal bundle, arcuate fibers). Projection tracts are cortico-subcortical bundles of fibers extending between the cortex and subcortical structures (i.e., thalamocortical and corticospinal tracts). The interhemispheric (or telencephalic) commissures are the anterior commissure, the corpus callosum (the most prominent of all in humans), and the hippocampal commissure (or psalterium). An anterior commissure and a hippocampal commissure are found in all vertebrates, whereas the corpus callosum is found in placental mammals only. Surrounded by these commissures, the septum pellucidum also carries commissural fibers. Because the commissures have developed close to the limbic structures, they have special anatomic relationships with the cingulum and the cingular gyrus, and obviously with the hippocampus. 3.2.1.1 Anterior Commissure

The anterior commissure extends from one hemisphere to the other in the depth of the anterior portion of the basal ganglia, crossing the midline at the upper end of the lamina terminalis of the third ventricle (which belongs to the telencephalon: telencephalon medium, or impar) [24]. It relates mostly to the olfac-

43

Malformations of the Telencephalic Commissures

a

b

d

Fig. 3.1a–d. Normal anatomy of the commissures. a Midsagittal STIR image. Normal appearance of the commissural plate. The most prominent commissure is the corpus callosum with, from c the front to the back, its genu (G), trunk (T), isthmus (I), and splenium (S). Below the genu, the rostrum (R) points backward toward the anterior commissure (AC), thus forming the lamina rostralis (open arrow). The anterior commissure (AC) forms a round-to-oval section at the upper end of the lamina terminalis; it is joined to the fornix (F) and, together with it, forms the lower anterior margin of the foramen of Monro. The fornix (F) itself forms the lower limit of the septum pellucidum (sp), then it joins the under-surface of the corpus callosum at the isthmus. The line of attachment is at the midline of the hippocampal commissure. b Coronal STIR image, anterior. The anterior commissure (arrows) is well shown extending from one amygdaloid nucleus to the other, through the lower part of the basal ganglia, below the heads of the caudates. c Coronal T1-weighted 3D RFT image, posterior. The hippocampal commissure is seen extending transversely (arrows). It is attached at the corpus callosum on the midline, and laterally extends to the fornical crura, above the thalamus and the choroid fissure. It forms the roof of the cistern of the velum interpositum, whose floor is the third ventricular tela choroidea. d Coronal STIR image, midportion of the hippocampus. Using an animal comparison, it forms the image of a duck. The beak is the fimbria (F). The head is the cornu ammonis (CA). The neck and the upper torso are the subiculum (S). The anterior and lower torso is the parahippocampal gyrus (PHG). It contains the white matter of the inferior cingulum, the great limbic association tract that runs also under the cingulate gyrus around the corpus callosum (see Fig. 3.1a). (a courtesy of P. Tortori-Donati, Genoa, Italy)

tory system (olfactory bulbs) and to the paleo-pallium: extending the data from the monkey to the man, this includes the amygdaloid complex (cortex and nuclei), some of the insula (anterior-inferior), some of the orbito-frontal cortex (posterior-medial), and some entorhinal-perirhinal cortex [25]. The anterior commissure contains about 3.5 million fibers. On a

sagittal midline section of the brain, it is seen as a rounded or oval thickening of the upper portion of the lamina terminalis, in front of the anterior column of the fornix, in front of and slightly below the interventricular foramen of Monro. Its diameter is never greater than 6 mm (personal data). On coronal cuts, it extends from one amygdaloid nucleus to the other,

44

C. Raybaud and N. Girard

forming an arch with an inferior concavity. On axial cuts, it sweeps from one hemisphere to the other through the anterior-inferior portion of the basal ganglia, looking like the handle-bar of a Dutch-style bicycle. It crosses the midline at the anterior end of the third ventricle, forming, together with the retrocommissural fibers of the fornix, the Greek letter π. 3.2.1.2 Corpus Callosum

The corpus callosum is the prominent commissure of advanced mammals with a large neocortex. In the human brain, it contains about 200 million fibers. It connects almost the entire neocortex of one side with the other side. Most fibers are homotopic reciprocal (two fibers are joining the cortex from one side to the other in a symmetric fashion). About 10% are homotopic and not reciprocal (a fiber joins two symmetric points in one way only). A few are heterotopic (a fiber joins two asymmetric parts of the brain, cortical or subcortical). Although the morphology of the corpus callosum varies among individuals, its general shape can be easily appreciated and described. Its size and thickness evolve in the first months of life [26], depending mostly on myelin content. It sweeps from the anterior commissure anteriorly to the hippocampal commissure posteriorly. It presents several segments, which seem to carry specific sets of commissural fibers. In spite of several studies in the monkey and in man [10, 27, 28], only broad rules of organization can be ascertained. Fibers are distributed from the front to the back according to the corresponding functional cortical areas, but irrespective of the strict lobar or gyral segmentation. ● The

lamina rostralis [29, 30] connects the anterior-inferior aspect of the corpus callosum to the lamina terminalis of the third ventricle and to the anterior commissure. This white matter lamina closes the space (real in fetuses and neonate, virtual afterwards) between the leaves of the septum pellucidum anteriorly. Although it is said not to contain any commissural fiber [28], its constituent white matter must be connected somewhere. It limits the paraterminal gyrus (of the septal area), to which the diagonal band of Broca is attached. ● The rostrum (the beak) is a thickening of the lamina rostralis where it joins the genu. It points backward to the lamina terminalis, below the callosal genu. It contains commissural fibers that are parallel, and contiguous to, the fibers of the anterior commissure [28], in front and below the ante-

rior putamen and nucleus accumbens. It connects the orbital aspect of the frontal lobes. ● The genu (the knee) is the anterior, rounded thickening of the corpus callosum, where it changes direction from inferior to posterior. Its fibers form the forceps minor (because of their shape as they come from the anterior frontal lobes); they connect the prefrontal cortices. The fibers connecting the ventro-medial portions of the prefrontal cortex are in the ventral genu; the fibers connecting the dorsolateral portions are in the dorsal genu [28]. The genu marks the anterior end of the septum pellucidum. Kier and Truwit [29] defined a “MAC line” (standing for mammillary body-anterior commissure-corpus callosum), which illustrates the anterior development of the corpus callosum during evolution. In the rat, rabbit, or cat, the genu is located behind or at the line. In the dog, monkey, and especially the normal human, it should be located in front of the line. If not, it is hypoplastic. ● The trunk (or body) is the portion of the corpus callosum that extends between the genu and the point where the fornix joins its inferior surface. It borders the septum pellucidum superiorly on the midline, and forms the roof of the body of the lateral ventricles laterally. It carries fibers from the precentral cortex, the supplementary motor area, the anterior cingulate gyrus, and the frontoparietal operculum, with the adjoining insular cortex. Laterally, the callosal fibers pass below (and around) the cingular bundle; then, as they diverge further, they cross and mingle with the other tracts of the centrum semiovale (superior longitudinal tract, occipitofrontal tract, corona radiata). ● The isthmus of the corpus callosum is a thinner portion where the corpus callosum is joined by the fornix. It carries fibers from the central, sensorimotor areas (except for the primary sensory representation of the hand and foot which are not interconnected). ● The splenium (the spleen) is the thickest part of the corpus callosum. It protrudes in the ambient cistern, and hangs above the collicular plate, with the great vein of Galen curving around it. Becoming more vertical, it marks a change in the general direction of the corpus callosum. The fibers of the splenium form the forceps major as they spread posteriorly. The upper portion of the splenium contains fibers from the inferior temporal cortex and parahippocampal gyrus, and from the peristriate cortex. Its lower portion contains fibers from the occipital striate areas. The primary visual cortex (area 17) is not connected.

Malformations of the Telencephalic Commissures

3.2.1.3 Hippocampal Commissure

3.2.1.4 Septum Pellucidum

The hippocampal commissure (or psalterium Davidi – the lyre of David) is part of the fornix (vault); it connects the hippocampi with one another. The fornix is made of longitudinal, ipsilateral fibers and of transverse, commissural fibers. The fimbria (the fringe) gathers the projection fibers of the hippocampus; these form a bundle that, coursing posteriorly and superiorly, becomes the crus, or posterior column, of the fornix, as it sweeps around the thalamus on each side. Both columns run toward the midline and meet under the isthmus of the corpus callosum, forming the trunk of the fornix by apposition. Then, the tracts continue forward along the lower margin of the septum pellucidum. In front of the interventricular foramen of Monro, they diverge again and form the anterior columns of the fornix, from which two tracts originate. One, the precommissural tract, continues forward above the anterior commissure to join the septal area. The other, the retrocommissural tract, turns down in front of the foramen of Monro, right behind the commissure (with which it forms the letter π, as described above), and courses within the wall of the third ventricle toward the mammillary body. Longitudinal fibers represent 80% of fornical fibers. Another 20% form the hippocampal commissure. This extends between the fornical crura (posterior columns) across the midline, under the corpus callosum and above the cistern of the velum interpositum. It has the shape of a trigone, with its tip behind the trunk of the fornix (below the callosal isthmus), its midline attached to the undersurface of the posterior corpus callosum, and two wings extending laterally toward the fornical crura. On the midline, it cannot be seen on a sagittal cut. It is best demonstrated on coronal cuts, posterior to the septum pellucidum. Its morphology gives the hippocampal commissure the appearance of a gothic “vault,” or “fornix” in Latin; the name of psalterium (lyre) comes from its striate, triangular shape. When the ventricles are enlarged, especially by hydrocephalus, these laminae tend to become vertical, pushed against the midline, prolonging posteriorly the appearance of the septum pellucidum. The hippocampal commissural fibers connect mostly the CA1 and CA3 sectors of the hippocampi in a symmetrical, homotopic fashion, but heterotopic fibers are also found, connecting CA3 on each side with the contralateral CA1 and dentate [31]. Together, the corpus callosum and the hippocampal commissure form what is commonly called the commissural plate.

The septum pellucidum is interposed between the corpus callosum and the body of the fornix. In the midline sagittal plane, its shape is roughly triangular, with the apex at the anterior commissure, the base at the trunk of the corpus callosum, one side at the lamina rostralis and rostrum, and another at the body of the fornix. Because of this close relationship with the commissures, it should be considered together with them. The septum pellucidum should not be confused with the septum granulosum, or septal area. The septal area is the posterior medio-basal part of the frontal cortex, associated with the septal nuclei. It is located anterior and inferior to, and is continuous with, the septum pellucidum; it prolongs the cingulate gyrus below the genu and the lamina rostralis, in front of the third ventricle. The septal area is part of the limbic structures, and is related to the olfactory tract. The normal septum pellucidum is made of the two apposed laminae of white matter which close medially the lateral ventricles, each belonging to the ipsilateral hemisphere. The upper and anterior margins of the septum pellucidum are attached to the trunk and genu, rostrum, and lamina rostralis of the corpus callosum, respectively. Its lower margin contains the fibers of the longitudinal fibers (columns) of the fornix, where they form the body of the fornix. It is covered with ependyma on its lateral, ventricular aspect, under which the septal veins can be identified, forming prominent subependymal afferents of the internal cerebral veins. The cellular lining of the medial aspect of the septal leaves is quite controversial, but its ventricular nature is no longer accepted. The septum pellucidum is devoid of any significant aggregates of neurons, and contains fibers only. Little is found in the literature about the connections of these fibers: hippocampal fibers diverging from the fornix [10], and/or medial subcallosal radiations of the cingulum, mostly toward and from the thalami [32]. The Fetal and Neonatal Septum Pellucidum

In fetuses and neonates, the leaves of the septum pellucidum are usually not apposed, and instead contain a cavum (Fig. 3.2). The meaning of this transitory cavity has been the subject of considerable controversy over the years [33]. Evidence from embryology and anatomy suggests that it is a remnant of the meningeal space that was isolated from the primitive interhemispheric fissure by the development of

45

46

C. Raybaud and N. Girard

a

b Fig. 3.2a,b. Transient midline cava in fetuses and neonates. a, b Coronal T2-weighted images. a Cavum septi pellucidi. CSF filled space limited by the corpus callosum above, the third ventricular tela choroidea, and the septal leaves with the longitudinal fornices laterally (asterisk). It is an anterior extension of the cistern of the velum interpositum. b Cavum Vergae. It is a blind posterior blind extension of the cavum septi pellucidi between the corpus callosum and the hippocampal commissure (arrowheads). Both cava, together or separately, may become encysted (see Fig. 3.16b)

the commissures. In fetuses and young infants, this cavum septi pellucidi (or Duncan’s cave) is limited laterally by the leaves of the septum pellucidum, and dorsally by the trunk and the genu of the corpus callosum; it communicates with the ventricular lumen through the interspace between the anterior columns of the fornix at the foramina of Monro, and with the anterior interhemispheric fissure when the lamina rostralis is not completed yet. It may present a posterior expansion between the corpus callosum above and the hippocampal commissure below, which has been called the cavum Vergae (or cavum psalterii). These are normal features, limited by normally developed commissures and septal leaves. As the brain matures and grows, the cava recede and disappear, leaving only a small triangular space adjacent to the corpus callosum. 3.2.1.5 Limbic Structures

The limbic structures are anatomically related to the commissures in two ways, either because these are directly connected with them (amygdala-anterior commissure, hippocampus-hippocampal commissure) or because they are in close proximity (cingulate gyrus-corpus callosum). The corpus callosum is covered by the indusium griseum (or hippocampal rudiment) laterally, and is surrounded by the cingulate gyrus, that sweeps around it in the continuity of the parahippocampal gyrus. It overlies a large bundle of white matter, the cingulum (Latin for “belt,” “girth”), which is an association tract linking the whole limbic lobe together, from the septal area to the entorhinal cortex. It forms a circle that is concentric with, and outside of, the circle of the fornix. The corpus callo-

sum is separated from the limbic lobe by the callosal sulcus. 3.2.1.6 Midline Cysts

In rare cases, after a cavum has closed, fluid may slowly accumulate in it, forming a cyst with bulging walls (see Fig. 3.16). All commissures are present and anatomically normal. Usually, the cyst concerns the whole space comprised between the commissures and the septal leaves. More rarely, it concerns only the cavum septi pellucidi, sparing the more posterior cavum Vergae. Exceptionally, it concerns the cavum Vergae only [34], above the hippocampal commissure. These cysts are often asymptomatic but they may be a potential cause for headaches and slowly progressing hydrocephalus. Surprisingly, they are found quite commonly in some types of rare metabolic brain diseases (Alexander’s disease, spongiform encephalopathies). They are not really malformative, but just accumulations of CSF in spaces where it should not accumulate. They should not be confused with midline cystic tumors or with suprasellar cysts, which tend to expand upward between the leaves of the septum because this, anatomically, is a natural pathway. Moreover, they should not be confused with midline malformations in which the septal leaves stay apart because part of the commissural system is missing; these never present bulging walls. All midline cysts described above develop between the corpus callosum and the fornix; therefore, they are different from the cysts of the velum interpositum, which develop as accumulations of fluid between the fornix and the roof of the third ventricle (see Chap. 21).

Malformations of the Telencephalic Commissures

3.3 The Development of the Telencephalic Commissures 3.3.1 Comparative Anatomy Data from comparative anatomy [35-37] show that, although the anterior and hippocampal commissures are common to all vertebrates, a corpus callosum is found in placental mammals only. The most rudimentary brain would be composed of an olfactory bulb with its primitive cortex and nuclei only (septal cortex and nuclei attached to the medial olfactory root; amygdaloid complex attached to the lateral olfactory root). This is referred to as the paleopallium, whose commissure is the anterior commissure. Further evolutionary advance introduces a more sophisticated cortex, the hippocampus with the entorhinal cortex. This forms the archipallium, whose commissure is the hippocampal commissure (with some of the perirhinal cortex still depending on the anterior commissure). The neocortex (neopallium) develops later in evolution, becoming huge in humans; it is interposed between the paleopallium and the archipallium, and its specific commissure is the corpus callosum, similarly interposed between the anterior commissure and the hippocampal commissure. Some species, such as the marsupials, have a neocortex without a corpus callosum, and the neocortical commissural fibers pass though the anterior commissure. In some marsupials [35] part of the neocortical fibers can also pass together with, and above, the hippocampal commissure.

3.3.2 Morphogenesis: The Lamina Reuniens Forms Two Commissural Sites After decades of controversies, Rakic and Yakovlev in 1968 provided a magisterial description of the events leading to commissuration of the cerebral hemispheres [17]. This description was subsequently refined and complemented, but not challenged by more recent studies. The basic fact is that the commissures form not from one, but from two distinct sites, i.e., the lamina reuniens (within the lamina terminalis) for the anterior commissure, and the massa commissuralis (by interhemispheric fusion) for both the hippocampal commissure and the corpus callosum. By weeks 6–8, the upper portion of the lamina terminalis (or lamina commissuralis) of the third ventricle thickens; its densely cellular dorsal portion

forms a lamina reuniens, whereas its loosely cellular ventral portion forms the area precommissuralis, or prospective septal area. Together, they form the telencephalon medium, or impar [24]. By week 10, fibers coming from the ganglionic eminences are seen to approach the midline and cross within the lamina reuniens; these are the pioneer fibers of the anterior commissure. By week 11, a thick, definitive anterior commissure is present. In the meantime, a deep sulcus has developed on the dorsal aspect of the lamina reuniens, called the sulcus medianus telencephali medii (SMTM). The loosely cellular banks of the SMTM, clearly demarcated from the densely cellular isocortical plate that overrides it, form the primordium hippocampi on each side, with the primordia of the longitudinal fornices appearing by week 9. Intense proliferation of cells takes place within each primordium hippocampi, and a juxtaposition of the banks of the SMTM develops. The proliferating cells break through the lamina limitans and invade the meninx primitiva between the banks. This results in fusion of the medial walls of the hemispheres. This area of fusion, developed in the dorsal part of the lamina reuniens, is called the massa commissuralis; it is distinct and independent from the ventral lamina reuniens. By week 10, the pioneer fibers of the hippocampal commissure cross the midline through the massa commissuralis, and the hippocampal commissure is clearly completed by weeks 11–12. A well-defined callosal plate is obvious by weeks 12–13, distinct from the anterior commissure. The corpus callosum is virtually complete by week 20. From the initial massa commissuralis, its development follows the development of the cerebral hemispheres, both caudally and cephalically, following the expansion of the frontal and parietotemporo-occipital lobes. The relative growth of the callosal body is even during maturation. The growth of the genu is more important in the weeks before birth, whereas that of the splenium takes place mostly after birth. The anterior development of the genu, and then the inferior development of the rostrum and lamina rostralis, encloses the subcallosal pocket of the SMTM, which becomes the cavum septi pellucidi. This cavum is found in most adult mammals. It is also found in immature human brains, but not in adult brains of primates (including man) and cetaceans.

47

48

C. Raybaud and N. Girard

3.3.3 The Role of the Specialized Glia: Glial Tunnel, Glial Sling, Glial Wedge, and Glia of the Indusium Griseum More recent experimental works on animal models have refined the understanding of the morphogenesis of forebrain commissures. It has been shown that four modalities exist to allow for crossing of the commissural fibers. ● through a “glial tunnel” in the upper lamina terminalis, for the anterior commissure; ● using the meninges of the interhemispheric fissure, for the hippocampal commissure; ● via an interhemispheric glial support (the “glial sling”), for the pioneer fibers of the corpus callosum in front of the hippocampal commissure; ● by fasciculation along the pioneer callosal fibers anteriorly and along the hippocampal fibers posteriorly, for the bulk of the corpus callosum. Silver et al. [38] and Katz et al. [39] demonstrated that in the mouse forebrain a “bridge-like” structure of specific glial cells, or “glial sling,” is required for commissuration. These cells originate from the germinal matrix medial to the lateral ventricles; they migrate medially through the fused walls of the cerebral hemisphere and form a bridge that extends from one hemisphere to the other across the medial interhemispheric meninx primitiva. The first callosal axons grow along its surface toward the contralateral hemisphere, above the anterior commissure. It has been shown that at least in rats, these pioneer fibers originate from the cingulate cortex [40]. This sling proceeds anteriorly and posteriorly as the brain grows. It should be noted that only the pioneer fibers of the corpus callosum need the sling to cross: later fibers follow and accumulate by a process of fasciculation, using early fibers as a support and mixing with them. This also occurs posteriorly, where the callosal fibers fasciculate along the hippocampal fibers to form the splenium. In a strain of acallosal mutant mice, other experiments have shown a delay in midline fusion and a failure of the sling to form. As a result, the would-be callosal axons form large “neuromas” adjacent to the interhemispheric fissure, similar to the longitudinal bundles of Probst. In other experiments on mouse embryos, surgical severing of the sling results in an acallosal brain that mimics the acallosal mutant; commissuration resumes if an artificial sling-like cellulose support is inserted. Wahlsten [41] and Livy and Wahlsten [42], in the same strain of acallosal mice, showed that the hippocampal fibers follow the pia at the bottom of the interhemispheric fissure without

using a glial sling. They confirmed that the posterior callosal fibers use the pre-existing hippocampal commissure as a support for their progression. More recently, Shu et al. [43, 44] further refined the understanding of the cellular processes of commissuration. Again in the mouse, they demonstrated a specialized “glial tunnel” guiding the anterior commissural fibers through the lamina reuniens of the lamina terminalis. They also demonstrated the repulsive role of two other glial structures to orient the callosal fibers toward the midline glial sling: one is a “glial wedge” located at the dorsomedial aspect of the lateral ventricles, the other is the glia of the indusium griseum which borders the lowermost cortex on the inner aspect of the hemisphere, at the margin of the primitive septum. These glial structures repel the commissural fibers toward the interhemispheric sling, at the level of the presumptive septum pellucidum. In knock-out mice lacking Nfia, the specialized glia of the glial wedge and of the indusium griseum is lacking; the cells of the glial sling migrate abnormally into the septum pellucidum instead of going into the interhemispheric fissure; therefore, the callosal fibers fail to cross, and course into the septum pellucidum instead. The hippocampal commissure is missing as well, whereas the anterior commissure is present but abnormally small, a fact that is noteworthy in that it reproduces the typical human malformation. The other fascinating point that was uncovered in recent years regards the role of the subplate. In the early embryo, the first neurons to migrate to the surface of the neural tube form a peripheral layer of gray matter, the preplate, which becomes split by the arrival of the neurons of the cortical plate into a superficial layer containing the Cajal-Retzius cells (future molecular layer or layer 1 of the cortex), and a deep, transient subcortical layer, the subplate. Mostly developed at the beginning of the third trimester, the subplate serves as a “waiting zone” for the incoming axons; temporary connections are established with associative and commissural axons, until the cortex is mature enough for permanent connections [45]. During the same period, many collateral branches of immature callosal axons are eliminated as the cortex matures [46, 47]. Other than these cellular guidance processes toward and along the sling and the fasciculation process, brain fiber pathfinding depends on many other factors. There is still much to be uncovered on the genic/molecular events controlling commissuration [48-51]. Factors involved in the formation of the cellular sling, the fasciculation process, the guiding of the axonal growth cones, and the control of midline crossing are complex and multiple. This is reflected in the fact that commissural disorders may be part of many different syndromes [52].

Malformations of the Telencephalic Commissures

Therefore, functional prognosis may depend not only on the anatomic disorganization but also on the failure of broader developmental processes.

3.3.4 Summary: A Practical Model of Commissuration From the morphogenetic data of classical embryology, using experimental observations in rodents and with the help of radiological evidence obtained from large clinical series in human patients, a model of commissuration can be proposed to classify the diverse disorders encountered in the clinical practice. Because the anterior commissure raises no real problem to this point, this model will only consider the other structures involved in the commissuration: the septum pellucidum, the hippocampal commissure, and the corpus callosum. The first step is a theoretical acommissural brain. No commissure is linking the two hemispheres and the interhemispheric fissure reaches the tela choroidea of the third ventricle. The medial walls of the lateral ventricles are composed of a medullary velum that extends from the medial cortical margin to the choroid fissure, behind the interventricular foramen of Monro, the primary septum pellucidum. Around the thalamus, it prolongs the fimbria hippocampi (which similarly closes the temporal horn between the hippocampus and the temporal choroid fissure). As the fimbria conveys the fibers leaving the hippocampus, the primary septum naturally also contains these fibers as they travel anteriorly toward the septal area and the mammillary body. It forms the longitudinal, ipsilateral columns of the fornix in the lower margin of the primary septum. The second step is hippocampal commissuration. To cross from one hemisphere to the other, both commissures use the massa commissuralis, representing the area of fusion of the medial walls of the hemispheres that corresponds to the primary septum, just below the indusium griseum. The hippocampal commissure first develops in the posterior part of the primary septum. The crossing fibers leave the crus fornicis at right angles, course within the leaves of the primary septum toward the fusion area, cross the midline, and continue symmetrically on the other side to join the contralateral crus, forming the transverse lamina of the hippocampal commissure. The third step is the initiation of the corpus callosum, starting more rostrally in the massa commissuralis. The fibers cross along the interhemispheric fusion line in the upper portions of the primary septum, below and adjacent to the indusium griseum. With the under-

lying portion of the primary septal leaves on each side, they isolate a space that forms the cavum septi pellucidi; eventually these leaves become apposed to form the mature septum pellucidum. The callosal commissuration process develops cranially as well as caudally [17], following the extension of the hemispheres. The fourth step is growth of the corpus callosum, with the later arriving fibers joining earlier ones by the process of fasciculation. Anteriorly, they follow the pioneer fibers above the septum pellucidum. Posteriorly, when reaching the earlier hippocampal commissure, they fasciculate along its fibers, forming with them the isthmus and the splenium. The consequences of this model are manifold: 1) Both the septum pellucidum and the wings of the hippocampal commissure derive from the same medullary velum that closes medially each lateral ventricle. If commissuration fails, the callosal fibers course longitudinally in the upper portion of this medullary velum (primary septum), forming the Probst’s bundle. Presumably, the hippocampal commissural fibers remain within the columns of the fornix in the inferior portion of the medullary velum. 2) The cavum septi pellucidi is a transitory space between the septal leaves and below the corpus callosum, which posteriorly is continuous with the space between the wings of the hippocampal commissure. Therefore, it is truly a portion of the interhemispheric fissure isolated by the development of the commissures. 3) If the anterior corpus callosum overrides posteriorly the hippocampal commissure before fusing with it, the cavum septi pellucidi extends between the two commissures as a cavity known as the cavum Vergae, enclosed by the septal leaves laterally, the corpus callosum above, and the hippocampal commissure below. 4) As the development of the commissure proceeds in different, successive interrelated stages, different types of malformations may result from the different possible combinations of developmental defects.

3.4 Imaging the Commissural Structures 3.4.1 Technical Issues Obviously, MR imaging is the only valuable method today to investigate the telencephalic commissural disorders. The goals of imaging are multiple: (i)

49

50

C. Raybaud and N. Girard

assessing the morphology of all three commissures; (ii) evaluating the general morphology of the brain, i.e., cortical pattern and white matter as can simply be approached through ventricular morphology (future developments of tract imaging, such as diffusion tensor imaging, may improve on this); (iii) looking for disorders of cortical development; (iv) appreciating the hydrodynamics of contingent fluid collections within the brain and in the meningeal spaces; and (v) hopefully, appreciating the functionality and, as much as possible, the functional organization of this malformed brain. Imaging sequences should therefore demonstrate the morphology of the brain as well as the structure of the tissues. The signal of commissural tracts is usually somewhat different from that of the surrounding white matter, with shorter T1 and T2; this can be used as an aid for identification. In infants and neonates, tissue identity is more difficult to assess because of the immaturity of the brain. It is reasonable to wait for the infant to mature before investigating the brain unless there is a therapeutic need, such as fluid collection to drain or a cranial cleft to repair. The following basic sequences can be used: ● T1-weighted images, preferably inversion recovery

for its superb anatomic definition; ● T2-weighted images, for the anatomy and the iden-

tification of structural abnormalities (i.e., migrational disorders); ● FLAIR is helpful for identifying brain dysplasias in the mature brain; ● T1W 3DFT volumetric acquisitions (MPRAGE, SPGR, etc.) are extremely useful as they allow to reformat images in any desirable plane, and because of thin slice acquisition (about one millimeter). This yields greater anatomic details and minimizes partial volume artifacts. It also allows surface rendering, which can be useful in depicting abnormal sulcation patterns; ● Contrast injection may be useful to image vascular structures, especially the venous tree; ● Special sequences for CSF flow studies are needed to characterize the kinetics of CSF into cysts and cavities, in the brain or in the surrounding meninges; ● For research purposes, functional imaging would allow a better understanding of the organization of the cortex in a brain devoid of commissures and with an abnormal connectivity [53]. The anatomic diffusion tensor imaging depiction of the different fiber tracts could be associated with this functional imaging.

The plane of reference is a problem. The best suited one would be the bicommissural plane, which is roughly parallel to the plane of the corpus callosum in normal brains. However, the anterior commissure may be absent, and the morphology of the hemispheres may be altered by the presence of cystic masses or associated malformations. A rule-ofthumb is to use the plane parallel to the longest axis of the hemispheres. We prefer to obtain T2-weighted and inversion recovery images in a coronal plane, and other T1-weighted images (such as MPRAGE) in axial, coronal, and sagittal planes. Because of the high incidence of associated malformations, MR imaging of the spine should be performed in case of pelvic disorders or of neurological deficit. CT remains a very useful complementary tool in case of malformations of the face or of the base of the skull.

3.4.2 The Common Form of Commissural Agenesis The so-called classical form of commissural agenesis is the most common and typical, and the one usually described in the medical literature. It may be complete or partial. When complete, the corpus callosum, the hippocampal commissure, and in 50% of cases the anterior commissure are absent. When partial, the anterior commissure is always present, and a variable portion of the posterior corpus callosum is missing, together with the adjoining hippocampal commissure. The meninges are normal. Associated malformations are frequent. 3.4.2.1 Classical Complete Commissural Agenesis

Complete commissural agenesis is the most frequent type, representing a third of all cases (Fig. 3.3). By definition, in this form both the corpus callosum and the hippocampal commissure are totally lacking. In half the cases, the anterior commissure is also absent, but this does not seem to further significantly affect the general morphology of the brain. When present, the anterior commissure is often small, much smaller than in normal individuals. While it has frequently been stated that the anterior commissure can be enlarged (because some callosal fibers would use it to cross the midline), a thicker anterior commissure was never observed in our material, and was not reported in mice with experimentally created or genetic agenesis [38, 41, 43, 44]. A rudimentary callosal genu might have been mistaken for an enlarged anterior commissure in old reports. Moreover, absence of thickening

Malformations of the Telencephalic Commissures

a

c

b

d

e

Fig. 3.3a–e. Classical, complete commissural agenesis. a Midsagittal T1-weighted 3D RFT image. Complete absence of the corpus callosum and of the hippocampal commissure. The anterior commissure in this case is present, albeit small (arrow). The roof of the third ventricle bulges upwards slightly. The usual pattern of the cingulate gyrus is not found. b Coronal T2-weighted image, anterior. There is no interhemispheric commissure. The medial cortex is rolled-in over the thalami. The lateral ventricles are away from each other, closed medially by the primordial septal leaves that contain the longitudinal bundles of Probst superiorly (thick arrow) (truly heterotopic, rerouted fibers of the corpus callosum) and the longitudinal fornix inferiorly (thin arrow). The temporal horns have an abnormal appearance, extending below the hippocampus into the parahippocampal gyrus (arrowhead), presumably because of the lack of cingulum. This lack of cingulum could also explain the loss of the normal pattern of the cingulate gyrus. c Axial T1-weighted 3D RFT image. The bodies of the lateral ventricles are wide apart, parallel to each other, separated by the Probst’s bundles (arrows), the rolled-in medial cortices and the interhemispheric fissure. d Coronal T2-weighted image, posterior. Major ventriculomegaly of the atrium, called colpocephaly as it has been assumed to be the persistence of the fetal anatomy. Yet, this ventriculomegaly is already present in young fetuses with the malformation, and is rather likely to reflect a lack of white matter fibers, especially under the medial and inferior occipital cortices. e Axial T2-weighted image (different patient). There is marked colpocephalic dilatation of the trigones. (e courtesy P. Tortori-Donati, Genoa, Italy)

of the anterior commissure is more consistent with the duality of the commissural beds for the anterior commissure on one hand, the corpus callosum and the hippocampal commissure on the other hand; as a consequence, the callosal fibers are not expected to change their path and mix with anterior commissural fibers. However, it is acceptable that a diffuse axonal guidance defect might affect both commissural beds.

Besides the absence of the commissures, the morphology of the medial aspect of the hemisphere is clearly abnormal. The interhemispheric fissure extends downwards to the roof of the third ventricle. The lower margin of the medial cortex is rolled in above the tela choroidea and the thalami. The cortical pattern is modified and no cingulate gyrus can be recognized; instead, there is an unusual arrangement of the median sulci, which more or less converge

51

52

C. Raybaud and N. Girard

towards the third ventricle in the parietal area, and are roughly concentric with an upward concavity in the frontal area. The parieto-occipital sulci are shallow. The roof of the third ventricle bulges upward gently, but a significant expansion into the interhemispheric fissure is rare. When it occurs, it develops as a single cavity continuous with the lumen of the third ventricle and/or the lateral ventricle (type 1 of the classification of the callosal agenesis with cysts, see below). No meningeal abnormality is observed in these patients. The vascular anatomy is abnormal, especially regarding the anterior cerebral artery, reflecting the abnormal morphology of the brain; however, typically no vascular dysplasia is seen. The morphology of the ventricular system is modified as well. The lateral ventricles are shifted laterally. They are closed medially by a lamina of white matter that extends from the rolled-in cortex of the medial aspect of the hemisphere, to the choroid fissure on the dorsal aspect of the thalami, between the temporal horn and the septal cortex on the medial aspect of the frontal lobe. It prolongs the fimbria hippocampi. This lamina of white matter is actually what should have become, on each side, the leaf of the septum pellucidum [16]. In its lower margin, this “primary septum pellucidum” (like a normal septum pellucidum) contains the longitudinal fibers of the fornix, coursing anteriorly around the thalamus from the fimbria toward the anterior third ventricle. In its upper margin, it contains a conspicuously large longitudinal bundle of fibers: it has been known for a century [11, 12] to represent the callosal fibers which failed to cross and instead are rerouted parasagittally, parallel to the midline. This abnormal bundle is commonly called the “longitudinal bundle of Probst.” On coronal cuts it is identified at the superior-medial aspect of the body of the lateral ventricles, showing short T1 and T2 signal in mature brains. It is interesting to note that these fibers, which should have connected symmetrical areas of the cortex, instead connect cortical regions that are not directly connected in normal individuals. For that reason, it has been stressed that the malformation, rather than a callosal agenesis, could be called a callosal heterotopia [11]. This implies a significant alteration of the anatomy of the hemispheric connectivity, which to our knowledge has not been specifically studied. Beside the callosal commissure, it has been reported that at least in knocked-out mice, the hippocampal commissural fibers follow the callosal fibers in their abnormal course within the septum pellucidum [44]. Because the Probst’s bundles encroach on the ventricular lumen, the inner walls of the lateral ventricles are concave medially. Together with the lumen of the

third ventricle, this makes up the typical appearance of a “bull’s head” on coronal images. In about 10% of patients, the frontal horns are hypogenetic, with no apparent lumen, apparently because their walls are apposed; this also can be related to an abnormal arrangement of white matter tracts. On axial images, the bodies of the lateral ventricles are parallel to each other, away from the midline. From lateral to medial, the ventricular lumen, the longitudinal bundles of Probst, the medial cortex of the hemispheres with its underlying white matter, and the interhemispheric fissure are found. The temporal horns also are abnormal. In 80% of the patients, the hippocampus looks unusually rounded on coronal images. Still more strikingly, the ventricular lumen surrounds it and extends within the core of the parahippocampal gyrus, which is practically devoid of white matter. This missing white matter corresponds to the inferior, parahippocampal portion of the cingulum, the limbic fiber bundle that normally courses from the entorhinal cortex to the septal area, in the medial aspect of the brain around the hilum of the hemisphere. As the superior portion of the cingulum normally underlies the cingulate gyrus, its absence is likely to explain the abnormal appearance of the medial cortex, with no clearly demarcated cingulate gyrus. In itself, the cingulate cortex may not necessarily be abnormal, although the lack of commissural and cingular fibers should somewhat modify its connections. Posteriorly, the ventricular atria are markedly enlarged, which is wrongly named a colpocephaly. This term describes the relative atrial dilatation in the normal fetus, but in fetuses with commissural agenesis atrial enlargement is exceedingly pronounced. This ventriculomegaly is significant, occasionally huge, and sometimes asymmetrical. It occurs mostly at the expense of the medial and inferior white matter, and is associated with shallow sulci. It is likely to be related to a defect of the intrinsic association bundles of the occipital lobe. On occasion, a significant enlargement may reflect some degree of hydrocephalus. Many associated abnormalities can be observed (Fig. 3.4; see also Fig. 3.14). Hemispheric disorders may include periventricular or subcortical heterotopia (15%), cortical dysplasia resembling polymicrogyria, or even clefting. The rarely seen fusion of the anterior basal ganglia may suggest some relation with holoprosencephalies. Cerebellar abnormalities (10%) include cystic malformations (Dandy-Walkerlike) and vermian clefts (i.e., rhombencephaloschisis). Visual system abnormalities are common (20%), and include colobomas, microphthalmia, strabismus, Joubert syndrome, and congenital nystagmus. Skull

Malformations of the Telencephalic Commissures

3.4.2.2 Classical Partial Posterior Commissural Agenesis

Fig. 3.4. Complete commissural agenesis: associated lesions. Coronal STIR image. Complete commissural agenesis with malformed hippocampi and a large left temporal subcortical heterotopia (arrowheads). See also Fig. 3.14 for an associated Dandy-Walker malformation

and face defects (25%) include facial clefts, encephaloceles, bifid nose, bifid uvula, acromio-mandibular dysplasia, and Apert syndrome. In simple cases, the face appears large, with hypertelorism that may at least in part reflect the separation. Other disorders affect the heart (9%), the pituitary gland (6%), the spine (6%), and the extremities (6%). Neurological (i.e., epilepsy) and/or intellectual deficits are common (54%); in our series of 32 patients, only 2 could be said to be entirely normal, the malformation having been found incidentally for unrelated reasons. This point should be considered when commissural agenesis is discovered antenatally. In summary, the classical complete commissural agenesis is a compound and extensive defect. Beyond the absence of the telencephalic commissures (corpus callosum and hippocampal commissure, anterior commissure in half the cases), major association bundles, such as the cingulum and the intrinsic bundles of the occipital lobe, are also absent. Abnormal organization of the white matter may be present; the callosal and hippocampal commissural fibers are rerouted along the medial aspect of the lateral ventricle, thus forming the “heterotopic” Probst’s bundle, with abnormal intrahemispheric connections. The longitudinal fornices are presumably normal. All abnormalities (failed crossing process, missing cingulum, and missing intrinsic occipital bundles) are located in the medial part of the hemisphere. As no abnormality is observed in the interhemispheric fissure, this points to a histogenetic defect there, that prevented the proper organization of the fibers.

What is meant by “partial commissural agenesis” is difficult to define, and is probably the source of enduring confusion (Fig. 3.5). These disorders probably form an heterogeneous group, with diverse but still uncertain mechanisms. This difficulty has been emphasized by Rubinstein et al. [21]. To overcome at least part of the problem, two groups of partial defects should be distinguished. The term “partial agenesis” can clearly be used to describe cases in which a variable posterior portion of the corpus callosum and of the associated hippocampal commissure is missing, as if it were amputated. The term “partial hypoplasia” should instead be used for the wide spectrum of cases in which the commissural plate is too short but normally shaped, or shows a posterior tapering. This is to assume that an agenesis can be explained by a local failure of the commissuration process, while an hypoplasia, either global or partial, might be due to a more diffuse disturbance of fiber production or guidance. However, this distinction is made more difficult by the possible association of both. Our series of partial defects includes 28 cases. Of these, 15 presented with a typical partial agenesis affecting the splenium/isthmus. A normally shaped but short commissural plate was found in 6. Posterior tapering with a normal anterior commissural plate was found in 5. An extreme hypoplasia with a missing posterior part was found in 2. In the typical partial posterior commissural agenesis, the defect concerns the posterior portion of the commissural plate. Normally, the posterior end of the corpus callosum is made of a more or less rounded splenium located above the collicular plate, surrounded inferiorly and posteriorly by the vein of Galen, right in front of the attachment of the falx cerebri on the tentorium cerebelli. The extent of the partial posterior commissural agenesis may vary in different patients, i.e., the splenium only or the splenium and the isthmus, also a portion of the trunk, or even the whole trunk. In extreme cases, only an hypoplastic genu remains (possibly misinterpreted as an enlarged anterior commissure in the older literature). Together with the posterior corpus callosum, the related portion of the hippocampal commissure is also missing, with a cavum septi pellucidi anteriorly. This may point to a possible morphogenetic mechanism. To build up the splenium, the caudal fibers of the corpus callosum fasciculate along the fibers of the hippocampal commissure. One may assume that the primary defect here would be a missing hippocampal commissure. However, when the agenesis extends also anterior to the isthmus (i.e.,

53

54

C. Raybaud and N. Girard

a

b

Fig. 3.5a–c. Classical, partial commissural agenesis. a Sagittal T1weighted image. The anterior portion of the commissural plate is present and appears essentially normal (arrowheads). The posterior portion, including the posterior corpus callosum and the hippocampal commissure, is lacking. b Coronal T1-weighted 3D RFT image, anterior. Normal appearance of the brain section. c Coronal T1-weighted 3D RFT image, posterior. Appearance typical for a commissural agenesis, including the Probst’s bundles (arrow)

anterior to the hippocampal commissure), one should assume that the defect also affected part of the sling. An unifying alternative would be the existence of common defects of the medial hemispheric wall or of the leptomeninges, preventing both the hippocampal fibers (and the splenial fibers as well) to cross, and the sling to form properly. On midsagittal MR imaging sections, absence of the posterior segment of the commissural plate is obvious. The more the agenesis extends to the trunk of the corpus callosum, the more the septum pellucidum is hypoplastic, often too thick, with the fornices lying close to the present callosal segment. Often, this segment itself is hypoplastic. The rostrum and the lamina rostralis may be hard to identify. The anterior commissure is usually seen, and is sometimes extremely small. On coronal cuts, the brain looks normal where the corpus callosum is present, with separated septal leaves; where the commissures are lacking, the brain looks like a typical agenesis with Probst’s bundles, separated septal leaves, abnormal cortical pattern

of the medial aspect of the hemisphere, and missing cingulum. Both ventricular and white matter abnormalities are less marked than in complete agenesis. Yet, when most of the corpus callosum is missing, the brain looks like it does in complete agenesis except for a narrow callosal bundle above the anterior end of the third ventricle, in front of the foramina of Monro. Associated abnormalities are of the same type as those associated with complete commissural agenesis: migrational disorders, hypothalamo-pituitary defects (20%), visual defects (12%), cerebellar abnormalities, cranial clefts, etc. Epilepsy is common, and the neurologic/intellectual condition of the patients is frequently impaired (92%).

3.4.3 Commissural Agenesis with Meningeal Dysplasia The meninges are closely related to the brain, both anatomically and developmentally. The telencephalic

c

Malformations of the Telencephalic Commissures

meninges are derivatives of the anterior neural crest, together with the membranous skull, among others [54]. In the embryo, the meninx primitiva forms a solid connective tissue that surrounds the brain. At the end of the embryonic period, the meninx primitiva progressively differentiates to produce the compact dura mater by condensation, and the fluid-filled leptomeninges by vacuolization, whereas the fourth ventricle opens in the cisterna magna. The process of vacuolization starts at the ventral mesencephalon and from there proceeds over the entire central nervous system. The last region to differentiate is in the dorsal midline, the region to become the interhemispheric fissure and the velum interpositum [55]. The meninges seem to play a significant role in the development of the central nervous system, first allowing diffusion of nutrients in the embryo and later providing support to the vessels. Together with the Cajal-Retzius cells of the cortical molecular layer, the pial lining of the cortex prevents overmigration of the neuroblasts. In the interhemispheric fissure, the leptomeninges allow the glial sling to form and the hippocampal fibers to cross [42]. The cortex adjacent to a meningeal lipoma developing over the convexity is usually dysplastic [56]. Therefore, dysplasia of the interhemispheric meninges may interfere with the development of the telencephalic commissures. Two types of such dysplasia are commonly encountered: the common multicystic dysplasia of the interhemispheric meninges, and the less frequent interhemispheric lipomas.

a

3.4.3.1 Commissural Agenesis with Interhemispheric Meningeal Cystic Dysplasia

The conspicuously frequent association of interhemispheric cysts with a commissural agenesis has been known since the first descriptions of the malformation [18, 57] (Fig. 3.6). It is not rare, accounting for 15% of cases in our series. In spite of the confusing terminology (i.e., arachnoid, neu-

Fig. 3.6a–c. Commissural agenesis with cystic meningeal dysplasia. a Sagittal T1-weighted image. No commissure is seen. There are multiple meningeal cystic, noncommunicating cavities; most are unrelated to the ventricles. Notice concurrent Chiari I malformation. b Axial T2-weighted image. There are multiple interhemispheric cavities. The right lateral ventricle is enlarged, whereas the left lateral ventricle is difficult to identify. The cortex of the medial aspect of the left cerebral hemisphere is dysplastic, with subcortical heterotopia. c Coronal T2-weighted image confirms cystic dysplasia of the interhemispheric fissure. (Case courtesy P. Tortori-Donati, Genoa, Italy)

b

c

55

56

C. Raybaud and N. Girard

roepithelial, diencephalic cysts), these cysts have always been assumed to represent dorsal expansions of roof of the third ventricle, unrestrained in the absence of the overlying commissural plate. Histology in selected cases confirmed this hypothesis by showing tufts of choroid plexus or patches of ependymal lining. All these data rested upon air contrast studies, surgery, or autopsy. With the advent of modern imaging modalities performed without skull opening in large series of patients, things seem to be more complex. To better understand the different situations, a classification has been recently proposed [58]. In all patients presenting with an interhemispheric cyst, commissural agenesis is found. In the vast majority of patients studied with transfontanellar ultrasonography or MR imaging (especially with CSF flow imaging sequences), the cyst appears to be composed of multiple, noncommunicating cavities. This corroborates the neurosurgical observation that a single cystoperitoneal shunt usually fails to evacuate the whole cystic complex. MR studies may also show associated abnormalities of the adjacent dural structures, particularly the falx. The anterior third ventricle may communicate with the nearest cyst, but this is not always the case. Sometimes, the cyst is apparently single and close to, but separate from, the ventricle. In selected cases with follow-up, these cysts enlarge progressively over the years. Fluid shunting is indicated to correct compression of the adjacent brain. The real nature of the cystic walls is uncertain, as no craniotomy is ever performed in these patients. Ancient reports stated that the cysts were of neuroepithelial origin, and at least partly choroidal. The presence of these neuroepithelial tissues in the subarachnoid space cannot be explained by normal embryology. It has been postulated that it could result from abnormal migration into the leptomeninges of the cells of the tela choroidea [22]. However, the ependymal and pial layers that constitute the tela choroidea do not migrate; therefore a differentiation disorder seems more likely. These multiloculated cysts are clearly different from the huge interhemispheric expansion of the tela choroidea of the third and/or of one lateral ventricle, that can rarely be observed in patients with a common commissural agenesis (Fig. 3.7) and, more typically, in rare cases of global septal and commissural agenesis (see below, septocommissural dysplasia). In these, the fluid collection forms a single cavity that is continuous with the ventricle, and draining the cyst produces a global loss of volume of the cystic cavity and of the ventricles.

a

b

c Fig. 3.7a–c. Commissural agenesis with upward expansion of the third ventricle. a Sagittal T1-weighted image shows complete agenesis of the commissural plate associated with expanded midline cerebrospinal fluid-filled space. b Axial T2-weighted image shows uniloculated interhemispheric cavity. The lateral ventricles have a typical parallel course. c Coronal T2-weighted image shows the “cyst” in fact results from upward expansion of the third ventricle. The appearance is otherwise typical for commissural agenesis. (Case courtesy P. Tortori-Donati, Genoa, Italy)

Malformations of the Telencephalic Commissures

Because the mass effect usually distorts the adjacent brain structure, the MR analysis of the hemispheric malformation is difficult. Agenesis of the commissures may be total or partial. The evaluation of the septal leaves, of the fornices, and even of the hippocampi is difficult, and their apparent condition varies from case to case. Shunting does not make this evaluation easier. The presence of the Probst bundles is hard to ascertain, and they may seem to be absent on the side of the cyst while present on the other side. Nevertheless, beyond the commissural agenesis, the brain is obviously dysplastic. The cortex lining the interhemispheric cysts is commonly dysplastic. Subcortical heterotopia are common, sometimes huge, as well as periventricular heterotopias. In some instances, cavities are found within the brain that seem to prolong the extracerebral cysts, but it is difficult to recognize whether such lesions are constitutive or, rather, they result from the deformity induced by the fluid masses. Affected children are usually referred for macrocephaly. Surprisingly, beside the problems with CSF dynamics, many seem to do better neurologically than patients affected with the common type of commissural agenesis. Some, however, are severely disabled, with mental deficits and/or severe epilepsy. Some present the features of the Aicardi syndrome (see below). In summary, it seems likely that interhemispheric multiloculated cysts are due to dysplasia of the interhemispheric meninges, and that this dysplasia might interfere with the development of the commissures and of the adjoining cortex. The mechanism of the frequent coexistence of subcortical gray matter heterotopia is uncertain. All these relate to different periods of the histogenesis, i.e., probably the 10th–12th week for the interhemispheric meninges, 8th–20th week for the gray matter heterotopia, and 18th–23rd week for the cortical dysplasia, although all abnormalities must obviously proceed from a common initial disorder. The Aicardi Syndrome

The Aicardi syndrome [59] was defined as an association of asymmetrical infantile spasms, agenesis of the corpus callosum, and chorioretinal lacunae, found in infant girls (Fig. 3.8). Since the first report, progresses with imaging have shown that, other than nonspecific commissural agenesis, the brains of affected patients always show a whole set of more characteristic, almost constant “migrational” disorders, including subcortical and periventricular nodular heterotopia, as well as cortical

polymicrogyria. Eye colobomas are also common. Single or multiple cyst formation in the choroid plexuses, especially in the ventricular atria and in the region of the posterior third ventricle, is characteristic, with a signal different from that of CSF expressing the absence of continuity with the subarachnoid space. These cysts have been reported to be of ependymal origin [59]. Other cysts may develop within the brain or in the posterior fossa. The incidence of the Aicardi syndrome is uncertain, perhaps accounting for 4% of infantile spasms. The prognosis is poor, with severe neurologic and mental impairment, persisting epilepsy, and a 40% survival rate at 15 years. Classification Proposals

Facing the obvious heterogeneity of these malformations, several authors have attempted to classify them. The simplest classification is that of Probst [18] who, using data from pneumoencephalography, simply distinguished between communicating and noncommunicating cysts. In 1998, using more detailed data from MR imaging, we proposed three types of malformations according to the nature of the meningeal and parenchymal dysplasias [1]: ● Type

1: the cysts are simple ventricular expansions, and therefore considered as hydrodynamic variants of the classical commissural agenesis. ● Type 2: a complex, multicystic meningeal dysplasia is associated with gray matter heterotopia and cortical dysplasias bordering the cystic cavity, together with the commissural agenesis. ● Type 3: the cysts are located not only in the meninges, but also within the parenchyma; they are dysgenetic or perhaps destructive, associated with gray matter heterotopia and cortical dysplasias (this type might include the Aicardi syndrome). In 2001, Barkovich et al. proposed a more detailed classification comprising two main groups, each with three subdivisions [58]: ● Type

1: the cyst is a diverticulation of the third and/or the lateral ventricle. This group is further subdivided into three subgroups: – Subtype 1a: macrocephaly, ventricular dilatation related to a simple communicating hydrocephalus, with or without Dandy-Walker malformation, parietal meningocele or falx hypoplasia; the corpus callosum is totally or partially absent. This group seems to concern boys only. – Subtype 1b: macrocephaly and hydrocephalus

57

58

C. Raybaud and N. Girard

a

b

c

d Fig. 3.8a–d. Aicardi syndrome. a Sagittal T1-weighted image. The commissural plate is absent. There is a thick interthalamic mass (IM) and an interhemispheric cavity (asterisk). b Axial T1-weighted image. There is right microphthalmos with congenital cataracts. The interhemispheric cavity is also displayed (asterisk). c Axial T2-weighted image. There is diffuse cortical dysplasia involving both frontal lobes, albeit more markedly to the left. This dysplasia shows a variable combination of subependymal heterotopia, subcortical heterotopia, and polymicrogyria. The ventricles are distorted. d Coronal T1-weighted image. The frontal horns have a crescentic shape due to absence of the corpus callosum. However, there is distortion of the ventricles due to concurrent cortical dysplasia. (Case courtesy M. Rutherford, London, UK)

related to thalamic fusion; callosal agenesis is complete or partial; the falx is absent, or deficient in relation to meningoceles. This group raises the problem of the differential diagnosis with the holoprosencephaly spectrum, although thalamic fusion was not felt to be specific to this malformation.

– Subtype 1c: microcephaly (instead of macrocephaly), complete callosal agenesis, inferior vermian hypogenesis and brainstem kinking, cyst involving the tela choroidea of both the third and the lateral ventricles, unilaterally. This category could be related to either cerebral dysgenesis or a prenatal destructive process.

Malformations of the Telencephalic Commissures

● Type 2: the cysts are multiloculated and intraparenchymal. Three subtypes again are observed: – Subtype 2a: macrocephaly, multiloculated cysts with no associated brain anomaly other than callosal agenesis. – Subtype 2b: macrocephaly, multilocular cysts, callosal agenesis, polymicrogyria, and subependymal heterotopia. All patients are female. The cysts presumably do not communicate with the free CSF (showing different signal). The morphologic abnormalities of the Aicardi syndrome may fit into this subtype, as does the fact that all patients are female. – Subtype 2c: Multilocular, noncommunicating cysts, callosal agenesis, subcortical heterotopias (cerebellar cortical dysplasia in one case). 3.4.3.2 Commissural Agenesis with Interhemispheric Lipomas

Interhemispheric lipomas associated with commissural agenesis have been known for a long time, and the causal relationship between these abnormalities has been much debated (Fig. 3.9). Verga in 1929 [60] was the first to assume that meningeal lipomas, instead of being dysraphic inclusions, fatty tumor, meningeal degeneration, or misplaced heterotopia, were a maldifferentiation of the meninx primitiva. Truwit and Barkovich [61] carefully reviewed the arguments in favor of Verga’s hypothesis: the subarachnoid location, the absence of other mesodermal derivatives, and the intralesional location of vessels (often dysplastic) and nerves. This early meningeal dysplasia may presumably interfere with the development of the brain, and especially of the commissures, but the mechanism is uncertain. The hypothesis of the lipoma being a mechanical obstacle to posterior growth of the commissural plate is unlikely, as there are anterior lipomas associated with posterior partial commissural agenesis, or lipomas intermingled within callosal fibers. Yet, there is a relationship between the location of the lipoma and the commissural defect. It appears that anterior, bulky (i.e., tubulonodular) lipomas are more often associated with gross agenesis whereas smaller, posterior ribbonlike (i.e., curvilinear) lipomas are associated with more subtle commissural hypoplasia. This may be related to the time when the dysplasia occurred, in relation to the timing of the commissural development. Clinically, patients with anterior, bulky lipomas are more severely disabled than those with more posterior curvilinear lipomas [62]. Concurrent lipomas or extensions of the main lipoma may involve the choroid plexuses. Calcifications are relatively common. Angiography would show arterial dysplasias in the lipomatous segment of

the arteries. The lipoma is not a tumor, and therefore it does not expand. However, it does enlarge during the first weeks of life together with the physiological development of the body fat of the young infant. These midline lipomas are not common, accounting for 3% of our whole series of commissural defects. The most usual presenting feature is epilepsy, typically partial; some mental deficit and, on occasion, a cranio-facial malformation may be observed.

3.4.4 Agenesis of a Single Commissure These infrequent forms of commissural agenesis are not always diagnosed as such (i.e., agenesis of the hippocampal commissure is read as “persistent cavum septi pellucidi”), or are not diagnosed with anatomic certainty. It may be assumed that the causal defect is specific to the commissure concerned, or to its own particular commissuration mechanisms, and not to the general commissuration processes. A careful assessment of the anatomic features should be obtained though appropriate imaging sequences and planes. 3.4.4.1 Isolated Agenesis of the Anterior Commissure

Isolated agenesis of the anterior commissure is rare (3% of all commissural ageneses) (Fig. 3.10). The abnormality is obvious on the midline sagittal plane, where the commissure is not seen in the upper part of the lamina terminalis. It should be confirmed in axial and coronal planes as well, with thin, contiguous slices. On axial cuts, the anterior columns of the fornix are visible, but are widely splayed and lack the transverse anterior commissural band (neither the handle-bar of the Dutch bicycle nor the letter π). On coronal cuts, the commissure arching between the amygdaloid nuclei is not seen, whereas the leaves of the septum pellucidum diverge, the columns of the fornices staying wide apart. The rostrum and the lamina rostralis of the corpus callosum may be also abnormal, with a deformity of the corpus callosum. The clinical picture associates epilepsy and developmental delay. 3.4.4.2 Isolated Agenesis of the Hippocampal Commissure

Careful reading of the anatomic features of the commissures shows that a significant number of patients diagnosed at a first look as having a “persisting cavum septi pellucidi” actually lack a hippocampal commissure (Fig. 3.11). The idea of a “normal”

59

60

C. Raybaud and N. Girard

a

b

d c

e Fig. 3.9a–e. Commissural agenesis with interhemispheric lipoma. Three different cases. a Curvilinear lipoma in a 17-year-old girl. Sagittal T1-weighted image. There is a thin stripe of fatty tissue surrounding a mildly hypoplastic corpus callosum (arrows). b–d Tubulonodular lipoma. b Sagittal T1-weighted image at age 3 months shows lipoma overlying a thin, dysplastic corpus callosum. c Sagittal T1-weighted image at age 3 years shows the lipoma has significantly enlarged. The dysplastic corpus callosum also is thicker after completion of myelination. d Coronal T1-weighted image shows extension of the lipoma into both lateral ventricles. e Tubulonodular lipoma. Sagittal STIR image. The lipoma was associated with an anterior cranial cleft. It likely represents a meningeal dysplasia, which in turn may explain the commissural agenesis. As the lipoma is anterior, its action in preventing the callosal development was not simply mechanical. (a–d courtesy P. Tortori-Donati, Genoa, Italy)

Malformations of the Telencephalic Commissures

a

b Fig. 3.10a,b. Isolated agenesis of the anterior commissure. a Midsagittal T1-weighted 3D RFT image. Absence of the anterior commissure is associated with defect of the anterior corpus callosum, appearing somewhat small. b Coronal T1-weighted image. No image of the anterior commissure is seen between the amygdaloid nuclei. Diverging septal leaves as the fornices have lost the usual relationship with the commissure

a

b

c

d Fig. 3.11a–d. Isolated agenesis of the hippocampal commissure. a Midsagittal T1-weighted 3D RFT image. Arched appearance of the corpus callosum; the fornix is not seen. b–d Consecutive coronal T1-weighted 3D RFT images cuts from the front to the back of the corpus callosum. Persisting space between the septal leaves (asterisk, b–d), which could be misread as a “persisting cavum septi pellucidi.” In fact, the fornices can be followed to the posterior part of the crura (arrows), and no transverse hippocampal commissure is found

61

62

C. Raybaud and N. Girard

persisting cavum stems from the common finding in neonates of a nonfusion of the septal leaves (see above). However, in these normal neonates a close analysis of the posterior part of the commissural plate shows the transverse velum of the hippocampal commissure crossing below the posterior corpus callosum. On the contrary, in a significant number of older patients in whom the septal leaves are widely apart, no hippocampal commissure can be identified. Anteriorly, the fornices are separated as the septal leaves are. Posteriorly, the same pattern is found, with the septal leaves extending as far back as to the splenium, away from each other and perpendicular to the lower surface of the corpus callosum. On a midline sagittal cut, the appearance is still more striking, as no image of the fornix is found interposed between the corpus callosum and the third ventricular roof. Among our six patients (6% of cases), neurodevelopmental deficits were observed in five. In one of them there also was diencephalo-mesencephalic dysplasia, and the patient suffered with growth delay. In another, there was associated cortical dysplasia. These considerations imply that, in opposition to the classical ideas dating back to pneumoencephalography era, this is neither a persistent cavum septi pellucidi nor a normal variant, but a true developmental defect. This corroborates the fact that such feature was commonly observed in mentally retarded or epileptic patients [63]. The agenesis of such an important structure is more likely to explain this type of clinical feature. Identifying the hippocampal commissure, or on the contrary ascertaining its absence, is therefore extremely important while evaluating these patients. From an embryological point of view, agenesis of the hippocampal commissure without agenesis of the posterior portion of the corpus callosum is problematic in that, according to normal embryological data, the posterior corpus callosum requires a pre-existing hippocampal commissure to fasciculate. Possibly, callosal fibers are able to cross without the support of the hippocampal fibers, perhaps because the glial sling would extend posteriorly. In two additional patients, we found a combination of agenesis of both the anterior and the hippocampal commissure. These patients also had midline defects, pelvic disorders, developmental delay, and convulsions. 3.4.4.3 Isolated Agenesis of the Corpus Callosum

There are instances in which only the corpus callosum is missing, whereas the hippocampal commissure is present [1, 64] (Fig. 3.12). Such cases are commonly

a

b Fig. 3.12a,b. Isolated agenesis of the corpus callosum. a Midsagittal T1-weighted 3D RFT image. An hypoplastic lamina of white matter (arrows) is seen extending backwards from the region of the foramen of Monro, above the roof of the third ventricle. It could be misread as an hypoplastic corpus callosum, but its location is too low with respect to the medial hemispheric cortex. b Coronal T1-weighted 3D RFT image. The appearance of the brain is close to that of a complete commissural agenesis, except that the transverse lamina of the hippocampal commissure (arrowheads) is seen extending between the two fornices (open arrow), and not above where a normal corpus callosum should be. The callosal bundles indeed are present but form typical Probst’s bundles (black arrow). Only this variety of commissural disorder truly deserves the name of callosal agenesis

misread as hypoplastic corpus callosum, but a careful analysis of the anatomy corrects the diagnosis. This malformation is not frequent: it accounts for 11% of our material, the corpus callosum being totally (3%) or only partially (8%) agenetic. In the complete isolated agenesis of the corpus callosum (viz. with normal anterior and hippocampal commissures), the interhemispheric fissure is large, and the lateral ventricles are away from each other, in a way similar to complete commissural agenesis. Yet, on a midline sagittal cut, a band of white matter extends posteriorly from the upper lamina terminalis. It is located too low to be a corpus callosum, just above the roof of the third ventricle (from which it is separated by the normal cistern of the velum interpositum), but not above the foramina of Monro: this cannot correspond to the proper location of the corpus

Malformations of the Telencephalic Commissures

callosum. The coronal cuts are more explicit, in that the lateral ventricles are closed by the usual medial leaf of septal white matter, containing the Probst’s bundle above and the longitudinal fornix below. The interhemispheric band of white matter of the hippocampal commissure crosses the midline from one fornix to the other, in its right position. Partial segmental agenesis of the corpus callosum is not so uncommon, although still undescribed in the classical literature to our knowledge. We have seen it in eight patients (8% of our cases). It affects the midportion of the corpus callosum, the splenium being present. As the corpus callosum is said to develop from the front to the back, this defect of its middle portion is usually assumed to be destructive: for instance, bilateral posterior frontal schizencephalic clefts, or bilateral anoxic necrosis of the sensorimotor area in the neonate, may produce a closely related appearance. However, the segmental agenesis of the corpus callosum with preservation of the splenium can be observed in patients showing no other brain abnormality, and especially without destructive lesions likely to have affected the callosal fibers. Moreover, we have observed it in two siblings presenting the same clinical features (isolated spasticity of the lower limbs with normal intelligence). This familial occurrence with quite identical brain images obviously suggests a developmental origin; both exhibited a defect of the median portion of the corpus callosum (Fig. 3.13) with a preserved genu and a present, albeit hypoplastic, splenium. On coro-

nal cuts, the images of the brain were normal anteriorly and posteriorly, but at the level of the defect, the appearance was the same as that of complete callosal agenesis: no callosal fibers but a patent hippocampal commissure extending from one fornix to the other. No associated parenchymal dysplasia (especially without schizencephalic cleft) or gliotic scar was found. Six other similar nonfamilial cases have been seen. In one, the hippocampal commissure was also interrupted. In two there also was a small corresponding defect in the septum pellucidum. It should be noted that the clinical conditions of these patients may not be good; neurodevelopmental, visual, or endocrine disorders were observed in seven patients out of eleven (obviously there may be a referral bias). In view of the modern embryological data, this type of malformation is not in contradiction to developmental rules. The normal corpus callosum indeed develops from front to back; however, this occurs in two portions, first over the glial sling, then by fasciculation along both the pioneer frontal fibers and the hippocampal fibers. Obviously, the developmental defect prevented a normal fasciculation over the anterior hippocampal commissure, whereas the posterior splenium did develop more appropriately. According to this scheme, even an isolated splenium, without any corpus callosum, is conceivable. Conversely, in the typical partial posterior commissural agenesis, there is no posterior corpus callosum because there are no hippocampal fibers along which to fasciculate.

3.4.5 Other Commissural Defects Hypoplasia, Complete or Partial, Isolated or Associated with Agenesis

Fig. 3.13. Isolated partial agenesis of the corpus callosum with the splenium present. Midsagittal T1-weighted 3D RFT image. There is a defect of the posterior part of the callosal trunk (arrows). It is typically assumed to result from a destructive lesion in the hemispheres, but nothing abnormal was found in this group of patients, and a sibling of the patient illustrated here presented an identical defect. This is presumed to result from a defective development of the posterior portion of the callosal segment that uses the glial sling to cross, while the splenial fibers could fasciculate normally along the hippocampal commissure

This condition, in which the commissures are exceedingly small, is found in patients with an otherwise normal-looking brain, and is commonly observed in children investigated for neurodevelopmental delay. It may also be observed in association with partial commissural agenesis (Fig. 3.14). This entity is probably developmental, and should not be confused with atrophy of the commissures which is part of a global cerebral atrophy, or a result of a periventricular destructive process such as a periventricular leukomalacia. Hyperplasia

A hyperplastic, or rather an ill-shaped, thick corpus callosum lacking its normally well-delineated seg-

63

64

C. Raybaud and N. Girard

sive telencephalic disorder. The corpus callosum is absent in some subgroups of lissencephalies. It may be thick and ill-shaped, or on the contrary thin, especially in cases of polymicrogyria (therefore likely to be destructive) or of microlissencephaly, of massive heterotopia, etc.

3.4.6 Septo-Commissural Dysplasia

Fig. 3.14. Global commissural hypoplasia. Sagittal T1-weighted 3D RFT image. Global hypoplasia of the commissures is here associated with posterior agenesis, affecting the corpus callosum more than the hippocampal commissure. Note associated Dandy-Walker malformation with communicating cyst

ments may be observed in some cases of neurofibromatosis type 1 (see Chap. 16). Its clinical significance is unclear. Also, a “fat,” often short corpus callosum is observed in some instances of pachygyria. This may possibly be attributed to a failure of the normal pruning process [39, 47] of the commissural fibers during the early childhood development. Finally, an enlarged corpus callosum is found in patients with Cohen syndrome, an autosomal recessive disorder showing nonprogressive psychomotor retardation, motor clumsiness and microcephaly, typical facial features, childhood hypotonia and hyperextensibility of the joints, ophthalmologic findings of retinochoroidal dystrophy and myopia in patients over 5 years of age, and granulocytopenia [65]. Dysplasia

Various types of dysplastic commissures are observed that have not been classified, in a variety of clinical conditions [22]: hypoplastic septum pellucidum with the fornix attached to the corpus callosum, thick septum pellucidum, either bilateral or unilateral. In hemimegalencephalies, a septal and commissural dysplasia associated with septal displacement toward the affected hemisphere is relatively common. Defects Associated with Gross Dysplasia of the Brain

Abnormalities of the corpus callosum are observed in association with massive dysplasias of the cerebral cortex [22]. They are not specifically studied in this chapter, as they are only part of a more mas-

It may seem surprising to include septo-optic dysplasia (or De Morsier syndrome) in a text on commissural agenesis. More usually, septo-optic dysplasia is assumed to be somewhat related to the groups of the holoprosencephalies [66] (see Chap. 4). Yet, there are reasons for doing so. Modern embryology demonstrates that commissural fibers cross at the level of the septum pellucidum [43, 44]. In our material, we found some anatomical and clinical overlap between septo-optic dysplasia and commissural agenesis. In our series of 18 cases of septo-optic dysplasia, total or partial commissural agenesis was observed in six. Moreover, both disorders have some of their associated features in common: among 99 cases of commissural agenesis, visual abnormalities were found in 12% (21% in case of complete agenesis), and endocrine disorders in 9% (personal data). The typical septo-optic dysplasia was first described histologically in 1956 by De Morsier [67], with more recent histological and radiological descriptions [66, 68, 69]. Its features include dysplasia of the hypothalamus, defective or absent septum pellucidum, and abnormal, atrophic or malformative anterior optic pathways. The MR imaging diagnosis is relatively easy (see Chap. 4); the septum pellucidum is entirely absent, or only a rudimentary portion is present anteriorly. Because the septum is absent, the fornix, often bulky, is not tethered to the inferior surface of the corpus callosum, and its course, straight or even concave superiorly, is well below it over the roof of the third ventricle, which itself is lowered. This is clearly visible on sagittal as well as on coronal cuts throughout the brain. In about half the cases, optic atrophy can be diagnosed from the small section of the optic chiasm on the midline sagittal cut. However, it should be confirmed with ophthalmoscopy. In four out of 18 cases from our series, an associated partial agenesis of the commissures (posterior corpus callosum and hippocampal commissure) was found. The splenium was missing in three, and the genu in one. This association does not seem to bear a poorer prognosis. In two cases (plus another one

Malformations of the Telencephalic Commissures

diagnosed in utero) there was a complete septo-commissural agenesis, represented by complete absence of commissures, as well as complete absence of the septal laminae (Fig. 3.15). The inner wall of the lateral ventricles is made of a membrane (tela choroidea?), which on one side may expand between the hemispheres and typically produces a huge interhemispheric cyst in continuity with the ventricular lumen at the atrium. This complex malformation has been initially reported by Sener in 1993 [70].

Fig. 3.15. Complete septo-commissural agenesis. a Coronal T1weighted image. No commissure is seen, no septal leaves, no Probst’s bundles, no fornix or hippocampal commissure. The right lateral ventricle is closed by a (choroidal?) membrane that extends from the lower margin of the medial cortex of the hemisphere to the dorsal aspect of the thalamus. On the left, large cystic expansion of the left ventricular atrium

3.4.7 The Case of the Commissure in Lobar Holoprosencephaly Holoprosencephaly is classically defined as a failure of development of the telencephalon. In the complete form (alobar holoprosencephaly), the single anteriorsuperior telencephalic vesicle remains undifferentiated from the diencephalon. The thalami and basal ganglia form a single basal mass. There is also a single ventricular cavity (instead of two lateral and a third ventricles), and therefore no septum pellucidum. However, in the same way as there is a posterior medullary velum bordering the posterior-inferior aspect of the cerebellum between the cerebellar cortex and the fourth ventricular choroid plexus, there may be a posterior medullary velum interposed between the holoprosencephalic cortex and the holoprosencephalic tela choroidea.

In the most differentiated form of holoprosencephaly (lobar holoprosencephaly), the two hemispheres have developed nicely except for a cortical continuity across the midline in their median portion. As in the more severe form, there is only one ventricle with no septum pellucidum and no apparent fornix associated to it; yet, the hippocampi, albeit abnormal, are present, well-differentiated, and have a fimbria. A bulky posterior bundle of white matter crosses the interhemispheric fissure from one side to the other behind the trans-hemispheric cortex, looking like a true splenium [69]; quite often, a similar commissure develops between the frontal lobes, but without continuity with the splenium. It has been hypothesized [71] that a well-developed interhemispheric fissure is necessary, and perhaps sufficient, for the development of a corpus callosum. In lobar holoprosencephaly with a normal division of most of the frontal lobes and of the occipito-temporal lobes, the normal interhemispheric meninges would initiate and/or allow the crossing of commissural fibers. It is certain that the meninges play a role for the crossing of the fibers and as a support for the glial sling (see above for embryology). In the frontal lobes, the specific leading glial cells of the medial portion of the frontal germinal matrix might differentiate, migrate, and build up the sling. However, this specific germinal matrix is located on the medial aspect of the frontal horns, which are fused even in the mildest form of holoprosencephaly. Therefore, other sites of origin have to be postulated. The mechanism of formation of the posterior splenium is possibly straightforward and simpler. It should be considered according to the peculiar anatomy of the holoprosencephalies. As indicated before, a medullary velum is interposed between the margin of the cortex and the tela choroidea, in the continuity of the fimbria which carries the outgoing fibers of the hippocampus. In an holoprosencephalic brain, this medullary velum follows the limit of the holoprosencephalic cortex, crossing the interhemispheric fissure behind it. The fibers it contains cannot be considered analogous to the columns of the fornix, as they are not directed toward the septal area or the mammillary bodies. However, they may well be considered to be the fibers of the hippocampal commissure. As in normal development, they may be used by the posterior callosal fibers to fasciculate toward the other side. Assuming that this model is valid, the “splenium-like” bundle of crossing white matter observed in lobar holoprosencephalies can be considered to be the true splenium of the corpus callosum.

65

66

C. Raybaud and N. Girard

3.5 Conclusions Commissural (so-called “callosal”) agenesis can be re-evaluated thanks to the large clinical series and precise anatomic assessment allowed by modern imaging modalities. Several groups of malformations, likely to result from different mechanisms, can be identified. Morphologically, they can be represented by relatively simple schemes (Fig. 3.16). 1. The group of the common commissural agenesis, the most typical one (50%–60% of the cases), is characterized by a defect involving part, or the whole, of several commissures (callosal and hippocampal commissures always, anterior commissure less frequently), as well as major white matter bundles in the medial wall of the cerebral hemispheres. These abnormalities concern the brain parenchyma only, and spare the meninges; however, malformations of the skull may be associated. 2. In contrast, in a second group of about 15% of cases, multiple commissural agenesis is associated with major meningeal abnormalities in the interhemispheric fissure. In most instances, these abnormalities represent a multicystic dysplasia of the meninges. Other than the complete or

a

b

d

e

partial commissural defect, the brain malformation includes a dysplasia of the cortex bordering the meningeal cysts, gray matter heterotopia, and occasionally apparent parenchymal cavitations. It seems reasonable to assume that, in this form, the causal disorder can be extraparenchymal, within the interhemispheric meninges. Although different, the association of interhemispheric dysplastic meningeal lipomas with commissural defects may imply a similar mechanism. 3. A third group is composed of cases in which only one commissure is missing; this obviously points to an intrinsic defect of that commissure. The defect may concern either the anterior commissure, or the hippocampal commissure, or the corpus callosum, totally or in part. 4. A fourth group includes cases in which the commissures are present, but have an abnormal morphology, without any other abnormality of the brain. Most commonly, they are globally too small, but in other instances only the posterior or the anterior part fails to develop. This is, by definition, developmental, and should be differentiated from atrophic lesions. In neurofibromatosis type 1, on the contrary, the commissures may appear too thick. Although major abnormalities of cortical develop-

c

Fig. 3.16a–e. Schematic summary of commissural and septal abnormalities. a Complete commissural agenesis: no corpus callosum, no hippocampal commissure. The callosal fibers form a thickening in the upper portion of the septal laminae (the Probst’s bundles), and the longitudinal fornices form their lower margin. b Encysted cavum septi pellucidi/cavum Vergae. Cerebrospinal fluid accumulates in the normally absent space enclosed between the septal leaves and the commissures. c Isolated agenesis of the corpus callosum with the hippocampal commissure present. d Isolated agenesis of the hippocampal commissure with the corpus callosum present. e Isolated agenesis of the septal leaves with the corpus callosum and the hippocampal commissure present (typical septo-optic dysplasia)

Malformations of the Telencephalic Commissures

ment have been excluded from this study, it should be mentioned that such instances of dysmorphic commissures are common in agyria/pachygyria and in polymicrogyria. 5. Finally, the possible association of partial or complete commissural agenesis in cases of septo-optic dysplasias raises the problem of a possible continuity between the two groups of malformations, especially as visual and hypothalamo-pituitary disorders are quite frequent in both. In the same way, this could also apply to holoprosencephalies, as a fusion of the anterior basal ganglia may be observed in commissural agenesis. All three types of malformations are concerned with the genetic processes of basal induction and dorsal interhemispheric differentiation. This classification is purely descriptive, and although it points to a general “location” for the causal disorders (intracerebral versus meningeal, multi- or unicommissural defect, isolated or associated with other cerebro-meningeal disorders), it does not assign a single specific, causal mechanism for each group. For example, the commissuration of the corpus callosum implies, among others, the formation of the sling anteriorly, the fasciculation with the hippocampal fibers posteriorly, and an anatomical relationship with the septum pellucidum. Yet, absence of the septum does not prevent the formation of normal commissures in septo-optic dysplasia. The classical types of commissural agenesis, either total or partial, may logically be related to a guidance defect intrinsic to the medial wall of the hemisphere, given that other intrahemispheric fiber tracts are concerned as well. Reciprocally, the presence of huge dysplastic abnormalities of the interhemispheric meninges may be sufficient to prevent the crossing of the commissures. Hippocampal and callosal fibers may cross independently (isolated callosal agenesis, and conversely, isolated agenesis of the hippocampal commissure; septo-optic dysplasia), although they normally fasciculate together. With the multiplicity of factors involved in the commissuration processes, a multiplicity of single or associated defects may explain such diverse phenotypes. Encouragingly, the accumulation of embryologic data in the last decades has, on the whole, confirmed the primordial paper of Rakic and Yakovlev [17], but also brought much additional information. Moreover, the links between commissural agenesis, septo-optic dysplasia, and at least some types of holoprosencephaly may become further elucidated with the accumulation of knowledge on the role of the genes controlling the complex development of the derivatives of the anterior neural plate.

References 1. Raybaud C, Girard N. Etude anatomique par IRM des agénésies et dysplasies commissurales télencéphaliques. Corrélations cliniques et interprétation morphogénétique. Neurochirurgie 1998; 44 (Suppl 1):38–60. 2. Girard N, Raybaud C, Gambarelli D, Figarella-Branger D. Fetal brain MR imaging. Magn Reson Clin N Am 2001; 9:19–56. 3. Raybaud C, Levrier O, Brunel H, Girard N, Farnarier P. MR Imaging of the fetal brain malformations. Childs Nerv Syst, in press. 4. Harner RN. Agenesis of the corpus callosum and associated defects. In: Goldensohn ES, Appel SH (eds) Scientific Approaches to Clinical Neurology. Philadelphia: Lea & Febiger, 1977:616–627. 5. Brun A, Probst F. The influence of associated sesions on the morphology of the acallosal brain. A pathologic and encephalographic study. Neuroradiology 1973; 6:121–131. 6. Parish ML, Roessmann U, Levinsohn MW. Agenesis of the corpus callosum: a study of the frequency of associated malformations. Ann Neurol 1979; 6:349–354. 7. Bianchi GB. Storia del mostro di due corpi. Torino, 1749 (quoted by Bull, 1967). 8. Bull J. The corpus callosum. Clin Radiol 1967; 18:2–18. 9. Reil J. Mangel des mittleren und freyen Theils des Balken im Menschengehirn. Arch. Physiol (Halle) 1812; 11:341–344 (quoted by Probst, 1973). 10. Déjerine J. Anatomie des Centres Nerveux. Paris : Rueff et Cie, 1895. Reprinted by Masson, Paris: 1980. 11. Sachs H. Das Hemisphärenmark des Menschlichen Grosshirns. I. der Hinterhauptlappen. Leipzig 1892 (quoted by Déjerine, 1895). 12. Probst M. Ueber den Bau des vollständigen balkenlosen Grosshirns sowie über Mikrogyrie und Heterotopie der Grauen Substanz. Arch Psychiat Nerventkr 1901; 34:709– 786 (quoted by Probst, 1973). 13. Davidoff LM, Dyke CG. Agenesis of the corpus callosum. Its diagnosis by encephalography. Report of three cases. AJR Am J Roentgenol 1934; 32:1–10. 14. Feld M. Les dysgénésies des commissures interhémisphériques, dysraphies téléncéphaliques. In : Heuyer G, Feld M, Gruner J (eds) Malformations Congénitales du Cerveau. Paris: Masson et Cie, 1959:303–327. 15. Gross H, Hoff H. Sur les malformations ventriculaires dépendantes des dysgénésies commissurales. In : Heuyer G, Feld M, Gruner J. Malformations Congénitales du Cerveau. Paris : Masson et Cie, 1959:329–351. 16. Loeser JD, Alvord EC. Agenesis of the corpus callosum. Brain 1968; 91:553–570. 17. Rakic P, Yakovlev PI. Development of the corpus callosum and cavum septi in Man. J Comp Neurol 1968; 132:45–72. 18. Probst FP. Congenital defects of the corpus callosum. Acta Radiol Suppl 1973; 331:1–152. 19. Kendall BE. Dysgenesis of the corpus callosum. Neuroradiology 1983; 25:239–256. 20. Jinkins JR, Whittemore AR, Bradley WG. MR imaging of callosal and corticocallosal dysgenesis. AJNR Am J Neuroradiol 1989; 10:339–344. 21. Rubinstein D, Youngman V, Hise JH, Damiano TR. Partial development of the corpus callosum. AJNR Am J Neuroradiol 1994; 15:869–875. 22. Utsunomiya H, Ogasawara T, Hayashi T, Hashimoto T, Okazaki M. Dysgenesis of the corpus callosum and associ-

67

68

C. Raybaud and N. Girard ated telencephalic anomalies: MRI. Neuroradiology 1997; 39:302–310. 23. Raybaud C, Girard N. The developmental disorders of the commissural plate of the telencephalon. MR imaging study and morphologic classification. Nervous System in Children (Japan) 1999; 24:348–357. 24. Yakovlev PI. Telencephalon “impar,”“semipar” and “totopar” (morphogenetic, tectogenetic and architectonic definitions). Int J Neurol 1968; 6:245–265. 25. Guénot M. Transfert interhémisphérique et agénésie du corps calleux. Capacités et limites de la commissure blanche antérieure. Neurochirurgie 1998; 44 (Suppl 1):113–115. 26. Barkovich AJ, Kjos BO. Normal postnatal development of the corpus callosum as demonstrated by MR imaging. AJNR Am J Neuroradiol 1988; 9:487–491. 27. De Lacoste MC. Topography of the human corpus callosum. J Neuropathol Exp Neurol 1985; 44:578–591. 28. Velut S, Destrieux C, Kakou M. Anatomie morphologique du corps calleux. Neurochirurgie 1998; 44 (Suppl 1):17–30. 29. Kier EL, Truwit CL. The normal and abnormal genu of the corpus callosum: an evolutionary, embryologic, anatomic, and MR analysis. AJNR Am J Neuroradiol 1996; 17:1631– 1641. 30. Kier EL, Truwit CL. The lamina rostralis: modification of concepts concerning the anatomy, embryology, and MR appearance of the rostrum of the corpus callosum. AJNR Am J Neuroradiol 1997; 18:715–722. 31. Knowles WD. Normal anatomy and neurophysiology of the hippocampal formation. J. Clin Neurophysiol 1992; 9:252– 263. 32. Yakovlev PI, Locke S. Limbic nuclei of thalamus and connections of limbic cortex. III. Cortico-cortical connections of the anterior cingulate gyrus, the cingulum, and the subcallosal bundle in the monkey. Arch Neurol 1961; 5:364–400. 33. Bruyn GW. Agenesis septi pellucidi, cavum septi pellucidi, cavum Vergae, and cavum veli interpositi. In: Vinken PJ, Bruyn GW (eds) Handbook of Clinical Neurology, vol. 30: Congenital Malformations of the Brain and Skull. Part 1. Amsterdam: North-Holland, 1977:299–336. 34. Donovan WD, Zimmerman RD, Deck MDF. MRI of cavum Vergae without cavum septi pellucidi. J Comp Assist Tomogr 1995; 19:1010–1011. 35. Ariëns Kappers CU, Huber GC, Crosby EC. The Comparative Anatomy of the Nervous System of Vertebrates including Man, vol. III. New York: Hafner, 1967. 36. Sarnat HB, Netsky MG. Evolution of the Nervous System. New York: Oxford University Press, 1974. 37. Romer AS, Parsons TS. The Vertebrate Body. Philadelphia: Saunders College, 1977. 38. Silver J, Lorenz SE, Wahlsten D, Coughlin J. Axonal guidance during development of the great cerebral commissures: descriptive and experimental studies, in vivo, on the role of preformed glial pathways. J Comp Neurol 1982; 210:10–29. 39. Katz MJ, Lasek RJ, Silver J. Ontophyletics of the nervous system: development of the corpus callosum and evolution of axons tracts. Proc Natl Acad Sci USA 1983; 80:5936– 5940. 40. Koester SE, O’Leary DDM. Axons of early generated neurons in cingulate cortex pioneer the corpus callosum. J Neurosci 1994; 14:6608–6620. 41. Wahlsten D. Defects of the fetal forebrain in mice with hereditary agenesis of the corpus callosum. J Comp Neurol 1987; 262:227–241. 42. Livy DG, Wahlsten D. Retarded formation of the hippocam-

pal commissure in embryos from mouse strains lacking a corpus callosum. Hippocampus 1997; 7:2–14. 43. Shu T, Richards LJ. Cortical axon guidance by the glial wedge during the development of the corpus callosum. J Neurosci 2001; 21:2749–2758. 44. Shu T, Butz KG, Plachez C, Gronostzjski RM, Richards LJ. Abnormal development of forebrain midline glia and commissural projections in Nfia knock-out mice. J Neurosci 2003; 23:203–212. 45. Kostovic I, Rakic P. Developmental history of the transient subplate zone in the visual and the somatosensory cortex of the macaque monkey and human brain. J Comp Neurol 1990; 297:411–470. 46. O’Leary DDM, Stanfield BB, Cowan WM. Evidence that the early postnatal restriction of the cells of origin of the callosal projection is due to the elimination of axonal collaterals rather than to the death of neurons. Develop Brain Res 1981; 1:607–617. 47. Kadhim HJ, Bhide PG, Frost DO. Transient axonal branching in the developing corpus callosum. Cerebral Cortex 1993; 3:551–566. 48. Serafini T, Colamarino SA, Leonardo ED, Wang H, Beddington R, Skarnes WC, Tessier-Lavigne M. Netrin-1 is required for commissural axon guidance in the developing vertebrate nervous system. Cell 1996; 87:1001–1014. 49. Qiu M, Anderson S, Chen S, Meneses JJ, Hevner R, Kuwana E, Pedersen RA, Rubenstein LR. Mutation of the Emx-1 Homeobox Gene Disrupts the Corpus Callosum. Dev Biol 1996; 178:174–178. 50. Flanagan JG, Van Vactor D. Through the looking glass: axon guidance at the midline choice point. Cell 1998; 92:429–432. 51. Lanier L, Gates MA, Witke W, Menzies S, Wehman AM, Macklis JD, Kwiatkowski D, Soriano P, Gertler FB. Mena is required for neurulation and commissure formation. Neuron 1999; 22:313–325. 52. Norman MG, McGillivray BC, Kalousek DK, Hill A, Poskitt KJ. Congenital Malformations of the Brain. New York: Oxford University Press, 1995:309–331. 53. Quigley M, Cordes D, Turski P, Moritz C, Haughton V, Seth R, Meyerand ME. Role of the corpus callosum in functional connectivity. AJNR Am J Neuroradiol 2003; 24:208–212. 54. Couly GF, Le Douarin NM. Mapping of the early neural primordium in quail-chick chimeras. II. The prosencephalic neural plate and neural folds: applications for the genesis of cephalic human congenital anomalies. Dev Biol 1987; 120:198–214. 55. Osaka K, Handa H, Matsumoto S, Yasuda M. Development of the cerebrospinal fluid pathways in the normal and abnormal embryos. Childs Brain 1980; 6:26–38. 56. Britt PM, Bindal AK, Balko MG, Yeh HS. Lipoma of the cerebral cortex: case report. Acta Neurochir (Wien) 1993; 121:88–92. 57. Zingesser L, Schechter M, Gonatas N, Levy A, Wisoff H. Agenesis of the corpus callosum associated with an interhemispheric cyst. Brit J Radiol 1964; 37:905–909. 58. Barkovich AJ, Simon EM, Walsh CA. Callosal agenesis with cysts. A better understanding and new classification. Neurology 2001; 56:220–227. 59. Aicardi J. Aicardi syndrome. In: Guerrini R, Andermann F, Canapicchi R, Roger J, Zifkin BG, Pfanner P (eds) Dysplasias of Cerebral Cortex and Epilepsy. Philadelphia: Lippincott Raven, 1996:211–216. 60. Verga P. Lipoma ed osteolipomi della piamadre. Tumori 1929; 15:321–357 (quoted by Truwit and Barkovich, 1990).

Malformations of the Telencephalic Commissures 61. Truwit CL, Barkovich AJ. Pathogenesis of intracranial lipomas: an MR study in 42 patients. AJNR Am J Neuroradiol 1990; 11:665–674. 62. Tart RP, Quisling RG. Curvilinear and tubulonodular varieties of lipoma of the corpus callosum: an MR and CT study. J Comput Assist Tomogr 1991; 15:805–810. 63. Bodensteiner JB, Schaefer GB. Wide cavum septum pellucidum: a marker of disturbed brain development. Pediatr Neurol 1990; 6:391–394. 64. Barkovich AJ. Apparent atypical callosal dysgenesis: analysis of MR findings in six cases and their relation to holoprosencephaly. AJNR Am J Neuroradiol 1989; 11:333–339. 65. Kivitie-Kallio S, Autti T, Salonen O, Norio R. MRI of the brain in the Cohen syndrome: a relatively large corpus callosum in patients with mental retardation and microcephaly. Neuropediatrics 1998; 29:298–301. 66. Fitz CR. Holoprosencephaly and septo-optic dysplasia. Neuroimag Clin N Am 1994; 4:263–281.

67. De Morsier G. Agénésie du septum lucidum avec malformation du tractus optique. La dysplasie septo-optique. Schweiz Arch Neurol Psychiat 1956; 77:267–292. 68. Roessmann U, Velasco ME, Small EJ, Hori A. Neuropathology of “septo-optic dysplasia” (de Morsier syndrome) with immunohistochemical studies of the hypothalamus and pituitary gland. J Neuropathol Exp Neurol 1987; 46:597– 608. 69. Barkovich AJ, Fram EK, Norman D. Septo-optic dysplasia: MR imaging. Radiology 1989; 171:189–192. 70. Sener RN. Septo-optic dysplasia associated with total absence of the corpus callosum: MRI and CT features. Eur Radiol 1993; 3:551–553. 71. Oba H, Barkovich AJ. Holoprosencephaly: an analysis of callosal formation and its relation to development of the interhemispheric fissure. AJNR Am J Neuroradiol 1995; 16:453–460.

69

Brain Malformations

4

Brain Malformations Paolo Tortori-Donati, Andrea Rossi, and Roberta Biancheri

CONTENTS Introduction

4.3.5

72

4.1

Cephaloceles 72

4.1.1 4.1.2 4.1.3 4.1.4 4.1.5

Background 72 Embryology 74 Clinical Features 74 Classification and Neuroradiological Features 75 Syndromes Associated with Cephaloceles 82

4.2

Defects of the Mediobasal Prosencephalon

Holoprosencephalies and Related Entities 4.2.1 4.2.1.1 4.2.1.2 4.2.1.3 4.2.1.4 4.2.1.5 4.2.2

86

4.2.2.1 4.2.2.2 4.2.2.3 4.2.3 4.2.3.1 4.2.3.2 4.2.3.3 4.2.3.4 4.2.4 4.2.4.1 4.2.4.2

Holoprosencephaly 86 Background 86 Embryogenesis and Pathogenetic Theories 87 Clinical Features 88 Associated Conditions 88 Imaging Studies 89 Syntelencephaly (Middle Interhemispheric Holoprosencephaly) 92 Background 92 Pathogenesis and Relationship with HPE 93 Imaging Findings 94 Septo-Optic Dysplasia 95 Background 95 Pathogenesis 95 Clinical Findings 96 Imaging Findings 96 Kallmann Syndrome 97 Background 97 Pathogenesis 99

4.3

Malformations of Cortical Development

4.3.1 4.3.2 4.3.3 4.3.4 4.3.4.1 4.3.4.2 4.3.4.3 4.3.4.4 4.3.4.5 4.3.4.6 4.3.4.7

100

Background 100 Normal Cortical Development 100 Classification of MCD 103 Malformations Due to Abnormal Neuronal/Glial Proliferation or Apoptosis 103 Microcephaly 103 Microlissencephaly 105 Megalencephaly 106 Tuberous Sclerosis Complex 108 Focal Transmantle Dysplasia (Taylor’s Focal Cortical Dysplasia) 108 Hemimegalencephaly 112 Neoplasms 116

4.3.7.1 4.3.7.2 4.3.7.3

Malformations Due to Abnormal Neuronal Migration 116 Lissencephalies 116 The Cobblestone Complex 121 Heterotopia 126 Malformations Due to Abnormal Cortical Organization 130 Polymicrogyria 130 Schizencephaly 134 Non-Taylor’s Focal Cortical Dysplasia (Architectural and Cytoarchitectural) 136 Malformations of Cortical Development, not Otherwise Classified 138 Zellweger Syndrome 138 Mitochondrial Disorders 138 Sublobar Dysplasia 138

4.4

Malformations of the Posterior Cranial Fossa

4.4.1 4.4.2 4.4.3 4.4.3.1 4.4.3.2 4.4.3.3 4.4.3.4 4.4.3.5 4.4.4 4.4.4.1 4.4.4.2 4.4.4.3

Embryology 138 Classification of Malformations 140 Cystic Malformations 142 Dandy-Walker Malformation 142 Dandy-Walker Variant 146 Persistent Blake’s Pouch 147 Mega Cisterna Magna 150 Arachnoid Cysts 152 Noncystic Malformations 155 Paleocerebellar Hypoplasia 156 Neocerebellar Aplasia and Hypoplasia 161 Isolated Brainstem Hypoplasia/Dysplasia 170

4.5

Chiari Malformations

172

4.5.1 4.5.2 4.5.3 4.5.4

Chiari I Malformation Chiari II Malformation Chiari III Malformation Chiari IV Malformation

172 178 184 186

4.3.5.1 4.3.5.2 4.3.5.3 4.3.6 4.3.6.1 4.3.6.2 4.3.6.3 4.3.7

References

186

138

71

72

P. Tortori-Donati, A. Rossi, and R. Biancheri

Introduction Brain malformations are extremely polymorphous, and individual cases very often escape rigid categorization. Moreover, several malformations are frequently associated with one another in individual patients, and many are comprised within complex multiorgan syndromes. Although some believe that a purely descriptive diagnostic approach is advantageous [1], classifications are useful to practicing clinicians to make sense of what they see and to approach their patients on the basis of a logical framework. Neuroradiologists have a wellknown propensity for classifications, basically because MRI has proved to be a very powerful and effective tool to attempt radiologic-pathologic correlations. However, classification schemes are continuously challenged by new advances in the understanding of the pathologies they attempt to describe. This has been the case with many disease processes involving the central nervous system (CNS), and especially with brain malformations. Knowledge of the basic molecular and genetic processes that direct normal brain development and, when deranged, result in congenital abnormalities has literally boomed in the past decade. For this reason, we are now witnessing a delicate transition from a purely morphological [2] to a molecular genetic (Table 4.1) [3, 4] approach to brain malformations. At present, both perspectives are probably equally unsatisfactory, basically because the molecular and genetic background is well known for a few entities, but is still unavailable for many others. In fact, causes of malformations can be divided into four groups, i.e., chromosomal abnormalities, single gene mutations, environmental agents, and unknown; unfortunately, the last category is the largest [1]. Therefore, neuroradiological classifications are still mostly based on morphology [5, 6] or on a combination of morphologic and biochemical data [7]. MRI is the single best modality for imaging brain malformations. There is little point in describing here the several advantages of MRI over other imaging modalities, as they are well known and have been widely described elsewhere. However, CT is still widely used in centers where MRI is not readily available. Advanced MRI techniques, such as diffusion-weighted imaging, perfusion-weighted imaging, and MR spectroscopy are not particularly useful. A notable exception is the role of diffusion-weighted imaging in the differentiation of arachnoid cysts from dysontogenetic masses [8]. Ultrasounds play a crucial role as a screening modality in neonates, and are credited for revealing many abnormalities that are subsequently investigated in detail with MRI (see Chap. 25). Before proceeding with the description of the various entities, a short premise on semantics is required [9].

Agenesis (synonym: aplasia) indicates failure of development of a whole organ or of a part of it, whereas hypoplasia (synonym: hypogenesis) involves incomplete formation. On the contrary, atrophy implies degeneration occurring after achievement of complete development; on this basis, atrophy can secondarily involve hypoplastic structures. Dysplasia indicates abnormal histogenesis, basically represented, in the CNS, by migrational anomalies (i.e., cortical dysplasia). The term dysgenesis is a generic synonym of malformation (i.e., corpus callosum dysgenesis). Use of semantically correct terms is important to develop a common language and to favor interaction between clinicians. In the following discussion, brain malformations have been somewhat arbitrarily categorized into five broad groups, i.e., cephaloceles, defects of the mediobasal prosencephalon, malformations of cortical development, malformations of the posterior cranial fossa, and Chiari malformations. Abnormalities of the telencephalic commissures, including corpus callosum abnormalities, have been discussed elsewhere in this book (see Chap. 3).

4.1 Cephaloceles 4.1.1 Background Cephaloceles are characterized by evagination of brain and/or leptomeninges through a congenital defect of the skull and dura, usually located on the midline. Cephaloceles affect 0.8–3:10,000 viable newborns [10, 11]. There are significant geographic variations in their location, incidence, and gender predilection. In Western populations, the majority of these lesions involve the occipital bone and occur in females, whereas anterior cephaloceles involving the frontal, nasal, and orbital bones are more frequent in male patients of Asian descent. These variations suggest that different genetic assets play a causal role. In cephaloceles, the meninges form the wall of a sac that is filled with CSF and, in some cases, nervous tissue. Depending on the contents of the herniation [10], cephaloceles are categorized as follows (Fig. 4.1): ● Meningocele: cerebrospinal fluid (CSF) collection lined by partially epithelized meningeal wall; ● Meningoencephalocele, or encephalocele: meningeal, variably epithelized sac containing brain; ● Meningoencephalocystocele: an encephalocele containing a portion of the ventricles;

Brain Malformations Table 4.1. Proposed new etiological classification of human nervous system malformations based upon patterns of genetic expression I. Genetic mutations expressed in the primitive streak or node A. Upregulation of organizer genes 1. Duplication of neural tube B. Downregulation of organizer genes 1. Agenesis of neural tube II. Disorders of ventralizing gradient in the neural tube A. Overexpression of ventralizing genes 1. Duplication of spinal central canal 2. Duplication of ventral horns of spinal cord 3. Diplomyelia (and diastematomyelia?) 4. Duplication of neural tube 5. Ventralizing induction of somite a. Segmental amyoplasia B. Underexpression of ventralizing genes 1. Fusion of ventral horns of spinal cord 2. Sacral (thoraco-lumbo-sacral) agenesis 3. Arrhinencephaly 4. Holoprosencephaly III. Disorders of dorsalizing gradient of the neural tube A. Overexpression of dorsalizing genes 1. Duplication of dorsal horns of spinal cord 2. Duplication of dorsal brainstem structures B. Underexpression of dorsalizing genes 1. Fusion of dorsal horns of spinal cord 2. Septo-optic dysplasia (?) IV. Disorders of the rostrocaudal gradient and/or segmentation A. Increased homeobox domains and/or ectopic expression 1. Chiari II malformation B. Decreased homeobox domains and/or neuromere deletion 1. Agenesis of mesencephalon and metencephalon 2. Global cerebellar aplasia or hypoplasia 3. Agenesis of basal telencephalic nuclei 4. Agenesis of the corpus callosum (some forms: Emx 1)

B. Neoplastic 1. Medullomyoblastoma 2. Dysembryoplastic neuroepithelial tumors 3. Gangliogliomas and other mixed neural tumors VI. Disorders of secretory molecules and genes that mediate migrations A. Neuroblast migration 1. Initial course of neuroblast migration a. Filamin-1 (X-linked dominant periventricular nodular heterotopia) 2. Middle course of neuroblast migration a. Doublecortin (DCX; X-linked dominant subcortical laminar heterotopia or band heterotopia) b. LIS1 (type I lissencephaly or Miller-Dieker syndrome) c. Fukutin (type II lissencephaly; Fukuyama muscular dystrophy) d. Empty spiracles (EMX2; schizencephaly) e. Astrotactin 3. Late course of neuroblast migration; architecture of cortical plate a. Reelin (pachygyria of late neuroblast migration and cerebellar hypoplasia) b. Disabled-1 (DAB1; also VLDL / Apoe2R?App recep tor defect; downstream of reelin, EMX2 and DCX) c. L1-NCAM (X-linked hydrocephalus and pachygyria with aqueductal stenosis) B. Glioblast migration C. Focal migratory disturbances due to acquired lesions of the fetal brain VII. Disorders of secretory molecules and genes that attract or repel axonal growth cones A. Netrin downregulation B. Keratan sulfate and other glycosaminoglycan downregulations C. S-100? protein downregulation or upregulation (?)

V. Aberrations in cell lineages by genetic mutation A. Non-neoplastic 1. Striated muscle in the central nervous system 2. Dysplastic gangliocytoma of the cerebellum (Lhermitte-Duclos) 3. Tuberous sclerosis 4. Agenesis of the corpus callosum (some forms: Emx 1) 5. Hemimegalencephaly (also VIII. Disorders of symmetry)

VIII. Disorders of symmetry A. Hemimegalencephaly (also see V. Aberrations of cellular lineages) 1. Isolated hemimegalencephaly 2. Syndromic hemimegalencephaly a. Epidermal nevus b. Proteus c. Klippel-Trénaunay-Weber d. Hypomelanosis of Ito B. Hemihyperplasia of the cerebellum

An additional variant is the atretic cephalocele, a rudimentary form composed by a small fibrolipomatous nodule. The degree of epithelization of the herniated sac is variable in individual cases. When the arachnoidal lining becomes exposed to the environment, ulceration and infection may ensue. As with other neural tube defects, the skin overlying the abnormality may exhibit birthmarks, such as hairy tufts, hemangiomas, dyschromic patches, and dimples. The neural tissue contained within the sac usually

is disorganized, and includes zones of necrosis, calcifications, and vascular proliferation [9]. The size of the cephalocele may vary from a small nodule to an enormous sac that may be larger than the head itself. However, size does not predict the severity of the dysgenesis of the portion of the brain remaining within the cranium [9]. Instead, there is a direct relationship between the amount of herniated neural tissue and the degree of microcephaly. This correlates with the eventual degree of mental retardation [12].

73

74

P. Tortori-Donati, A. Rossi, and R. Biancheri

a

b

c

Fig. 4.1a–c. Schematic drawing of the different types of cephaloceles. a Meningocele: herniation of the meninges only. b Meningoencephalocele (or encephalocele): herniation of the meninges and nervous tissue through a skull defect. c Meningoencephalocystocele: a portion of the ventricular system herniates into the sac

4.1.2 Embryology The etiology of cephaloceles remains incompletely determined. Multiple defects of brain maturation, such as neuronal migration, axonal projection, and synaptogenesis, may be found, but none appears to play a key pathogenetic role. The relationship between cephaloceles and neural tube closure defects is the key point in the etiopathogenetic debate. Based on the traditional bidirectional, “zipper-like” neural tube closure model, cephaloceles have been related to defects of closure of the anterior neuropore [13]. However, intermediate defects, such as occipital or frontal cephaloceles, remained difficult to explain. Recently, a multisite neural tube closure model was proposed [14]. The multisite model suggests that cephaloceles and other neural tube closure defects, such as myelomeningoceles, anencephaly, and craniorachischisis, can be explained by failure of fusion of one of the closures or their contiguous neuropore [14]. A striking correlation between the neural tube defects observed in clinical cases and the location of the closure sites and their neuropores is believed to exist by these authors [14]. However, other authors [15] believe that cephaloceles should be considered postneurulation defects. The primary event should be traced to failure in the bony covering of the neural tube, and cephaloceles would be caused by events not associated with neurulation, but with distortion and crowding of a rapidly growing brain through the defect. It is still controversial whether cephaloceles result from primary neural tube nonclosure or from secondary reopening. The absence of a membrane coverage to the malformation is probably the single most significant indicator of primary failure of

neural tube closure. Conversely, it is difficult to determine whether skin-covered cephaloceles result from incomplete neural tube closure or secondary reopening. Lateral (i.e., off-midline) cephaloceles most likely result from secondary neural tube reopening, and are usually seen in association with disruptive disorders [14].

4.1.3 Clinical Features The degree of neurological and developmental damage associated with cephaloceles is variable, and depends on the amount of herniated neural tissue, the presence of hydrocephalus, and the association with hindbrain lesions or cerebral hemisphere dysplasias which result from the associated disorder of cellular migration and organization [13]. Some patients are only mildly impaired. Mental retardation, spastic diplegia, and cognitive disorders may not become evident until childhood. Patients with meningoceles or atretic cephaloceles have an essentially normal neurological picture, and their prognosis is favorable provided there is neither hydrocephalus nor associated congenital abnormalities. Conversely, children with encephaloceles show a complex clinical picture that depends on variably combined factors such as the size, contents, and location of the herniation, as well as the presence of hydrocephalus, perinatal injury, apneic crises, seizures with consequent hypoventilation, hypoxia, and acidosis [16]. In general, occipital cephaloceles are associated with a unfavorable prognosis compared to that of skull base cephaloceles. Severe occipital cephaloceles are often associated with distortion of the brainstem

Brain Malformations

and cranial nerves leading to visual disturbances, neurovegetative disorders (apnea, bradycardia, sialorrhea), and cranial nerve deficit (strabismus, nystagmus, dysphagia). Hydrocephalus is often progressive and manifests with increased intracranial pressure, opisthotonus, and hypertonic limbs. Ulceration of the sac may be complicated by local infection or meningoencephalitis. Associated anomalies are especially important, as they often influence the clinical picture more severely than the cephalocele itself. Agenesis of the corpus callosum is found in 80% of cases. Occipital cephaloceles are also often part of complex abnormalities such as the Chiari III malformation and tectocerebellar dysraphia, and are found in a host of inherited syndromes (see below). Associated abnormalities account for a variable degree of developmental and psychomotor delay, blindness, superior functions compromise (symbolic functions, language, learning, memory), motor incoordination, and seizures [16]. Skull base cephaloceles usually are characterized by severe cranio-facial dysmorphism with cleft lip and palate, coloboma, and microphthalmos. Sphenoidal cephaloceles usually become manifest either with signs of upper airway obstruction or with endocrine and visual dysfunction, due to prolapse of the pituitary gland and optic chiasm into the herniated sac. Visual disturbances may also be due to associated abnormalities, such as holoprosencephaly and septo-optic dysplasia. Anosmia may result from compromise of the olfactory nerves and bulbs in patients with sincipital cephaloceles. Ascending infections of the CNS with recurrent meningitides may be caused by occult connections between unsuspected skull base cephaloceles and the nasal and paranasal cavities, rhinopharynx, and inner ear [16].

Cephaloceles of the Convexity

Occipital Cephaloceles

Depending on the relationship of the calvarial defect with anatomic landmarks, such as the external occipital protuberance and the posterior lip of the foramen magnum, occipital cephaloceles are categorized into three subsets: (i) occipito-cervical: the defect involves the occipital squama, foramen magnum, posterior arch of the atlas, and often the neural arches of the upper cervical vertebrae; (ii) inferior occipital: the defect lies at or below the external occipital protuberance and above the posterior lip of the foramen magnum (Figs. 4.2–4); and (iii) superior occipital: the defect lies above the external occipital protuberance (Fig. 4.5). Occipital cephaloceles are obvious at birth and are usually diagnosed antenatally. In newborns, the differential diagnosis includes cephalohematomas, caput succedaneum, and cystic lymphangiomas. The majority (80%) are meningoencephaloceles that contain either cerebral tissue, cerebellar tissue, or both [12, 19]. However, the content of the sac is not predictable based on the site of the cranial defect [10]. The occipital lobes are the single most common portion of brain that is included within the herniation. Dural sinuses and other venous structures may also lie close to or within the sac, generating a risk for hemorrhage and subsequent brain damage during surgery. The overall size of the cephalocele is usually large, exceeding 5 mm in diameter in 72% of cases [10].

4.1.4 Classification and Neuroradiological Features Cephaloceles are usually categorized according to their location into convexity, skull base, and internal cephaloceles [10, 17]. Their neuroradiological evaluation aims to assess: (i) their exact location; (ii) the morphology, size, and content of the sac; (iii) possible associated abnormalities, such as callosal dysgenesis [18] and the Chiari II malformation, that portend a poorer prognosis; and (iv) the anatomic relationship with vascular structures, especially the dural sinuses, whose inadvertent damage during surgery may cause profuse bleeding. MR angiography is valuable to identify the location and course of the major venous structures.

Fig. 4.2. Inferior occipital meningocele. Sagittal T1-weighted image. Small occipital meningocele covered by dystrophic skin (arrow)

75

76

P. Tortori-Donati, A. Rossi, and R. Biancheri

a

b Fig. 4.3a,b. Occipital meningocele. a. Sagittal T1-weighted image; b. Three-dimensional CT scan. A complex malformation is associated with an occipital meningocele (asterisk, a). CT scan clearly shows the skull defect (arrows, b)

a Fig. 4.4a,b. Occipital meningoencephalocystocele. a Sagittal T1weighted image; b Axial T1-weighted image. Large occipital cephalocele, characterized by herniation of meninges (arrow, a), nervous tissue (arrowheads, a, b), and a portion of the ventricular system (asterisk, a, b) that communicates with the left atrium. The herniated ventricle is well separated from the adjacent herniated subarachnoid space (M, b). (Case courtesy Dr. Z. Patay, Riyadh, Saudi Arabia)

b

Brain Malformations

Simple meningoceles are usually isolated (Fig. 4.2). Conversely, meningoencephaloceles can be associated with other malformations (Figs. 4.3, 4.4), such as holoprosencephaly, callosal dysgenesis, myelomeningocele, Klippel-Feil syndrome, and hydrocephalus. Some of these associations are considered as autonomous entities. The presence of signs of Chiari II malformation (small posterior fossa, tectal beaking, enlarged interthalamic mass) in a child with an occipital cephalocele is designated “Chiari III malformation.” The cephalocele is occipito-cervical or, more rarely, inferior cervical [20]. The coexistence of an occipital cephalocele, agenesis of the vermis, and marked deformation of the tectum is called “tectocerebellar dysraphia.” These entities are described in detail elsewhere in this chapter. a

Sagittal (Interparietal) Cephaloceles

These rare cephaloceles are located between the parietal bones in the midline (Fig. 4.6), and may be associated with callosal dysgenesis and arachnoid cysts. Lateral Cephaloceles

Lateral (i.e., off-midline) cephaloceles are very rare (Fig. 4.7). They are found along the coronal and lambdoid sutures, and at the pterion or asterion. They are believed to result from secondary reopening of the neural tube. Therefore, they are considered secondary, disruptive disorders [14]. Bregmatic Cephaloceles

Cephaloceles located at the anterior fontanel must be differentiated from dermoids, which are much more common. Computerized tomography (CT) may be useful to demonstrate the calvarial defect that is associated with cephaloceles.

b

Interfrontal Cephaloceles

They are found between the glabella and the bregma, and the defect lies along the metopic suture. Portions of the frontal lobes may herniate into the sac. Skull Base Cephaloceles

c Fig. 4.5a–c. Superior occipital meningocele. a Sagittal T1weighted image; b Gd-enhanced sagittal T1-weighted image; c 2D TOF MRA. Huge occipital meningocele (a, b). The venous sinuses are not herniated into the sac (c). A falcine sinus is associated (arrow, a, b)

Anterior cephaloceles (Fig. 4.8) usually are found in children with little or no neurological impairment. However, they may be associated with callosal agenesis, interhemispheric cysts, and interhemispheric lipomas; the latter usually belong to the tubulonodular variety and can be partially calcified. Cortical dysplasia and holoprosencephaly may also be associated, whereas hydrocephalus, albeit possible, is rare. Cephaloceles of the skull base may be accompanied by midline facial defects, such as hypertelorism, cleft

77

78

P. Tortori-Donati, A. Rossi, and R. Biancheri

b a Fig. 4.6a,b. Interparietal cephalocele. a Sagittal T2-weighted image; b 3D TOF MR angiography. Small posterior interparietal meningocele (arrow, a). The venous sinuses are not herniated into the sac (b). A large falcine sinus is recognizable (arrowheads, a, b)

a

b Fig. 4.7a,b. Parietal meningocele. a Coronal T2-weighted image; b Axial T2-weighted image. Large meningocele containing septa in the right parietal region. Widespread cortical malformations, represented by subependymal heterotopia (arrowhead, a) and multiple bilateral polymicrogyric infoldings (arrows, a, b) are recognizable. (Case courtesy Dr. Z. Patay, Riyadh, Saudi Arabia)

lip, and cleft palate [21]. Contrary to occipital cephaloceles, they may be clinically occult [22] and discovered later in life. Sincipital Cephaloceles

Sincipital or fronto-ethmoidal cephaloceles are categorized according to their relationship with the frontal bone, nasal bones, and ethmoid.

Nasofrontal or glabellar cephaloceles: The defect lies between the frontal and nasal bones at level of the “fonticulus nasofrontalis,” a small transient fontanel corresponding externally to the glabella [23, 24]. These lesions may be in the midline or lie slightly off the midline, and can be associated with other abnormalities such as callosal dysgenesis, lipomas, holoprosencephaly, and cortical malformations (Fig. 4.9).

Brain Malformations

a

b

d

e

c

f

Fig. 4.8a–f. Schematic representation of sincipital (a–c) and naso-pharyngeal (d–f) cephaloceles. Asterisks or arrows indicate the bony schisis. E, ethmoid bone; F, frontal bone; N, nasal bones; S, sphenoid bone; V, ventricles. a Naso-frontal cephalocele: The skull defect involves the glabella, being located between the frontal bone superiorly and the ethmoid and nasal bones inferiorly. b Nasoethmoidal cephalocele: The skull defect is located between the frontal and nasal bones superiorly and the ethmoid inferiorly. The cephalocele creeps into the nose between the nasal bones and cartilage. c Naso-orbital cephalocele: The skull defect involves the ethmoidal lamina papyracea, lacrimal bone, and frontal process of the maxillary bone. The facial extremity of the defect is localized at level of the medial orbital wall, and the ostium involves the internal canthus. d Trans-ethmoidal cephalocele: Skull defect involves the ethmoid, with resulting intranasal mass. e Spheno-ethmoidal cephalocele: Skull defect is between the ethmoid and sphenoid bones, with resulting mass in the posterior nasal cavity. f Spheno-nasopharyngeal cephalocele: Sphenoidal bone defect with nasopharyngeal mass

Nasoethmoidal cephaloceles: The congenital defect is located between the nasal bones and the cartilage. The cephalocele can develop either on the surface of the nasal pyramid or in the inner canthus of the eye. Nasoorbital cephaloceles: The defect lies at the junction between the frontal, ethmoidal, lacrimal, and maxillary bones. The cephalocele develops either externally on the face or internally in the orbit. These cephaloceles are believed to result from incomplete regression of a dural outpouching that normally crosses the fonticulus nasofrontalis during fetal life, and is subsequently obliterated [23, 24], resulting in the foramen caecum. The enlargement of this foramen is a hallmark of these lesions, and is well demonstrated by CT [23].

Sincipital cephaloceles must be differentiated from nasal gliomas and dermoids, which share a common embryological mechanism. Nasal gliomas (Fig. 4.10) are also called nasal brain heterotopia [24, 25]. The term “nasal glioma” is a dreadful, confusing misnomer, as it implies a neoplastic condition with malignant potential, which it is not. Nasal glioma is a rare developmental abnormality, and should be differentiated from glioma, which is a tumor of the brain, and from a primary encephalocele, which is herniation of the cranial contents through a bone defect in the skull, through which it retains an intact connection with the CNS [25]. Nasal gliomas are composed of dysplastic brain tissue located into the nasal fossae or in the subcutaneous tissues of the glabella. They are not connected with the CNS. Embryologically, nasal gliomas are similar to sincipital cephaloceles in that both entities are believed to result from nervous

79

80

P. Tortori-Donati, A. Rossi, and R. Biancheri

a

b

c

Fig. 4.9a–d. Sincipital encephalocele. a. Photograph of the patient; b. Sagittal T2-weighted image; c. Axial T1-weighted d image; d. 3D CT scan. Patient has large, slightly left-sided subcutaneous mass at the nasal root, and is hyperteloric (a). MRI shows herniation of dysplastic nervous tissue, originating from the left frontal lobe, through the fonticulus frontalis (arrow, b, c). A falcine sinus is associated (arrowhead, b). 3D CT scan shows location of bony defect, involving the caudal portion of the frontal bone at the glabella (arrow, d)

tissue herniating through the foramen caecum along a dural outpouching. If the proximal portion of the dural stalk regresses, sequestration of distal nervous tissue produces a nasal glioma, whereas preservation of the connection at the level of the foramen caecum results in a cephalocele. The differentiation between the two entities is obviously important in view of the inherently different surgical strategy. Dermoids are dysontogenetic masses that usually are associated with a dermal sinus tract (Fig. 4.11). The latter arises at an external ostium situated along the midline of the nose and extends deeply for a variable distance, sometimes crossing the

foramen caecum to reach the intradural intracranial space [21]. The differentiation of dermoids from nasal gliomas is based on their density on CT, which parallels that of fat, and signal behavior on MRI. Sphenoidal Cephaloceles

Sphenoidal cephaloceles usually are not clinically evident at birth, and require time to become manifest with either exophthalmos or upper airway obstruction. Therefore, they are generally recognized later in infancy, or even in adulthood. Physical examination in these children often reveals a pulsatile rhinopha-

81

Brain Malformations

a

c

b

Fig. 4.10a–c. Nasal glioma. a Sagittal T1-weighted image; b Gd-enhanced sagittal T1-weighted image; c Axial T2-weighted image. Large mass involving both the nasal cavity (asterisk, a, b) and the subcutaneous tissue of the nasal pyramid (arrow, a–c). The lesion is isointense on T1-weighted images (a), hyperintense on T2-weighted images (c), and shows inhomogeneous enhancement (b)

b

a Fig. 4.11a,b. Nasal dermoid. a Sagittal T1-weighted image; b Sagittal T2-weighted image. A small mass (arrowhead, a, b) extends intracranially through the foramen caecum (white arrow, a, b). A dependent fluid-fluid level is recognizable (black arrow, a)

82

P. Tortori-Donati, A. Rossi, and R. Biancheri

ryngeal mass covered by nasal mucosa that expands with a Valsalva maneuver. Spontaneous fistulas are rare, but may be complicated by ascending meningitis. Sphenoidal cephaloceles are further divided into several groups. Spheno-orbital cephaloceles: The defect may involve the superior orbital fissure, the great sphenoidal wing [26], the optic canal, and the walls of the orbit. Exophthalmos may result from chronic mass effect within the rigid orbital cavity [27]. Spheno-maxillary cephaloceles: The cephalocele herniates through the inferior orbital fissure into the pterygopalatine fossa, and usually contains portions of the temporal lobe. Naso-pharyngeal cephaloceles: This category includes several forms, depending on the relationship between the bony defect and the ethmoid, sphenoid, and basiocciput [28]. Transethmoidal cephaloceles (Fig. 4.12) cross a defect in the cribroid plate and extend into the ethmoidal cavities, which are deformed and remodeled; sometimes the sac protrudes into the nasal cavity, vestibule, and rhinopharynx. Spheno-ethmoidal cephaloceles pass between the body of the sphenoid and the ethmoid, and occupy the posterior nasal cavity or the rhinopharynx [29]. Spheno-nasopharyngeal cephaloceles develop through the body of the sphenoid bone and represent the more typical form of sphenoidal cephalocele, classically presenting as a submucosal pharyngeal mass (Fig. 4.13). When the defect involves the floor of the pituitary fossa, the pituitary gland, hypothalamus, and optic chiasm may herniate into the sac [30]. Basioccipital-nasopharyngeal cephaloceles are exceedingly rare and develop through a midline clival defect at the level of the spheno-occipital synchondrosis. The sac may contain a portion of the prepontine cistern, brainstem, and even fourth ventricle [31].

a

b

Internal Cephaloceles

These very rare cephaloceles are different from both calvarial and skull base cephaloceles in that they are acquired forms that result from positional rearrangement of the nervous structures secondary to major trauma or surgery [10, 16]. c

4.1.5 Syndromes Associated with Cephaloceles Cephaloceles may be part of malformative complexes, whose pattern of genetic transmission has been estab-

Fig. 4.12a–c. Transethmoidal encephalocele. A Reconstructed sagittal image from T1-weighted 3D sequence; b Coronal T1weighted image; c Coronal T2-weighted image. A large portion of the basal portion of the right frontal lobe herniates through an ethmoidal defect into the nasal cavities (asterisk, a–c). (Case courtesy Dr. Z. Patay, Riyadh, Saudi Arabia)

83

Brain Malformations

a

b Fig. 4.13a,b. Spheno-nasopharyngeal encephalocystocele. a. Sagittal T1-weighted image; b. Coronal T1-weighted image. The anterior third ventricle herniates through the sellar floor in the sphenoidal region (asterisk, a, b). The two portions of the pituitary gland (anterior and posterior lobe) are recognizable, although the gland is flattened against the clivus (arrowheads, a)

lished in some cases [32]. Cephaloceles may represent either a constant or a possible finding, depending on the nature of the syndrome (Table 4.2). The association with other congenital CNS anomalies accounts for the often severe psychomotor compromise exhibited by affected children. Genetic counseling is obviously very important in view of the severe prognosis which is often associated with these syndromes.

underdeveloped premaxilla, median cleft of the secondary palate, widely separated nasal halves, extreme asymmetrical orbital hypertelorism with strabismus, widow’s peak, enlarged ethmoidal air cells, low cribriform plate, thickened nasal septum, duplication of the frenulum, nose or teeth, and intracranial-associated anomalies, such as holoprosencephaly, lipomas of the interhemispheric fissure, and callosal agenesis.

Frontonasal Dysplasia

Tecto-Cerebellar Dysraphia

Frontonasal dysplasia (also called frontonasal dysostosis or median cleft face syndrome) [33] comprises ocular hypertelorism, median facial cleft affecting nose and/or upper lip, unilateral or bilateral cleft of the alae nasi, anterior cranium bifidum occultum, or a widow’s peak (the strip of hairline at the top of the forehead sticks out, as opposed to being horizontal). It usually is a sporadic disorder, although a few familial cases have been reported, suggesting either autosomal dominant or X-linked dominant transmission [34]. The embryological origin of this syndrome is in the period prior to the 28 mm stage. It is due to deficient remodeling/differentiation of the nasal capsule which causes the future fronto-naso-ethmoidal complex to freeze in the fetal form [35]. This syndrome has a range of severity in its manifestations. There is an increased incidence of basal/ fronto-ethmoidal/transsphenoidal encephalocele with pituitary herniation. Other deformities that may be present are [21] a bifid nose tip/dorsum with or without a median notch/cleft of the upper lip, bifid or

This abnormality is characterized by absent vermis, kinking of the brainstem, marked tectal beaking, and herniation of the cerebellum into an occipital encephalocele (see below, malformations of the posterior cranial fossa). Chiari III Malformation

This rare abnormality is defined by the association of a high cervical/low occipital encephalocele and signs of Chiari II malformation, that may include a small posterior fossa, tectal beaking, and low bulbo-medullary junction. Other features include hypoplasia of the low and midline aspects of the parietal bones, petrous and clival scalloping, cerebellar hemisphere overgrowth, cerebellar tonsillar herniation, hydrocephalus, dysgenesis of the corpus callosum, posterior cervical vertebral agenesis, and spinal cord syringes. The encephalocele may contain varying amounts of brain (either cerebellum and occipital lobes or cerebellum only), ventricles (fourth and lateral), cisterns, and brainstem [20].

84

P. Tortori-Donati, A. Rossi, and R. Biancheri Table 4.2. Syndromes associated with cephalocele Syndrome

Cephalocele location

Frequency of asso ciated cephalocele

Transmission modality

Frontonasal dysplasia

Frontal

100%

Heterogeneous

Tecto-cerebellar dysraphia

Occipital

100%

Sporadic

Chiari III malformation

Occipital, occipito-cervical

100%

Sporadic

Von Voss-Cherstvoy

Occipital

100%

Autosomal recessive

Iniencephaly

Occipital

Frequent

Sporadic

Meckel-Gruber

Occipital

80%

Autosomal recessive

Knobloch

Occipital

80%

Autosomal recessive (presumed)

Joubert

Occipital

30%

Autosomal recessive

Walker-Warburg

Occipital

30%

Autosomal recessive

Dyssegmental dwarfism

Occipital

20%

Autosomal recessive

Cryptophthalmos

Occipital

10%

Autosomal recessive

Klippel-Feil

Occipital

Rare

Heterogeneous

Pseudo-Meckel

Occipital

Rare

Chromosomal; t(3p+)

Roberts

Frontal

Rare

Autosomal recessive

Goldenhar

Occipital

Rare

Heterogeneous

Early amniotic rupture

Anterior (other)

Rare

Sporadic

Fetal warfarin

Occipital

Rare

Iatrogenic

Apert

Frontonasal

Rare

Autosomal dominant

von Voss-Cherstvoy Syndrome

Iniencephaly

This entity, also known as DK-phocomelia syndrome, is characterized by radial ray defects, occipital encephalocele, and urogenital abnormalities [36, 37]. Other possible CNS findings include callosal agenesis, partial vermian agenesis, and hypoplasia of the olivary nuclei and pyramids [16]. Both sexes are affected, and parental age is not increased. The inheritance pattern may be heterogeneous, but autosomal recessive inheritance has been suggested in at least some instances [36].

Iniencephaly is a rare neural tube defect of the craniocervical junction [1:1000–1:100,000 births] [39] involving the occiput and inion, combined with rachischisis of the cervical and thoracic spine. It is characterized by marked, fixed retroflection of the skull on the cervical spine due to thick fibrous bands, representing hypoplastic trapezius muscles and a thick ligamentum nuchae [39]. Occipital bone defect and spina bifida of the cervical vertebrae are other distinguishing features of this abnormality [40]. The neck is often absent, and the skin is continuous with the scalp, face, and chest. The abnormal and retroflexed occiput may join the vertebrae of the spine. The cervical spine is always severely deformed and shows multiple anomalies, such as a block or deficiencies. Posterior midline schises may include cleft occipital bone, wide foramen magnum, and vertebral anomalies affecting the cervical and thoracic spine with anterior/posterior spina bifida. The pedicles may be short and the laminae absent. The overlying skin and dorsal soft tissues may be intact or completely deficient, exposing the neural tissue to the environment [39]. Although the cause of this abnormality is unknown, low parity and socioeconomic status probably play a role [41]. There is a marked female predilection (90% of cases). Most iniencephalic fetuses are stillborn;

Joubert Syndrome

Occipital cephaloceles are found in 30% of patients with Joubert syndrome. Clinically, hyperpneic/ apneic spells, ataxia, nystagmus, and psychomotor delay represent the classical tetrad. Radiologically,the “molar tooth” malformation is a typical feature (see below). Walker-Warburg Syndrome

Walker-Warburg Syndrome is an autosomal recessive disorder showing characteristic brain and eye malformations (cobblestone complex, cerebellar and brainstem malformations, retinal abnormalities) in patients with congenital muscular dystrophy. Occipital cephaloceles occur in 24% of cases [38].

Brain Malformations

however, a limited number of viable newborns has been described in whom surgery allowed correction of the abnormal position of the head. Iniencephaly is subdivided into iniencephaly clausus and iniencephaly apertus, depending on the presence of an associated occipital cephalocele that may contain the occipital lobes, cerebellum, and brainstem. Although the morphological appearance of iniencephaly may resemble a large occipital cephalocele, the huge occipital bone defect and rachischisis of the cervical vertebrae are distinctive features [40]. Associated malformations of the CNS include anencephaly, hydrocephalus, callosal dysgenesis, microcephaly, holoprosencephaly, atresia of the ventricular system, and malformations of the cerebral cortex [16, 39, 40]. Systemic malformations include cyclopia, cleft lip and palate, bifid uvula, deformed ears, fusion or deficiency of ribs, abnormal lobation of the lungs, diaphragmatic hernia, single umbilical artery, omphalocele, situs inversus, horseshoe or polycystic kidneys, imperforate anus, club feet, and congenital heart disease. Meckel-Gruber Syndrome

Also called dysencephalia splanchnocystica or simply Meckel syndrome, it is a rare, lethal syndrome characterized by occipital cephalocele, postaxial polydactyly, and dysplastic cystic kidneys. It can be associated with several other conditions, including fibrotic lesions of the liver. The incidence varies from 0.07 to 0.7 per 10,000 births, but in Finland the disorder is unusually frequent and reaches 1.1 per 10,000 births [42]. It is also estimated that this syndrome corresponds to 5% of all neural tube defects [43]. Inheritance is autosomal recessive, and the genetic locus has been mapped on the long arm of chromosome 17 [44]. Occipital cephaloceles are present in 60%–80% of cases. Other possible CNS abnormalities include microcephaly, holoprosencephaly, cerebral and cerebellar hypoplasia, hypoplasia of pituitary gland, and the Dandy-Walker malformation. Eye anomalies (microphthalmos, cataract, coloboma), cleft lip and palate, and facial abnormalities (Potter-like face) can also be found [16]. Pseudo-Meckel Syndrome

Pseudo-Meckel syndrome is characterized by congenital absence of the rhinencephalon, callosal agenesis, Chiari I malformation and a variable association of cardiac defects, cleft palate, club feet, and hammer toes; occipital cephalocele occurs rarely [16]. Knobloch Syndrome

Knobloch syndrome is an autosomal recessive syndrome of hereditary vitreoretinal degeneration with

retinal detachment, severe myopia, and occipital encephalocele in children with normal intelligence [45]. However, congenital occipital scalp defects, rather than true encephaloceles, may accompany Knobloch syndrome in some cases [46]. Dyssegmental Dwarfism

Dyssegmental dwarfism is an autosomal, recessively inherited, lethal, generalized chondrodysplasia, characterized by micromelia, cleft palate, and variable limited mobility at the elbow, wrist, hip, knee, and ankle joints. In some cases, occipital encephalocele, inguinal hernia, hydronephrosis, hydrocephalus, and patent ductus arteriosus are found [47]. Cryptophthalmos

Cryptophthalmos is a condition that results in failure of eyelid formation. It is divided into three types: the complete, incomplete, and symblepharon variety. The complete variety is the most common; the eyelids do not form and the eyelid skin grows continuously from the forehead to the cheek, covering the underlying globe which is usually abnormal. Cryptophthalmos is associated with several other congenital anomalies, including abnormal hairline, syndactyly, and an occipital cephalocele in 10% of cases [16, 48]. Klippel-Feil Syndrome

The Klippel-Feil deformity is a complex of osseous and visceral anomalies that includes low hair line, platybasia, fused cervical vertebrae with short neck, and deafness. The classical clinical triad consist of short neck, limitation of head and neck movements, and low set posterior hairline. Bony malformations may entrap and damage the brain and spinal cord. The disorders of the lower vertebral region may become symptomatic in adolescence or adult life. The pathogenesis has been related to anomalous somitic segmentation between gestational weeks 4 and 8 [16]. Associated CNS abnormalities include occipital cephalocele, Chiari I malformation, syringes, microcephaly, and hydrocephalus [16]. Several associated abnormalities, such as scoliosis, posterior bony spina bifida, absence of ribs, conductive hearing loss, mirror movements, unilateral renal ectopia with dilated collecting system, microtia, and preaxial polydactyly have also been reported. The pattern of bony fusion may involve more than one level, producing the “wasp waist sign” when two adjacent levels are involved [49]. Cervical spondylosis, disk herniation, and secondary degenerative changes are more common at levels adjacent to fused vertebrae [50]. Spontaneous and progressive neurological sequelae and neurological injury may follow minor neck trauma.

85

86

P. Tortori-Donati, A. Rossi, and R. Biancheri

Roberts Syndrome

Roberts syndrome (also known as pseudo-thalidomidic syndrome) is a rare genetic disorder characterized by pre- and postnatal growth retardation, limb defects, and craniofacial anomalies. Affected individuals have variable malformations that involve symmetric reduction in the number of digits and length or presence of bones in the arms and legs. Craniofacial malformations include hypertelorism, hypoplastic nasal alae, and a high incidence of cleft lip and palate. Familial and sporadic cases have been reported, consistent with an autosomal recessive mode of inheritance [51]. Cephaloceles occur occasionally in the frontal region. Goldenhar Syndrome

The main features of this condition are unilateral underdevelopment of one ear associated with underdevelopment of the jaw and cheek on the ipsilateral side of the face (hemifacial microsomia), possibly associated with vertebral anomalies and an epibulbar dermoid. The muscles of the affected side of the face are underdeveloped and there often are skin tags or pits in front of the ear, or in a line between the ear and the corner of the mouth. Often, there are abnormalities of the middle ear, and the ear canal may be completely absent. Unilateral deafness is extremely common. Cephaloceles (both anterior and posterior), plagiocephaly, and intracranial dermoids are occasionally associated [16, 52, 53]. Amniotic Band Syndrome

Severe fetal malformations, including neural tube defects, may be secondary to early amniotic rupture followed by formation of fibrous bands that cause fetal malformation, deformation, and compression. Most of the craniofacial defects (encephaloceles and/ or facial clefts) occurring in these fetuses are believed to result from a vascular disruption sequence, with or without cephalo-amniotic adhesion [54]. Fetal Warfarin Syndrome

In utero exposure to warfarin between gestational weeks 6 and 9 produces a constellation of nasal hypoplasia, punctate epiphyses, optic atrophy, mental retardation, hydrocephalus, and occasionally occipital cephalocele [16]. Apert Syndrome

It is characterized by craniosynostosis and syndactyly, and may be associated with hydrocephalus and naso-frontal cephalocele. The inheritance is autosomal dominant, although a number of sporadic cases have been reported [16]. (see Chap. 30).

4.2 Defects of the Mediobasal Prosencephalon Holoprosencephalies and Related Entities

4.2.1 Holoprosencephaly 4.2.1.1 Background

Holoprosencephaly (HPE) was defined by Yakovlev [55] as “a median holosphere with a single ventricular cavity instead of two hemispheres with symmetrical lateral ventricles.” HPE is a complex developmental abnormality of the forebrain that is related to anomalous differentiation and separation of the prosencephalic vesicle, the cranialmost of the three primitive vesicles that appear at the anterior end of the neural tube shortly after the end of primary neurulation. Anatomically, the basic defect involves incomplete separation of the eye fields and developing forebrain into distinct left and right halves [56]. Therefore, HPE may be viewed as a developmental field defect, with impaired cleavage of the embryonic forebrain as its cardinal feature. The prevalence is about 1 in 16,000 live births [57], but the incidence in conceptuses obtained through induced abortion is much higher (approximately 1 in 250) [58], reflecting early spontaneous loss of the majority of the abnormal fetuses. In most cases, craniofacial abnormalities are associated (Fig. 4.14), and reflect in 80% of cases the degree of severity. Their severity is variable, ranging from cyclopia to minimal craniofacial dysmorphism, such as mild microcephaly with a single central incisor [59]. HPE is characterized by hypoplasia of the most rostral portions of the neural tube, resulting in fused cerebral hemispheres and absence of midline structures, such as the rhinencephalon, corpus callosum, and septum pellucidum. In the past, HPE was confused with “arhinencephaly,” thereby stressing the absence of olfactory structures. Subsequently, DeMyer and Zeman [60] introduced the term HPE, which identifies more clearly the lack of cleavage and differentiation of the prosencephalon as the main causal factor. HPE is classically divided into alobar, semilobar, and lobar types according to the severity of the malformation, which follows an anterior-to-posterior gradient. In general, the basal forebrain, in the region of the hypothalamus and subcallosal cortex, is most severely affected. As one goes posteriorly in the cerebrum, the involvement becomes progressively less [61]. However, a precise boundary among the three groups does not exist, and intermediate cases may

Brain Malformations

be identified. Therefore, HPE should be viewed as a continuous spectrum ranging from the most severe alobar forms, in which the fetus is usually nonviable, to the milder lobar forms, in which the abnormality may be discovered incidentally [62]. An exception to this anterior-to-posterior rule is syntelencephaly, or middle interhemispheric holoprosencephaly (MIH) [63], which could represent a separate form [61]. 4.2.1.2 Embryogenesis and Pathogenetic Theories

a

b

c

Traditionally, HPE has been related to absent cleavage of the prosencephalon, which results in fusion of the cerebral hemispheres along the midline associated with a number of midline abnormalities, such as absence of the olfactory system, septum pellucidum, and corpus callosum. The etiology of HPE is very heterogeneous and comprises environmental factors (i.e., maternal diabetes mellitus) [64], teratogens, and genetic causes. HPE can occur as apparently sporadic or as familial cases. Approximately 50% of cases are associated with a cytogenetic abnormality, of which trisomy 13 is the most common. Other chromosomal aberrations include partial deletion of the long arm of chromosome 13, ring chromosome 13, trisomy 18, partial deletion of the short arm of chromosome 18, ring chromosome 18, and partial trisomy 7. Based on recurrent cytogenetic abnormalities, there are at least 12 genetic loci that likely contain genes implicated in the pathogenesis of HPE. Currently, four human HPE genes are known: HPE1 on chromosome 21, HPE2 on chromosome 2, HPE3 on chromosome 7, and HPE 4 on chromosome 18 [56, 59, 65]. Because of such heterogeneity, it probably is incorrect to think of a gene “for” HPE, but rather, it is the action of a gene or genes with the environment of the developing embryo that determines the outcome [56]. The HPE3, or sonic hedgehog (shh) gene, maps on human chromosome 7q36 and is responsible for an autosomal dominant form of HPE that has been well described. Penetrance in families with HPE3 mutations approaches 88%. Shh encodes proteins that are implicated in ventral embryonic patterning, suggesting that HPE is a disorder of ventral induction. Shh is expressed during gastrulation in

Fig. 4.14a–c. Facial dysmorphism in holoprosencephaly. a Cyclopia: complete fusion of orbits and eyes. The nose is absent, while there is a midline proboscis above the single orbital cavity. b. Ethmocephaly: Separation of the orbits, albeit with marked hypotelorism. A single or double proboscis originates between the eyes. The nose is absent. c. Cebocephaly: Small, rudimentary nose with a single nostril. The orbits are small and hypoteloric

87

88

P. Tortori-Donati, A. Rossi, and R. Biancheri

several locations. In the spinal cord, the main source of ventralizing signals is the notochord, whereas in the forebrain ventralization is chiefly regulated by the prechordal plate, a region of foregut mesoderm underlying the embryonic hypothalamus and thalamus that lies cephalad to, and is induced by, the cranial extremity of the notochord. In normal embryonic development, an interplay of dorsalizing molecules (emanating from the roof plate) and ventralizing molecules (emanating from the prechordal plate and floor plate) modulates regional identity of tissues along the dorsoventral axis of the neural tube. Either a lack of production of ventralizing factors or an overproduction of dorsalizing factors can result in noncleavage (commonly called fusion) of structures that normally lie just lateral to the midline. This is the presumed mechanism by which HPE develops [66, 67]. Accurate characterization of the longitudinal (i.e., segmental) organization of the forebrain along the anteroposterior axis indicates that the hypothalamus is not part of the diencephalon, as classically conceived, but is grouped with the telencephalon in an anterior domain known as the secondary prosencephalon [66]. Thus, the structures most consistently malformed in HPE (i.e., the cortex, striatum, and hypothalamus) all belong to the same embryonic anteroposterior domain—the secondary prosencephalon—but occupy different dorsoventral positions within it. In this light, HPE may be best viewed as a disorder of dorsoventral patterning affecting the secondary prosencephalon [66]. Interestingly, cases of associated HPE and caudal agenesis, also known as the caudal regression syndrome, have been reported in the recent literature [68, 69]. This association could perhaps indicate a common embryogenetic background for these abnormalities. Caudal agenesis results from a disorder of notochordal formation, and has been thought to relate with cellular depletion at the level of the Hensen’s node (see Chap. 38). The neural tube that forms in chick embryos in which the caudal part of the Hensen’s node has been removed does not express ventral markers such as Shh, and is subsequently removed by apoptosis; however, grafting with Shhsecreting fibroblast rescues the caudal neural tube [70]. The effect of Shh on cell survival in the early neuroectoderm could reveal a common pathogenetic ground for caudal agenesis and HPE.

crine dysfunction [60]; however, symptoms may vary depending on the severity of the malformation. Also, HPE is typically associated with facial abnormalities (Fig. 4.14) that vary according to the severity of the abnormality. This occurs because the paired nasal and optic placodes are induced by the prosencephalon, and these in turn induce the maturation of the frontonasal and maxillary processes, which are ultimately responsible for the formation of the face [71]. Cyclopia, ethmocephaly, and cebocephaly are the most severe forms of facial derangement, and are usually associated with alobar HPE. Cyclopia is characterized by complete fusion of the ocular globes and orbits [72]. There is absence of a nose, with a midline appendage or “proboscis” usually present above the single orbit, and trigonocephaly is usually associated. Ethmocephaly exhibits marked hypotelorism. A bulky, single or double proboscis grows in between the two orbits along the midline, whereas the nose is absent. Cebocephaly is characterized by a short, flat nose with a single nostril. The orbits are small and hypoteloric. Less severely affected individuals may show agenesis of the premaxillary segment of the face, cleft lips and palate, hypotelorism, or trigonocephaly. In mild (i.e., lobar) forms, facial abnormalities usually are absent, and the clinical picture is one of mild to moderate psychomotor delay, hypothalamic-pituitary dysfunction, and visual disturbances [61]. The occurrence of a solitary median maxillary central incisor is a very rare condition, and may be a sign of a mild degree of HPE [73].

4.2.1.3 Clinical Features

The Genoa syndrome involves HPE, craniosynostosis, and multiple congenital anomalies including cloverleaf skull, Dandy-Walker malformation, bilateral microphthalmia, cleft soft palate, congenital scoliosis, hypoplastic nails, and coarctation of aorta [76].

The neurologic picture includes mental retardation, spastic quadriparesis, athetosis, epilepsy, and endo-

4.2.1.4 Associated Conditions

HPE may represent a part of several complex syndromic associations. Neural tube defects [74] and anterior cephaloceles [16] are much greater associations than is commonly thought. Smith-Lemli-Opitz syndrome is an autosomal recessive disorder of cholesterol biosynthesis. The syndrome involves poor growth, developmental delay, and a common pattern of congenital malformations including cleft palate, genital malformations, polydactyly, and HPE [75]. Because cholesterol is required for proper spatial restriction of Shh signaling [56], mutations of cholesterol synthesis may result in HPE.

Brain Malformations

CHARGE (Coloboma, Heart Anomaly, Retardation, Genital and Ear anomalies) has been rarely described in association with HPE [77]. 4.2.1.5 Imaging Studies

The MRI diagnosis of HPE is based on a careful assessment of the telencephalic structures, particularly in the basal regions, for noncleavage [61]. This can be difficult, especially in microcephalic newborns and when superimposed hydrocephalus further distorts the anatomy. Thin coronal sections, both on T1- and T2-weighted images, are especially important, and reassessment following decompression of the ventricular system may improve diagnostic accuracy [61]. The traditional classification into alobar, semilobar, and lobar HPE [60] reflects the graded severity of the malformation with respect to both the brain and face. Although this categorization is useful to give an idea of the severity of the picture, it has become apparent that several important features, such as cortical malformations and variations in separation of deep gray nuclei, cannot be incorporated into this classification scheme, and neurodevelopmental outcome cannot always be predicted accurately by this rather inflexible system [66]. As was previously stated, HPE should be viewed as a continuous spectrum of abnormality ranging from mild (i.e., lobar) to severe (i.e., alobar) forms, without precise boundaries among the three varieties. This concept is reflected clinically in a continuous phenotypic spectrum ranging from cyclopia at one extreme to clinically unaffected carriers of an HPE mutation at the other [56]. Several specific features have been used to assess the severity of the abnormality using MRI. These have included, traditionally, the degree of anterior extension of the corpus callosum [78, 79] which, however, is prone to subjective evaluation. Barkovich et al. [80] have proposed using the determination of the so-called sylvian angle to evaluate the severity of HPE. The sylvian angle is defined as the anterior intersection of two lines drawn to connect the anterior and posterior extremity of the insula in the axial plane. The sylvian angle progressively increases with the severity of the malformation; the mean value is 15° in normal individuals, 39° in lobar HPE, 63° in mild semilobar HPE, 101° in severe semilobar HPE, and 122° in alobar HPE [80]; in the most severe forms, no sylvian fissures are developed, thus identifying a category of “asylvian” HPE, which represents the most severe end of the HPE spectrum. The same authors found that as the severity of the malformation increases, the sylvian fissures are located more and more anteriorly with respect to the front of

the brain, and the amount of cerebral tissue that is comprised between them progressively diminishes [80]. More recently, noncleavage of the deep gray matter nuclei has been extensively analyzed by Simon et al. [66]. Contrary to other reports [81], these authors found that the degree of noncleavage of the deep gray matter nuclei is not always in direct proportion to the severity of the cerebral hemispheric malformation. Hypothalamic noncleavage was present in all cases, caudate noncleavage was found in 96% of cases, and 85% of cases showed some noncleavage of the lentiform nuclei. Noncleavage of the thalami occurred in 67% of patients in the same study, and abnormalities in orientation of the long axis of the thalamus were associated with the degree of thalamic noncleavage. Some degree of mesencephalic noncleavage, associated with the presence of thalamic noncleavage, was found in 27% of cases. These results show that the frequency of noncleavage in HPE is dependent on proximity to the midline and to the ventricular system. Structures normally located adjacent to the midline and/or ventricles (such as the cingulate cortex, caudate nucleus, and medial hypothalamus) are more frequently noncleaved than are structures normally located further from the midline or ventricles (such as the globus pallidus). Also, the involvement of more posterior brain structures, such as the thalamus (diencephalon) and midbrain (mesencephalon), decreases with distance from the anterior pole. This gradient of involvement in HPE may be related to the relative effectiveness of ventralizing signals from the notochord and prechordal plate [66]. In conclusion, it seems likely that the traditional classification into alobar, semilobar, and lobar HPE will sooner or later be dropped in favor of other more valuable classification schemes; however, we still believe it is useful to maintain the traditional classification in the present textbook, for didactic purposes. Alobar Holoprosencephaly

Alobar HPE traditionally designates complete failure of cleavage of the two cerebral hemispheres. The diagnosis of alobar HPE is usually made antenatally by ultrasounds (see Chap. 26), and may be confirmed by fetal MRI (see Chap. 27). Neonates are usually stillborn or have a very short life-span. Therefore, they are rarely imaged by MRI. Survivors can present with neonatal seizures and/or infantile spasms, profound mental retardation, rigidity, apnea, and temperature imbalance. Generalized epilepsy may develop during childhood [82]. Facial anomalies usually are severe and result from absence or hypoplasia of the premaxillary segment of the face [71].

89

90

P. Tortori-Donati, A. Rossi, and R. Biancheri

a Fig. 4.15a,b. Alobar holoprosencephaly. a Sagittal T1-weighted image; b. Axial T2-weighted image. There is a single rudimentary b ventricle that is continuous with a large dorsal CSF-filled cavity, the dorsal sac, which extends cranially to the calvarium (a, b). The brain is extremely rudimentary and uncleaved, i.e., midline structures such as the interhemispheric fissure and falx cerebri are thoroughly absent, while the deep white matter is continuous across the midline (b). The septum pellucidum is also absent. The posterior fossa is normal. (Case courtesy Dr. E. Simon, Philadelphia, PA, United States)

a

b

d

c

Fig. 4.16a–d. Alobar holoprosencephaly. a Sagittal T2-weighted image; b Axial T2-weighted image; c Coronal T2-weighted image; d 3D TOF MRA, coronal MIP. The brain is reduced to a flattened pancake of tissue anteriorly (a). The holoventricle (asterisk, a, b) communicates widely with a large dorsal cyst (DC, a, b). The fused thalami are visible on coronal image (arrows, c). MRA shows agenesis of both anterior cerebral arteries (d)

Brain Malformations

The MRI picture reflects the pathological appearance of the malformation (Figs. 4.15, 4.16). The definition of alobar HPE involves absence of distinct temporal lobes and temporal horns of the lateral ventricles, and absent interhemispheric fissure [80]. The cerebrum is rudimentary, may be shaped like a pancake, cup, or ball [60], and is located in the cranialmost aspect of the calvarium. There is continuation of gray and white matter across the midline with absence of both interhemispheric fissure and falx cerebri. The brain surface may be smooth or have a few, broad gyri with shallow, abnormally oriented sulci; however, cortical thickness is normal. The only cortical malformation that has been found in alobar HPE in a large series was subcortical heterotopia [80]. Midline structures, such as the superior sagittal sinus, septum pellucidum, corpus callosum, third ventricle, pituitary gland, and olfactory bulbs, are absent; the optic nerves and chiasm are usually hypoplastic and poorly myelinated. The thalami, hypothalamus, and basal ganglia are not separated, resulting into absence of the third ventricle; usually, these rudimentary nuclei form a bilobed mass that abuts a single, rudimentary, crescent-shaped ventricle (“holoventricle”). The dorsal-most part of the rudimentary brain usually encompasses the holoventricle. When the dorsal lip of the brain does not roll over the holoventricle, the ependymal lining bulges dorsally and extends upward to the calvarium. This so-called “dorsal sac” or “dorsal cyst” is really the holoventricle itself. The presence of a dorsal cyst correlates with the presence and degree of thalamic noncleavage; it has been speculated that the unseparated thalamus mechanically blocks egress of cerebrospinal fluid from the third ventricle, resulting in ballooning of the posterior third ventricle to form the cyst [83]. The brainstem is usually grossly normal, except for absence of the pyramidal tracts. The cerebellum may be normal or small. If normal, it appears relatively large compared to the small cerebrum. The vascular system is commonly composed of a network of vessels arising from the internal carotid and basilar arteries [61]. The main differential diagnosis is with agenesis of the corpus callosum associated with large interhemispheric cysts. The diagnosis is mainly based on the assessment of the falx cerebri, which is absent in alobar holoprosencephaly and present, albeit usually displaced, in callosal agenesis with interhemispheric cysts.

Semilobar Holoprosencephaly

In semilobar HPE (Fig. 4.17), facial abnormalities are milder or even absent. The interhemispheric fissure and falx cerebri are present in the posterior regions of the brain, whereas there is failure of cleavage of the frontal and parietal lobes anteriorly. The brain is less dysmorphic than in alobar HPE, but is invariably smaller than normal. The holoventricle shows development of rudimentary temporal horns, although the hippocampus is usually incompletely formed; the frontal horns are not present. The thalami may be partially separated and a rudimentary third ventricle. The dorsal sac usually is absent. When present, it is usually smaller than in alobar HPE, and may be difficult to differentiate from an enlarged interhemispheric fissure. The corpus callosum is absent in the uncleaved regions, whereas a rudimentary commissure is present where the two cerebral hemispheres are separated. This so-called pseudosplenium gradually extends anteriorly as the severity of the forebrain malformation lessens, and has been considered a marker of the severity of the abnormality [78, 79]. It represents a notable exception to the rule of anterior-to-posterior development of the corpus callosum. HPE is the only anomaly in which isolated agenesis of the anterior segments of the corpus callosum is present, whereas it is the posterior aspect of the corpus callosum that is malformed in partial callosal agenesis [71]. Lobar Holoprosencephaly

Lobar HPE is the mildest form, usually found in asymptomatic or mildly retarded individuals. Lobar HPE (Fig. 4.18) implies the presence of the interhemispheric fissure with midline continuity of the neocortex [84] and some degree of formation of the frontal horns. Failure of cleavage usually involves the anterior and basal regions of the frontal lobes, other than the basal ganglia. The thalami usually are separate or may be joined by a broad interthalamic mass. The lateral ventricles are rather well-formed as compared with semilobar HPE. Abnormal vertical orientation of the hippocampal formation is a possible associated finding. An arrest of the normal process of hippocampal inversion was found in all patients with HPE in one study [85]. An azygous anterior cerebral artery is usually found. The septum pellucidum is constantly absent and the frontal horns are uncleaved, resulting in a rudimentary box-

91

92

P. Tortori-Donati, A. Rossi, and R. Biancheri

a

b

Fig. 4.17a–c. Semilobar holoprosencephaly. a Axial T2-weighted image; b Coronal STIR image; c Sagittal T2-weighted image. There is complete uncleavage of the frontal lobes (a, b) with absence of both the interhemispheric fissure and falx cerebri. Notice that both the interhemispheric fissure and falx are formed posteriorly (arrows, a), consistent with normal cleavage of the posterior regions of the brain. Correspondingly, the trigones of the lateral ventricles, albeit rudimentary, are formed, whereas there is agenesis of the frontal horns (a, b). There is an azygous anterior cerebral artery (arrowheads, a, c) that courses abnormally into a deeper median sulcus over the surface of the brain. The corpus callosum is absent except for the splenium (arrow, c). Notice concurrent Dandy-Walker variant (c)

c

like shape on coronal sections, a feature in common with septo-optic dysplasia. Isolated absence of the septum pellucidum probably represents the mildest form of lobar HPE, and must be differentiated from septo-optic dysplasia. Assessment of the optic nerves does not reveal abnormalities in lobar HPE, unlike in septo-optic dysplasia. Differentiation between semilobar and lobar HPE can be difficult, as no clear-cut boundary exists between the two abnormalities. The degree of anterior extension of the “pseudosplenium” may be used as an approximate indicator of the severity of the noncleavage. If the third ventricle is fully formed, some frontal horn formation is present, and the splenium and posterior half of the callosal trunk are formed, the abnormality can be classified as lobar HPE [61].

4.2.2 Syntelencephaly (Middle Interhemispheric Holoprosencephaly) 4.2.2.1 Background

Syntelencephaly, or middle interhemispheric holoprosencephaly (MIH), was first described by Barkovich and Quint in 1993 [63]. Although initially considered a variant of semilobar HPE, MIH is typically characterized by failure of cleavage of the posterior frontal and parietal regions of the brain, in spite of separation of the rostrobasal forebrain, with presence of an interhemispheric fissure anteriorly. This represents a crucial difference from classical HPE, in

Brain Malformations

a

b

Fig. 4.18a–c. Lobar holoprosencephaly. a Axial STIR image; b Coronal STIR image; c Sagittal T1-weighted image. Cleavage of the cerebral hemispheres is greater than in semilobar holoprosencephaly. In this case, there are rudimentary frontal horns. The anterior portions of the frontal lobes are uncleaved (asterisks, a, b), and the anterior interhemispheric fissure is absent; the bottom of an aberrant sulcus can be seen in this axial image (arrowhead, a). Notice that the septum pellucidum is absent (a), a constant feature of all variants of holoprosencephaly. On the midsagittal image, the corpus callosum is more developed and extends far more anteriorly than in semilobar forms (compare with Fig. 4.17); the degree of anterior extension of the corpus callosum (arrow, c) can be used as a rough indicator of the severity of the malformation.

c

which the rostrobasal forebrain typically is involved. Another difference is the possible presence of the septum pellucidum [86]. MIH currently is viewed as a unique form of HPE [61], but could in fact represent a different malformation. Affected patients are usually severely developmentally retarded, but they have no facial deformities and either normal interorbital distance or, even, hypertelorism, contrary to typical HPE patients. However, we saw a patient with MIH and normal intelligence (IQ = 102). 4.2.2.2 Pathogenesis and Relationship with HPE

Although MIH and classical HPE share a number of similarities, including noncleavage of a substantial

portion of the cerebral hemispheres, abnormal anterior cerebral arterial circulation, and abnormal orientation of the sylvian fissures, significant differences exist between the two entities. This raises the question whether MIH should be regarded as a variant of HPE or, rather, as a separate entity. As was previously stated, classical HPE is regarded as a disorder of dorsoventral patterning due to excessive ventralization of the neural tube at level of the floor plate. Instead, MIH can be viewed as excessive dorsalization of the neural tube at level of the roof plate. Interestingly, a mutation of the ZIC2 gene has been found in a patient with MIH [87]. ZIC2 plays an important role in differentiation of the roof plate and in the development of the interhemispheric fissure, thereby explaining the prevalence of abnormalities at the level of the

93

94

P. Tortori-Donati, A. Rossi, and R. Biancheri

dorsal portions of the brain in MIH. These data seem to support the hypothesis that MIH is an abnormality of dorsal induction, contrary to classical HPE, which is produced by abnormal ventral induction. As such, a future genetic classification scheme could lead to a separation of these entities into different categories. 4.2.2.3 Imaging Findings

MIH is characterized by defective cleavage of the two cerebral hemispheres at the level of the posterior frontal and parietal lobes, with resulting “fusion” across the midline at the level of the oval centers and absence of the corresponding part of the interhemispheric fis-

sure. The corpus callosum is typically hypoplastic. A thinned genu and anterior trunk are present in some cases (Fig. 4.19). In others, the corpus callosum is separated in two distinct anterior and posterior portions (Fig. 4.20). The nervous tissue that bridges the midline may be represented by a heterotopic nodule or by cortical dysplasia [88]; in the majority of cases, the sylvian fissures are abnormally connected across the midline [87]. In MIH, cleavage of deep gray nuclei occurs differently than in classical HPE; significantly, the hypothalamus (which is consistently involved in classical HPE) is always cleaved with intervening formation of the third ventricle [87]; the lentiform nuclei are also consistently cleaved, whereas the caudate nuclei are separated in the majority of cases. Failure

a

c

b

d Fig. 4.19a–d. Syntelencephaly. a Sagittal T1-weighted image; b Sagittal T2-weighted image; c, d Coronal T2-weighted images. Cerebral hemispheres are uncleaved in the parietal region, where cortical dysplasia is seen to connect the two hemispheres (arrowheads, a–d). Anteriorly, the corpus callosum is reduced to a thin lamina (arrows, a, b). An azygous anterior cerebral artery is present (open arrows, a, b)

Brain Malformations

a

b Fig. 4.20a–c. Syntelencephaly in a mildly hypotonic patient with normal intelligence (IQ = 102). a Axial T2-weighted image; b Coronal STIR image; c Sagittal T1-weighted image. On axial images, lateral ventricles have a rudimentary shape and the septum pellucidum is absent; however, normally developed interhemispheric fissure both anteriorly and posteriorly (arrows, a) would not suggest holoprosencephaly at first glance. However, coronal images clearly show a thin sling of normal-appearing cortex bridging the midline just below the callosomarginal sulcus (arrowheads, b); notice that cortex should never be found between the commissural plate and the callosomarginal sulcus in normal individuals. This bridging cortex represents the site of uncleavage of the two cerebral hemispheres. On sagittal images, partial intermediate agenesis of the corpus callosum is found, with both the genu (G) and splenium (S); and an intervening thin hyperintense commissural structure (arrows, c). One could speculate that the developmental defect prevented normal fasciculation of the callosal fibers over the hippocampal commissure, whereas both the genu (whose development follows the glial sling) and posterior splenium did develop more appropriately (see Chap. 3 for thorough discussion)

c

of cleavage involves more commonly the thalami, and the mesencephalon in a minority of patients [87]. The arterial tree is nearly normal, although an azygous anterior cerebral artery is usually present [61].

4.2.3 Septo-Optic Dysplasia 4.2.3.1 Background

Septo-optic dysplasia (SOD), first described by De Morsier in 1956 [89], is characterized by the association of optic nerve hypoplasia and absence or hypoplasia of the septum pellucidum. As knowledge

evolved, it gradually became clear that SOD is, in fact, a constellation of lesions that can also include pituitary-hypothalamic dysfunction, olfactory aplasia, and brain abnormalities, such as schizencephaly, cortical dysplasia, and agenesis of the corpus callosum. This apparent heterogeneity has resulted in some disagreement as to whether SOD should be regarded as a single precisely defined entity or, rather, a group of heterogeneous cases [1]. 4.2.3.2 Pathogenesis

Recent studies by Dattani et al. [90] led to the identification of a homeobox gene, Hesx 1, whose absence is believed to be related with SOD. Hesx1, which encodes

95

96

P. Tortori-Donati, A. Rossi, and R. Biancheri

a pituitary transcription factor, is first expressed during gastrulation in the mouse embryo. Hesx1 expression begins in prospective forebrain tissue but later becomes restricted to the Rathke’s pouch, the primordium of the anterior pituitary gland. Transgenic mice lacking Hesx1 exhibit a phenotype comprising variable anterior CNS defects, such as a reduced prosencephalon, abnormalities in the corpus callosum and septum pellucidum, anophthalmia or microphthalmia, defective olfactory development, and bifurcations of the Rathke’s pouch with pituitary dysplasia. A comparable and highly variable phenotype in humans is SOD. In the Dattani et al. study, two siblings with SOD were homozygous for a missense mutation within the HESX1 homeobox, suggesting a vital role for Hesx1 in forebrain and pituitary development, and hence in some cases of SOD, in humans [90]. Other authors [91] postulated that SOD could result from vascular disruption, possibly involving the proximal trunk of the anterior cerebral artery. SOD was also described in association with maternal exposure to valproic acid throughout pregnancy [92]. 4.2.3.3 Clinical Findings

The clinical picture exhibited by children with SOD is highly variable. Visual disturbances may range from blindness to nystagmus to normal vision [93]. Hypothalamic-pituitary dysfunction, mainly represented by growth deficit, is seen in approximately two thirds of patients [62, 79]. About half of patients have schizencephaly, and usually present with seizures [62, 79, 93]. Spastic motor deficits may be related to other malformations of cortical development, such as polymicrogyria. Developmental delay is also found in patients with associated cortical malformations. Anosmia is present in cases with associated abnormalities of the olfactory pathways. On the basis of the spectrum of associated abnormalities and the corresponding clinical features, SOD can be categorized into [94]: (i) isolated SOD, presenting with endocrine dysfunction, and (ii) SOD-plus, in which schizencephaly or other cortical malformations are associated; these patients usually present with seizures and/or spastic motor deficit associated with neurodevelopmental delay.

4.2.3.4 Imaging Findings Isolated SOD

Absence of the septum pellucidum is easily recognized in both axial and coronal MR images (Fig. 4.21). In the latter, the frontal horns may display a box-like squared shape that resembles that seen in mild forms of holoprosencephaly. The septum pellucidum is not confidently assessed in the sagittal plane; however, an indirect sign of its absence is a lowering of the fornix, which adjoins the corpus callosum only at the level of the inferior surface of the splenium. Moreover, the anterior columns of the fornix may be fused along the midline. The septum pellucidum is usually thoroughly absent in patients with isolated SOD [93]. One should note that septal agenesis may seldom be isolated, in which case patients usually are asymptomatic. Therefore, all patients with an apparently isolated agenesis of the septum pellucidum should be actively scrutinized for other abnormalities of the SOD spectrum. Hypoplasia of the optic nerves is readily assessed with high-resolution, thin-slice MR images, and is confirmed at ophthalmological examination. Only one nerve may be involved in some cases. Hypoplasia of the optic chiasm causes enlargement of the anterior recesses of the third ventricle, and may be associated with a small pituitary gland. White matter hypoplasia of the cerebral hemispheres may be restricted to the optic radiations or may be diffuse, in which case ventriculomegaly results. Diffuse white matter hypoplasia is characteristically associated with isolated SOD (83% of cases), whereas it typically is absent in SOD-plus [93]. SOD-Plus

Unlike isolated SOD, a remnant of the septum may be found in patients with SOD-plus [93] (Fig. 4.22). The spectrum of brain malformations that are found in SOD-plus is probably wider than was initially suspected. Schizencephaly is the most consistent among these malformations; gray-matter lined holohemispheric clefts may be unilateral or bilateral, and may be multiple in a minority of cases. Clefts may vary in size from fused to open lips. The corpus callosum may show focal thinning correlating to cleft location [93], and may rarely be thoroughly absent [95].

97

Brain Malformations Fig. 4.21a–c. Isolated septo-optic dysplasia. a Sagittal T1weighted image; b Coronal T1-weighted image; c Axial T2weighted image. There is moderate hypoplasia of the optic chiasm (arrow, a–c) and both optic nerves (arrowheads, c). Notice marked hypertelorism (c). There is agenesis of the septum pellucidum (asterisk, b). Notice that the fornix (open arrows, a) adjoins the corpus callosum at the inferior surface of the splenium, a typical indicator of septal agenesis

a

b

c

Polymicrogyria has been recently recognized as a possible component of the SOD-plus spectrum [94, 96]; it can be isolated or coexist with schizencephaly, possibly involving the contralateral hemisphere in case of unilateral clefts.

4.2.4 Kallmann Syndrome 4.2.4.1 Background

Kallmann syndrome (KS) is a form of congenital hypogonadotropic hypogonadism associated with

hyposmia or anosmia [97]. The olfactory disturbance is caused by hypoplasia or aplasia of the olfactory bulbs and tracts, whereas hypogonadism is related to deficient incretion of gonadotropin-releasing hormone (GnRH), resulting in failure of adenohypophyseal stimulation to synthesize and secrete luteinizing and folliculo-stimulating hormones. KS is more frequent in males, and can be transmitted with X-linked, autosomal recessive, or autosomal dominant modes of inheritance. Mutations of the KAL gene, located at Xp22.3, have been demonstrated to cause X-linked KS [98, 99]. The Kal protein is involved in both neuronal migration of the olfactory placode and axonal pathfinding. Moreover, KAL expression is also required for the normal development of the kidney in the first

98

P. Tortori-Donati, A. Rossi, and R. Biancheri

d

a

b

c

e

Fig. 4.22a–e. Septo-optic dysplasia plus. a Sagittal T1-weighted image; b, c Coronal STIR images; d Axial T2-weighted image, e Axial STIR image. Marked hypoplasia of the optic chiasm (white arrow, a) and both optic nerves (arrowheads, c, d), associated with agenesis of the septum pellucidum (asterisk, b, e). Also in this case, the fornix adjoins the corpus callosum at the inferior surface of the splenium (open arrows, a). There is a schizencephalic cleft in the right fronto-basal region (arrow, c). Diffuse abnormalities of cortical organization are also visible (arrowheads, b, e)

99

Brain Malformations

trimester of pregnancy, accounting for frequent concurrent renal abnormalities in X-linked KS [100]. 4.2.4.2 Pathogenesis

KS is considered a migrational abnormality involving the olfactory placode. In normal conditions [101], cells and fibers originating from the olfactory placodes in the cephalic nasal fossa migrate towards the telencephalic vesicles to form the olfactory nerves laterally, and the terminal and vomeronasal nerves medially. In response, the telencephalic vesicles are induced to form the olfactory bulbs, which represent ventricular diverticula around which the olfactory nerve fibers wrap. Thus, neuronal migration from the olfactory placode to the telencephalon is required to induce formation of the olfactory bulbs. In addition, the terminal and vomeronasal nerves form a scaffold along which GnRH-secreting neurons originating from the medial olfactory placode migrate into the forebrain to their eventual hypothalamic and septal locations. In KS, projections from the lateral olfactory placode to the brain are insufficient to induce adequate formation of the olfactory bulbs [101]. Aplasia of the olfactory bulbs results in lack of formation of the olfactory sulci over the inferomedial surface of the frontal lobes. Moreover, there is failed migration of GnRHsecreting neurons to their eventual location in the hypothalamus [102], resulting into inability of the adenohypophysis to synthesize and secrete luteinizing and folliculo-stimulating hormones.

a

Imaging Findings

b

High resolution, thin-slice coronal MRI sections covering the anterior cranial fossa (i.e., from the back of the frontal sinus to the hypothalamus) on both T1and T2-weighted images must be obtained in order to adequately display the abnormality of the olfactory system. Sagittal T1-weighted images are also mandatory to study the sella turcica. The MRI diagnosis of KS involves [101, 103, 104] (Fig. 4.23):

Fig. 4.23a,b. Kallmann syndrome. a Coronal T2-weighted image, b Sagittal T1-weighted image. Absence of the olfactory sulci (arrows, a) and bulbi (arrowheads, b) is shown on coronal planes. Global hypoplasia of the pituitary gland is recognizable on sagittal image (b). The pons is also hypoplastic

1) olfactory bulb, tract, and sulcus aplasia/hypoplasia. There is absent visualization of the olfactory bulbs associated with flattened gyri recti due to lack of formation of the corresponding sulci. In case of olfactory bulb hypoplasia, smaller than normal bulbs associated with rudimentary sulcal formation are visualized [101]. Asymmetry of the two sides is possible, and correlates with asymmetric olfactory function [101].

2) Hypothalamic-hypophyseal abnormalities. Because of the small contribution to the overall size of the hypothalamus by GnRH-releasing cells, the hypothalamus is macroscopically normal [101]. However, a small anterior pituitary lobe can result from insufficient adenohypophyseal stimulation. The posterior pituitary lobe is normal.

100

P. Tortori-Donati, A. Rossi, and R. Biancheri

4.3 Malformations of Cortical Development 4.3.1 Background Malformations of cortical development (MCD) encompass a large amount of abnormalities that result from interruption of the orderly process of generation and maturation of the cerebral cortex. Until recently, the concept of MCD basically implied that neurons failed to complete their journey from their site of origin in the periventricular germinal matrix to their final resting place in the cortical mantle, and therefore remained somewhere along their migratory route [1]; therefore, MCD were usually referred to as “disorders of neuronal migration.” However, neuronal migration is one of a host of extremely complex, partially overlapping maturational processes that can be basically grouped into three major steps, i.e., (i) neuronal/ glial proliferation, (ii) neuronal migration, and (iii) cortical organization. As a consequence, current classification of MCD basically rests on the attribution of individual malformations to malfunction of one of the above processes [7, 105]. MCD are the single category of CNS malformation whose understanding has changed more dramatically in the past decade. A substantial part of this development must be credited to MRI, which has allowed identification of MCD with increasing frequency in patients with apparently heterogeneous conditions such as epilepsy, developmental delay, cognitive dysfunction, and congenital neurologic disorders; recent evidence suggests that greater than 50% of children referred for intractable seizures to epilepsy centers have different forms of developmental pathology [105]. Moreover, biochemical and genetic research has made it possible to accumulate knowledge regarding both normal and abnormal cortical development, to identify new clinico-pathologic entities, and to elucidate specific genetic and acquired etiologies [106, 107]. A detailed description of this immense body of knowledge is beyond the scope of this paper. Here, a primer on normal cortical development will be followed by a discussion of the main entities, based on the framework of the updated classification of MCD published by Barkovich et al. in 2001 [7].

4.3.2 Normal Cortical Development At least a basic understanding of the complex events that ultimately lead to the formation of the six-layered human cerebral cortex is essential in order to understand the disorders of its development [108]. The normal cerebral neocortex is a thin ribbon whose thickness does not exceed 4.5 mm, organized into six layers [109]. The main events that characterize cortical development can be organized in three basic steps, as was previously stated. However, one should not forget that cortical development is, in fact, a continuous process. Its separation into three periods, albeit somewhat artificial, is useful to understand both normal and abnormal development. Knowledge of the factors involved in these mechanisms is far from complete. To date, over a dozen molecules that are peculiar to cortical development have been reported [110, 111]. Mutations of genes encoding some of these molecules have been demonstrated to cause MCD in humans (Table 4.3). Cell Proliferation and Apoptosis

Neurons that will eventually rest in the cortical mantle are generated in the germinal matrix along the ventricular margins of the embryonic cerebral wall. Stem cells of the germinal matrix are common progenitors to neurons and glia. Programmed cell death, or apoptosis, plays an important role in forebrain growth by modulating the production of certain stem cell populations while sparing others, i.e., modulating the number of early progenitors before the onset of neurogenesis; about 25%–50% of all generated neuroblasts are eliminated by apoptosis [112]. Stem cells in the germinal matrix undergo rapid mitosis which results in exponential increase of the founder cell population [113]. In the human neocortex at week 6 of gestation, proliferation is confined to the ventricular zone, or neuroepithelium. Cell division is symmetric, with both daughter cells re-entering mitosis. At week 7, the subventricular zone, a secondary proliferative zone, appears. It mainly gives rise to local circuit neurons and glial cells. Around week 12, the ventricular and subventricular zones are thickest, indicating that proliferation peaks at this stage [114]. Thereafter, asymmetric division becomes the predominant mode of proliferation, with one daughter cell re-entering mitosis and the other one migrating out into the cortical mantle. As neuronal proliferation proceeds, the germinal matrix progressively recedes. Eventually, it is transformed into a germinal source for ependymal cells [108]. Generation of neurons basi-

Brain Malformations Table 4.3. Genetic basis of human MCD (modified from references # 4 and 111) Gene

Locus

Encoded molecule

DISORDERS OF NEURONAL/GLIAL PROLIFERATION/APOPTOSIS TSC1 9q34 Hamartin

TSC2

16p13.3

Tuberin

DISORDERS OF NEURONAL MIGRATION (initial course) FLM1 Xq28 Filamin-1 DISORDERS OF NEURONAL MIGRATION (middle course) LIS1 17p13.3 Platelet activating factor acetylhydrolase 1bl (Pafah1bl)

Suspected role

Human mutation phenotype

Growth inhibitory (tumor suppressor) protein

•Tuberous sclerosis complex 1 •Focal cortical dysplasia, Taylor balloon cell type

Growth inhibitory (tumor suppressor) protein

Tuberous sclerosis complex 2

Actin crosslinking phosphoprotein

Bilateral periventricular nodular heterotopia

Signaling/tubulin cytoskeletal dynamics

•Miller-Dieker syndrome •Lissencephaly (isolated lissencephaly sequence) (class LISa1-4) •Lissencephaly with cerebellar hypoplasia (LCH) (class LCHa)

DCX/XLIS

Xq22.3-q23

Doublecortin

Microtubule associated protein-cytoskeletal dynamics

•Double cortex or subcortical band heterotopia •X-linked lissencephaly (class LISb1-4) •LCH (class LCHa)

FCMD

9q31

Fukutin

Putative extracellular matrix molecule

Fukuyama congenital muscular dystrophy

POMGNT1

1p34-p33

Protein 0-mannose β-1, 2-N-acetylglucosaminyltransferase

Putative extracellular matrix molecule

Muscle-eye-brain disease

POMT1

9q34.1

Protein 0-mannosyltransferase 1

Endoplasmic reticulum molecule – required for cell integrity and cell wall rigidity

Walker-Warburg syndrome

ASTN

1q25

Astrotactin

Neuronal adhesion molecule

LCH? (mouse model)

KAL1

Xp22.3

Anosmin

Extracellular matrix molecule

Kallmann syndrome

Extracellular matrix molecule

LCH (class LCHb)

DISORDERS OF NEURONAL MIGRATION (late course) RELN 7q22 Reelin DAB1

1p32-p31

Disabled

Docking protein for cAbl tyrosine kinase, intracellular signaling

LCH? (predicted, based on similarity of RELN and DAB mutant mice)

L1-NCAM

Xq28

Neural cell adhesion molecule L1

Neural recognition molecule

X-linked hydrocephalus and pachygyria

Transcription factor

Schizencephaly (rare cases)

DISORDERS OF CORTICAL ORGANIZATION EMX2 10q26.1 Empty spiracles

MALFORMATIONS OF CORTICAL DEVELOPMENT, NOT OTHERWISE SPECIFIED PEX1, 2, 3, 5, 6

7q21 (PEX1) 8q (PEX2) 6q (PEX3) 12 (PEX5) 6p (PEX6)

Peroxisomal proteins

Zellweger syndrome

101

102

P. Tortori-Donati, A. Rossi, and R. Biancheri

cally occurs in two bursts, the first from 8 to 10 weeks and the next from 12 to 14 weeks [115]. By 16 weeks, most of the neurons have been generated and have started their migration into the cortex [1]. Neuronal Migration

Migrating cell Nucleus

Neuronal migration basically consists of two different mechanisms according to the involved cell type. Excitatory glutamatergic projection neurons are generated from the germinal matrix in the ventricular zone by a process of radial migration, whereas most inhibitory gamma-aminobutyric acid (GABA)ergic interneurons (as well as oligodendrocytes) are elaborated from ventral forebrain stem cells that undergo tangential migration [116].

Leading process Radial fiber

Trailing process

Radial Migration

Radial migration is the better known and most widely studied of the two mechanisms. According to the radial unit hypothesis, the ependymal layer of the embryonic cerebral ventricle consists of proliferative units that provide a proto-map of prospective cytoarchitectonic areas [115]; the columnar organization of the cortex arises through large numbers of neuronal precursors using a “point to point” radial migration from the ventricular zone to the cortical plate [117]. Neuronal migration is initially simple, and involves cell elongation with displacement of the nucleus to the cell end farthest from the ventricular lining, followed by detachment of the cell from the ventricular surface. As the thickness of the cerebrum increases, migration becomes more complex. Subsequent waves of migrating neurons migrate to their final destination along radial glial cells that extend long, unbranched processes from the ventricular zone lining the lateral ventricle out to the cortical pial surface (Fig. 4.24). Groups of 4–10 radial glial cells form glial fascicles that provide a scaffold and supply metabolites for the migrating neurons, and organize the vertical lamination of the developing neocortical plate [118]. Regulatory mechanisms for neuronal migration probably involve [111]: (i) a “go” signal; (ii) regulation of adhesive and contractile elements producing a net cell movement when the cytoskeleton contracts; (iii) determination of direction or vector of movement; and (iv) a “stop” signal when the final destination into the cortex has been reached. Development of the neocortex starts with the establishment of a primordial plexiform layer (PPL), representing a primitive cortical organization which is shared by amphibians, reptiles, and mammals [119, 120]. The molecular layer (layer 1) and the transient

Fig. 4.24. Schematic of radial neuronal migration. Neurons migrate from the germinal matrix to their final cortical destination along radial glial cells that extend long, unbranched processes from the ventricular zone lining the lateral ventricle out to the cortical pial surface. (Modified from [1])

layer 7, or subplate, derive from the PPL. The formation of the PPL is required for subsequent formation of the cortical plate, from which the remaining layers of the neocortex derive. Specialized layer 1 neurons, called Cajal-Retzius (CR) cells, secrete reelin, a glycoprotein which attracts the migrating neurons toward layer 1 and determines their detachment from the radial glia. All migrating neurons, guided by the radial glia, must reach layer 1, establish contacts with CR cells, develop an apical dendrite, and become pyramidal cells [119, 120]. Without losing either their original contact with layer 1 or their cortical level, each neuron elongates its apical dendrite to accommodate the arrival of subsequent neurons [119, 120], so that each newly generated cohort of neurons migrates past the previously formed neurons. This accounts for the orderly “inside-out” formation of subsequent layers, beginning from layer 6 and ending with layer 2 [108]. In the meantime, proliferation occurring in the subplate results in formation of glial cells [114]; other astrocytes develop from the radial glia. Tangential Migration

Recent data on the development of the mammalian neocortex show that cells originating from the gan-

Brain Malformations

glionic eminences in the subpallium (future basal ganglia) migrate tangentially into the pallium. These tangentially migrating cells are a significant source of cortical inhibitory GABAergic interneurons, accounting for approximately 15%–30% of the total neuronal population, and possibly other cell types such as oligodendrocytes [117, 121]. Cells undergoing tangential migration must travel long distances to their sites of residence. While some attach to the existing neocortical radial glial scaffold after having entered the cortical plate [122], others probably interact with their migrating neighbors to provide a foothold for their movement [123]. Alternatively, these migrating cells may use as yet uncharacterized modes of migration to facilitate their movement [117]. In contrast to radial migration, where genetic analyses have provided a framework for the underlying molecular mechanisms, much remains unknown about the cues that guide nonradial migration, although a role for the same guidance molecules used to guide the outgrowth of neuronal growth cones is suspected [117]. Although many questions regarding tangential migration in the developing telencephalon are still unresolved, it is already apparent that this type of migration is extremely widespread, and that there probably are further pathways of migration yet to be discovered [117]. Cortical Organization

After proliferation and migration have occurred, local intracortical growth of cells and processes will determine the final appearance of the cortex [1]. The laminated cortical pattern becomes evident around 23–25 weeks of gestation, the age at which large neurons can first be seen within layer 5 [1]. Subsequent events involve growth in size of the neurons, axonal sprouting, formation of the dendrites, and development of synapses. Growth of neuron cell bodies and their processes, as well as of glia, is related to formation of the gyri and sulci. Gyrus formation is also associated with the arrival of thalamo-cortical and cortico-cortical axons into the cortex [124]. The molecular forces that underlie the invasion by axons into specific cortical layers followed by activity-dependent maturation of synapses are poorly understood. An important role in cortical target selection and ingrowth is played by subplate neurons. These are the first postmitotic neurons in the neocortex, located in the transient layer 7, or subplate. Afferent axons accumulate in the subplate before selecting their final destination in the cortical plate, and their close association to subplate neurons is crucial for this process [125].

Subplate neurons are also likely to provide important cues that aid the process by which cortical axons grow toward, select, and invade their subcortical targets [126].

4.3.3 Classification of MCD Barkovich et al. [105] are credited for conceiving the first modern classification scheme of MCD. This scheme is based upon categorization into three main groups according to the developmental step that is preferentially involved, plus a fourth group of unknown etiologies. These authors introduced an updated version in 2001 [7], whose framework will be used here to describe individual malformations (Table 4.4). Because new data are rapidly added to the already consistent body of knowledge regarding normal and abnormal cortical development, it is likely that further revisions will be introduced in the near future. Controversies in classification still exists, and will also tentatively be addressed here.

4.3.4 Malformations Due to Abnormal Neuronal/Glial Proliferation or Apoptosis

4.3.4.1 Microcephaly

Microcephaly is defined as head circumference 2 standard deviations or more below the age-matched norm [7]. At present, classification of microcephalies is, to say the least, nebulous. Differentiation from microlissencephaly (MLIS) is based on cortical thickness, which is normal to reduced in microcephaly and increased in MLIS. Microcephaly is basically classified in three categories [7], i.e., (i) primary microcephaly, or microcephalia vera (Fig. 4.25); (ii) extreme microcephaly with simplified gyral pattern (defined as occipitofrontal circumference of at least 3 standard deviations below the mean at birth) (Fig. 4.26); and (iii) extreme microcephaly associated with either polymicrogyria (Fig. 4.27) or corpus callosum agenesis and cortical dysplasia. This categorization is probably rather arbitrary, and overlap among the various entities is likely to exist [7]. Furthermore, it should also be considered that, in the event of an association with polymicrogyria, it may be difficult to assess whether microcephaly or polymicrogyria

103

104

P. Tortori-Donati, A. Rossi, and R. Biancheri Table 4.4. Classification scheme for MCD [7] I. Malformations due to abnormal neuronal/glial proliferation or apoptosis A. Decreased proliferation/Increased apoptosis: Microcephalies 1. Microcephaly with normal to thin cortex 2. Microlissencephaly (extreme microcephaly with thick cortex) 3. Microcephaly with polymicrogyria/cortical dysplasia B. Increased proliferation/decreased apoptosis (normal cell types): 1. Megalencephalies C. Abnormal proliferation (abnormal cell types): 1. Non-neoplastic a. Cortical hamartomas of tuberous sclerosis b. Cortical dysplasia with balloon cells c. Hemimegalencephaly 2. Neoplastic a. DNT (dysembryoplastic neuroepithelial tumor) b. Ganglioglioma c. Gangliocytoma II. Malformations due to abnormal neuronal migration A. Lissencephalies 1. Classical lissencephaly-subcortical band heterotopia spectrum 2. X-linked lissencephaly with agenesis of the corpus callosum 3. Lissencephaly with cerebellar hypoplasia 4. Lissencephaly, not otherwise classified B. Cobblestone complex 1. Congenital muscular dystrophy syndromes a. Walker-Warburg syndrome b. Fukuyama congenital muscular dystrophy c. Muscle-eye-brain disease 2. Isolated cobblestone complex C. Heterotopia 1. Subependymal (periventricular) 2. Subcortical (other than Band Heterotopia) 3. Marginal glioneuronal III. Malformations due to abnormal cortical organization (including late neuronal migration) A. Polymicrogyria and schizencephaly 1. Bilateral polymicrogyria syndromes 2. Schizencephaly (polymicrogyria with clefts) 3. Polymicrogyria with other brain malformations or abnormalities 4. Polymicrogyria or schizencephaly as part of multiple congenital anomaly/mental retardation syndromes B. Cortical dysplasia without balloon cells C. Microdysgenesis IV. Malformations of cortical development, not otherwise classified A. Malformations secondary to inborn errors of metabolism 1. Mitochondrial and pyruvate metabolic disorders 2. Peroxisomal disorders B. Other unclassified malformations 1. Sublobar dysplasia 2. Others

a

b

c Fig. 4.25a–c. Primary microcephaly (microcephalia vera). a. Sagittal T1-weighted image; b, c. Axial STIR images. The brain is smaller than normal, but shows normal convolutional pattern and cortical thickness. Small frontal horns and slightly receding frontal convexity indicate underdevelopment of the frontal lobes

105

Brain Malformations

are the primary event in the determination of the eventual picture. Microcephaly is believed to result from early exhaustion of the germinal matrices or abnormal apoptotic events [7]. Primitive microcephaly must be differentiated from microcephaly secondary to congenital TORCH infection [127], in which polymicrogyria also may be associated (Fig. 4.28). Clinically, patients with microcephalia vera are less severely affected than those with extreme microcephaly. The former have a normal prenatal course and present with developmental delay and mild corticospinal tract signs in infancy. Normal intelligence has been rarely reported [128]. The latter are severely hypotonic at birth and rapidly develop myoclonic seizures [129]. Early death is common in this patient group. On neuroimaging studies, microcephaly is characterized by a markedly small brain with a normal to thin, smooth cortex, surrounded by enlarged CSF spaces [128]. The gyral pattern is normal in primary microcephaly (Fig. 4.25), and simplified in extreme microcephaly (i.e., oligogyria) [129] (Fig. 4.26). The degree of cerebral gyration and sulcation can be difficult to assess; the van der Knaap score [130] may be helpful. Associated polymicrogyria (Fig. 4.27) or callosal dysgenesis can be found.

a

4.3.4.2 Microlissencephaly

The definition of MLIS [7] involves extreme microcephaly (i.e., occipitofrontal circumference of at least 3 standard deviations below the age-matched norm) associated with thick cortex. As such, cortical thickness is the main determinant in the differentiation between microcephaly and MLIS. Thus, the term MLIS is presently used less loosely than in the recent past [7, 131]. As with other microcephalies, MLIS is believed to result from early exhaustion of the germinal matrices, resulting in reduced progenitor cell proliferation, possibly associated with increased or abnormal apoptotic events. Clinically, affected patients seem to be very similar to those with extreme microcephaly belonging to the microcephaly group. The neonatal course is typically abnormal, with either severe hypotonia or hypertonia, myoclonic seizures, and possible early death. Delayed myelination can be associated in some patients [131]. A categorization of MLIS into five patient groups based on combined clinical and imaging features [131] was published before the reorganization of microcephalies and microlissencephaly into different groups [7]. Therefore, further investigations are awaited

b Fig. 4.26a,b. Extreme microcephaly with simplified gyral pattern. a Sagittal T1-weighted image; b Axial T2-weighted image. There is marked decrease in head circumference; notice the slanted frontal convexity with size preponderance of the face with respect to the calvarium (a). The gyral pattern is simplified with too few, rudimentary convolutions; however, cortical thickness is normal (b). Notice marked thinning, but not dysgenesis, of the corpus callosum (arrows, a). (Case courtesy of Dr. Z. Patay, Riyadh, Saudi Arabia)

to shed light on clinical and prognostic differences between the two entities. Clinical entities showing a microlissencephalic appearance of the brain include the Norman-Roberts syndrome, characterized by microcephaly, bitemporal hollowing, low sloping forehead, slightly prominent occiput, widely set eyes, broad and prominent nasal bridge, and severe postnatal growth deficiency. Hypertonia, hyperreflexia, seizures, and profound mental retardation are also present [132].

106

P. Tortori-Donati, A. Rossi, and R. Biancheri

a Fig. 4.28. Microcephaly secondary to congenital TORCH infection. Axial CT scan. There is diffuse bilateral polymicrogyria associated with widespread calcified foci in the deep white matter

MRI detects a markedly small brain; the gyral pattern is simplified, with few gyri and shallow sulci or even complete agyria [131, 133]. The corpus callosum may be absent, and the cortical mantle is variably thickened (Fig. 4.29). MLIS can coexist with cerebellar hypoplasia within the heterogeneous group of lissencephalies with cerebellar hypoplasia (LCH). According to current classifications, six subgroups of LCH are recognized, called a to f [134]. All these groups are characterized by microcephaly, that is extreme (greater than 3 standard deviations from norm) in LCH types c, d, and f. According to Barkovich’s definition [7], these types meet the criteria for MLIS. In these types, MLIS is associated with moderate to severe cerebellar hypoplasia and mild to moderate brainstem hypoplasia. Cortical thickness is increased. The corpus callosum is absent in LCH type f. LCH is discussed in greater detail in a following section.

b

c Fig. 4.27a–c. Extreme microcephaly with polymicrogyria. a Sagittal T1-weighted image; b Axial T2-weighted image; c Coronal STIR image. There is extreme microcephaly. The brain is small and shows diffuse perisylvian and parietal polymicrogyria (b, c). The corpus callosum is macroscopically normal; however, it apparently lies more posteriorly than normal on sagittal views (a), consistent with underdevelopment of the posterior brain as opposed to some preservation of the frontal lobes. The posterior fossa is grossly normal. The extreme degree of microcephaly differentiates this entity from syndromic bilateral polymicrogyria

4.3.4.3 Megalencephaly

Megalencephaly is defined as a brain volume which exceeds the mean by more than twice the standard deviation or is above the 98th percentile [135]. Primary (i.e., anatomic) megalencephaly results from increased proliferation of normal cell populations [7]. As such, it must be differentiated from other condi-

Brain Malformations

tions in which megalencephaly represents a secondary condition, such as metabolic megalencephalies [7, 136]. Megalencephaly affects the brain as a whole. This represents a notable difference from hemimegalencephaly (HME), in which volume increase is unilateral. Classification of anatomic megalencephaly is difficult. Familial and sporadic forms exist; in the familial forms, a benign, autosomal dominant form and a symptomatic, autosomal dominant or recessive form are identified [1]. Clinical syndromes with megalencephaly are numerous. Some will be briefly discussed here. Sotos Syndrome (Cerebral Gigantism)

This autosomal dominant syndrome is characterized by excessively rapid growth, acromegalic features, advanced bone age, and a nonprogressive cerebral disorder with mental retardation [137]. A 5q35 microdeletion, encompassing the NSD1 gene, was reported as the major cause [138, 139]. As with other overgrowth syndromes, children with Sotos syndrome are at risk for subsequent development of cancer [140]. MRI shows a large brain with normal gyration (Fig. 4.30) except for rare gray matter heterotopia. Prominence of the whole, or parts of, the ventricular system is common, as well as thinning of the posterior portions of the corpus callosum, possibly as a result of deficient or inadequate development of the posterior cerebral white matter.

a

Megalencephaly with Megalic Corpus Callosum b

Gohlich et al. [141] described three patients with head circumference above the 97th percentile at birth, who completely lacked motor and speech development and showed very little intellectual progress. Frontal bossing, low nose bridge, and large eyes resulted in a characteristic facies. MRI showed bilateral megalencephaly with a broad corpus callosum, enlarged white matter, and focally thick gray matter, resulting in a pachygyric appearance of the cortex. Sylvian opercularization was incomplete. Megalencephaly-Polymicrogyria-Hydrocephalus Syndrome

c Fig. 4.29a–c. Microlissencephaly. a Sagittal T1-weighted image; b Axial T2-weighted image; c Coronal T2-weighted image. Head circumference is markedly reduced. The cerebral cortex is thickened, with a pachygyric appearance (b, c). There is associated agenesis of the corpus callosum (a). (Case courtesy of Dr. Z. Patay, Riyadh, Saudi Arabia)

Hayashi et al. [142] described the autopsy finding of hydrocephalus, brain overweight after complete CSF removal, and bilateral frontoparietal polymicrogyria in a child with severe growth failure, frequent hypoglycemia, and cardiac malformations.

107

108

P. Tortori-Donati, A. Rossi, and R. Biancheri

lesions, usually angiomyolipomas, and cardiac rhabdomyomas. Two genetic forms of the disease exist, TSC1 and TSC2, caused by mutations in the TSC1 gene on chromosome 9 and the TSC2 gene on chromosome 16, respectively. The products of these genes, hamartin and tuberin, are growth inhibitory proteins playing a crucial role in tumor suppression [143, 144]. Brain manifestations of TSC include cortical tubers, subependymal nodules, white matter abnormalities, and subependymal giant cell astrocytomas. Histologically, cortical tubers are characterized by the presence of balloon cells, representing a common histologic feature with other malformations due to abnormal proliferation, such as Taylor’s focal cortical dysplasia and HME. Therefore, the cortical tubers of TSC are included in this category of malformation [7]. TSC is discussed in greater detail in Chap. 16. a

4.3.4.5 Focal Transmantle Dysplasia (Taylor’s Focal Cortical Dysplasia) Background

b Fig. 4.30a,b. Megalencephaly in a 12-month-old boy. a Sagittal T1-weighted image; b Axial T2-weighted image. Diffuse, homogeneous increase in brain volume. Frontal bossing is prominent, contrary to microcephaly. The cortex is normal, and myelination is normal for age

4.3.4.4 Tuberous Sclerosis Complex

Tuberous sclerosis complex (TSC), or Bourneville disease, is a dominantly inherited disease of high penetrance, characterized pathologically by the presence of hamartoma in multiple organ systems. The typical clinical triad is represented by (i) skin lesions, such as depigmented spots and adenoma sebaceum; (ii) seizures; and (iii) mental deficiency. However, this triad is not present in all patients. Many patients have renal

Focal cortical dysplasias (FCDs) display a broad spectrum of structural changes, reflecting abnormal proliferation, migration, differentiation, and apoptosis of neuronal precursors and neurons during cortical development. In the past, the term “cortical dysplasia” was used loosely by neuroradiologists [145], because it was difficult to differentiate among the various types of MCD with low-field MRI units and histologic correlation was not always available. Therefore, radiologic classification of cortical dysplasias has traditionally been problematic and largely unsatisfactory. The advent of new technology and the progressive development of epilepsy surgery have enabled researchers to attempt histologic-radiologic correlations of cortical dysplasia and to identify new subtypes [146, 147]. A subtype linked to chronic intractable epilepsy, showing distinct clinical and phenotypic features and characterized histologically by abnormally large dysplastic cells, has been called Taylor’s focal cortical dysplasia because the first description of this entity was made by Taylor et al. in 1971 [148]. Because the histologic abnormality, which is reflected in corresponding MRI features (see below) involves not only the cortex, but rather the whole thickness of the cerebral mantle (i.e., from the ventricular to the pial surface), this entity has also been called focal transmantle dysplasia (FTD) [149]. Clinically, affected patients present with intractable partial epilepsy, whose onset is usually in childhood or young adulthood. Intellectual impairment may be present [146].

Brain Malformations

Pathology

Histopathologic analysis shows a glioneuronal malformation with striking similarities to the cortical tubers of patients with TSC. The close relationship between FTD and TSC is reinforced by the presence of sequence alterations of the TSC1 gene product in both the pathologic tissue and adjacent normal cells of patients with histologically documented FTD [150]. However, patients with FTD lack additional features of neurocutaneous syndromes, especially regarding development of CNS or systemic tumor. Therefore, FTD is sometimes regarded as a forma frusta of TSC. Histologically, FTD is characterized by the association of heterotopic neurons in white matter, derangement of cortical lamination, giant neurons, and dysmorphic neurons; the possible presence of balloon cells, i.e., large cells containing a large cytoplasm volume that may express neuronal markers, glial markers, or both, identifies two subtypes of FTD (i.e., FTD with or without balloon cells) [147]. Balloon cells coexist with megalic and dysmorphic neurons within a focally enlarged cortical region that shows subverted lamination and organization (“cortical chaos”). Both cell types are also disseminated along a diffuse front throughout the subjacent white matter, resulting in loss of demarcation between gray and white matter [144]. The gray-white matter junction contains the larger proportion of balloon cells and typically is blurred. The subcortical white matter shows hypomyelination with astrogliosis. Unlike cortical tubers, calcification is exceptional [151, 152]. Cortical epileptogenicity is probably accounted for by an increased number of excitatory neurons and a decreased number of inhibitory neurons [153, 154]. Imaging Findings

Although cortical dysplasia may occur anywhere in the cerebrum, the frontal lobe is affected more frequently in case of FTD [155, 156]. This represents a notable difference from nonballoon-cell FCDs, which tend to be more common in the temporal lobes. The sensitivity of MRI in the detection of FTD is not 100% [146, 147, 155], even when an adequate technique, including highresolution volumetric techniques with thin partition size and curvilinear reconstructions, is used [147, 157, 158]. Because identification of subtle dysplastic areas with this technique is time-consuming, prior identification of the seizure focus with both clinical and electrical criteria is helpful to concentrate one’s attention to the likely lesion site [158]. Phased array surface

coils may improve lesion detection [159]. As will be described in a following section, FLAIR images are of crucial importance in order to detect the subtle signal abnormalities typical of FTD, whereas lack of delineation of the gray-white matter junction is best appreciated with high resolution T2-weighted fast inversion recovery sequences [160]. The site of the lesion often correlates with the stereo-electroencephalographic focus, although the size of the epileptogenic zone is frequently larger than the anatomic abnormality detected by MRI, probably as a result of pathologic connections or membrane hyperexcitability of normally differentiated neurons [144]. MRI detects a highly characteristic triad of signs, as follows [144, 147, 152, 156, 161, 162] (Figs. 4.31, 4.32): 1) Focal cortical thickening: Enlargement of the affected gyri can be best demonstrated on 3D surface rendering. There usually is mild to moderate focal thickening of the cortex, generally in the range of 5–8 mm. Careful assessment of cortical thickness with respect to adjacent normal cortical areas is crucial, because the malformation may be subtle. The signal from the cortex can be variably increased, mixed, or normal on T2-weighted images [156, 161], whereas it is slightly increased on FLAIR images. Spin-echo T1-weighted images show isointense cortical signal with respect to normal gray matter [156]. However, slight signal increase can be detected by means of thin-slice gradient-echo imaging. This sign has been regarded as a potentially specific feature of FTD, and could result from increased cortical cellularity [161]. 2) Blurred gray-white matter junction: The gray-white matter junction is blurred with disappearance of subcortical white matter digitations [144, 161, 163]. Loss of gray-white matter demarcation has been attributed to the presence of ectopic neurons and bizarre glial cells, dysmyelination, and a reduction in the number of myelinated fibers [163]. 3) Funnel-shaped high T2/FLAIR signal intensity in the subcortical white matter: The subcortical white matter is homogeneously hyperintense on T2weighted and FLAIR images, and slightly hypointense on T1-weighted images, possibly reflecting spongiolytic changes [162, 164, 165]. High signal intensity tapers as it extends inward from under the thickened cortical ribbon to the lateral ventricle [147, 152], reflecting involvement of the entire radial glial-neuronal unit. The distinct three-dimensional geometry of this lesion is best appreciated on FLAIR images, and is thought to represent a neuroradiological hallmark of FTD [156].

109

110

P. Tortori-Donati, A. Rossi, and R. Biancheri

Differential Diagnosis

a

b

c Fig. 4.31a–c. Focal transmantle dysplasia (Taylor’s focal cortical dysplasia). a Coronal STIR image; b and c Coronal FLAIR images. Notice gyral enlargement (arrow, a) and hyperintense signal on FLAIR images (arrow, b) that tapers towards the lateral ventricle (arrows, c)

Owing to the common histological background, differentiation of FTD from isolated cortical tubers of TSC may be problematic (Fig. 4.32). In our experience, the single most useful differential feature has been calcification, which is extremely common in tubers and very uncommon, if at all existent, in FTD. Two consequences derive from this observation, i.e., (i) calcification involves a differential diagnosis between tubers and tumor, whereas it makes FTD a very unlikely diagnosis; and (ii) CT plays a key role in the differential diagnosis of tumoral and nontumoral conditions affecting the cortex of children, and should therefore be employed as an useful adjunct to MRI in the diagnostic workup. Another point that has been advocated by some authors is that identification of an apparently single FTD should prompt a Wood light skin examination and abdominal ultrasounds in search of possible occult manifestations of TSC. Distinguishing FTD from tumors, such as dysembryoplastic neuroepithelial tumors (DNTs) or gangliogliomas, is usually not particularly problematic based on MRI findings. However, cortical dysplasia and tumor often coexist [166]. Contrast enhancement, while more common with tumor, has rarely been reported also in FTD [144, 152, 161]. In one investigation [152], statistically significant MRI findings suggesting FTD rather than tumor were frontal lobe involvement, gray matter thickening, homogeneous T2 hyperintense signal in the subcortical white matter, and ventricular extension of signal; on the other hand, temporal lobe involvement, especially when medial, significantly favored the diagnosis of tumor. Although in that study calvarial remodeling did not reach the level of statistical significance, we nevertheless believe it also strongly suggests tumor rather than cortical dysplasia. To date, no large studies have ultimately assessed the value of MR spectroscopy for differentiating FTD from tumor. Decreased NAA/Cr ratio in the affected cortex, possibly reflecting the presence of dysfunctional cells with abnormal synaptic activity and connectivity, has been described in at least three major studies [167–169]. However, the spectral findings alone are nonspecific, because NAA decrease may also be observed in low grade neoplasms. In a recent study, greater diffusion abnormalities and more marked NAA decrease were found in a patient harboring a low grade neoplasm within an area of cortical dysplasia than in one with cortical dysplasia alone [170].

111

Brain Malformations

b

a

d

c Fig. 4.32a–d. Focal transmantle dysplasia (Taylor’s focal cortical dysplasia) versus forma frusta of tuberous sclerosis. a Coronal T2-weighted image; b Axial T2-weighted image; c. Axial T1-weighted image; d Axial CT scan. There is focal cortical thickening (arrow, a, b). The involved convolution is slightly hyperintense on T1-weighted image (arrow, c), with blurred gray-white matter junction. There is corresponding slight hyperdensity on CT (arrow, d), that may indicate calcification

Classification Issues

Balloon cells are a common feature of FTD, TSC, and hemimegalencephalies. Failure of these cells to commit to or differentiate into a specific phenotype suggests they originate from abnormal stem cell differentiation [105, 152]. This assumption led Barkovich et al. to classify FTD among disorders of cell proliferation [7, 105]. Thereby, FTD was separated from nonballoon cell FCDs, which are characterized histologically by clear-cut neuronal differentiation and are separately categorized by the same authors among

disorders of cortical organization [7]. On the other hand, histopathologic studies support the view that FTD represents the severe end of a spectrum of abnormality that also includes other forms of FCD (i.e., architectural and cytoarchitectural FCDs) [146, 147, 153] (Table 4.5). The latter classification is valuable in that it correlates with epilepsy surgery outcome; in fact, an inverse correlation between histologic grade and percentage of seizure-free patients after one year of follow-up has been found [146, 156]. These consid-

112

P. Tortori-Donati, A. Rossi, and R. Biancheri Table 4.5. Histopathologic classification of focal cortical dysplasia [147] Type

Histology

Architectural FCD (microdysgenesis)

•Heterotopic neurons in white matter •Derangement of cortical lamination

Cytoarchitectural FCD

•Heterotopic neurons in white matter •Derangement of cortical lamination •Giant neurons

Taylor’s FCD without balloon cells

•Heterotopic neurons in white matter •Derangement of cortical lamination •Giant neurons •Dysmorphic neurons

Taylor’s FCD with balloon cells

•Heterotopic neurons in white matter •Derangement of cortical lamination •Giant neurons •Dysmorphic neurons •Balloon cells

FCD: focal cortical dysplasia

erations illustrate the controversies that still exist in the classification of FCD, the rapid evolution of classification schemes in the recent years, and especially the difficulties one finds when attempting to correlate histological diagnoses with MRI findings. Hopefully, disclosure of the genetic background will shed further light on the true nature of this elusive group of disorders and improve our approach to them, both from a diagnostic and a therapeutic perspective. 4.3.4.6 Hemimegalencephaly Background

HME, or unilateral megalencephaly, is an uncommon condition characterized by hamartomatous growth of one cerebral hemisphere. This condition is presently categorized among disorders of neuronal/glial proliferation and apoptosis, together with FTD and TSC, on the basis of the common presence of balloon cells in this group of disorders (see above). It has been recently suggested that HME could be a genetically programmed disorder related to cellular lineage and/or establishment of symmetry [171]. However, these genetic abnormalities have not yet been demonstrated. Affected patients present during the first days or weeks of life with medically intractable seizures; there also is severe psychomotor delay and contralateral hemiparesis [171, 172]. Management of these children is difficult. Seizure control may not be achieved by pharmacologic treatment in all cases, and adverse side effects may complicate the picture. Hemispherectomy has been advocated as a treatment method for

patients with frequent, refractory, and intractable seizures [173]. Pathology

On pathologic examination, all or part of the involved hemisphere is larger and heavier than the contralateral one. Concomitant ipsilateral enlargement of a cerebral hemisphere, half of the brainstem, and cerebellar hemisphere is called total hemimegalencephaly [174]. Histologically, the cortex shows lack of alignment in the horizontal layers and an indistinct demarcation from the underlying white matter. Giant, dysplastic neurons and balloon cells identical to those seen in FTD (see above) are scattered throughout the cortex and subcortical white matter. Striking demyelination of the centrum semiovale is seen in some cases [175]. Imaging Findings

HME occurs on either side of the brain. There is gross asymmetry with enlargement of all or part of one hemisphere and possible concomitant enlargement of the ipsilateral half of the brainstem and cerebellar hemisphere (Fig. 4.33). Midline shift can be present, either total or limited to the posterior region, a condition called the occipital sign [171]. The cerebral cortex is thickened and shows a variable spectrum of abnormalities, including lissencephaly, pachygyria, polymicrogyria, and, rarely, schizencephaly [171], probably reflecting the fact that what we call “hemimegalencephaly” probably represents a spectrum of, rather than a single, disorder. The most severe end of this spectrum has been termed hemilis-

Brain Malformations

a

b

Fig. 4.33a–c. Hemimegalencephaly. a Coronal T2-weighted image; b, c Axial T2-weighted images. There is global increase in size of the whole right cerebral hemisphere (a, b) and homolateral cerebellar hemisphere (c). Diffuse cortical dysplasia involves the frontoinsulo-temporal regions (arrowheads, a,b). There also is abnormally increased signal intensity within the white matter. The right lateral ventricle is dysmorphic, although size is not significantly different than contralaterally. Infratentorially, abnormal foliation of the right cerebellar hemisphere (arrowheads, c) is associated with dilatation of the homolateral recess of the fourth ventricle

c

a

b Fig. 4.34a,b. Hemimegalencephaly in a newborn. a Coronal T1-weighted image; b Axial T1-weighted image. General increase in size of the right cerebral hemisphere, showing cortical thickening with a lissencephalic appearance and hypermyelination of the residual white matter (arrowheads, a, b). Notice marked dilatation of the lateral ventricle. The homolateral cerebellar hemisphere is also larger than the contralateral one (a). (Case courtesy of Dr M. Rutherford, London, UK)

113

114

P. Tortori-Donati, A. Rossi, and R. Biancheri

sencephaly [176] (Fig. 4.34). The gyri are broad and flat, and the sulci are shallow. The gray-white matter junction is blurred [164]. The affected white matter shows abnormal signal intensity that varies with patient age. In neonates and small infants, these signal changes are characterized by T1 and T2 shortening, reflecting abnormally advanced myelination for the age [177] (Fig. 4.34). In older children, the white matter of the involved hemisphere becomes hyperintense on T2weighted images (Fig. 4.33). This correlates histologically with lack of myelin [164]. White matter volume is usually increased. Calcification may involve both the cortex and the white matter, and is better demonstrated by CT. There is a roughly inverse relationship between the severity of the malformation and the size of the involved hemisphere, with polymicrogyric

hemispheres appearing larger than lissencephalic ones [178]. The lateral ventricle in the affected hemisphere shows variable abnormalities, including a straightened or collapsed frontal horn, colpocephalic dilatation of the occipital horns, and global dilatation of the ventricle [171]. In some cases, only a portion of one cerebral hemisphere, usually corresponding to one lobe, is enlarged (Fig. 4.35). Serial imaging studies may show progressive atrophy of the involved hemisphere [179], so that the “megalic” hemisphere may actually become smaller than the unaffected one in the long run. A grading system for HME has been proposed based on the severity of the detected abnormalities [171] (Table 4.6). Proton MR spectroscopy shows marked decrease of NAA/Cr ratio in the involved white matter [180].

a

b

c

d Fig. 4.35a–d. Focal megalencephaly. a Axial CT scan; b Axial T1-weighted image; c Sagittal T2-weighted image; d Coronal STIR image. The left occipital lobe is diffusely slightly hyperdense on CT scan (a). The occipital horn is larger than the contralateral (a, b). The occipital lobe is slightly hyperintense on T1-weighted images (b) and markedly hypointense on T2-weighted images (c); its morphology is abnormal, with a lissencephalic appearance. On STIR images (d), the involved regions show mild signal intensity changes

Brain Malformations Table 4.6. Grading system for hemimegalencephaly [172]

Klippel-Trenaunay-Weber Syndrome (KTWS)

Grade I

KTWS is one of the hemihypertrophy syndromes. It is characterized by a constellation of anomalies that includes cutaneous capillary malformation, usually affecting one limb, abnormal development of the deep and superficial veins resulting in multiple varicosities, and limb asymmetry due to osseous and soft tissue hypertrophy [185]. Mixed vascular malformations may include capillary, venous, arterial, and lymphatic systems [186]. Affected patients are at risk for thromboembolic episodes and bleeding from multiple sites. HME is the most frequently reported brain abnormality in patients with KTWS [187]. Other reported CNS findings include brain atrophy, calcifications, enlarged choroid plexus, and leptomeningeal angiomatosis, similar to Sturge-Weber syndrome [185].

• Mild hemispheric enlargement • Straight midline/minimal displacement • Ventricular asymmetry with straightened frontal horn • Hyperintense white matter • No apparent cortical dysplasia Grade II • Moderate hemispheric enlargement • Moderate dilatation or reduction of lateral ventricle, or colpocephaly • Slight/moderate midline displacement • Moderate focal cortical dysplasia Grade III • Marked hemispheric enlargement • Marked midline displacement/occipital sign • Marked dilatation and distortion of lateral ventricle • Severe, extensive cortical dysplasia including hemilissencephaly

Associated Conditions

HME can be isolated or associated with several hemihypertrophic syndromes or phakomatoses. There are no imaging differences between isolated and syndromic HME. The main entities will be briefly discussed here; some are discussed in greater detail in Chaps. 16 and 17. Organoid Nevus Syndrome (ONS)

ONS, also called epidermal nevus syndrome, consists of sebaceous nevus of the face or scalp (nevus sebaceous of Jadassohn) associated with ipsilateral defects of the brain, eye, and connective tissue [181]. A lipoma of the cheek is usually found in association with HME. It has been suggested that ONS may be caused by mosaic mutation of a gene which would be lethal if expressed in all cells [181]. CNS abnormalities described in the setting of ONS include HME, DandyWalker malformation, and corpus callosum agenesis [182].

Proteus Syndrome (PS)

PS is named after the Greek god Proteus, “the polymorphous”, who could change his shape at will to avoid capture. PS involves partial gigantism of the hands and/or feet, pigmented nevi, hemihypertrophy, subcutaneous hamartomatous tumors and macrocephaly, and/or other skull anomalies [188]. The suggested cause is an as yet unknown dominant lethal gene surviving by mosaicism [189]. Mandatory diagnostic criteria include mosaic distribution of lesions, progressive course, and sporadic occurrence. When present, connective tissue nevi are almost pathognomonic for PS [190]. HME is the typical intracranial abnormality seen in patients with PS [191, 192]. Other CNS abnormalities include migrational disorders, Dandy-Walker malformation, callosal anomalies, and calcified subependymal nodules [191, 192]. Meningiomas, pinealomas, and optic nerve tumors have also been reported [193]. Neurofibromatosis Type 1 (NF1)

HME has rarely been reported in association with NF1 [194, 195]. The prognosis is apparently better than with isolated HME, with later onset and better therapeutic control of seizures and absence of focal neurologic signs [195].

Hypomelanosis of Ito (HI)

Tuberous Sclerosis Complex (TSC)

HI does not represent a distinct entity, but is rather a symptom of many different states of mosaicism [183]. Affected patients have unilateral or bilateral macular hypopigmented whorls, streaks, and patches with variably associated eye, musculoskeletal, and CNS abnormalities, such as HME and other migration disorders [184].

There is a close histologic relationship between TSC and HME in that balloon cells are found in both conditions. Hemispheric enlargement may be diffuse [196–198] or focal [196]. When a large calcification is found within a hemimegalencephalic cerebral hemisphere, associated TSC must be suspected [198].

115

116

P. Tortori-Donati, A. Rossi, and R. Biancheri

4.3.4.7 Neoplasms

perspective, in view of the significant differences in their neuroimaging features.

The inclusion of tumor conditions, such as DNT, gangliocytoma, and ganglioglioma, in a classification of MCD must be considered judiciously. The possible presence of dual pathology, i.e., association of cortical dysplasia with tumor, is commonly found in histological studies of resected cortical specimens, both from the temporal lobe [199] and other locations [199, 200]. While there is basically no question concerning the neoplastic nature of gangliocytomas and gangliogliomas, controversies still exist concerning DNTs. These entities are described in Chap. 10.

Genotype-Phenotype Correlations

4.3.5 Malformations Due to Abnormal Neuronal Migration 4.3.5.1 Lissencephalies The Lissencephaly-Subcortical Band Heterotopia Spectrum Background

The term “lissencephaly” refers to a smooth brain with total absence of cortical convolutions. As such, it is synonymous with agyria. Pachygyria refers to a simplified convolutional pattern with few, broadened gyri and shallow sulci. There is a spectrum of variability in the phenotypic expression of lissencephalies, accounting for the denomination “agyriapachygyria,” which is still widely used in the everyday practice. Subcortical band heterotopia (SBH), or double cortex, is characterized by a poorly organized band of arrested neurons residing beneath a relatively normal cortex. Lissencephaly and SBH were once categorized separately. However, there is etiological intersection between the two conditions, represented by the fact that subcortical band heterotopia (SBH) can be caused by less severe mutation in the same genes, LIS1 and DCX, that are responsible for classical lissencephaly (formerly known as type 1 lissencephaly) [7]. Strictly speaking, all malformations in this group are heterotopia in that there is arrested neuronal migration with failure of achievement of the cortical plate, resulting in a band of arrested neurons beneath the cortical mantle. Separation from other entities, such as subependymal and subcortical heterotopia, is justified from a neuroradiological

The LIS1 gene [201] is located on 17p13.3 and encodes a noncatalytic subunit of platelet activating factor acetylhydrolase 1bl (Pafah1b1), or Lis1 protein, which in turn regulates platelet activating factor (PAF). PAF is involved in a variety of biologic and pathologic processes; however, it is not yet proven whether LIS1 influences neuronal migration through regulation of PAF [111]. More importantly, the Lis1 protein has been implicated in the organization of microtubule dynamics, which is needed for neurite extension and nuclear translocation [202, 203]. Mouse models inactivating LIS1 display slowed neuronal migration [111]. The DCX/XLIS gene [204, 205] is on the X chromosome (Xq22.3-q23) and encodes a protein called doublecortin. DCX is expressed in migrating neurons throughout the central and peripheral nervous system during embryonic and postnatal development, and is believed to direct neuronal migration by regulating the organization and stability of microtubules [206]. Miller-Dieker Syndrome (MDS)

MDS is manifested as lissencephaly with characteristic facial abnormalities and profound psychomotor retardation [207, 208]. MDS is a contiguous gene deletion syndrome; deletion of or mutation in the LIS1 gene causes the lissencephaly, whereas facial dysmorphism and other anomalies result from deletion of additional genes distal to LIS1. Affected patients display profound psychomotor retardation, intractable epilepsy generally presenting before age 6 months, and progressive spastic paraplegia [202]. The MDS facial phenotype includes bitemporal hollowing, prominent forehead, short nose with upturned nares, prominent upper lip, thin vermilion border of the upper lip, and small jaw [209]. Isolated Lissencephaly Sequence (ILS)

ILS corresponds to lissencephaly without facial abnormalities. Mutations in both LIS1 and DCX/XLIS genes may cause ILS. Together, these account for 76% of classical lissencephaly [210]. LIS1 mutations are generally represented by point mutations or intragenic deletions; they are less extensive than those causing MDS, and do not involve contiguous genes. LIS1-related ILS has no gender predilection. DCX/XLIS-related ILS, also called X-linked lissencephaly, is typically found in hemizygote males, whereas the same mutation causes SBH in heterozygous females. Failure of lyonization

Brain Malformations

(i.e., random X-inactivation) accounts for exceptional female cases of X-linked lissencephaly [211]. The clinical manifestations of the two forms of ILS are not remarkably different, and basically comprise seizures, mental retardation, difficult feeding, and aspiration. Differentiation from MDS is easy on the basis of the facial phenotype, which only involves mild bitemporal hollowing and slightly small jaw [212]. On the other hand, there are remarkable differences in the pathological features, in that the brain malformation due to LIS1 mutations is more severe over the parietal and occipital regions, whereas DCX mutations produce more severe involvement of the frontal cortex [212]. Subcortical Band Heterotopia (SBH)

SBH is the least severe end of the malformation spectrum. It typically is found in females harboring mutations in DCX. This phenotype arises because lyonization, i.e., random inactivation of one X chromosome early in embryogenesis, results in roughly half of neurons expressing the mutated allele. SBH has also exceptionally been demonstrated in males. In these cases, it occurs as a consequence of minor missense mutations of DCX/XLIS or LIS1 [213, 214]. Affected patients are far less severely affected than those with classical lissencephaly. Although they may manifest seizures and a variable degree of cognitive impairment, about 25% possess normal or near normal intelligence [111]. There is a relationship between the degree of disability and the thickness of the heterotopic band of neurons [215]. Pathology

Lissencephaly is characterized by a smooth hemispheric surface with absent primary fissures. As such, it is synonymous with agyria. Complete lissencephaly is very rare. More often, at least a few broad, coarse gyri with shallow sulci are present, in which case the term pachygyria is more correct. The cortex is much thicker than normal, and there is a corresponding reduction of white matter volume. Microscopically, there is a four-layered cortex composed of [17, 108]: (i) a molecular layer with an excess of nerve fibers; (ii) a disorganized outer cellular layer corresponding to the would-be cortex, containing the few neurons that have completed their migration; (iii) a sparse-cell layer, formed by a tangential plexus of myelinated fibers containing few neurons; and (iv) a thick inner cell layer, composed of heterotopic neurons whose migration has been arrested. As such, this malformed cortex is remarkably similar to the normal migration pattern seen in

the 11th–13th week fetus. The ventricles are variably enlarged. In SBH, the heterotopic band consists of a superficial zone of disorganized neurons, an intermediate zone of small neurons with tentative columnar organization, and a deep zone of nodular conglomeration of the heterotopic neurons [106]. Imaging Findings

From a neuroimaging perspective, there is a continuous spectrum of variability that ranges from complete lissencephaly to localized forms of SBH. A six-tiered grading system for lissencephaly and SBH has been proposed [212] (Table 4.7). For practical purposes, only the basic pictures will be discussed here. Table 4.7. Grading system for classical lissencephaly and subcortical band heterotopia [213] 1. Diffuse agyria 2. Diffuse agyria with a few shallow sulci 3. Mixed agyria and pachygyria 4. Diffuse or partial pachygyria only 5. Mixed pachygyria and subcortical band heterotopia 6. Subcortical band heterotopia only

Lissencephaly (Agyria) (Dobyns Grade 1)

The size of the brain is smaller than normal, and the brain surface is smooth with complete lack of sulci; the sole exception is relative sparing of the orbitofrontal and anterior temporal zones. The brain has a figure-of-eight shape resulting from wide, vertically oriented Sylvian fissures [108] (Fig. 4.36). Cortical thickness is greater than 10 mm. MRI detects a three-layered pattern. The outer stripe corresponds to the molecular and outer cellular layers. The median stripe is characterized by T1 and T2 prolongation and corresponds to the sparse-cell layer [108, 216]. The deep stripe is the thickest, and corresponds to the inner layer of arrested neurons. The gray-white matter junction is linear, and the white matter is greatly reduced in volume. The lateral ventricles are enlarged. Other signs include hypoplasia of the pyramidal tracts or of the brainstem as a whole, corpus callosum dysgenesis, and eversion of the hippocampi [108, 144]. Lissencephaly associated with cerebellar hypoplasia is presently categorized separately from classical lissencephaly (see below). Agyria-Pachygyria (Dobyns Grades 2-4)

Complete lissencephaly is rare. More often, there is diffuse or localized presence of shallow sulci sub-

117

118

P. Tortori-Donati, A. Rossi, and R. Biancheri

a

b Fig. 4.36a,b. Classical lissencephaly. a Axial T2-weighted image; b Coronal T2-weighted image. On axial sections, the brain has a “figure of eight” appearance (a) due to shallow, vertically oriented Sylvian fissures, whereas on coronal planes it resembles a “chicken brain” (b). The malformation is almost complete except for minimal, rudimentary sulcation in the frontal region. The brain surface is totally smooth. Thickness of the cortex is much greater than that of white matter. The sparse cells layer, separating the thin outer cortical layer from the thick layer of arrested neurons, is clearly recognizable as a hyperintense stripe (arrowheads, a, b). Dilated perivascular spaces are visible in the periventricular regions as well as below the sparse cell layer (arrows, b)

dividing the cortical mantle into broad, coarse gyri (Figs. 4.37, 4.38). In the less severe cases, no completely agyric portions are found. There is a clearcut correspondence between the gyral pattern and the gene mutation which is responsible for the malformation. Mutations of LIS1 are associated with a posterior-to-anterior gradient of lissencephaly, with a more severe malformation in the parieto-occipital regions of the brain. Conversely, mutations of DCX/ XLIS are associated with an anterior-to-posterior gradient, with regional prevalence over the frontal regions of the brain [212]. In the LIS1 group, patients with MDS are more likely to have a high-grade picture than those with ILS [212]. SBH (Dobyns Grades 5-6)

There is a symmetric, circumferential band of heterotopic gray matter deep to the cortical mantle, separated by well-defined, smoothly marginated layers of normal-appearing white matter from both the overlying cerebral cortex and the underlying ventricle [217] (Fig. 4.39). The overlying gyral pattern can be normal, may have normal thickness with shallow sulci, or can be frankly pachygyric. Patients with pachygyria have thicker heterotopic bands [215]. SBH may coexist with other regions of frank pachygyria in the most severe cases, whereas involvement may be incomplete with sparing of either frontal or posterior regions in less severe cases, depending on the involved gene [212]. Asymmetric involvement of the two hemispheres has also been described [218].

X-Linked Lissencephaly with Abnormal Genitalia (XLAG)

XLAG is found in genotypic males with neonatalonset intractable epilepsy, hypothalamic dysfunction including deficient temperature regulation, and ambiguous or underdeveloped genitalia. Brain manifestations of this syndrome involve a consistent association of lissencephaly and agenesis of the corpus callosum [219, 220]. Additional clinical features include severe hypotonia, poor responsiveness, and early death [106, 221]. There is X-linked inheritance; the involved gene, ARX, has been mapped on Xp22.13 [222]. Carrier female relatives of affected patients have isolated, complete or partial agenesis of the corpus callosum; however, they do not harbor cortical abnormalities [221]. Histologically, there is a 3-layered cortex lacking a sparse-cell zone [221]. MRI findings [221] include a posterior-to-anterior gradient of severity consisting of posterior agyria and frontal pachygyria, associated with corpus callosum agenesis. Remarkably, this pattern of lissencephaly is the opposite of DBX/XLIS lissencephaly, and similar to that of LIS1-related lissencephaly. Cortical thickness is in the range of 6–7 mm, i.e., thinner than in classical lissencephaly. Additional findings include poor delineation of the nucleocapsular regions, ventricular dilatation, and an abnormal signal intensity in the periventricular white matter, probably related to gliosis and richness of heterotopic neurons [221].

119

Brain Malformations

a

a

b

b Fig. 4.38a,b. Focal pachygyria. a Axial STIR image; b Sagittal T1weighted image. Focal pachygyria in the fronto-rolandic regions (arrows, a), associated with vertically oriented, shallow Sylvian fissure. An associated persistent Blake’s pouch, causing tetraventricular hydrocephalus, is recognizable (b). The brainstem is hypoplastic

c

Fig. 4.37a–c. Classic lissencephaly: incomplete form. a Axial T2-weighted image; b Coronal STIR image; c Sagittal T1weighted image. The brain is totally lissencephalic in the posterior regions, whereas it is pachygyric in the fronto-temporal regions (a). In the most severely abnormal areas, there is a complete loss of arborization of the white matter, with a smooth gray-white matter interface (b). The corpus collosum is mildly dysmorphic, and thickened anteriorly (arrows, e). Perivascular spaces are dilated in the para-artrial regions (arrowheads, a). The brainstem is hypoplastic, and there is concurrent mega zisterna magna (c)

120

P. Tortori-Donati, A. Rossi, and R. Biancheri

a

b Fig. 4.39a,b. Subcortical band heterotopia: two different cases. a, b Axial STIR images. a In this case, a thick layer of heterotopic neurons (arrowheads, a), located deep to a grossly normal-appearing cortex, extends continuously along both hemispheres bilaterally. The band is separated from the cortex by a layer of myelinated white matter. Notice that the convolutions are somewhat coarse, and the gyral pattern is probably simplified. b In another case, heterotopia is incomplete, forming a thin band (arrowheads, b) that involves the hemispheres asymmetrically. (b courtesy Dr R. Devescovi, Trieste, Italy)

Lissencephaly with Cerebellar Hypoplasia (LCH)

LCHb

The association between lissencephaly and cerebellar hypoplasia has recently emerged as a separate class of malformations [134] that, for the largest part, result from different genetic abnormalities rather than classical lissencephaly. The affected cortex shows variably increased cortical thickness. The convolutional pattern varies from frank agyria to gyral simplification. In some cases, the cortical malformation may be better defined as microlissencephaly, because of associated marked brain size decrease [133]. Cerebellar involvement ranges from mild vermian hypoplasia to diffuse cerebellar underdevelopment. Six subtypes of LCH (LCH a-f) have been identified based on phenotype [134]. Causative gene mutations have been identified for two of them, whereas they remain undetermined for the remaining four.

LCHb is associated with mutations of the RELN gene, located on 7q22 [223]. RELN encodes reelin, a protein implicated in the process of detachment of migrating neurons from the radial glia, which terminates neuronal migration. The LCHb phenotype is characterized by pachygyria with mild cortical thickening and anterior predominance, hippocampal abnormalities, and severe cerebellar hypoplasia with absent foliation.

LCHa

LCHd

LCHa is, for all purposes, a subtype of classical lissencephaly characterized by associated midline cerebellar hypoplasia [134]. The genetic background is mutation of either LIS1 or DCX/XLIS. Accordingly, the severity of the cortical malformation follows a posterior-toanterior or an anterior-to-posterior gradient, respectively. Cortical thickness is in the 10–20 mm range.

LCHd is characterized by variably severe lissencephaly with microcephaly (i.e., microlissencephaly) associated with moderate to severe hypoplasia of both the hemispheres and the vermis. Cerebellar foliation, albeit simplified, is preserved. Cortical thickening is in the 10–20 mm range.

LCHc

Microlissencephaly with associated severe cerebellar hypoplasia and cleft palate has been described in a family [224]. Histologically, there is a thickened, unlayered cortex. MRI studies of this entity have not been published yet.

Brain Malformations

LCHe

In LCHe, there is a distinctive cortical pattern represented by an abrupt transition from frontal agyria with 10–20 mm cortical thickening to a simplified gyral pattern with near normal cortical thickness posteriorly [134]. Cerebellar hypoplasia predominately involves the vermis. LCHf

This entity displays a characteristic association of agyria-pachygyria, cerebellar hypoplasia with greater involvement of the vermis than the hemispheres, and agenesis of the corpus callosum. There is mild brainstem hypoplasia.

4.3.5.2 The Cobblestone Complex Background

Abnormalities belonging to this group were once considered to be part of the lissencephaly spectrum [105]. The typical cortical malformation was once called lissencephaly type II [225]. However, the surface of the brain is not always smooth in this group of pathologies; rather, it usually appears irregularly bumpy and knobbed. Therefore, the descriptive term “cobblestone complex” was introduced [7]. Separation from lissencephalies is also justified in view of the different pathogenetic mechanism, involving neuronal migration beyond layer 1 into the leptomeninges through gaps in the glial limiting membrane [7, 111], and the characteristically disorganized, unlayered appearance of the cortex on histopathologic examination. Imaging Findings

On imaging studies, the cobblestone complex is characterized by cortical thickening with an irregular, pebbled external surface comprising an irregular mixture of agyria, pachygyria, and polymicrogyria, with severely reduced sulcation. The white matter is abnormal due to dysmyelination or edema, resulting in a bright signal on T2-weighted MR images. The ventricles can be enlarged either because of dysmorphic ventriculomegaly (i.e., colpocephaly) or true hydrocephalus. The posterior fossa is consistently abnormal in the cobblestone complex. The brainstem can be hypoplastic and deformed. The cerebellum is also hypoplastic, usually with a greater degree of involvement of the vermis than the

hemispheres, and/or dysplastic (cerebellar polymicrogyria) [226]. Classification

The cobblestone complex is typically found in patients with congenital muscular dystrophy (CMD). CMDs are a heterogeneous group of disorders characterized by hypotonia of prenatal onset, weakness, and frequent congenital contractures, associated with histological findings of muscular dystrophy [227, 228]. CMDs may show isolated muscular involvement or be associated with brain and ocular abnormalities. Three conditions share the combination of CMD and brain abnormalities: Walker-Warburg syndrome (WWS), Fukuyama congenital muscular dystrophy (FCMD), and muscleeye-brain disease (MEB). Ocular abnormalities are typically associated in WWS and in MEB, whereas they are distinctly uncommon in FCMD. Recently, the cobblestone complex has been reported as an isolated finding in patients without muscle or ocular involvement [229]. These entities must be differentiated from merosin-deficient CMD, in which brain involvement is restricted to the white matter without cortical abnormalities [230]. There are striking similarities between these entities both on pathological and imaging studies. From a broad perspective, one can conceive a spectrum of severity ranging from the more severe WWS to the milder MEB, with FCMD showing intermediate severity and partially overlapping features with both WWS and MEB. MRI can play an important role in the differentiation among these entities, but genetic studies are required for an ultimate diagnosis. A brief discussion of these entities follows. Walker-Warburg Syndrome (WWS)

WWS is a lethal autosomal recessive developmental disorder characterized by CMD and complex brain and eye abnormalities, previously known as HARD±E (hydrocephalus, agyria, retinal dysplasia, with or without encephalocele). WWS is caused by mutation in the gene encoding protein O-mannosyltransferase (POMT1), located on 9q34.1 [231]. WWS is the most severe entity among CMD with coexistent CNS involvement, with a life span that is usually not longer than a few months. Rare survivors display profound mental retardation, intractable seizures, failure to thrive, and hypotonia [232]. In WWS, the cobblestone pattern prevailingly comprises diffuse agyria blended with bumpy-surfaced pachygyria (Fig. 4.40). Cortical thickening is in the order of 7–10 mm, i.e., thinner than in clas-

121

122

P. Tortori-Donati, A. Rossi, and R. Biancheri

sical lissencephaly. However, the cortex may be secondarily thinned when hydrocephalus is present. Irregular subcortical heterotopic foci are often present. Obliteration of the interhemispheric fissure, often detected on MRI, results from leptomeningeal thickening due to neuronal overmigration [232]. Diffuse T1 and T2 prolongation in the white matter is a consistent feature of WWS, and correlates histologically with hypomyelination and edema. There can be associated callosal dysgenesis [228]. Hydrocephalus, albeit frequent, is not a necessary diagnostic criterion; when present, it is caused either by fusion of the cortical surfaces with the overlying leptomeninges, as a result of neuronal overmigra-

tion causing obstruction to CSF circulation [233], or by concurrent aqueductal stenosis [230]. Cerebellar hypoplasia is usually greater in the vermis than in the hemispheres. A full-blown Dandy-Walker malformation is found in 50% of cases [111, 232]. The cerebellar cortex is dysplastic. An occipital cephalocele is associated in 25%–50% of cases [232]. The brainstem is hypoplastic, with a striking size reduction of the pons. The brainstem also shows a typical posterior kink, resulting in a broad inverted S shape on sagittal MR images [228]. Microphthalmia, cataracts, congenital glaucoma, persistent hyperplastic primitive vitreous, retinal detachment, and optic nerve atrophy are seen [228,

Fig. 4.40a–c. Walker-Warburg syndrome. a. Sagittal T1-weighted image; b, c. Axial T2-weighted images. There is marked supratentorial hydrocephalus associated with corpus callosum agenesis. Notice Dandy-Walker malformation (c), with rotation of the hypoplastic vermis (arrow, a). The brainstem is dysmorphic with an italic S appearance (arrowheads, a). The quadrigeminal plate is hypertrophic with aqueductal stenosis. The cerebral cortex is flat and markedly thinned, with shallow, verticalized Sylvian fissures (arrow, b). Multiple subependymal neuronal heterotopic nodules are visible (arrowheads, b, c). There is left microphthalmos with persistent hyperplastic primitive vitreous (c). (Case courtesy Dr. F. Triulzi, Milan, Italy)

a

b

c

Brain Malformations

232]. These abnormalities are readily depicted by MRI. Retinal malformations are universally present [232]. Fukuyama Congenital Muscular Dystrophy (FCMD)

FCMD has been reported almost exclusively in the Japanese. Although a diagnosis of FCMD in non-Japanese patients still requires caution [230, 234], there is increasing evidence that the distribution of FCMD may be more widespread. FCMD is an autosomal recessive entity caused by mutation of a gene located on 9q31 [235] and encoding fukutin, a putative extracellular matrix molecule. FCMD differs from WWS and MEB basically because eye abnormalities are usually absent [236]. Affected children show progressive facial and limb weakness, congenital and progressive joint contractures, and markedly elevated serum creatine kinase [237]. Clinical presentation is usually in the neonatal period with hypotonia and diminished tendon reflexes. The clinical course is less severe than in WWS, with most patients surviving beyond infancy and remaining relatively stable afterwards. Brain MRI shows two basic patterns of cortical involvement (Fig. 4.41). A cobblestone appearance prevails in the temporal and occipital lobes, whereas MRI imaging characteristics consistent with polymicrogyria are preferentially found in the frontal and parietal lobes [227, 228]. This peculiar distribution may assist in the differentiation from WWS or MEB. In the vast majority of cases, there is either diffuse or patchy T1 and T2 prolongation in the central white matter of the cerebral hemispheres that may lessen with age, possibly in relation to an extreme form of delayed myelination [227]. The cerebellum is usually not severely hypoplastic; however, there is disorganized cerebellar foliation corresponding to cerebellar polymicrogyria, with associated small, multiple subcortical “cysts” that are isointense with CSF in all sequences (Fig. 4.41). These “cysts” are believed to result from entrapment of the subarachnoid space from superficial fusion of the dysplastic cortex and leptomeninges in the boundary between the normal and polymicrogyric cortex [238]. The brainstem may be normal; however, a variable degree of pontine hypoplasia with flattened anterior surface, similar to MEB, has been described in some cases [227, 228] (Fig. 4.41). Muscle-Eye-Brain Disease (MEB)

MEB is the least severe entity in the cobblestone complex spectrum. This autosomal recessive entity is caused by mutations in the O-mannose β-1, 2-Nacetylglucosaminyltransferase gene (POMGNT1), located on 1p34-p33 [239]. The protein encoded by

POMGNT1 participates in O-mannosyl glycan synthesis. As such, MEB is closely related to WWS, although it is less severe from both clinical and neuroimaging standpoints. MEB is distinctly frequent in Finland [240]. Affected patients present with congenital hypotonia, muscle weakness, and elevated serum creatine kinase. Ophthalmologic findings include severe visual failure, uncontrolled eye movements associated with severe myopia, and progressive retinal degeneration associated with giant visual evoked potentials [240]. Most patients have severe psychomotor delay. On the whole, pathologic and imaging findings are less severe than in WWS regarding both the severity and extent of cerebral cortical dysplasia and the extent of white matter abnormalities [230]. MRI shows a less severe cobblestone pattern than in the other entities, with remarkable differences of severity and extent in individual cases (Fig. 4.42). The appearance of the cortex is prevailingly polymicrogyric with a variable pachygyric appearance [230]. Complete agyria is not encountered [241]. Cortical involvement prevails in the frontal, temporal, and parietal lobes, and thickness of the cortical mantle is normal or slightly increased [241], whereas the posterior regions of the brain usually show a normal cortical pattern. Unlike WWS, the gray-white matter interface shows interdigitations. The white matter is either normal or shows patchy areas of T1 and T2 prolongation, lacking the diffuse involvement found in WWS [240]. When present, ventricular dilatation is basically ex-vacuo, not hydrocephalic [230], and may be associated with absence of the septum pellucidum. The corpus callosum is thinned or dysplastic. Typically, there is cerebellar polymicrogyria with multiple subcortical cysts [241], similar to those found in FCMD and, rarely, in WWS. Isolated vermian hypoplasia is also present. There is a striking hypoplasia of the pons, with a flat brainstem in sagittal images due to absence of the anterior pontine bulge [241, 242] (Fig. 4.42). Isolated Cobblestone Complex (ICC)

ICC is believed to be an allelic variant of WWS and MEB which does not fulfill the diagnostic criteria for these syndromes [229]. ICC is characterized by a cobblestone complex with normal eyes and muscles. Affected patients have moderate to severe psychomotor delay. MRI shows generalized pachygyria with a cobblestone surface, flattened gray-white matter interface, and ventriculomegaly. The white matter shows scattered areas of bright signal on T2-weighted images. The brainstem and cerebellum are hypoplastic, and there may be an associated retrocerebellar cyst [229].

123

124

P. Tortori-Donati, A. Rossi, and R. Biancheri

a

c

b

c

d

e

Fig. 4.41a–e. Fukuyama congenital muscular dystrophy. a,c,e Axial T2-weighted images; b Coronal T2-weighted image; d Sagittal T1-weighted image. Diffuse, bilateral frontal cortical dysplasia without signal changes of the underlying white matter (a, b). There is marked ventriculomegaly with absence of the septum pellucidum (asterisk, b) and fusion of the anterior pillars of the fornix (arrow, b). The pons is hypoplastic (arrows, d), there is mega cisterna magna (d), and diffuse subcortical cerebellar cysts, indicating cerebellar polymicrogyria (arrowheads, c). There also is right microphthalmos with persistent hyperplastic primitive vitreous (arrowhead, e)

Brain Malformations

a

b

c

d Fig. 4.42a–d. Muscle-eye-brain disease. a and c Axial T2-weighted images; b Sagittal T1-weighted image; d Axial STIR image. There is diffuse cortical dysplasia involving both frontal lobes (a); with coarse interdigitation between the thickened cortex and the underlying white matter, which is diffusely T2 hyperintense (a). Hyperintensity involves also the paratrigonal regions bilaterally. On sagittal images, a Dandy-Walker variant is associated with hypoplasia of the pons (arrows, b). Diffuse cortico-subcortical cerebellar microcysts, indicating polymicrogyria (arrowheads, c), result in a spongy appearance of the cerebellar surface (arrowheads, d). Left buphthalmos (asterisk, d)

125

126

P. Tortori-Donati, A. Rossi, and R. Biancheri

4.3.5.3 Heterotopia

of FLN1 [249]. These observations indicate that several syndromes with BPNH exist [247].

The term heterotopia refers to normal cells occupying an abnormal position within their organ of origin [4]. In the CNS, heterotopia implies arrested migration of normal neurons along the radial glial path between the ependyma of the lateral ventricles and the cortex. Heterotopia are best classified according to location into subependymal and subcortical heterotopia. A third group, marginal glioneuronal heterotopia, implies overmigration of neurons and glial cells into the leptomeninges through gaps in the pial-glial border. Because marginal glioneuronal heterotopia are usually detected only microscopically [144], these will not be further dealt with. SBH, or double cortex, was once classified together with other heterotopia, but has been recently reclassified to the lissencephaly category based on a common genetic background [7] (see above, lissencephalies). On imaging studies, heterotopia are isointense with normal gray matter on all MRI sequences and lack enhancement on postcontrast scans [108].

Pathology

Subependymal Heterotopia Background

Subependymal heterotopia are the most common form of heterotopia [243]. They comprise subependymal nodules of gray matter that can be unilateral or bilateral, localized or diffuse. While unilateral and localized, heterotopia are usually sporadic and show no sex predilection. Bilateral periventricular nodular heterotopia (BPNH) is an X-linked dominant disorder with marked predominance in females due to near total male lethality during early embryogenesis [244]. The affected gene in BPNH, FLN1, is located on Xq28 [245] and encodes filamin-1, an actin-binding protein that promotes cell motility [245, 246]. Filamin-1 also plays a role in vascular development and blood coagulation, which may account for male lethality during early embryogenesis [245]. In females, random inactivation of one chromosome X, or lyonization, results in a milder phenotype that is compatible with life. Affected patients are often grouped into pedigrees with multiple miscarriages [106]. Heterozygous females usually have normal intelligence to borderline mental retardation, epilepsy of variable severity, and coagulopathy [244]. Presentation is usually in the second to third decade, accounting for the distinct rarity of BPNH in the pediatric age group. Recently, BPNH has been observed in unusual settings, such as in male patients with mental retardation [247, 248]. Some of these cases are probably explained by partial loss-of-function mutations

Pathologically, subependymal nodular collections of well differentiated neurons, often bulging into the ventricular cavity, are found. Neurons show random orientation and size distribution, and may assemble into laminar patterns [250]. In BPNH, these neurons represent those cells expressing the X chromosome that carries the mutation, and are located symmetrically along the ventricular body. Imaging Features

MRI shows nodular, subependymal masses that are isointense with gray matter in all sequences and do not enhance with gadolinium administration [251]. These nodules are typically located along the lateral wall of the lateral ventricles. Their number and distribution is variable, ranging from one (Fig. 4.43) or few scattered, prevailingly peritrigonal nodules in sporadic cases to bilateral near-continuous heterotopic nodules in FLN1related BPNH (Fig. 4.44) [109, 144, 252]. The surrounding white matter shows normal signal intensity, and the overall appearance of the cerebral hemispheres is not significantly distorted, with a normal thickness and gyral configuration of the cortical mantle. We have also seen a case of “laminar” subependymal heterotopia, in which the heterotopia has a flat, rather than bumpy, appearance, and was associated with radially oriented transmantle heterotopia (Fig. 4.45). The main differential diagnosis is with the subependymal nodules of TSC. However, the latter usually have an elongated shape, are not isointense with gray matter due to calcification, and may enhance with gadolinium [108]; moreover, other brain manifestations of TSC are typically present. Subcortical Heterotopia

Subcortical heterotopia are irregular lobulated masses of gray matter located in the subcortical white matter of the cerebral hemispheres [250]. Unlike SBH, subcortical heterotopia are essentially always contiguous to the overlying cortex and the underlying ventricular surface, i.e., without intervening well-defined white matter layers [253]. Most reported cases have been males [254], suggesting an as yet unproven X-linked mechanism of inheritance, as is the case with several other migrational disorders. The etiology of subcortical heterotopia is unknown, but it could be related to a secondary, heterotopic germinal zone abnormally

127

Brain Malformations

b Fig. 4.43a,b. Isolated subependymal heterotopia. a Axial T2weighted image; b Axial STIR-weighted image. Small heterotopic nodule in the lateral wall of the left atrium (arrow, a, b) is isointense with gray matter

a

a

b Fig. 4.44a,b. Diffuse subependymal nodular heterotopia. a Axial T1-weighted image; b Axial T2-weighted image. Numerous, diffuse subependymal heterotopic nodules (arrowheads, a, b) are evident. A further small heterotopic nodule is recognizable in the left frontal region (arrow, a, b). All heterotopia are isointense with gray matter on both T1- and T2-weighted images. (Case courtesy Dr. C. Hoffmann, Tel-Hashomer, Israel)

located deep in the cerebral hemisphere, as shown in some rat models [67, 253]. Affected patients have hemiparesis in the contralateral body half, developmental delay, normal or near-normal intelligence, and simple partial motor seizures presenting any time between birth and the second decade [254]. MRI shows irregularly lobulated conglomerations of gray matter extending from the ventricular surface into the hemispheric white matter and, sometimes, to

the cerebral cortex (Fig. 4.46) [251, 253, 254]. The size of the heterotopia is variable. Some cases are relatively localized and are better defined as nodular, whereas in other cases there is a curvilinear appearance that extends to a large part of, or the whole, cerebral hemisphere [253]. The cortex overlying the heterotopia is thin with abnormally shallow sulci. Sometimes the cortex appears continuous with the heterotopia, which may contain engulfed vessels and CSF [251]. In these

128

P. Tortori-Donati, A. Rossi, and R. Biancheri

a

b Fig. 4.45a,b. Laminar subependymal heterotopia. a Sagittal T2-weighted image; b Axial T2-weighted image. There is diffuse laminar arrangement of neurons in the right periventricular occipital region (arrowheads, a, b). The lateral ventricle is sectorially dilated. Radial stripes of transmantle heterotopia (arrows, a, b) are recognizable. The corresponding cortical mantle is somewhat thickened and flat. This patient also had persistent hyperplastic primitive vitreous in his right eye

a

b Fig. 4.46a,b. Subcortical heterotopia in a newborn with Aicardi syndrome. a Axial T2-weighted image; b Sagittal T1-weighted image. This newborn has gross heterotopic islets in the left frontal subcortical region (thin arrows, a), surrounded by abnormally myelinated white matter. In the corresponding frontal areas, the cortical mantle appears abnormally flattened and perhaps thinned. Subcortical heterotopia is also visible in the contralateral frontal lobe (arrowhead, a). Subependymal heterotopia is also visible along the right atrium (open arrow, a). Colpocephaly and agenesis of the corpus callosum are also present. Notice multiple associated arachnoid cysts (asterisks, a, b). Congenital pharyngeal tumor (T, b) caused demise a few months after this examination; histology remained unascertained

cases, subcortical heterotopia may resemble cortical infoldings. There is a high prevalence of associated corpus callosum dysgenesis, probably resulting from failure of the heterotopic neurons to extend normal interhemispheric commissural axons [67, 243]; some cases are actually consistent with the definition of Aicardi syndrome (Fig. 4.46) (see Chap. 3). The basal

ganglia homolateral to the heterotopia are small and dysplastic, and the white matter is reduced, probably as a result of a reduction of intrahemispheric association axons [67]. As a consequence, the affected hemisphere is smaller than the contralateral one [251, 254]. In rare instances, giant heterotopic nodules may replace a whole cerebral hemisphere (Fig. 4.47) [255].

Brain Malformations

a

b

d

c Fig. 4.47a–d. Giant subcortical heterotopia. Two different cases. Case #1 (a, b): a Coronal T1-weighted image; b Sagittal T1-weighted image. There is a huge heterotopic nodule in the right temporo-parieto-occipital region (asterisk, a, b). The frontal cortex anterior to the heterotopia is also dysplastic. A second subependymal heterotopic nodule is recognizable al the level of the temporal horn (arrowhead, b). (Case courtesy of Dr. G. Fariello, Rome, Italy.) Case #2 (c, d): c Axial T2-weighted image; d Coronal T1-weighted image. There is a severe abnormality of the migrational process, resulting in subverted organization of the right hemisphere. Individual lobes are no longer recognizable. The whole width of the hemisphere is occupied by a single-nodule giant subcortical heterotopia. CSF spaces are trapped within the lesion (arrow, c), suggesting the lesion may be at least partly explained as a large cortical infolding. Only on coronal images, a thin white matter layer separating the heterotopia from the normal cortex is recognizable (arrow, d). (Case courtesy of Prof. G. Scotti and Dr. A. Falini, Milan, Italy)

129

130

P. Tortori-Donati, A. Rossi, and R. Biancheri

Transmantle Heterotopia

This descriptive term, not to be confused with focal transmantle dysplasia (see above), indicates a column of heterotopic gray matter that extends through the full thickness of the cerebral mantle, i.e., from the subependymal to the pial surfaces (Fig. 4.48). Careful analysis must be performed to differentiate this abnormality from fused lips schizencephaly; absence of a ventricular dimple is an useful clue. Furthermore, the abnormality is differentiated from a polymicrogyric cortical infolding because the superficial sulcation, while usually abnormal, is shallow.

4.3.6 Malformations Due to Abnormal Cortical Organization 4.3.6.1 Polymicrogyria Background

Polymicrogyria (PMG) refers to an abnormal cortical pattern characterized by an excess of small, irregular convolutions without intervening sulci, with obliterated sulci due to fusion of the molecular layers, or with intervening shallow sulci [108, 256]. PMG is thought to result from events occurring during late neuronal migration and early cortical organization [7]. Causes of PMG involve both genetic and acquired events. Several bilateral PMG syndromes have been described in which either X-linked or autosomal modes of inheritance are contemplated [257]. In some cases, the causal genetic mutations have been identified [258, 259]. Environmental causes of PMG include prenatal cytomegalovirus infection [260] and in utero ischemic events [1, 261]. Remarkably, cytomegalovirus-related brain damage is caused by cerebral ischemia through angiitis or transient systemic perfusion failure, rather than by a direct cytopathic effect [262]. Pathology

a

PMG can be unilateral or bilateral. Owing to the common pathogenesis, PMG can be associated with schizencephaly (see below). The cortex is characterized by a flat external surface that results from adhesion of multiple small adjacent gyri, often with complete fusion of the superficial layers and obliteration of the intervening sulci. As a result, the cortical mantle appears macroscopically thicker, even though microscopically there is, in fact, cortical thinning, owing to reduction of the neuronal population. Histologically, two major forms of PMG exist, i.e., unlayered PMG and four-layered PMG. However, because the two forms cannot be distinguished by means of imaging, this categorization is not particularly useful for neuroradiologists.

b Fig. 4.48a,b. Transmantle heterotopia. a Axial STIR-weighted image; b Coronal STIR image. Cortical dysplasia involves the whole thickness of the right cerebral hemisphere in the temporo-parietal lobes (black arrowheads, a, b), extending from the pia to the ependymal lining. On coronal image, association with diffuse cortical dysplasia of the mesial aspect of the hemisphere (white arrowhead, b) is clearly recognizable. Notice there is neither cleft nor ventricular dimple, thereby ruling out schizencephaly

Unlayered PMG is characterized by a disorganized cortex with absent lamination, and results from earlier, genetically determined or exogenous events [10th–18th gestational week]. Four-layered PMG arises later [13th–24th gestational week] and is basically associated with perfusion fail-

Brain Malformations

ure causing laminar cortical necrosis that preferentially involves cortical layer 5 [202, 256, 263, 264]. Clinical Findings

Clinical presentation varies significantly among patients, depending on the extent and location of cortical involvement as well as on associated abnormalities. The spectrum of clinical manifestations ranges from severe encephalopathy with intractable epilepsy to near-normal patients with limited cognitive impairment [202]. Definite clinical syndromes with PMG have been described, as detailed below. Regardless of the morphologic abnormality, involvement of the central sulcus, operculum, or sylvian fissure is more likely to be associated with neurologic signs than abnormalities in other cortical regions [109]. Imaging Findings

PMG can be uni- or bilateral, localized or diffuse, and results in an irregular contour of the hemispheric surface. The cortex has an irregular appearance with multiple, tiny gyri lining a folded region. The flattened or bumpy appearance of the cortical surface results from the individual gyri merging with one another as a result of fusion of the superficial layers. This results in an apparent thickening of the cortical ribbon, simulating pachygyria on thick MRI sections [265]. The gray-white matter interface is typically corrugated [256], reflecting the multiple subcortical white matter interdigitations in the depth of the microgyri. The extent of the cortical abnormality is variable in individual cases. Bilateral, symmetric PMG is typically found in genetic syndromes, and the perisylvian regions are most commonly involved (see below). In these cases, there is failure of sylvian opercularization with abnormally shallow, verticalized sylvian fissures that extend posteriorly into the parietal regions. Unilateral PMG (Fig. 4.49) can be variably severe. The involved hemisphere is usually smaller than the contralateral one, and there is concurrent dysmorphic ventriculomegaly. Localized polymicrogyric areas may involve the cortex surrounding abnormally deep, large sulci, forming cortical infoldings that may reach the ependymal surface (Fig. 4.50). Volumetric MRI techniques with thin partition size and curvilinear and surface reconstructions are especially useful to fully depict the cortical abnormality. CT plays an important role in detecting associated cerebral calcification, suggesting underlying congenital cytomegalovirus infection. Prominent vascularity is typically detected in the enlarged sub-

arachnoid space superficial to the dysplastic cortex, and represents abnormal venous drainage related to persistence of tributaries of the embryonic dural plexus [256]. Syndromes with PMG

Unilateral PMG

The severity of the clinical picture correlates with the extent of the abnormal cortex, its location, and the presence of associated anomalies. The affected hemisphere is invariably smaller than the contralateral one. The regions of the brain around the sylvian fissure are the most common location (Fig. 4.49). Affected patients usually present with congenital hemiparesis or hemiplegia that involves the contralateral body. Seizures, often complex partial, developmental delay, and neurologic signs are other well-known presenting signs [244, 266]. The contralateral hemisphere can be normal or show associated abnormalities, most notably schizencephaly. Although most cases of unilateral PMG have been sporadic, a genetic basis has also been suspected (A.J. Barkovich, personal communication). Bilateral PMG

Bilateral, symmetric hemispheric involvement of the cerebral hemispheres with PMG has been increasingly recognized in the past few years [257]. Distribution of PMG regionally involves discrete, nonoverlapping brain portions. Bilateral perisylvian PMG is the most common of these syndromes. In this disorder, PMG involves the perisylvian regions with variable extent among patients [257, 267]; typically, the sylvian fissures have an abnormally verticalized orientation and extend far more posteriorly than normal into the parietal regions. Affected patients display a typical clinical syndrome consisting of facial, lingual, and masticatory diplegia (developmental Foix-ChavanyMarie syndrome), associated with mental retardation and seizures [268-270]. Bilateral perisylvian PMG is probably genetically heterogeneous [271]. However, X-linked inheritance is predominant [272]. A locus for bilateral perisylvian PMG has been mapped to Xq28 [259]. Bilateral parasagittal parieto-occipital PMG. The cortical abnormality is centered around the mesial parieto-occipital regions bilaterally (Fig. 4.51). Affected patients basically present with seizures, often complex partial, and significant developmental delay. Age

131

132

P. Tortori-Donati, A. Rossi, and R. Biancheri

a

b Fig. 4.49a,b. Unilateral polymicrogyria. Two different patients, both presenting with congenital hemiplegia involving the contralateral body half. Case #1: a Axial STIR image. The right hemisphere is smaller than the contralateral, unaffected hemisphere. The cortical mantle is apparently thickened, resulting from fusion of multiple adjacent small gyri. Such condition becomes apparent when one considers the corrugated gray-white matter interface, resulting from subcortical white matter interdigitations within the core of the microgyri. The sylvian fissure is shallow, verticalized, and somewhat displaced posteriorly from its normal location; it contains a few large veins (arrow, a). Case #2: b Sagittal STIR image. Diffuse polymicrogyria involves the fronto-temporo-insular regions. Note the smooth outer cortical surface, whereas the gray-white matter interface is irregular due to multiple fine interdigitations resulting from tightly packed abnormal convolutions. The malformative condition has caused defective sylvian opercularization. Notice large veins within the sylvian fissure (arrow, b). Case courtesy of Dr. C. Capellini, La Spezia, Italy

a

b Fig. 4.50a,b. Cortical infoldings. a Axial STIR image; b Coronal STIR image. a cortical infolding is recognizable in the anterior left frontal region (white arrowheads, a). The infolding reaches the surface of the lateral ventricle (black arrowhead, a). Coronal image shows a large infolding of thickened cortex (black arrows, b), associated with subcortical heterotopia (white arrow, b). It is difficult to determine the nature of the dysplasia involving the cortex, that appears thickened and smoothly marginated

Brain Malformations

a

b

c

d

e

Fig. 4.51a–e. Diffuse polymicrogyria. Case #1: a Axial STIR image, b Sagittal STIR image. Bilateral parieto-rolandic polymicrogyria. Both the inner and outer surfaces are irregular, although the outer is more gross. Adjacent subarachnoid spaces and ventricular portions close to the malformation are dilated. Case #2: c Axial T1-weighted image; d, e. Axial T2-weighted images. There is absent opercularization of both sylvian fissures. Polymicrogyria involves the perisylvian regions, extending cranially to both the frontal and parietal lobes

at presentation is highly variable. Although reported cases have been sporadic, a genetic basis is suspected [273]. Bilateral posterior parietal PMG. The abnormality involves the lateral portions of the parietal lobes over the convexity. Affected patients display minor speech difficulties; seizures are not a typical manifestation in this patient group [274]. Bilateral frontal PMG. There is involvement of the frontal lobes, from the frontal poles anteriorly to the precentral sulcus posteriorly and the frontal operculus inferiorly. Affected patients present in early childhood with delayed development and mild spastic quadriparesis, whereas seizures usually appear later [263]. Possible autosomal recessive inheritance has been suggested [263].

Diffuse PMG

Not surprisingly, affected patients display a more severe clinical compromise that presents earlier than in children with more localized pathology. In many cases, bilateral extensive involvement of the cerebral hemispheres likely results from a combination of the discrete regions described above (Fig. 4.51) [257]. However, PMG involving widespread cortical areas more likely has epigenetic causes, such as exposure of the fetus to viral or toxic agents. Diffuse PMG has also been described in several syndromes, including the Aicardi syndrome [275] (see Chap. 3), Zellweger disease (see below), and in association with hydrocephalus, craniosynostosis, severe mental retardation, and facial and genital anomalies [276].

133

134

P. Tortori-Donati, A. Rossi, and R. Biancheri

4.3.6.2 Schizencephaly Background

The term schizencephaly refers to a uni- or bilateral, full-thickness cleft in the brain that connects the lateral ventricles to the subarachnoid space through the cerebral mantle. The definition of schizencephaly has long been controversial, and pathologists have often disagreed on its real significance, basically because of uncertainties regarding its etiology [1, 277, 278]. Differentiation from purely clastic lesions (i.e., porencephaly) is based on the fact that the schizencephalic cleft is consistently lined by gray matter in the form of unlayered PMG [108]. Thus, PMG and schizencephaly are presently grouped together (“polymicrogyriaschizencephaly spectrum”) [7]. In fact, there probably is a spectrum of severity in the causal event that correlates with the extent of the resulting abnormality, which can vary from flat PMG to cortical infoldings lined by PMG, to full thickness schizencephaly [108, 279]. This categorization is reinforced by the observation that PMG and schizencephaly may coexist in individual patients, for instance with schizencephaly in one hemisphere and PMG affecting the other (Fig. 4.52).

a

As is the case with PMG, schizencephaly could be the end result of a variety of insults resulting in abnormal cortical organization. In utero ischemic events probably account for the majority of cases [107, 280]. Mutations of the EMX2 homeobox gene, located on 10q26.1, have been rarely found in patients with large schizencephalic clefts [281]. The role of the Emx2 protein has not been fully established, but it may be related with cell fate determination. Classification

Schizencephaly is categorized on the basis of the degree of separation of the cleft lips, which grossly correlates with the severity of the clinical picture. Open lips schizencephaly is characterized by a variable degree of separation of the cleft walls, resulting in a well-defined holohemispheric cleft filled with CSF (Figs. 4.52, 4.53). Patients with open lips schizencephaly are more severely impaired, with moderate to severe developmental delay, seizures, hypotonia, spasticity, inability to walk or speak, and blindness. The incidence of contralateral hemiparesis and the severity of developmental delay and mental retardation correlate significantly with the size of the cleft, a

b Fig. 4.52a,b. Unilateral open lips schizencephaly. a Axial STIR image; b Coronal STIR image. In the left frontal region, a deep fissure involves the whole thickness of the cerebral hemisphere, allowing direct communication between the pial surface and the lateral ventricle, that is sectorially dilated. Polymicrogyric cortex surrounds both lips of the fissure, that are widely separated from one another (arrowheads, a, b). The septum pellucidum is absent (b). Polymicrogyria also involves a portion of the contralateral hemisphere (black arrow, a). Large drainage veins are visible into the dilated CSF spaces corresponding to the cortical malformations on both sides (white arrows, a, b)

Brain Malformations

a

b Fig. 4.53a,b. Bilateral schizencephaly in a patient with septo-optic dysplasia. a Coronal STIR image; b Axial STIR image. A large fissure is recognizable in the left hemisphere, surrounded by dysplastic cortex (white arrowheads, a, b). The large CSF cavity (white asterisk, b) causes asymmetric macrocrania. A narrower fissure is visible contralaterally (black arrowheads, a), and a third is seen in the left temporal region that allows communication between the hemispheric surface and the temporal horn (black arrowheads, a). Also these narrower fissures are surrounded by dysplastic cortex. Transmantle cortical heterotopia involves the posterior portion of the right frontal lobe (black asterisk, b). Notice absence of the septum pellucidum

location involving the rolandic region, and the presence of bilateral defects [278, 279]. Seizures appear earlier and are more likely to be intractable in open lips schizencephaly than in closed lips schizencephaly [202]. Closed lips schizencephaly is characterized by apposition of the walls of the cleft with a variable degree of fusion of the opposed cortical surfaces, resulting in a narrow, variably obliterated furrow that extends from the pial surface to a dimple located along the lateral ventricular wall (Fig. 4.54). Patients with unilateral closed lips schizencephaly can be almost normal, although they may have seizures and spasticity. Imaging Findings

MRI depicts schizencephaly as a CSF-filled cleft bordered by gray matter, extending from the pial surface to the ventricular wall [108] (Figs. 4.52–54). Schizencephalic clefts may be unilateral or bilateral; in the latter case, they are usually almost symmetric in location, but not necessarily in size [202, 279, 280]. Unilateral and bilateral clefts probably occur with a similar incidence in the general population [280]. Multiple clefts are exceptional; however, we observed cases with multiple bilateral clefts and associated septooptic dysplasia.

Open-lipped defects are more readily appreciated than closed-lipped ones. When large, open-lips schizencephaly may cause enlargement of the skull as a result of CSF pulsations (Fig. 4.53) [279, 280]. Differentiation from porencephalic cavities is based on the presence of gray matter lining both sides of the cleft. Large vessels are often seen within the CSF into open-lipped clefts [108]. A single case report has described open lips schizencephaly in conjunction with an intracranial arteriovenous malformation [282]. An important clue to the presence of a fused cleft is a dimple along the lateral ventricular wall, corresponding to the site where the cleft enters the ventricle [108] (Fig. 4.54). This sign can be particularly useful with CT, as recognition of a fused-lipped cleft may not be easy due to insufficient contrast resolution. Associated cortical malformations are common. These typically include PMG and subependymal heterotopia extending for variable distances from the superficial and deep extremities of the cleft, respectively. Moreover, cortical malformations (typically PMG) often involve the contralateral hemisphere in cases of unilateral schizencephaly, usually in a symmetric location (Fig. 4.52) [279]. The septum pellucidum is absent in the majority of patients with schizencephaly [280, 283]. Some patients also have associated optic nerve hypoplasia and are best categorized as having “septo-optic dys-

135

136

P. Tortori-Donati, A. Rossi, and R. Biancheri

4.3.6.3 Non-Taylor’s Focal Cortical Dysplasia (Architectural and Cytoarchitectural) Background

Non-Taylor’s focal cortical dysplasia (NFCD) is believed to result from a localized insult to the developing cortex occurring during the process of cortical organization [7]. As such, it is currently separated from FTD (see discussion, Taylor’s FCD). However, such separation is questionable on the basis of histological findings. In fact, there is a histological spectrum of severity ranging from mild forms of NFCD with essentially normal cell populations to full-blown FTD with balloon cells [146, 147]. According to this classification (Table 4.5), NFCD is categorized into two subsets [146, 147]:

a

[1] Architectural FCD (formerly microdysgenesis): The affected cortex shows mildly to moderately abnormal lamination without malformed or giant neurons, associated with heterotopic neurons in the underlying white matter; [2] Cytoarchitectural FCD: Moderately to severely abnormal cortical lamination is associated with giant neurons distributed throughout the cortex and heterotopic neurons in the white matter.

Clinical Findings b Fig. 4.54a,b. Closed lips schizencephaly. a Axial T2-weighted image; b Axial STIR image. Two thin, full-thickness clefts are visible in the occipital regions bilaterally (arrowheads, a, b). An ependymal dimple, corresponding to the orifice of the cleft, is clearly visible on the left (thin arrow, a, b), whereas a cortical depression, corresponding to the outer orifice of the cleft, is detected to the right (thick arrow, a, b). Both sequences, but especially STIR, clearly show both clefts are surrounded by abnormal cortex

plasia plus” (Fig. 4.53) [94] (see “septo-optic dysplasia”). There is a relationship between location of the cleft and involvement of septum, in that absence of the septum correlates with frontal clefts, whereas presence of the septum is found with clefts involving the parietal, occipital, and temporal lobes; however, clefts involving the rolandic regions do not show an equally well-defined relationship with septal defects [284].

Clinically, patients basically present with seizures. Epilepsy onset is generally in the first decade. Seizure frequency is high, typically with one or more episodes a day, and increases with the histological grade [146]. Mild mental retardation and abnormal neurological examination are found in a minority of patients. Imaging Findings

The role of MRI in the identification of NFCD is not completely known. A significant proportion of cases (probably as many as one third) remains undetected even with high resolution MRI, and is only detected by postsurgical histological examination. Moreover, the boundaries of the lesion identified with MRI are often maldefined, thereby providing an imperfect guide to surgical resection [146]. Architectural FCD is characterized on MRI by normal cortical thickness and pattern of gyration

137

Brain Malformations

[147]. This represents a notable exception from most other MCD. Significant findings [147] are focal brain hypoplasia, white matter core atrophy, and moderate white matter hyperintensity on T2weighted images. Typical location is in the temporal lobe (Fig. 4.55) [147]. A blurred gray-white matter junction, best appreciated on fast inversion recovery sequences, may also be seen, although it is less marked than in FTD [147, 160]. An association with the typical MRI features of hippocampal

sclerosis is found in about 55% of cases (Fig. 4.55) [147]. Cytoarchitectural FCD was only rarely commented upon in the literature. MRI is normal in about 50% of patients. In the remainder, MRI are not characteristic, being either undistinguishable from FTD (Fig. 4.56) or showing isolated frontotemporal hypoplasia without signal abnormalities. The diagnosis was histological in all these cases [147].

a

b Fig. 4.55a,b. Architectural focal cortical dysplasia associated with hippocampal sclerosis (dual abnormality). a Coronal STIR image; b Coronal T2-weighted image. There is hypoplasia of the left temporal pole, with blurred gray-white matter junction (arrow, a). Hippocampal sclerosis is also detected (arrow, b)

a Fig. 4.56a,b. Cytoarchitectural focal cortical dysplasia. a Coronal STIR image; b Axial FLAIR image. There is blurring of the gray-white matter junction in the left frontal region (arrow, a). Notice that cortical thickness is normal. Heterotopic neurons deepen into the white matter without reaching the ventricular wall, causing signal hyperintensity on the FLAIR image (arrow, b). Findings are similar to those of focal transmantle dysplasia

b

138

P. Tortori-Donati, A. Rossi, and R. Biancheri

4.3.7 Malformations of Cortical Development, not Otherwise Classified 4.3.7.1 Zellweger Syndrome

Zellweger syndrome (ZS), or cerebro-hepato-renal syndrome, is a lethal disorder affecting one in every 50,000 live newborns [285]. ZS is a peroxisomal disorder characterized by a host of biochemical abnormalities, basically involving marked plasma elevations of very long chain fatty acids, pipecolic acid, and bile acid intermediates [286]. Presentation is typically neonatal. Affected newborns are profoundly hypotonic and areflexic, have seizures, show a characteristic craniofacial appearance including turribrachycephaly, and display multiorgan abnormalities including cortical dysgenesis, white matter hypomyelination, hepatomegaly, polycystic kidneys, and calcific stippling of patellae, hips, and other epiphyses [286]. Life span is short, usually in the range of a few months [285]. Brain MRI displays a characteristic association of cortical dysgenesis and severely diminished myelination [285, 287]. The entirety of the cerebral cortex is abnormal. However, the pattern of cortical involvement varies regionally and among patients, with a mixture of polymicrogyria, pachygyria, and gyral pattern simplification [287]. Because affected patients are newborns or small infants, the degree of myelination is best assessed by evaluating the internal capsules. In ZS, there is lack of physiologic low T2 signal intensity, and diminished or complete absence of high T1 signal intensity, in the posterior portions of the posterior limbs of the internal capsules [287]. Zellweger syndrome is described in greater detail in Chap. 13. 4.3.7.2 Mitochondrial Disorders

Diffuse polymicrogyria has been rarely described in patients with pyruvate dehydrogenase complex deficiency [7, 288]. 4.3.7.3 Sublobar Dysplasia

This recently described abnormality [289] is characterized by localized dysplasia of a portion of one cerebral hemisphere, separated from the remainder of the hemisphere by one or more deep cortical infoldings. Presentation is with generalized seizures that

may progress to become partial complex; the neurologic examination is otherwise normal. The affected cortex is thickened, with shallow sulci and an abnormal sulcal pattern. Separation from polymicrogyria is based on the use of thin-slice volumetric imaging sequences. Associated abnormalities prominently include callosal dysgenesis [289].

4.4 Malformations of the Posterior Cranial Fossa 4.4.1 Embryology Knowledge of the main embryological steps leading to the formation of the cerebellum is required in order to understand its malformations. The development of the cerebellum differs from that of the supratentorial brain in many ways. Most importantly, the cerebellum forms from midline fusion of two paired rudiments, whereas the telencephalon cleaves along the midline to form the two cerebral hemispheres. Around the 28th gestational day, that is roughly by the end of primary neurulation, the anterior end of the neural tube becomes expanded to form the three primary brain vesicles, of which the rhombencephalon, or hindbrain, is the caudalmost. The cavity of the hindbrain will become the fourth ventricle. The upper portion of the rhombencephalon exhibits a marked constriction, the isthmus rhombencephali, which divides it from the mesencephalon, or midbrain. As the result of unequal growth of the primary cerebral vesicles, three flexures are formed and the embryonic brain becomes bent on itself in zigzag-like fashion. The two earliest flexures (cephalic and cervical flexure) appear shortly after the institution of the primary cerebral vesicles, and are concave ventrally so that the neural tube forms a wide, upside-down U shape [1]. The third bend (pontine flexure) appears during the middle of the second gestational month and is concave dorsally, thereby converting the shape of the cephalic extremity into a broad M. The rhombencephalon further divides into two secondary vesicles: an upper, called the metencephalon, and a lower, the myelencephalon. The pons develops from a thickening in the floor and lateral walls of the metencephalon. The floor and lateral walls of the myelencephalon are thickened to form the medulla oblongata, which is continuous inferiorly with the spinal cord. The cerebellum, once thought to derive

Brain Malformations

solely from the hindbrain, has in fact a dual origin from the alar plates of the caudal third of the mesencephalon (basically forming the vermis) and the alar plates of the metencephalon (basically forming the cerebellar hemispheres) [290–292]. Anterolateral migration of cells from the alar plates of the metencephalon forms the pontine nuclei. The isthmus plays a crucial role in the control of early patterning of the prospective midbrain and anterior hindbrain (“isthmic organizer”). The isthmic organizer lies at the interface of the expression of two homeotic genes, OTX2 and GBX2. Additionally, signaling molecules encoded by FGF8 and WNT1 are expressed within the isthmic organizer, and are necessary to maintain expression of EN1, EN2, and Pax 2 [293]. Moreover, the transcription factor Lmx 1b maintains isthmic WNT1 expression. Perturbations of genes involved in early cerebellar patterning have been shown to produce a wide spectrum of abnormal phenotypes, ranging from agenesis of the entire midbrain-cerebellar region to minimal abnormalities of cerebellar foliation [293]. Because of the appearance of the pontine flexure, the cranial and caudal extremities of the fourth ventricle are progressively approximated dorsally and the roof is folded inward toward the cavity of the fourth ventricle, while the alar (dorsal) laminae are splayed laterally as a consequence of the bending of the pons, and eventually lie dorsolateral to the basal laminae. Therefore, the roof plate (corresponding to the roof of the developing fourth ventricle) remains thin and is markedly expanded transversally; as seen dorsally, the roof of the fourth ventricle assumes a somewhat lozenge-shaped appearance, with the margins angling obliquely toward the midline both cephalically and caudally starting from their widest separation. Mesenchyme insinuating itself into the fold at level of the fourth ventricular roof forms the plica choroidalis, that is the precursor of the choroid plexus. This plica starts in the extremity of each lateral recess, passes mesially, and then extends caudally to the extremity of the fourth ventricle [294]. The roof of the fourth ventricle is divided by the plica choroidalis into two areas: one cranial, called the anterior membranous area (AMA), and one caudal, the posterior membranous area (PMA) (Fig. 4.57). Owing to neuroblastic proliferation, reflecting production of the neurons of the cerebellar nuclei, Purkinje cells, and radial glia, the alar laminae along the lateral margins of the AMA become thickened to form two lateral plates, called rhombic lips. This process begins at approximately 28–44 days [295], and causes the rhombic lips to enlarge and to approach each other. As the rhombic lips grow, the AMA

regresses. Eventually, the rhombic lips fuse in the midline, producing a thick lamina which constitutes the rudiment of the cerebellum, the outer surface of which is originally smooth and convex [296]. At this stage, the AMA has been completely incorporated into the developing choroid plexus. Growth and backward extension of the cerebellum pushes the choroid plexus inferiorly, while the PMA greatly diminishes in relative size as compared to the overgrowing cerebellum. Subsequently, a marked caudal protrusion of the fourth ventricle appears, causing the PMA to expand as the finger of a glove [294]. This transient protrusion has been called Blake’s pouch (Fig. 4.57). The Blake’s pouch consists of the ventricular ependyma surrounded by a condensation of the mesenchymal tissues [294]; it is a closed cavity, that is, it initially does not communicate with the surrounding subarachnoid space of the cisterna magna, whose canalization from the original leptomeningeal meshwork presumably commences at about 44 days of gestation [297]. The choroid plexus extends in the roof of the pouch a short distance caudad of the extremity of the vermis and along its under surface. The network between the vermis and the pouch progressively becomes condensed, whereas the other portions about the evagination become rarified [294]; thus, permeabilization of Blake’s pouch institutes the foramen of Magendie. The precise time of opening of the foramen of Magendie is not established [295]; however, persistence of the Blake’s pouch has been demonstrated into the fourth gestational month [298]. The foramina of Luschka also probably appear late into the fourth month of gestation [295]. During the first two gestational months the cerebellum lies within the rhombencephalic vesicle. Beginning in the third gestational month, intense neuroblastic proliferation causes it to enlarge extraventricularly. The fissures of the cerebellum appear first in the vermis and floccular region, and traces of them are found during the third month, whereas the fissures on the cerebellar hemispheres do not appear until the fifth month [296]. The groove produced by the bending over of the rhombic lip is known as the floccular fissure; when the two lateral walls fuse, the right and left floccular fissures join in the midline and their central part becomes the postnodular fissure. Because it is the first cerebellar portion to become recognizable as a separate portion, the flocculonodular lobule is also designated archicerebellum. The formation of the vermis antedates that of the hemispheres by 30–60 days [295] and is complete by the 15th gestational week; therefore, the vermis (with the exception of the nodulus) is also referred to as paleocerebellum, whereas the hemispheres (with the exception of the

139

140

P. Tortori-Donati, A. Rossi, and R. Biancheri

Isthmus rhombencephali AMA Choroid plexus PMA a

Blake’s pouch

Cerebellar rudiment

Foramen of Magendie

PMA b

c

Isthmus rhombencephali

d

Cerebellar hemisphere

Rhombic lips AMA Cloroid fold

Vermis PMA (Blake’s pouch)

PMA e

g

Foramen of Magendie

f

Fig. 4.57a–g. Schematic depiction of normal fourth ventricular development. a–d Sagittal views; e–g. Posterior views. Abbreviations: 4 v, fourth ventricle; AMA, anterior membranous area; PMA, posterior membranous area. a and e The initially flat roof, lozengeshaped of the developing fourth ventricle is divided by the choroid fold (the precursor of the choroid plexus) into two areas: a cranial area, called AMA, and a caudal area, called PMA. b and f: Cellular proliferation occurs at the edges of the AMA, forming the rhombic lips. These, together with the isthmus rhombencephali, contribute to the development of the cerebellar rudiment. As the rhombic lips grow, the size of the interposed AMA decreases. Eventually, the AMA completely disappears, its remnants being incorporated into the choroid fold. Notice that, as the AMA is regressing, the PMA is persisting. c, d, and g: The PMA expands as a finger of a glove, forming an ependyma-lined protrusion that is continuous with the fourth ventricle and is caudal to the choroids plexus. This structure is called Blake’s pouch. It occupies a space comprised between the vermis superiorly, and the developing occipital squama inferiorly and posteriorly. Permeabilization of the Blake’s pouch institutes the foramen of Magendie, and allows for cerebrospinal egression from the fourth ventricle into the developing subarachnoid spaces of the cisterna magna

flocculi) constitute the neocerebellum [17]. The best marked of the early fissures are the primitive fissure between the developing culmen and declive, and the secondary fissure between the future pyramid and uvula. Anatomically, the portion of cerebellum that lies anterior to the primary fissure is called anterior lobe, whereas the posterior lobe lies between the primary and flocculonodular fissure. Therefore, there is only a limited correspondence between the phylogenetic and anatomic subdivisions of the cerebellum. The cerebellar cortex consists of three layers: an external gray molecular layer; an intermediate, incomplete stratum of Purkinje cells; and an internal granular layer. Cells of the intermediate Purkinje layer and of the deep cerebellar nuclei arise from the germinal matrix in the fourth ventricular wall from 11–13 weeks and migrate radially. Neurons that will form the molecular and inner granular layer arise

from the lateral portion of the rhombic lips and, by 11–13 weeks, migrate tangentially over the surface of the cerebellum to form a transient external granular layer. Proliferation of external granular cells produces daughter cells that migrate inward to form the basket cells of the molecular layer and the internal granular layer. The external granular layer eventually disappears by the end of the first postnatal year [295].

4.4.2 Classification of Malformations No fully satisfactory classification of cerebellar malformations has been devised so far, in spite of many attempts. Most authors have tried to classify cerebellar malformations on the basis of their embryogenesis from the rhombencephalon [17, 295, 299]. Sidman and

Brain Malformations

Rakic [300] considered that the key to morphogenesis of the cerebellum lies in understanding the differential growth rates of the cerebellar primordium. Therefore, embryology-based classification schemes mainly discriminate between paleocerebellar malformations, which have defects of the vermis as their key feature, and neocerebellar malformations, which are characterized by predominantly hemispheric abnormality. The major drawback of a purely embryological approach is that our knowledge of human neuroembryogenesis is far from complete, and rests basically on animal experimentation; moreover, it may be impossible to exactly determine the timing of a given malformation [301]. Recently, Patel and Barkovich [5] proposed an alternative, imaging-based classification scheme (Table 4.8). This approach categorizes cerebellar malformations into hypoplasias (defined as a small or incompletely formed, but otherwise normalappearing cerebellum) and dysplasias (defined as a disorganized pattern with abnormal folial pattern or presence of heterotopic gray matter nodules). Hypoplasias and dysplasias are further classified into generalized (involving both cerebellar hemispheres and Table 4.8. Classification scheme for cerebellar malformations (Patel and Barkovich) [5] I. Cerebellar hypoplasia A. Focal hypoplasia 1. Isolated vermis 2. One hemisphere hypoplasia B. Generalized hypoplasia 1. With enlarged fourth ventricle (“cyst”), Dandy-Walker continuum 2. Normal fourth ventricle (no “cyst”) a. With normal pons b. With small pons 3. Normal foliation a. Pontocerebellar hypoplasias of Barth, types I and II b. Cerebellar hypoplasias, not otherwise specified II. Cerebellar dysplasia A. Focal dysplasia 1. Isolated vermian dysplasia a. Molar tooth malformations (associated with brain stem dysplasia) b. Rhombencephalosynapsis 2. Isolated hemispheric dysplasia a. Focal cerebellar cortical dysplasias/heterotopia b. Lhermitte-Duclos-Cowden syndrome B. Generalized dysplasia 1. Congenital muscular dystrophies 2. Cytomegalovirus 3. Lissencephaly with RELN mutation 4. Lissencephaly with agenesis of corpus callosum and cerebellar dysplasia 5. Associated with diffuse cerebral polymicrogyria 6. Diffusely abnormal foliation

the vermis) and focal (localized to either a single hemisphere or the vermis). The pros and cons of this new classification scheme still require time to be fully evaluated. Based on our experience, we have decided to basically retain a morphological approach that distinguishes posterior fossa abnormalities into two broad categories: (i) cystic malformations, characterized by the presence of a substantial CSF collection resulting from active expansion of CSF spaces, and (ii) noncystic malformations, in which either there is no CSF collection or expansion of CSF spaces is passive, i.e., it results from defective cerebellar development (Table 4.9). Moreover, these malformations can also be classified into focal (i.e., with isolated vermian or hemispheric involvement), diffuse (involving both the vermin and hemispheres), and combined (with involvement of the cerebellum and brainstem).

Table 4.9. Classification scheme for posterior fossa malformations (Tortori-Donati) A) Cystic malformations Malformations of the fourth ventricular roof 1. Anomalies of the anterior membranous area (Dandy-Walker continuum) a. Dandy-Walker malformation b. Dandy-Walker variant 2. Anomalies of the posterior membranous area (Blake continuum) a. Persistent Blake’s pouch b. Mega cisterna magna Malformations of the arachnoid Arachnoid cysts B) Noncystic malformations 1. Paleocerebellar (vermian) hypoplasia a. Rhombencephaloschisis/molar tooth malformations b. Rhombencephalosynapsis c. Tectocerebellar dysraphia 2. Neocerebellar (hemispheric-vermian) hypoplasia a. Cerebellar agenesis (total and subtotal) b. Pontocerebellar hypoplasia c. Unilateral hemispheric aplasia/hypoplasia 3. Cerebellar cortical malformations a. cortical dysplasias b. granular layer aplasia c. cortical hyperplasias (Lhermitte-Duclos, macrocerebellum) 4. Isolated brainstem hypoplasia/dysplasia

141

142

P. Tortori-Donati, A. Rossi, and R. Biancheri

4.4.3 Cystic Malformations Cystic malformations of the posterior cranial fossa are characterized by the presence of substantial retrocerebellar and/or infracerebellar CSF collections that result from “active” (i.e., not ex-vacuo) expansion of CSF spaces. Thereby, they are distinguished from noncystic malformations, in which CSF collections are either absent or result from ex-vacuo dilatation of CSF spaces. One should be aware of the fact that the nature of posterior fossa “cysts” may be extremely variable. Some “cysts” are in fact related to massive dilatation of the fourth ventricle, others to persistence of embryonic structures such as the Blake’s pouch, others to malformative dilatation of subarachnoid spaces, others to true arachnoid loculations. Because histological verification of the nature of these CSF collections is usually not obtained, classifications, including ours (Table 4.9), may be subject to criticism. However, an important point is that some of these abnormalities are amenable to CSF shunting, which may radically improve the outcome. Therefore, an effort is required to establish diagnostic criteria that may help in treatment decision-making. In the past, CT cisternograms were performed to assess communication between these CSF collections and the subarachnoid spaces, thereby allowing differentiation of noncommunicating arachnoid cysts from other entities. In the MRI era, CT cisternography is no longer performed. MR studies of CSF flow have not been conclusively proved to be helpful in the differential diagnosis of these malformations. Therefore, the mainstay of the diagnosis rests in the assessment of a number of indirect signs, i.e., the morphology of the “cyst,” fourth ventricle, and subarachnoid spaces; the morphology of the vermis and cerebellar hemispheres (normal, hypoplastic, or compressed); the size of the posterior cranial fossa; the position of the torcular; and the presence of hydrocephalus. Two issues appear to be particularly relevant, i.e., differentiation of the Dandy-Walker continuum from global cerebellar hypoplasias, and nature identification of infraretrocerebellar CSF collections. The following discussion is an attempt to establish a number of standpoints concerning the differential diagnosis of the various malformative conditions. 4.4.3.1 Dandy-Walker Malformation Background

The term ”Dandy-Walker malformation” (DWM) was suggested in 1954 by Benda [302] to describe a

malformation consisting of a cystic enlargement of the fourth ventricle associated with agenesis or, more frequently, hypoplasia of the vermis. This malformation had originally been described by Dandy and Blackfan in 1914 [303], and subsequently studied in deeper detail by Taggart and Walker in 1942 [304]. The characteristic triad of the full-blown DWM is: (i) complete or partial agenesis of the vermis; (ii) cystic dilatation of the fourth ventricle; and (iii) enlarged posterior fossa with upward displacement of transverse sinuses, tentorium, and torcular. Hydrocephalus, albeit common [80% of cases], is not an essential criterion [9, 17]. Subsequent observations of cases with partial agenesis and rotation of the vermis and cystic enlargement of the fourth ventricle, but no substantial enlargement of the posterior fossa, led to the introduction of the term “Dandy-Walker variant” (DWV) [305]. Unification of DWM and DWV into a spectrum of variability of rhombencephalic roof development (“The Dandy-Walker continuum”) was proposed by Barkovich et al. in 1989 [306]. However, further investigations from the same group subsequently questioned whether DWV is a mild form of DWM or, rather, represents a generalized form of cerebellar hypoplasia [5]. Pathogenesis

Although pathogenetic explanations initially focused on nonpatency of the fourth ventricular foramina [303, 304], it subsequently became clear that these foramina are patent in the majority of cases [17]. Moreover, formation of the vermis is known to occur several weeks earlier than permeabilization of the foramina of Magendie and Luschka. The pathogenesis of the malformation is presently believed to be related to a developmental arrest in the hindbrain. Failure to incorporate the AMA into the choroid plexus leads to persistence of the AMA between the caudal edge of the developing vermis and the cranial edge of the developing choroid plexus. CSF pulsations cause the AMA to balloon out into a cyst that displaces the hypoplastic vermis superiorly, so that the hypoplastic vermis appears to be “rotated” in a counter-clockwise fashion. The PMA can persist unopened or become patent, accounting for the reportedly variable patency of the foramen of Magendie and association with hydrocephalus. Global enlargement of the posterior fossa may result from arrested development of the tentorium, straight sinus, and torcular, with failure of migration of the straight sinus from the vertex to the lambda, possibly as a consequence of the abnormal distention of the fourth ventricle. A genetic predisposition probably plays a role in the pathogenesis,

Brain Malformations

although the involved genes remain undetermined [5]. Clinical Findings

DWM causes 2%–4% of hydrocephalus in children. Incidence is one in 25–35,000 live births. Hydrocephalus often is not evident at birth, but generally appears by 3 months of age [295]. Stigmata of hydrocephalus are usually represented by an enlarged head, bulging fontanelle, and developmental delay. Cerebellar findings are uncommon. Because growth of the cerebellum by apposition of cells from the external granular layer continues well into the first postnatal year, CSF shunting is required in newborns to reduce CSF pressure into the posterior fossa and favor this physiological process. In our experience, placement of the shunt catheter into the dilated fourth ventricle has been particularly effective in favoring re-expansion and further growth of the cerebellar hemispheres, which typically occurs by 6–12 months of surgery. This has been the preferred modality of treatment in patients with a patent cerebral aqueduct in our experience as well as in many other centers [299, 307, 308], and has resulted in better neurodevelopmental outcomes. Aqueductal stenosis is present in approximately 20% of cases, and affected children require additional ventriculoperitoneal shunting. Subsequent development of affected children may be variable. While some are severely retarded, as many as 47% have normal intellect with little or no motor deficit. Cognitive outcome is basically related to the presence of other brain or systemic abnormalities; patients with isolated DWM have a significantly better prognosis than those with associated CNS or systemic malformations [309]. Associated Conditions

In most instances, DWM is an isolated malformation with a low recurrence rate in siblings. However, DWM is known to occur in many conditions, including the Walker-Warburg syndrome, Coffin-Siris syndrome, Meckel-Gruber syndrome, Fraser cryptophthalmos, Aicardi syndrome [295], and hydrolethalus syndrome [310], other than in chromosomal abnormalities such as trisomy 9, 13, and 18 [295]. DWM also is a component of the Ritscher-Schinzel cranio-cerebellocardiac [3C] syndrome (Table 4.10) [311] as well as of several phakomatoses, such as neurocutaneous melanosis [312–314]. Association of posterior fossa malformations, including DWM, with facial capillary hemangiomas and arterial, cardiac, and ophthalmologic abnormalities is designated PHACE syndrome [315, 316]; this is thought to represent a vascular

Table 4.10. Ritscher-Schinzel cranio-cerebello-cardiac (3C) syndrome Inheritance

Autosomal recessive

CNS manifestations

Dandy-Walker malformation or variant

Craniofacial manifestations

Low-set ears Hypertelorism Down-slanting palpebral fissures Depressed nasal bridge Prominent occiput Cleft palate Micrognathia Ocular coloboma

Cardiac manifestations

Septal defects Valvular defects Cono-truncal abnormalities

phakomatosis (see Chap. 17). Familial MDW has also been described to occur in association with neurometabolic disorders, including congenital disorders of glycosylation (CDG) [317] and a supratentorial leukodystrophic pattern [318]. Imaging Findings

Since both DWM and DWV are radiological diagnoses, the goal of imaging is to separate a dilated fourth ventricle from extraventricular cysts, distinguish among a variety of cerebellar vermian dysgenesis syndromes, and assess degree and type of hydrocephalus. This effort requires evaluation of a number of elements, as follows (Figs. 4.58, 4.59). Osteodural coatings. The skull is enlarged, resulting in a dolichocephalic conformation in the most severe cases. The widened posterior cranial fossa results in thinning and prominence (“scalloping”) of the occipital squama and petrous ridges. The straight sinus and tentorium are elevated, and the torcular is located above the lambda (”torcular-lambdoid inversion”). The falx cerebelli typically is absent [299]. Hydrocephalus causes diastasis of cranial sutures. Brain. The hallmark of DWM is vermian hypoplasia with verticalization (i.e., counter-clockwise rotation of the vermis); as a consequence, the vermis lies behind the quadrigeminal plate (Figs. 4.58, 4.59). Foliation of the vermis is rudimentary [319], and the vermis itself is thoroughly absent in a minority of cases. While some cases of DWM show a hypoplastic vermis, in others the vermis may be better defined as dysplastic (i.e., showing abnormal foliation) [5]. The cerebellar hemispheres are winged outward against the petrous ridges due to massive dilatation

143

144

Ver mi s

P. Tortori-Donati, A. Rossi, and R. Biancheri

IV Ventricle

b

sas a

IV Ventricle

c

d

Fig. 4.58a–d. Dandy-Walker malformation. a Schematic drawing, sagittal projection; b Sagittal T2-weighted image; c Schematic drawing, axial projection; d Axial T2-weighted image. The posterior fossa is enlarged and filled with a huge CSF cavity, corresponding to an enlarged fourth ventricle (a,b) Subarachnoid spaces (sas, a, c)are compressed between the ependyma-lined cyst and the occipital bone. The sagittal image shows hypoplastic vermis with verticalization (i.e., counter-clockwise rotation of the vermis) (arrowhead, b) and elevated tentorial insertion. Hydrocephalus is associated, with a patent cerebral aqueduct. Axial image shows two “hypoplastic” cerebellar hemispheres (arrowheads, d) that are winged outwards against the petrous ridges. The cerebellar falx is normal (arrow, d)

a

b

Fig. 4.59a,b. Dandy-Walker malformation in a neonate. a Sagittal T1-weighted image; b Axial T1-weighted image. The posterior fossa is enlarged and the occipital squama is scalloped. The cerebellar vermis is hypoplastic and rotated counter-clockwise (arrowhead, a), with marked elevation of the tentorium (arrows, A). The cerebellar hemispheres are hypoplastic as well (arrowheads, b). Notice associated agenesis of the corpus callosum and hypoplasia of the brainstem (a)

145

Brain Malformations

a

b

Fig. 4.60a–c. Dandy-Walker malformation: postshunting evolution. a Axial CT scan; b Axial T1-weighted image. c Sagittal T1weighted image. CT scan obtained 1 day after cystoperitoneal shunt surgery shows catheter in the fourth ventricle (a); notice hydrocephalic dilatation of the temporal horns supratentorially (asterisks, a). After one year, MRI (b) shows total re-expansion of both cerebellar hemispheres and resolution of hydrocephalus. Notice, however, that rotation of the vermis persists after shunt surgery (arrowheads, c)

c

of the fourth ventricle. Placement of a shunt catheter into the dilated fourth ventricle permits expansion and growth of the cerebellar hemispheres, indicating that compression, rather than hypoplasia, basically accounts for their appearance; however, rotation of the vermis is not affected by CSF shunting (Fig. 4.60). The brainstem generally is thin, mainly due to hypoplasia of the pons. The midbrain may display a butterfly configuration in axial views, possibly related to absent decussation of the superior cerebellar peduncles. In a minority of cases, the brainstem has a normal conformation. Fourth ventricle. The hugely dilated posterior fossa “cyst” is, in fact, represented by a markedly dilated fourth ventricle. This has been confirmed in histo-

logical studies of the cyst wall, showing an internal ependymal layer coated by a stretched neuroglial layer (representing the would-be inferior vermis) and, externally, by the leptomeninges [17]. The fourth ventricle may extend upward through a congenital dehiscence of the tentorium, thereby occupying a space between the occipital lobes. Inferiorly, the cyst may bulge into the foramen magnum. Laterally, the extension of the fourth ventricle is limited by the reflection of the pia mater over the posterior surface of the cerebellar hemispheres; this sort of “lateral hinge point” is a useful differential sign from other posterior fossa cystic malformations in the axial plane, especially lateral arachnoid cysts. The choroid plexus lies at the level of the medullary insertion of the cyst. The foramen of Magendie is usually (but not necessarily) not patent, unlike the foramina of Luschka, which may be

146

P. Tortori-Donati, A. Rossi, and R. Biancheri

patent uni- or bilaterally. This is a traditional finding at pneumoencephalography [320] and may account for the variable incidence of hydrocephalus. 4.4.3.2 Dandy-Walker Variant

DVW is characterized by the same pathological and neuroradiological features of DWM, with the sole exception being the size of the fourth ventricle, which is not sufficiently large to produce enlargement of the posterior fossa. However, counter-clockwise rotation of the vermis is consistently present, indicating an original abnormality of the AMA (Fig. 4.61). Vermian rotation is a key diagnostic feature [299], allowing one to differentiate DWM from other cystic malformations, especially a mega cisterna magna. The vermis is usually also hypoplastic. Although the inferior portion of the vermis appears to be absent,

Verm

is

it may be difficult to determine which vermian folia are actually absent [5]. Hydrocephalus is an unusual feature of DWV, and is generally caused by associated malformations, such as aqueductal stenosis. Questions have arisen as to whether the denomination of DWV must be retained or should instead be abandoned in favor of a more general definition of cerebellar hypoplasia. Indeed, it can be difficult to establish when a posterior fossa is sufficiently large for an attribution to full-blown DWM to be applied in individual cases [5]. As a matter of fact, the full-blown DWM and DWV actually represent two aspects of a continuous spectrum of variability. We still believe that rotation of the vermis is the key feature that allows one to differentiate DWVs from other forms of cerebellar hypoplasia. Accordingly, we have applied this concept also to cases in which the degree of inferior vermian hypoplasia was macroscopically minor (Fig. 4.62).

Fig. 4.61a–c. Dandy-Walker variant. a Schematic drawing; b Sagittal T1-weighted image; c Axial T1-weighted image. The cerebellar vermis is hypoplastic and rotated counter-clockwise (arrowhead, b). However, different to the full-blown Dandy-Walker malformation, posterior fossa size is normal. Most of the posterior fossa is filled with a huge CSF cavity, corresponding to a dilated fourth ventricle (b,c). Both cerebellar hemispheres are hypoplastic (c). In this case, hydrocephalus is secondary to aqueductal stenosis (arrow, b)

IV Ventricle

a

b

c

147

Brain Malformations

a

b Fig. 4.62a,b. Minor form of Dandy-Walker variant. a Sagittal T1-weighted image; b Axial T2-weighted image. Both the cerebellar hemispheres and the vermis are grossly well-formed. However, the vermis is rotated counter-clockwise and probably slightly hypoplastic at level of its inferior portions, with absence of well-defined fissuration (open arrows, a). The fourth ventricle freely communicates with the CSF spaces. The size of the posterior fossa is essentially normal, with mild elevation of the tentorium (thin arrows, a)

4.4.3.3 Persistent Blake’s Pouch Background

As was previously described, the Blake’s pouch is a marked caudal protrusion of the fourth ventricle that results from finger-like expansion of the PMA (Fig. 4.63) [294]. The Blake’s pouch is a normal, albeit transient, embryological structure, that initially does not communicate with the surrounding subarachnoid space. Subsequent permeabilization of the pouch institutes the foramen of Magendie. The timing of such permeabilization has not been conclusively established [296], although persistence of the Blake’s pouch has been demonstrated into the fourth gestational month [298, 321]. Failure of permeabilization with postnatal persistence of the Blake’s pouch has been suggested as an explanation for the occurrence of infra- and retrocerebellar CSF collections associated with normal cerebellum and tetraventricular hydrocephalus [322, 323]. These conditions are characterized by absence of communication between the fourth ventricle and the subarachnoid spaces in the midline, and therefore do not fit the definitions of DWM, mega cisterna magna, or arachnoid cyst. In fact, the normal appearance of the cerebellum rules out both DWM, in which the vermis is hypoplastic and rotated, and arachnoid cysts, in which the cerebellum is compressed by mass

effect. Moreover, hydrocephalus is not found in cases of mega cisterna magna because the enlarged cisterna magna freely communicates with surrounding subarachnoid spaces. Finally, resolution of hydrocephalus and cyst size decrease occur with both ventriculoperitoneal and cystoperitoneal shunting, suggesting the “cyst” likely represents a fourth ventricular outpouching that does not communicate with the surrounding subarachnoid spaces [323]. Clinical Findings

Clinically, persistence of the Blake’s pouch becomes symptomatic early in life (usually in the neonatal period) with macrocrania and hydrocephalus. However, in some cases the normal function of the foramina of Luschka may help to maintain CSF flow between intraventricular and subarachnoid spaces, thereby establishing a precarious equilibrium characterized by a compensatory ventriculomegaly; in such cases, the abnormality may become manifest in older children or even in adults [324]. Imaging Findings

Radiologically, persistence of the Blake’s pouch can produce two subsets of findings [323]. In the first category (Fig. 4.64), the cyst is purely infravermian, the fourth ventricle is markedly enlarged, and posterior fossa size is normal. In the second category (Figs. 4.65,

148

P. Tortori-Donati, A. Rossi, and R. Biancheri

4v

4v

Choroid plexus

Blake’s pouch a

Cisterna magna

Blake’s pouch

Developing occipital bone b

Fig. 4.63a,b. The Blake’s pouch. a Normal fetal development: the Blake’s pouch is a finger-like caudal expansion of the fourth ventricle (4 v) that protrudes into the developing cisterna magna. b 130-day-old human fetus: ballooning of the Blake’s pouch, which retains its continuity with the fourth ventricle but is clearly not permeabilized with respect to the surrounding subarachnoid space. Notice the choroid plexus is interposed between the vermis and the pouch, indicating that the pouch develops as an expansion of the posterior membranous area. The vermis itself is normal. (Modified from [295])

a

b Fig. 4.64a,b. Persistent Blake’s pouch: purely infracerebellar (i.e., category 1). a Sagittal T2-weighted image; b Coronal T2-weighted image. There is tetraventricular hydrocephalus. The cerebellar vermis is normally formed and not rotated. Posterior fossa size is normal. The fourth ventricle is markedly enlarged, and communicates with an infravermian cystic formation that effaces the cisterna magna. Such collection represents persistence of the Blake‘s pouch (BP, a). Compare this case with Fig. 4.63b

149

Brain Malformations

Vermis 4v Blake’s pouch

b

a

c

d Fig. 4.65a–d. Persistent Blake’s pouch: infra-retrocerebellar (i.e., category 2), before and after shunting. a Schematic drawing; b Sagittal T2-weighted image and c Axial T2-weighted image at presentation; d Sagittal T2-weighted image 6 months after shunting. There is tetraventricular hydrocephalus. The vermis is well-formed and not rotated. Posterior fossa size is mildly increased. The fourth ventricle is enlarged and communicates with a cystic formation extending into the infra- and retrocerebellar space (BP) whose postero-superior wall is recognizable (arrowhead, b). The retrovermian CSF collection above the pouch is consistent with entrapped subarachnoid cisternal spaces (asterisk, b). Following shunting of the right lateral ventricle (arrowhead, d), there is marked reduction in size of both the ventricular system and the Blake’s pouch

4.66), extension of the pouch to the retrocerebellar space probably interferes with meningeal development causing elevation of the tentorium, splitting of the falx cerebelli, scalloping of the occipital squama, and increased posterior fossa size. In the latter group, the fourth ventricle is less markedly enlarged, and the brainstem can be displaced anteriorly against the clivus (Fig. 4.66). In both categories, the cerebellum is normal, as confirmed by postshunting MRI studies; particularly, absence of cerebellar hypoplasia is crucial in order to categorize the anomaly as a persistent Blake’s pouch, as vermian hypoplasia would reflect an abnormality of the AMA, rather than of the PMA. Associated brain abnormalities are consistently

absent; this could reflect a relatively late timing for this abnormality. Persistence of the Blake’s pouch in otherwise normal fourth gestational month fetuses [298] could support this theory. Recognition of this entity and differentiation from other cystic abnormalities of the posterior fossa is relevant for treatment decisions. While the preferred treatment modality for either DWM with patent cerebral aqueduct or arachnoid cysts is cystoperitoneal shunt surgery, ventriculoperitoneal derivation is a safe and effective procedure in patients with a persistent Blake’s pouch (Fig. 4.65). Incorporation of persistent Blake’s pouch within the Dandy-Walker continuum is justified by the common histological

150

P. Tortori-Donati, A. Rossi, and R. Biancheri

a

b Fig. 4.66a,b. Persistent Blake‘s pouch: infra-retrocerebellar (i.e., category 2), extreme form. a Sagittal T2-weighted image; b Axial T1-weighted image. There is very severe tetraventricular hydrocephalus. The size of the cerebellum is reduced because of increased pressure exerted by the persistent Blake’s pouch (BP, a, b), that fills almost the whole posterior fossa. The postero-superior wall of the pouch is recognizable (arrow, a). The vermis is normally formed and not rotated. The brainstem is pushed against the quadrilateral plate with deformation of the anterior aspect of the pons (a)

nature of the cavity, representing an ependyma-lined expansion of the fourth ventricle, just as the “cyst” found in Dandy-Walker malformations and variants [321, 322]. 4.4.3.4 Mega Cisterna Magna

The cisterna magna is a CSF-filled space at the base of the brain, located below the cerebellum and behind the medulla. The embryological origin of the cisterna magna is debated. According to one theory (C. Raybaud, personal communication), it originates from permeabilization of the Blake’s pouch, with establishment of a permeable foramen of Magendie that allows egress of CSF from the fourth ventricle. As such, the anatomy of the cisterna magna reflects that of the embryologic Blake’s pouch. The posterior extent of the cisterna magna is limited by an arachnoid membrane that bridges the inferior surface of the cerebellar hemispheres, thereby separating the cisterna magna from the superior vermian cistern. The cisterna magna communicates superiorly with the fourth ventricle through the foramen of Magendie, and inferiorly with the perimedullary subarachnoid spaces. Its size has been the subject of debate in the past. Gonsette et al. [325], who first introduced the term “mega cisterna magna” to describe an enlarge-

ment of the cisterna magna that was often detected by means of ventriculography, considered a normal cisterna magna to be 15 mm long, 5 mm high, and 20 mm wide on average. The term mega cisterna magna (MCM) refers to a cystic posterior fossa malformation characterized by an intact vermis, an enlarged cisterna magna, and absence of hydrocephalus (Fig. 4.67) [295, 299, 319, 322]. As such, the appearance is similar to that previously described of persistent Blake’s pouch, except for the consistent absence of hydrocephalus. Owing to the common embryological mechanism, MCM could be viewed as a “variant” of the persistent Blake’s pouch, in which eventual permeabilization of an enlarged pouch permits CSF communication, without generating obstructive hydrocephalus. Consequently, the two malformations could be viewed as a spectrum of variability of a single disorder, that we propose to name “Blake continuum”. Similarly to the persistent Blake’s pouch, the extent of the abnormal CSF collection is variable; while it may be purely infravermian in some cases, in others it extends far beyond the normal borders of the cisterna magna laterally, posteriorly, and superiorly, sometimes reaching the quadrigeminal plate cistern (Fig. 4.67) [295]. The collection may extend superiorly into a focal dehiscence of the falcotentorial junction, which is usually associated with fenestration of

151

Brain Malformations

MCM

a

b

c

e

d Fig. 4.67a–e. Mega cisterna magna. Two different cases, both referred for MRI due to unrelated complaints. a Schematic drawing. There is free communication between the mega cisterna magna (MCM), the fourth ventricle, and the spinal subarachnoid spaces. Case #1. b Sagittal T1-weighted image; c Axial T2-weighted image. The subarachnoid spaces of the posterior fossa are enlarged (MCM, b, c), with scalloping of the occipital squama (arrowheads, b). Notice there is no hydrocephalus. Both the cerebellar hemispheres and the vermis are normally formed. Posteriorly, the mega cisterna magna insinuates itself into a small fenestration of the tentorial insertion (arrow, b). Splitting of the falx cerebelli, a typical finding in mega cisterna magna, is also visible (arrows, c). Case #2. d Sagittal T1-weighted image; e Axial T2-weighted image. The subarachnoid spaces of the posterior fossa are enlarged (MCM, d, e), with scalloping of the occipital squama (arrowheads, d). Posteriorly, the mega cisterna magna insinuates itself into a large fenestration of the tentorial insertion (arrow, d). The right cerebellar hemisphere appears to be reduced in size compared to the left one (e), and the vermis also shows morphological changes. It is difficult to determine whether such findings are due to hypoplasia, CSF pressure, or both. However, notice that neither signs of hydrocephalus nor compression on the ventricular structures are present

152

P. Tortori-Donati, A. Rossi, and R. Biancheri

the straight sinus [295]. The falx cerebelli is seen in approximately two thirds of cases, and is frequently bi- or tripartitioned. Developmental abnormalities of the tentorium and falx cerebelli significantly indicate interference with normal meningeal development. The posterior fossa is often enlarged due to scalloping of the occipital squama, and the torcular is displaced cranially in 12.5% of cases [322]. The cerebellum and brainstem are typically normal. The presence of intact vermis and cerebellar hemispheres strongly favors a diagnosis of MCM versus an arachnoid cyst, which typically compresses the surrounding brain. However, slight compression of the inferior or posterior vermian folia may sometimes be seen when the size of the CSF collection is large. An important concept that must be stressed is that patients with MCM do not complain of neurological signs of involvement of the posterior cranial fossa. Mental retardation, which has often been advocated as a presenting sign of MCM, is actually seen in patients with associated supratentorial conditions, such as corpus callosum dysgenesis. In itself, MCM is an asymptomatic condition that is usually discovered incidentally [299, 322, 325]. However, it is a very common abnormality, representing over 50% of all cystic posterior fossa malformations [295]. Moreover, hydrocephalus is not found in patients with MCM, because the enlarged cisterna magna freely communicates with the surrounding CSF spaces and does not obstruct CSF circulation. Important consequences of these concepts

are that (i) the presence of hydrocephalus rules out MCM and should direct the diagnosis towards a persistent Blake’s pouch; and (ii) shunt surgery has no indications and should be avoided, even in case of large CSF collections. Finally, extreme caution should be employed in newborns before labeling a posterior fossa CSF collection as MCM. We have observed an apparent neonatal MCM that developed, in a few months, into a retrocerebellar cyst that caused significant cerebellar compression and supratentorial hydrocephalus (Fig. 4.68). One should remember that arachnoid cysts may take time before generating significant compression over surrounding nervous tissues; moreover, slowly expanding arachnoid cysts may initially cause scalloping of the plastic neonatal skull, before progressive ossification forces them to expand at the expense of the surrounding brain, thereby becoming clinically manifest. 4.4.3.5 Arachnoid Cysts

Arachnoid cysts are CSF-filled cysts that do not communicate with the surrounding subarachnoid space (Fig. 4.69) and ventricular system, and usually are not associated with brain maldevelopment. Posterior fossa arachnoid cysts may be located below or posterior to the vermis in a midsagittal location (Fig. 4.70), lateral to the cerebellar hemispheres, cranial to the vermis in the tentorial hiatus, anterior

a

b Fig. 4.68a,b. Mega cisterna magna versus arachnoid cyst in a neonate. a Sagittal T2-weighted image at age 28 days; b Sagittal T2weighted image at age 7 months. Initial MR imaging is consistent with a mega cisterna magna. However, follow-up MRI clearly shows compression of the posterior aspect of the vermis (arrows, b), indicating an arachnoid cyst; also notice enlargement of the ventricular system

153

Brain Malformations

cyst

a

b

Fig. 4.69a,b. Arachnoid cyst. a Schematic drawing; b Axial CT cisternography. Intrathecal iodized contrast medium administration shows posterior fossa CSF collection (cyst, b) is excluded from the surrounding subarachnoid spaces, that are filled by the contrast medium

Fig. 4.70a–c. Arachnoid cyst. a Sagittal T1-weighted image; b Axial T2-weighted image; c Coronal T1-weighted image. A large posterior fossa CSF cystic collection (ac, a–c) compresses and elevates the vermis (arrows, a) and compresses both the cerebellar hemispheres (b, c). The fourth ventricle is deformed and compressed (arrowhead, a). Supratentorial hydrocephalus is clearly seen

a

b

c

154

P. Tortori-Donati, A. Rossi, and R. Biancheri

to the brainstem (Fig. 4.71), or anterolaterally in the cerebellopontine angle cistern (Fig. 4.72). While they can be found incidentally on imaging, patients with posterior fossa arachnoid cysts are more likely to be symptomatic than those harboring cysts in other locations. Although controversy surrounds the treatment of arachnoid cysts, cystoperitoneal shunting is used for decompression of symptomatic cysts at our institution. Pathologically, arachnoid cysts are classified by the histologic composition of the cyst wall, which is either arachnoid connective or glioependymal tissue. Retrocerebellar cysts may be either arachnoid or glioependymal, whereas supratentorial ones are more often purely arachnoid in composition. However, histological classification has little relevance in terms of prognosis. Cyst walls result from splitting of the arachnoid membrane, with an inner and outer leaflet surrounding the cyst cavity. Expansion of noncommunicating cysts may occur either from a ball-valve mechanism or because of fluid secretion by the cyst wall or, more likely, along an osmotic gradient. On CT, arachnoid cysts are isodense with CSF on CT. On MRI images, arachnoid cysts appear as welldefined nonenhancing intracranial masses that are isointense to CSF in all sequences, although slight signal increase may be caused by proteinaceous fluid content. Differential diagnosis include (epi)dermoid cysts and other cystic malformations, such as MCM and persistent Blake’s pouch. It is important to underline that a retrocerebellar CSF collection can be conclusively labeled as an arachnoid cyst only with CT cisternography (Fig. 4.69). In fact, “cysts” that fill with

contrast immediately are regarded as diverticula of the subarachnoid space, whereas only those showing delayed contrast accumulation or no filling can be regarded as true arachnoid cysts. However, CT cisternography is no longer performed in the MRI era. Therefore, diagnosis must rely on indirect signs, such as the condition of the surrounding brain and the presence of hydrocephalus [300, 320, 322]. A large MCM occasionally may be confused with an arachnoid cyst. However, while an arachnoid cyst may demonstrate mass effect with displacement of the cerebellum and may cause supratentorial hydrocephalus, MCM demonstrates no mass effect, and the cerebellum and vermis are intact. Differentiation from a persistent Blake’s pouch is immediate when one considers that arachnoid cysts are not associated with an enlarged fourth ventricle. Arachnoid cysts must be differentiated from (epi)dermoid cysts. Both masses may have similar characteristics on T1-weighted and T2-weighted images, and neither show enhancement with gadolinium. On FLAIR images, arachnoid cysts give low signal because of their CSF content, whereas signal is higher in epidermoid cysts. Diffusion-weighted MRI enables performance of such differentiation easily; arachnoid cysts are hypointense in diffusionweighted images, whereas epidermoids are hyperintense because of slow water diffusion [8]. Differential diagnosis from other CSF collections is crucial for optimal treatment decisions. Although most arachnoid cysts are an incidental finding and patients are asymptomatic, arachnoid cysts may show progressive enlargement. Therefore, follow-up MR imaging is essential in order to identify candidates

b

a

Fig. 4.71a,b. Premedullary arachnoid cyst. a Sagittal T2-weighted image; b Axial T2-weighted image. Small arachnoid cyst at the cranio-cervical junction (ac, a, b) whose walls are recognizable (arrowheads, a, b). The cyst deforms and compresses the cervico-medullary junction

155

Brain Malformations

a

b

c

d Fig. 4.72a–d. Antero-lateral arachnoid cysts of two different cases. Case #1: a Axial T2-weighted image; b Axial FLAIR image. Small arachnoid cyst of the right cerebellopontine angle (ac, a, b) is isointense with CSF on both T2-weighted (a) and FLAIR (b) images. There is no significant compression of the surrounding nervous tissue, although the posterior wall of the petrous bone is scalloped (arrowheads, a). Case #2: c Coronal T2-weighted image; d Axial FLAIR image. This larger arachnoid cyst causes significant compression of the homolateral cerebellar hemisphere

for surgery that, unlike other conditions such as a persistent Blake’s pouch, must involve cystoperitoneal derivation also in patients with a supratentorial hydrocephalus.

4.4.4 Noncystic Malformations This heterogeneous category comprises very diverse entities, whose common background is basically represented by the absence of indications for CSF derivation surgery. In short, these malformations are characterized by either the absence of CSF collections or the presence of CSF collections that result from passive (i.e., ex-vacuo) enlargement of CSF spaces as a consequence of defective development of cerebellar portions.

From a neuroradiological standpoint, these ab normalities can be conveniently classified into: (i) paleocerebellar hypoplasias, in which defective development basically involves the vermis; (ii) neocerebellar hypoplasias, in which there usually is a combination of vermian and hemispheric hypoplasia or, more rarely, isolated hemispheric hypoplasia; and (iii) cortical dysplasias, in which the abnormality rests primarily into an anomalous development of the cerebellar cortex, either with or without global abnormality of cerebellar size. It is, however, apparent that this classification scheme, as well as others, probably oversimplifies a much more complex reality and gathers into somewhat loosely defined categories entities that are probably significantly different from one another from both a pathological and a clinical perspective.

156

P. Tortori-Donati, A. Rossi, and R. Biancheri

4.4.4.1 Paleocerebellar Hypoplasia

Cerebellar hypoplasia is a broad category that includes very diverse entities. Semantically, hypoplasia indicates underdevelopment of a tissue or of an organ usually due to a decrease in the number of cells. Usually, the surface of a hypoplastic cerebellum is grossly smooth and the volume is reduced; however, the relative size of the fissures and folia is preserved. Conversely, the term dysplasia indicates abnormal, disorganized development, such as presence of abnormal folial pattern and gray matter heterotopia [5]. As such, dysplasia can affect a hypoplastic cerebellum. Moreover, atrophy implies that the organ reaches normal size and maturity, but shows a progressive volume loss and shrinkage due to an acquired insult or from a genetic disease not expressed during fetal development; therefore, atrophy can affect a hypoplastic or dysplastic cerebellum. It is beyond the scope of the present paper to discuss inherent differences between cerebellar hypoplasias and dysplasias, which remain the subject of debate among authors and a major obstacle to a fully acceptable classification scheme for cerebellar malformations. The general heading of hypoplasia has been used here to indicate cases with primitively small, underdeveloped cerebella as opposed to secondary atrophic changes. Paleocerebellar hypoplasia is defined as defective development of the vermis with basically normal formation of the hemispheres and absence of substantial posterior fossa cysts. Entities involving paleocerebellar hypoplasia basically include rhombencephaloschisis, rhombencephalosynapsis, and tectocerebellar dysraphia.

not exclusively found in JS, but can be present in several other apparently distinct syndromes belonging to the group of cerebello-oculo-renal syndromes, all of which are also characterized by cerebellar vermian hypoplasia and similar clinical features to those of JS. These entities include Arima, Senior-Löken, and COACH syndrome [328, 329] (Table 4.11). Little is known about the genetic background of JS. Because JS is clinically heterogeneous, it has been suggested that there may also be genetic heterogeneity [330]. Although members of the Engrailed, Wnt, Hox, and Pax gene families have been found to regulate pattern formation and segmentation in the developing brainstem, none of them has been demonstrated to play a causal role [331], and some, such as EN1, EN2, FGF8, and BAHRL1, have positively been excluded [330]. Table 4.11. Joubert-related cerebello-oculo-renal syndromes [329] Arima syndrome (1) Vermian hypoplasia (2) Congenital retinopathy (3) Cystic dysplastic kidneys (4) Autosomal recessive inheritance Senior-Löken syndromes (1) Vermian hypoplasia (2) Retinopathy (3) Juvenile-onset nephronophthisis COACH syndrome (1) Vermian hypoplasia (2) Colobomas (3) Nephronophthisis (4) Hepatic fibrosis

Pathology Rombencephaloschisis Joubert Syndrome Background

Joubert syndrome (JS) [326] is an autosomal recessive syndrome characterized by hyperpneic/apneic spells, hypotonia, ataxia, abnormal ocular movements, and psychomotor delay that is found in association with a characteristic posterior fossa abnormality, involving vermian hypoplasia and abnormal pontomesencephalic junction, which results in the “molar tooth sign” on neuroradiologic examination [327]. Although there has been a strong tendency in the literature to call JS all malformations with this neuroradiologic appearance [5], it has become increasingly clear that JS is an integrated clinical-neuroradiological diagnosis [299]. In itself, the molar tooth sign is

Only very little documentation of the neuropathologic changes in JS exists [292, 332]. Available data indicate there is malformation of multiple brainstem structures, including lack of decussation of the superior cerebellar peduncles, central pontine tracts, and corticospinal tracts. There also is marked dysplasia of caudal medullary nuclei and tracts, including the inferior olivary nuclei, solitary nuclei and tracts, and the nucleus and spinal tracts of the trigeminal nerve, associated with reduction of the nuclei of the basis pontis and reticular formation. Additionally, there is vermian hypoplasia or aplasia, with abnormal foliation, poor demarcation of cortex and white matter associated with fragmentation of the dentate nuclei [292], and separation of the two cerebellar hemispheres. Therefore, probably the term “rhombencephaloschisis” could be proposed as the best descriptor of

Brain Malformations

the complex pathological abnormality of the posterior fossa that is found in these patients (R. Canapicchi, personal communication). Clinical Picture

The classical clinical findings of JS include hypotonia, ataxia, psychomotor retardation, abnormal breathing, and nystagmus [326]. These findings can be present in variable proportions in individual cases. Absence or underdevelopment of the vermis impairs control of balance and coordination, basically resulting in truncal ataxia and decreased muscle tone that can be marked in the neonatal period and in infancy. On their own, ataxia and hypotonia are of limited diagnostic value, as they can be present in a wide range of congenital and acquired disease of the cerebellum or brainstem [333]. Vermian hypo-dysplasia probably also interferes with normal cognition, resulting in psychomotor retardation. This can be the only clinical sign in patients with the molar tooth sign, and is usually severe. No positive correlation has been found between the severity of neuropsychologic dysfunction and the composite score of posterior fossa abnormalities, although there is a correlation between increasing age and worsening psychomotor delay [334]. Caudal brainstem dysplasia, particularly abnormalities of the solitary fascicles and nuclei gracili that control afferent respiratory impulses, probably account for the abnormal breathing pattern, characterized by hyperpneic spells, in which affected infants pant, and sudden cessation of breathing that typically occurs during sleep [292, 295]. Breathing abnormalities, once believed to be characteristic [326], are in fact present in fewer than 75% of patients [333]. Their onset is typically in the neonatal period, and they tend to improve as the child grows [299, 333]. Nystagmus can be associated with oculomotor apraxia and abnormal supranuclear eye movement control, and probably results from abnormal development of the pontine nuclei, reticular formation, and vermian lobules [292]. Ocular and oculomotor disturbances are present in greater than 70% of children with JS [333]. Seizures are a less typical manifestation of the syndrome, and have not received a conclusive neuropathologic correlation. Young infants with JS often have a characteristic facial appearance, including a large head with prominent forehead, high rounded eyebrows, epicanthal folds, upturned nose with evident nostrils, open mouth with tongue protrusion, and low-set, tilted

ears [333]. The facial phenotype tends to become less severe as the child grows. Prognosis is variable, but generally poor. A number of patients have died within age 3 years in situations compatible with sudden infant death syndrome; however, it is not known whether these deaths were caused by prolonged apnea, and no satisfactory investigation has so far allowed to individuate those infants who are at increased risk for early death [333]. Survivors generally show global profound developmental delay. Breathing tends to progressively normalize as the child grows. There are a few cases of affected individuals with normal intelligence or learning abilities [335]. Imaging Findings

The neuroradiological picture is characterized by a combination of vermian and midbrain hypoplasia, resulting in the typical “molar tooth sign” [331, 336]. The molar tooth sign is seen on axial MR sections through the malformed pontomesencephalic junction (Fig. 4.73); it consists of the following triad [336]: (i) dysgenesis of the isthmic portion of the brainstem, manifest as elongation and thinning of the pontomesencephalic junction, elongation of the interpeduncular fossa, and deepening of the posterior foramen caecum; (ii) thickened superior cerebellar peduncles, which project straight back, coursing perpendicular to the brainstem; and (iii) hypoplasia of the vermis with incomplete lobulation and sulcal formation, enlargement of the fourth ventricle, and rostral shift of the fastigium [331, 336]. The molar tooth sign is seen in 85% of patients with JS, and usually represents its sole neuroradiologic manifestation [331]. Varying degrees of severity in the key abnormalities can be found, and MRI can be used to assess severity of the dysgenesis [336, 337]. The superior cerebellar peduncles are enlarged and do not decussate in the isthmus portion of the brainstem. The isthmus is therefore thin, and the interpeduncular fossa is secondarily enlarged. The vermis can be either thoroughly absent or significantly hypoplastic. In the latter case, the superior vermis usually is present, but is invariably dysplastic with abnormal foliation, whereas absence of the inferior vermis produces a midline cleft that connects the fourth ventricle to the cisterna magna, resulting in a “bat wing”, “lamp”, or “umbrella” shaped fourth ventricle in axial and coronal planes (Fig. 4.73) [299]. On sagittal planes, the fourth ventricle displays a convex roof with a high riding fastigium [336]. In the vast majority of cases, the fourth ventricle is only mildly enlarged without formation of a posterior fossa cyst or hydrocephalus, and the posterior fossa is not expanded.

157

158

P. Tortori-Donati, A. Rossi, and R. Biancheri

b

a

c

d Fig. 4.73a–d. Molar tooth malformation in a patient with Joubert syndrome. a and b. Axial T2-weighted images; c Coronal T2weighted image; d Sagittal T1-weighted image. The typical “molar tooth sign” results from midbrain hypoplasia with elongation and thinning of the superior cerebellar peduncles (arrowheads, a). Agenesis of the cerebellar vermis is associated, resulting in an interhemispheric cleft (arrows, b, c). As a consequence, the fourth ventricle has a lamp shape on axial (b) and an umbrella shape coronal images (c); on sagittal planes, it displays a convex shape (“ballooning”) with high-riding fastigium (d)

MRI fails to demonstrate the abnormalities of the nuclei and tracts in the caudal medulla that are believed to account for the respiratory problems exhibited by affected patients [331]; however, there usually is lack of the bulges produced over the surface of the medulla by the pyramids anteriorly and olives posteriorly. The supratentorial brain is generally normal [336]; mild to severe atrophy, corpus callosum abnormalities, delayed myelination, or lissencephaly can be seen in a minority of cases [331, 336]. A minority of patients with a greater degree of inferior vermian hypoplasia can have associated tectocerebellar dysraphia or Dandy-Walker malformation;

these patients have been classified as having “Joubert syndrome plus” [331, 336]. An occipital cephalocele is found in about one third of cases [299]. Rhombencephalosynapsis

Rhombencephalosynapsis is a rare entity characterized by vermian agenesis and fusion of the cerebellar hemispheres, middle cerebellar peduncles, dentate nuclei, and inferior colliculi along the midline, resulting in a single-lobed, hypoplastic cerebellum [299, 338]. Although usually sporadic, this condition has been observed in children born from

159

Brain Malformations

consanguineous parents, indicating a likely autosomal recessive mode of inheritance [339]. Affected children primarily present with signs of cerebellar dysfunction, which represents a notable distinction among posterior fossa malformations, and delayed achievement of developmental milestones. Cognitive impairment is variable, and apparently correlates with the degree of cerebellar hypoplasia [338]. The prognosis is poor, with the majority of affected patients dying in infancy or childhood; however, a few cases have been described in young adults [293, 340]. The MRI picture closely reproduces the pathological appearance (Fig. 4.74) [338]. The single-lobed cerebellum is consistently hypoplastic, although its

size may be variable. On axial views, the cerebellum is flat-based with absence of the vallecula. There is continuation of the folia and fissures through the midline; their orientation is usually transverse, but asymmetrical hypoplasia may result in angulation of the folia and fissure across the midline [295, 338]. The fused or closely apposed dentate nuclei form a horseshoe-shaped arc across the midline [17, 295]. Vermian maldevelopment involves absence of the anterior vermis and deficiency of the posterior vermis; lack of definition of the normal vermian lobules results in a cerebellar–not vermian–configuration of the midline on sagittal views [295, 299]. However, identification of a protuberance below the fastigium of the fourth ventricle indicates the presence of the nodulus [341],

a

b

c

d Fig. 4.74a–d. Rhombencephalosynapsis. a Axial STIR image; b Coronal T2-weighted image; c Sagittal T1-weighted image. d Pathological specimens. The cerebellar white matter is completely fused across the midline at the level of the dentate nuclei (arrow, a, b), with total absence of the vermis. On sagittal images, this results in a hemispheric, not vermian, arrangement of the fissures; notice that the normal vermian folia are absent (c). The vallecula is absent (arrowheads, b). The fourth ventricle has a key-hole appearance on axial planes (a). Additionally, the fornix is fused on the midline (arrowheads, c). Pathology (d) shows fusion of the severely hypoplastic hemispheres without vermis, resulting in a single dentate nucleus with a characteristic horseshoe profile. (a–c, courtesy of Prof. U. Salvolini, Ancona, Italy; d. Reproduced from [17], with permission from Springer Verlag, Heidelberg)

160

P. Tortori-Donati, A. Rossi, and R. Biancheri

suggesting that development of the archicerebellum is not affected. Rhombencephalosynapsis is commonly believed to be a rare, isolated malformation. However, the widespread use of MRI has evidenced a higher incidence than was commonly believed, and a common association with other CNS abnormalities including abnormal gyration, fused thalami and fornices, absence of the septum pellucidum, hypoplastic temporal lobes, optic chiasm and anterior commissure, and corpus callosum dysgenesis [17, 295, 338, 341, 342]. Hydrocephalus may result from concurrent aqueductal stenosis [341]. A single case was reported in which rhombencephalosynapsis, holoprosencephaly, and absence of the ventricular system were present [343]. Systemic abnormalities are uncommon; minor hand malformations have been reported [339, 344]. Embryologically, rhombencephalosynapsis is believed to represent underexpression of a dorsalizing organizer gene [3] affecting the “isthmic organizer” at the mesencephalic-metencephalic border [293]. A candidate molecule has been identified in the Dreher Lmx 1a gene mutant mouse. This mouse harbors an autosomal recessive mutation of the LIM homeobox gene and shows agenesis of the vermis with fused cerebellar hemispheres, closely resembling human rhombencephalosynapsis. The gene affected in the dreher mutant is responsible for correct patterning of the dorsal-most cell types of the neural tube, that is, the neural crest and the roof plate, in the hindbrain region [345]; in the absence of a normally functioning gene, the roof plate fails to develop, resulting in defective dorsalization of the neural tube [346]. Associated heterotopia is found both in the cerebral and cerebellar cortex, probably due to disrupted glial limiting membranes resulting in abnormal guidance of neuronal migration [347]. This mechanism could account for the occurrence of both forebrain and hindbrain abnormalities in rhombencephalosynapsis [342]. Tectocerebellar Dysraphia

Tectocerebellar dysraphia is an uncommon abnormality consisting of occipital cephalocele, agenesis of the cerebellar vermis, and deformity of the tectum (Figs. 4.75, 4.76) [17]. Extreme traction of the brainstem towards the site of the cephalocele results in marked tectal beaking, resembling the analogous deformity seen in Chiari II malformation, and in marked dorsal angulation of the brainstem. As a result of this deformity and the absence of the vermis, the cerebellar hemispheres

rotate ventrally and may come to lie ventral and lateral to the brainstem, a condition that has been termed “inverse cerebellum” [17, 295]. Agenesis of the vermis can be difficult to demonstrate due to the extreme distortion and traction of nervous structures in the posterior fossa; the CSF-filled cleft separating the two cerebellar hemispheres can be more easily visualized on coronal planes (Fig. 4.75). The cephalocele typically contains cerebellar cortex [17]; frequently, a lipoma is located in close vicinity to the calvarial defect (Fig. 4.76). Hydrocephalus is common and likely results from aqueductal deformation and stenosis. Less consistent components are aplasia of mammillary bodies, fusion of thalami, anomalies of cerebral gyral patterns, bifid atlas or bifid occipital squama, elevation of torcular, and cervical hydromyelia [17]. Differentiation from simple traction deformities secondary to cerebellar tissue dislocation into an occipital cephalocele is based on the greater severity of the deformities, their complexity, and the severe hypoplasia or aplasia of the vermis [295, 348]. Clinically, affected newborns present with an occipital cephalocele, and are usually microcephalic [295]. Hydrocephalus requires treatment with shunt surgery. Despite medical efforts, most children die before age 10 years. Embryologically, tectocerebellar dysraphia is believed to result from an insult occurring during early embryogenesis, resulting in a cephalocele that causes extreme traction of the brainstem towards the calvarial defect. As a consequence, the tectum and cranial nerves are markedly elongated, while the cerebellum tilts lateral and ventral to the brainstem. Vermian agenesis is thought to result from the anatomical deformation preventing the cerebellar hemispheres to fuse in the midline to form the vermis [295]. Because of this, tectocerebellar dysraphia is believed to be more pertinent to cephaloceles than to other forms of vermian hypoplasias [5]. Morphologically, tectocerebellar dysraphia bears some resemblance to Dandy-Walker malformation, basically in that there is hypoplasia of the cerebellar vermis, high position of the transverse sinuses, and large posterior fossa size; tectal beaking and kinking of the brainstem are similar to those seen in the Chiari II and, more typically, Chiari III malformation. Kinking of the brainstem and vermian agenesis also figure prominently among the pathological features of Walker-Warburg syndrome. Despite these apparent similarities, the theory that tectocerebellar dysraphia is a tandem malformation linking the Chiari II and Dandy-Walker malformation [348] has not been proven.

161

Brain Malformations

a

b

c

d Fig. 4.75a–d. Tectocerebellar dysraphia in a child with prior surgery for occipital cephalocele. a Sagittal T1-weighted image; b Axial T1-weighted image; c Coronal T1-weighted image. d Pathological specimen. There is extreme deformation of the midbrain, that is attracted posteriorly resulting in marked angulation of the ponto-mesencephalic junction (black arrows, a). There also is extreme deformation of the tectum, resulting in marked tectal beaking (arrowheads, a,b) that points to the site of cephalocele surgery (white arrow, a). Associated agenesis of the vermis, resulting in an interhemispheric cleft (black arrows, c), is well recognizable. There is hydrocephalus. The subarachnoid spaces ventral to the brainstem are markedly enlarged (asterisk, a). In a different case, pathology confirms midbrain deformation and marked tectal beaking (arrowheads, d). (d Reproduced from [17], with permission from Springer Verlag, Heidelberg)

4.4.4.2 Neocerebellar Aplasia and Hypoplasia

Contrary to the well-defined syndromes of vermian hypoplasia, there still is unsatisfactory nosologic definition of neocerebellar aplasia and hypoplasia [17]. Aplasias include primary nonformation of the neocerebellum and lesions that result from damage to the cerebellar anlage early during embryogenesis,

whereas in the hypoplasia the entire cerebellum, or part of it, is smaller than normal, but the individual folia are grossly or microscopically normal in appearance [295]. While cerebellar agenesis is usually isolated, cerebellar hypoplasias may be associated with supratentorial abnormalities.

162

P. Tortori-Donati, A. Rossi, and R. Biancheri

a

b Fig. 4.76a,b. Tectocerebellar dysraphia. a Sagittal T1-weighted image; b Axial T1-weighted image. Complex malformation characterized by a large occipital lipoma (asterisks, a, b) and by nervous tissue (whose origin is not easily recognizable) anterior to it (arrow, a). The posterior fossa is mostly filled with a large CSF collection. The vermis is absent, whereas the cerebellar hemispheres are hypoplastic (H, b) and seem to be attached to the lipoma; an additional small lipoma is visible anterior to the left cerebellar hemisphere (arrow, b). The tectum of the midbrain is splayed, elongated, and elevated (arrowhead, a). The midbrain is hypoplastic and attracted posteriorly, resulting in an angulation of the ponto-mesencephalic junction. The subarachnoid spaces anterior to the brainstem are markedly enlarged. (Case courtesy of Dr. B. Bernardi, Detroit, United States)

Cerebellar Agenesis

Cerebellar agenesis (synonym: neocerebellar aplasia) is an exceedingly rare abnormality involving absence of the vermis and cerebellar hemispheres. The posterior fossa is of normal or slightly reduced size and basically contains the brainstem, which lacks the normal prominence of the pons, and a CSF collection which passively fills the space normally occupied by the cerebellum (Fig. 4.77). In some instances, pea-like masses of disorganized white matter or slightly larger amounts of cerebellar tissue can be observed [349]. In these cases, the term subtotal agenesis is probably more suitable. The cerebellar remnants generally occupy the antero-superior regions in the would-be location of the quadrangular lobule, in a typically asymmetric fashion [349, 350]. The ventral pons is underdeveloped. Rarely, the size of the posterior fossa may actually be increased, probably as a result of CSF pulsation (Fig. 4.77). Supratentorial abnormalities that have been observed in association with cerebellar agenesis include hydrocephalus, agenesis of corpus callosum, and arhinencephaly [17]. The clinical picture can be surprisingly mild, and involve slight intellectual handicap or developmental delay and rather mild cerebellar motor signs [349, 350]. The paucity of clinical signs has been explained

in terms of plasticity of the fetal brain, which could permit “cerebellization” of a part of the cerebrum [349]. While the neuroradiologic diagnosis of complete cerebellar agenesis is straightforward, differentiation of subtotal cerebellar agenesis from pontocerebellar hypoplasia basically rests on both clinical grounds and the appearance of the cerebellar remnants, which typically is asymmetric in the former and symmetric in the latter [349]. The embryological origin of cerebellar agenesis is still controversial. According to traditional theories, it derives from destruction of normally developed tissue, i.e., a congenital form of cerebellar atrophy [17, 295]. This theory basically rests on the speculation that, because the posterior fossa forms, prior development of the hindbrain must have occurred [295]. However, the concept of basicranial induction of the neuroectoderm is still incompletely understood; actually, the cranial base and calvarium are fully capable of forming even when the brain fails to develop [9]. More recent theories invoke an intrinsic arrest of cerebellar development at an early stage, i.e., a severe form of hypoplasia. Mutation of homeobox genes at the level of the isthmic organizer could lead to underexpression of the dorsalizing gradient [3], similar to what probably occurs in rhombencephalosynapsis.

163

Brain Malformations

a

b

c

d Fig. 4.77a–d. Cerebellar aplasia: Two different cases. Case #1: a Sagittal T1-weighted image; b Coronal T1-weighted image. The size of the posterior fossa is reduced. Cerebellar structures are completely absent, except for a rudimental structure located in the upper portion of the fossa (arrows, b). The pons is markedly hypoplastic (a).Case #2: c Sagittal T1-weighted image; d Axial T1-weighted image. The posterior fossa is markedly enlarged. Cerebellar structures are completely absent. The only recognizable structures are a thinned quadrigeminal plate, the superior medullary velum (arrows, c, d), and rudimentary middle cerebellar peduncles (MP, d). The brainstem is hypoplastic, mainly due to pontine hypoplasia. Notice that there is no hydrocephalus, and that the CSF cavity filling the posterior fossa does not displace the brainstem against the clivus, ruling out an arachnoid cyst

Pontocerebellar Hypoplasia Background

Pontocerebellar hypoplasia (PCH) (synonym: pontoneocerebellar hypoplasia) is a rare entity among posterior fossa abnormalities. The basic anatomic

abnormality involves marked size reduction of the pons, cerebellar hemispheres, and vermis. The question whether this condition results from primary hypoplasia or secondary destruction remains poorly settled, as is the case with cerebellar agenesis (see above); however, the term “hypoplasia” is preferable, at least from an imaging perspective.

164

P. Tortori-Donati, A. Rossi, and R. Biancheri

Clinical Syndromes

PCH represents a heterogeneous disorder with similar neuroradiologic features. Clinical and pathological features may differ significantly among patients with similar MRI pictures. PCH can be isolated or associated with supratentorial abnormalities (Table 4.12). PCH Type 1 of Barth (PCH-1)

PCH Type 1 of Barth is an autosomal recessive disorder characterized by neonatal onset, congenital contractures, ventilatory insufficiency, and early death [351, 352]. The pathological hallmark of PCH-1 is spinal anterior horn involvement with similarities to spinal muscular atrophy [352].

Table 4.12. Causes of cerebellar hypoplasia (modified from [357]) Viral

Cytomegalovirus

Chromosomal

Trisomy 18

Familial cerebellar Autosomal recessive hypoplasia X-linked PEHO (infantile cerebello-optic atrophy) Cerebral calcifications and cerebellar hypoplasia Biochemically defined causes

Carbohydrate deficient glycoprotein (CDG) syndromes Mitochondrial respiratory chain defects Glutaric aciduria type I

Part of a complex malformation

Lissencephaly with cerebellar hypoplasia (LCH) Congenital muscular dystrophies (WalkerWarburg, Fukuyama, muscle-eye-brain syndromes) Marden-Walker syndrome Oto-palato-digital syndrome Oral-facial-digital syndrome

PCH Type 2 of Barth (PCH-2)

PCH Type 2 of Barth is an autosomal recessive disorder, so far described only in European children. The disease features progressive microcephaly, severe lack of mental and motor development, and characteristic dyskinesia/dystonia that usually starts in infancy [351, 352]. There is no anterior horn involvement. Carbohydrate-Deficient Glycoprotein (CDG) Syndromes

CDG are a newly discovered group of inherited disorders, also called congenital disorders of glycosylation. These result from abnormal synthesis of sugar groups that are parts of glycoproteins. Most affected patients have a special physiognomy, neurological problems that include psychomotor retardation, ataxia, hypotonia, little or no head control, failure to thrive, lethargy and unresponsiveness, and liver and/or intestinal problems. PCH has been described in autopsy reports [353]. Increased serum sialotransferrin biologically marks these syndromes, thereby allowing for verification of suspected diagnoses. Mitochondrial Respiratory Chain Defects

de Koning et al. [354] described a lethal case of PCH showing clinical features similar to those of PCH-1 but without spinal anterior horn involvement, in which there was elevated lactate concentration in plasma and urine; biochemical analysis demonstrated multiple deficiencies of respiratory chain enzymes in skin fibroblast. MRI demonstrated diffuse PCH associated with abnormal signal intensity in the supratentorial white matter. Lissencephaly with Cerebellar Hypoplasia (LCH)

Although classical lissencephaly involves primarily the cerebral cortex, malformations in this spectrum can be associated with significant cerebellar

Toxic causes

i.e., phenytoin

Pontocerebellar hypoplasia (of Barth)

PCH type I PCH type II

underdevelopment, and have recently been referred to as lissencephaly with cerebellar hypoplasia (LCH) [134] (see above). Mutations in the LIS1, DCX, and RELN genes have been found to cause some LCHs, but additional classes of genes remain undetermined. The clinical spectrum involves small head circumference and cortical malformation, ranging from agyria to simplification of the gyral pattern, and from near normal cortical thickness to marked thickening of the cortical gray matter [134]. One of the LCH subtypes also displays corpus callosum agenesis [134]. Cerebellar manifestations range from midline hypoplasia to diffuse volume reduction and disturbed foliation; therefore, some cases may be better defined as having cerebellar dysplasia. Cases associated with RELN mutation display particular severity of cerebellar and hippocampal involvement [134]. Extreme cerebellar hypoplasia has been noted to occur in microlissencephalic patients [133]. Congenital Muscular Dystrophies (CMDs)

CMDs associated with structural brain abnormalities include Walker-Warburg, Fukuyama, and muscleeye-brain diseases (see above). These entities most likely represent different degree of severity along a malformation spectrum. Supratentorial abnormalities are basically represented by the cobblestone cortex, which is thought to result from failure to block neuronal migration at the pial barrier [355]. Poste-

165

Brain Malformations

rior fossa abnormalities include PCH, Dandy-Walker malformations, cerebellar cortical dysgenesis, and a characteristic posterior kink of the brainstem that is exclusively seen in the Walker-Warburg syndrome (Fig. 4.40). Neuroradiology

Hypoplasia of the vermis and cerebellar hemispheres involves marked reduction in size with a reduced number of folia (Fig. 4.78), as opposed to cerebellar atrophy, in which there is a normal number of thin folia [356], and to partial agenesis, in which there is a reduced number of folia but the present folia are normal [357]. However, such distinction may not be straightforward on imaging. On midsagittal sections, the vermis is small, but retains a relatively normal shape, with a larger posterior lobe and a smaller anterior lobe separated by a shallow primary fissure. The vermis is not rotated, and the shape of the fourth ventricle is normal, establishing a crucial differen-

tial diagnostic sign with Dandy-Walker variants [299, 319]. The cerebellar hemispheres are symmetrically small, flattened, and located close to the tentorium [356, 357]. The pons is small, with a decreased or absent bulge of its anterior surface on sagittal images [356]. Owing to the reduction in size of the cerebellum, the subarachnoid spaces are passively enlarged [357]. Thereby, there is no “cyst” in the posterior cranial fossa, allowing differentiation from MCM in which, additionally, there is a normal pons. The overall posterior fossa size is normal or slightly small, with normal or mildly verticalized tentorium. In PCH-1 and PCH-2, the abnormality is restricted to the posterior fossa with a substantial lack of supratentorial anomalies. However, all patients with PCH must be scrutinized for associated supratentorial abnormalities, basically involving lissencephalic cortices, in order to identify different etiologies. One must also remember that PCH-1 and PCH-2 are integrated clinical-neuroradiological diagnoses. Recognition of these entities is important in view of the

b a

Fig. 4.78a–c. Pontocerebellar hypoplasia. a Sagittal T2weighted image; b Coronal T2-weighted image; c Axial T2weighted image. The size of both the cerebellar hemispheres and of the vermis is diffusely reduced. Notice there is a reduced number of folia, as opposed to cerebellar atrophy, in which there is a normal number of thin folia, and to partial agenesis, in which there is a reduced number of folia but the present folia are normal. On midsagittal sections (a), the vermis is small, but retains a relatively normal shape. The vermis is not rotated, unlike in Dandy-Walker variants. The subarachnoid spaces and the fourth ventricle are dilated exvacuo. The pons is also hypoplastic

c

166

P. Tortori-Donati, A. Rossi, and R. Biancheri

possible autosomal recessive transmission, requiring genetic counseling [358]. Unilateral Hemispheric Aplasia/Hypoplasia

Unilateral cerebellar hypoplasia is probably more common than bilateral hypoplasia [295], whereas unilateral aplasia is exceptional [358]. Although it is usually a sporadic anomaly, it has been described in association with ipsilateral cutaneous organoid nevus [359], Aicardi syndrome [360], and facial hemangioma; the latter association is a part of the

PHACE syndrome spectrum [315, 316]. There often is associated inferior vermian hypoplasia, resulting in a cleft that connects the fourth ventricle with the subarachnoid spaces (Fig. 4.79). The ipsilateral cerebellar peduncles also show reduced size. The homolateral dentate nucleus and contralateral inferior olivary and pontine nuclei are hypoplastic. Contralateral underdevelopment of the brainstem is found in case of unilateral aplasia [361]. The pathogenesis is unknown and may not be uniform, with some cases possibly resulting from prenatal destructive lesions [361].

a

c

b

d Fig. 4.79a–d. Cerebellar hypo-aplasia. Two different cases. Case #1: a Sagittal T1-weighted image; b Coronal T2-weighted image. Diffuse form of cerebellar hypoplasia involving both the inferior vermis and the left cerebellar hemisphere (b). Notice that interpretation of the sagittal image may be tricky. The cleft that appears to connect the fourth ventricle with the cisterna magna (arrowhead, a) is due to hypoplasia of the inferior vermis; below that, the right cerebellar hemisphere fills the gap on the midline (CH, b). Coronal planes clear the view, showing that the inferior vermis is absent (arrows, b) and the left cerebellar hemisphere is hypoplastic. Case #2: c Axial T1-weighted image; d Coronal T1-weighted image. Partial aplasia of the right cerebellar hemisphere: the right cerebellar hemisphere is partially absent with secondary enlargement of the subarachnoid spaces. Notice that the CSF cavity does not exert mass effect; in fact, the inferior portion of the fourth ventricle is attracted to the right (arrow, c). This rules out an arachnoid cyst (compare with Fig. 4.72)

Brain Malformations

Cerebellar Cortical Malformations

Until recently, abnormalities of cerebellar foliation and fissuration were believed to be microscopic and basically detectable by histology, with a notable exception being Lhermitte-Duclos disease. However, the widespread use of high-field MR equipment has provided evidence that the incidence of cerebellar cortical malformations is higher than was previously thought. Although minor cerebellar dysplasias are commonly found on histological examination within the cerebellar white matter or vermian nodulus of otherwise normal newborns [17], larger dysplasias are obviously pathologic. They are found either as isolated entities or, more commonly, in patients with congenital muscular dystrophies, intrauterine infection, or other supra- and infratentorial malformations. Cerebellar cortical malformations have also been observed in the vicinity of posterior fossa tumors [199, 362], raising the suspicion of a possible pathogenetic association. Classification of cerebellar cortical malformations from a neuroimaging perspective is problematic. This depends basically on the fact that correlation with pathologic data is usually lacking. Three broad categories can be identified for classification purposes, i.e., (i) cortical dysplasia, corresponding to the pathological categories of heterotopia, caused by disturbed migration of neuroblasts from the rhombic lip, and dysplasia, resulting from disturbances of cortical layering and gyrus formation [17]; (ii) granule layer aplasia, which results in congenital cerebellar atrophy; and (iii) hyperplasias, a still controversial category including very diverse entities such as the LhermitteDuclos disease and macrocerebellum. Still today, there is an enduring debate as to whether Lhermitte-Duclos disease is a neoplastic, malformative, or hamartomatous condition [363]. Because the Lhermitte-Duclos disease has been found to be consistently associated with Cowden syndrome, a genetic disease characterized by multiple hamartomas and malignancies, the two conditions are presently considered to represent a single phakomatosis [364, 365]. The brain MRI manifestations of this entity are described in Chap. 17. Unilateral hemispheric enlargement can be found in hemimegalencephaly as well as in several conditions associated with somatic hemihypertrophy, including Beckwith, Klippel-Trenaunay-Weber, and RussellSilver syndromes. Finally, cases of isolated cerebellar hemihypertrophy have rarely been reported [366].

Cerebellar Cortical Dysplasia

Background

This category comprises anomalies resulting morphologically in abnormal cerebellar foliation and fissuration. Pathologically, heterotopia form abnormal clusters of gray matter in the cerebellar white matter, whose size is often microscopic: therefore, they may remain undetected on MRI [362]. True cortical dysplasia corresponds to areas of smooth, thick cortex showing interlacing nests of granular and molecular layers with irregularly arranged Purkinje cells [17]. Often, the two types of malformation are associated. Although the term cerebellar polymicrogyria is often used [362], its appropriateness has been questioned, basically from a semantic perspective [367]. In one histologically proven case, there was deep folding of the cerebellar surface, with fusion of sulci in the upper parts and entrapment of the deep sulcal portions [368], consistent with a pathological definition of polymicrogyria. Embryologically, cerebellar cortical dysplasias are believed to occur from a pathologic process involving genetically or environmentally disturbed migration of Purkinje and granule cells, improper function of cellular and interstitial signaling that guide developing cells and axons, or both [5]. An important role in the foliation and lamination of the cerebellar cortex is also played by the leptomeninges; selective destruction of the leptomeninges has been shown to result in defective foliation, granule cell ectopia, and reduction in cerebellar size [17]. Cerebellar cortical dysplasias can be isolated, but more often they are associated with other malformations, such as vermian abnormalities, corpus callosum dysgenesis, and cerebral cortical malformations [362]; as a consequence, correlation with neurologic focal symptoms or developmental/cognitive deficits is often problematic [362]. They have also been described in chromosomal abnormalities, congenital muscular dystrophies, intrauterine infection, phakomatoses (i.e., PHACE syndrome) [316], and as a result of exposure to gamma radiation and ethanol [368]. We also observed a case of isolated rostral vermian dysplasia in a child with Cohen syndrome, whose main radiologic feature is a megalic corpus callosum. When present, supratentorial abnormalities typically dominate the clinical picture [366]. Isolated dysplasias may be incidental findings, or can be found in patients with developmental delay or cognitive disorders. A clear-cut clinical-radiological correlation is difficult to establish. However, patients with isolated vermian dysplasia are more likely to be asymptomatic than those with diffuse involvement of the hemispheres.

167

168

P. Tortori-Donati, A. Rossi, and R. Biancheri

Neuroradiology

Classification of cerebellar cortical dysplasia into subtypes from a neuroradiological standpoint has been attempted by Demaerel [366] (Table 4.13). Basically, type 1 dysplasias involve the vermis and type 2 dysplasias involve the cerebellar hemispheres. Although lack of histological verification and the presence of overlapping features are substantial drawbacks, this classification offers a rationale to approach these malformations from a radiological perspective. The MRI appearance of cerebellar cortical dysplasias is variable. Basically, MRI shows abnormal foliation and fissuration. MRI signs [362] can be reconducted to three main categories, as follows. It is noteworthy that the three groups of abnormality may coexist in individual patients. Abnormal foliation/fissuration: these are the most common MR abnormalities in patients with cerebellar cortical dysplasia. MRI signs include defective, large, or vertical fissures; blurred gray-white matter interface; abnormal arborization of the white matter; and heterotopia within the cerebellar white matter (Fig. 4.80). Localized fissural malorientation, with individual folia running vertically rather than horizontally, may involve the rostral vermis [369] or one cerebellar hemisphere [370]. This probably represents a mild form of cerebellar dysplasia in patients without elicitable cerebellar neurological deficit. Cortical thickening: It often coexists with abnormally oriented folia and fissures, and results in an irregularly smooth, bumpy cerebellar surface (Fig. 4.80). This pattern may selectively involve certain areas or diffusely surround a poorly digitated white matter in both hemispheres. Diffuse cortical thickening is frequent in case of cerebellar involvement from intra-

uterine cytomegalovirus or toxoplasma infection, and may be associated with subcortical calcification. However, we have seen a case of diffuse cortical thickening involving both cerebellar hemispheres, which were also markedly hypoplastic. Subcortical cysts: A characteristic, albeit rare, MRI finding is the presence of multiple, variably sized subcortical cyst-like inclusions (Figs. 4.42, 4.43). These are isointense with CSF in all MRI sequences, and likely represent subarachnoid spaces engulfed by the fusion of disorganized folia at the boundary between normal and polymicrogyric cortices [238]. They preferentially involve the posterior portions of the cerebellar hemispheres and are associated with absence of normal folia and fissures. Subcortical cysts have been reported in congenital muscular dystrophies of the Fukuyama [238] and muscle-eye-brain type [241]; in the latter, hypoplasia of the pons is consistently associated. Recently, they have also been identified in patients without clinical or laboratory evidence of muscular dystrophy [371]. Granular Layer Aplasia

Granular layer aplasia (also called primary degeneration of the granular layer) is a rare autosomal recessive disorder that has not yet been linked to a specific etiology [17]. The original damage is destruction or depletion of precursor cells in the external granule layer, a transient germinal layer located over the surface of the fetal and neonatal cerebellum from which the granule, basket, and stellate cells of the granular layer develop as a result of inward migration. As a consequence of such destruction, the granular layer fails to develop. This is different from secondary destruction of an already developed granular layer, although such differentiation may be difficult to establish even on histological examination [17].

Table 4.13. Classification of cortical dysplasia according to Demaerel [367] Type

Radiological features

1a

– Malorientation of the fissures in the anterior lobe of the vermis Cerebellar hemispheric extension of the vermian abnormalities is uncommon

1b

– Malorientation of the fissures in the anterior lobe of the vermis – Irregular foliation of the anterior lobe and part of the posterior lobe of the vermis Cerebellar hemispheric extension of the vermian abnormalities is common; associated cerebral abnormalities may occasionally be seen

2

Cerebellar hemispheric abnormalities consisting of one or more of the following: – Cortical dysgenesis (bumpy gray/white matter junction) with or without small included cysts – Cortical hypertrophy – Aberrant orientation of the folia Type 1b changes in the vermis can be seen in more than 60% of the patients; associated cerebral abnormalities can be seen in more than 50% of the patients

169

Brain Malformations

a

b

c

d Fig. 4.80a–d. Cerebellar cortical dysplasia: two different cases. Case #1: a Sagittal STIR image; b Axial T2-weighted image. Diffuse abnormality of foliation of both cerebellar hemispheres and the inferior vermis (arrow, a). Case #2: c Sagittal T1-weighted image; d Coronal T2-weighted image. Diffuse cerebellar dysplasia involving both the cerebellar hemispheres, whose cortex shows a bumpy, thickened appearance similar to cerebral polymicrogyria (d). An islet of heterotopia is recognizable deep within the right cerebellar white matter (arrow, d). Foliation of the vermis is absent (c). The cerebellum is globally hypoplastic

Affected children exhibit characteristic clinical features, represented by hypotonia, strabismus, delayed motor development, nonprogressive ataxia, delayed language development with dysarthria, and mental retardation [372]. Symptoms are present from early childhood and remain stationary [17]. MRI detects diffuse cerebellar atrophy [372], characterized by a small cerebellum with shrunken folia and large fissures (Fig. 4.81). As such, atrophy is clearly distinguishable from cerebellar hypoplasia, in which the fissures are of normal size compared with the folia [5]. The brainstem and cerebrum are normal. Because cerebellar atrophy can be the end result of progressive

metabolic injury of different etiologies and, in several instances, its cause remains undetermined, granular layer hypoplasia can be reasonably suspected only if cerebellar atrophy is detected in a newborn. Macrocerebellum

This uncommon condition was reported by Bodensteiner in 1997 [373], and is characterized by abnormally large cerebellum with preservation of the overall shape. Cerebellar volume exceeds normal by at least 2 standard deviations. Clinically, affected patients show hypotonia, delayed motor and cogni-

170

P. Tortori-Donati, A. Rossi, and R. Biancheri

b Fig. 4.81a,b. Granular layer aplasia. a Sagittal T2-weighted image; b Coronal T2-weighted image. Shrunken appearance of the cerebellar vermis (a) and hemispheres (b). Marked reduction in cerebellar size, dilatation of the subarachnoid spaces, and defective myelination all contribute to the globally high signal intensity of the cerebellum on T2-weighted images (b). The brainstem is normal

a

b Fig. 4.82a,b. Macrocerebellum. a Sagittal T1-weighted image; b Coronal T2-weighted image. Marked increase in size of the whole cerebellum, without evidence of focal cortical dysplasia. Notice that the size of the posterior cranial fossa is normal. The fourth ventricle is small

a

tive development, and oculomotor apraxia. On MRI [374], the cerebellum is disproportionately large but appears to have normal imaging features in terms of tissue characteristics (Fig. 4.82). There is consistent association with delayed myelination of the supratentorial white matter. The etiology is unknown, but it could be related to the cerebellum responding to the elaboration of growth factors intended to augment the slow development of cerebral structures [374].

4.4.4.3 Isolated Brainstem Hypoplasia/Dysplasia

Pathological categorization of brainstem abnormalities includes dysplasias of the inferior olivary nuclei, anomalies of crossing of the corticospinal tracts, aplasia of the corticospinal tracts, hypertrophy of the corticospinal tracts, aplasia of the dorsal spinal tracts, and congenital facio- and ophthalmoplegia (Moebius syndrome) [17]. Our experience has been that iso-

171

Brain Malformations

lated brainstem abnormalities are encountered only exceptionally. What follows is a brief account of this limited experience. Brainstem Hypoplasia in Horizontal Gaze Palsy with Scoliosis

Horizontal gaze palsy with scoliosis is a rare autosomal recessive disorder characterized by congenital absence of conjugate horizontal eye movement with progressive scoliosis developing in childhood or adolescence. Affected patients compensate for their deficit by turning their heads in the desired direction, thereby obtaining regular binocular vision. Horizontal gaze palsy has been related to dysfunction of abducens nuclei and medial longitudinal fasciculus

(MLF), whereas scoliosis is thought to be caused by lower brainstem nuclear abnormalities. On MRI, the prominence normally produced by the abducens nuclei on the fourth ventricular floor is absent; the medulla is also hypoplastic and shows a butterfly configuration. In a case we observed, MRI also revealed a hypoplastic pons whose posterior two thirds were split into two halves by a midsagittal cleft extending ventrally from the fourth ventricular floor (Fig. 4.83) [374]. Embryologically, the pons is developed from the ventrolateral wall of the metencephalon approximately between gestational weeks 5 and 8. During this period, the developing fourth ventricle shows a ventral furrow that deeply indents the posterior aspect of the metencephalon. Failed development of

Fig. 4.83a–c. Brainstem hypoplasia in horizontal gaze palsy with scoliosis. a Sagittal T1-weighted image; b, c. Axial T2-weighted image. Patient is a 13-year-old girl with congenital horizontal gaze palsy and progressive idiopathic scoliosis. On sagittal planes, hypoplasia of the brainstem is evident; plus, there is a depression of the fourth ventricular floor (arrowhead, a). On axial planes at level of the medulla, the pyramids (P, b) are not prominent with respect to the inferior olives (IO, b), and the floor of the fourth ventricle is tent-shaped (arrows, b). At level of the pons, the normal prominence of the abducens nuclei is absent (arrows, c); additionally, the posterior two thirds of the pons are split into two halves by a midsagittal cleft extending ventrally from the fourth ventricular floor (“split pons sign”) (arrowhead, c)

a

b

c

172

P. Tortori-Donati, A. Rossi, and R. Biancheri

medial pontine structures including the abducens nuclei and MLF could result into persistence of an abnormally deep ventral fourth ventricular furrow. A butterfly configuration of the medulla is further evidence of maldevelopment of dorsomedial brainstem structures [374]. Segmental Brainstem Agenesis

Congenital absence of a segment of the brainstem has been reported exceptionally, and has involved the inferior pons with pontomedullary disconnection [375] or the upper pons and mesencephalon [376]. Because severe cerebellar hypoplasia was consistently associated, these cases cannot be regarded as isolated brainstem malformations; however, the pathologic and neuroimaging picture is dominated by segmental brainstem aplasia. Affected patients had severely impaired neurological conditions with prominent central respiratory disturbances, and died in early infancy. A pathogenetic role for a mutation or deletion in the EN2 gene has been speculated to exist [376]. Perhaps more frequently, one may encounter isolated hypoplasia of portions of the brainstem. In a case we observed, unilateral pontomesencephalic hypoplasia was found in a patient with congenital third cranial nerve palsy (Fig. 4.84). Hypertrophy of the Corticospinal Tracts

Unilateral hypertrophy of the corticospinal tract is usually associated with a prior destructive lesion

in the contralateral cerebral hemisphere [17, 377]. The volume of the pyramid is increased and causes it to bulge with posterior dislocation of the inferior olivary nucleus. The bulging medullary pyramid is well depicted on axial MRI sections (Fig. 4.85). The main differential concern is an intrinsic glioma. Although benign hypertrophy should theoretically yield isointense signal with normal brain, abnormal myelination has been reported to occur within the hypertrophied tract [378], and could possibly cause T2 prolongation. Absence of contrast enhancement is not sufficient to rule out low-grade tumors. Therefore, we believe close follow-up of apparent pyramidal hypertrophy with MRI is mandatory.

4.5 Chiari Malformations 4.5.1 Chiari I Malformation Background

The Chiari I malformation is a congenital hindbrain abnormality characterized by downward displacement of elongated, peg-like cerebellar tonsils through the foramen magnum into the upper cervical spinal canal (Fig. 4.86), sometimes associated with descent and distortion of the bulbo-medullary junction and often complicated by hydrosyringomyelia (HSM) and

a

b Fig. 4.84a,b. Ponto-mesencephalic hypoplasia in a 2-month-old infant with congenital left third nerve palsy. a, b. Axial T2-weighted images. The left anterolateral portion of the pons (arrow, a) and the homolateral cerebral peduncle (arrow, b) are hypoplastic

173

Brain Malformations

a

b Fig. 4.85a,b. Brainstem hyperplasia. a Axial T1-weighted image; b Axial T2-weighted image. The right half of the medulla oblongata is hypertrophic (arrow, a, b). This finding has been stable for 5 years, without evidence of MR signal modifications

hydrocephalus. Associated anomalies include osseous abnormalities at the craniovertebral junction and basilar invagination. Although this abnormality was initially described in autopsy observations, the extensive use of MRI has made clear that a condition of caudal tonsillar ectopia is much more common than was previously believed, and that such condition is often associated with a broad spectrum of clinical and neuroradiological features. However, the increase of reported cases has emphasized the need for greater understanding of the pathogenesis and clinical manifestations of this abnormality [379]; often, the questions raised by the various published articles have been more numerous than the answers given. From a terminological perspective, it should be stressed that the term “Arnold-Chiari malformation type I” is often misused to refer to caudal tonsillar displacement [380]. In fact, “Arnold-Chiari malformation” is an abandoned synonym of the Chiari II malformation, that is, malformation of the posterior cranial fossa with caudal vermian displacement associated with myelomeningocele. Pathogenesis

Contrary to what the term might suggest, this anomaly is inherently different from Chiari II or III malformations, except for the name of the Austrian pathologist H. Chiari, who first described these entities in 1891 [381]. In fact, unlike Chiari II or III malformations, the Chiari I malformation is not related with neural tube closure defects. It is believed that the anomaly may be primarily related to a disorder of

the paraxial mesoderm and, particularly, to hypoplasia of the occipital bone due to underdevelopment of the occipital somites, with reduced volume and overcrowding of the posterior cranial fossa, which contains the normally developed hindbrain [382]. The possible association with other skeletal developmental abnormalities, such as basilar invagination, further reduces the size of the posterior fossa, thereby increasing overcrowding and downward herniation of nervous structures. Although the vast majority of cases appear to occur sporadically, recent reports of increased incidence in monozygotic twins or triplets [383, 384] suggest a possible underlying genetic disorder in some cases; however, the possible risk of inheritance has not been determined yet. As was previously stated, there is a significant incidence of HSM and hydrocephalus in patients with Chiari I malformation. Hydrocephalus is found in 6%–25% of cases (Fig. 4.87) [381, 385], whereas the incidence of HSM is variable according to the various series, ranging from 37% to 75% [381, 385–387]. CSF cavities may extend over variable distances along the spinal cord (see Chap. 38). They involve more often the cervical spinal cord, but may also be holocord. They involve the medulla (i.e., syringobulbia) in a minority of patients. Significantly, hydrocephalus tends to be more common in patients who also have HSM [381]. This is due to the fact that both hydrocephalus and HSM result from impaired CSF flow at the foramen magnum due to abnormal pulsatile motion of the cerebellar tonsils, producing a selective obstruction of CSF flow from the cranial cavity to the

174

P. Tortori-Donati, A. Rossi, and R. Biancheri

a

b

c

d Fig. 4.86a–d. Chiari I malformation. a Sagittal T1-weighted image; b Sagittal T2-weighted image; c Coronal T1-weighted image; d Axial T1-weighted image. The posterior fossa is small. The peg-like cerebellar tonsils (T, a, b) herniate into the foramen magnum (arrows, a, b). Tonsillar ectopia is greater to the right (arrows, c). Note that the vermis (V, a, b) is located intracranially, and is raised by the underlying tonsils. Axial image shows crowding of the foramen magnum due to the presence of the tonsils (T, d) behind the medulla oblongata. In this case, crowding of the posterior fossa causes abnormal CSF circulation leading to hydrocephalus

Fig. 4.87. Chiari I malformation in a craniosynostotic patient, associated with hydrocephalus. Sagittal T1-weighted images. Severe ectopia of the cerebellar tonsils (T). Notice the vermis is in its normal position (V). There is associated hydrocephalus

Brain Malformations

spine during systole [388]. It has been proposed that ball-valve obstruction at the foramen magnum produces increased systolic CSF waves in the spinal compartment, thereby promoting bulk flow of CSF into the spinal cord through dilated perivascular spaces, resulting in HSM [381, 389]. Thickening, hyalinization, and calcification of the arachnoid and dura mater, possibly resulting from chronic friction of the cerebellar tonsils against bone, may further constrict the herniated hindbrain at the foramen magnum, thereby contributing to clinical symptoms and the development of HSM and hydrocephalus [390]. Rupture of these dural bands could account for the cases of spontaneous regression of a Chiari I malformation that have been, albeit rarely, reported in the literature [391, 392]. The issue of the so-called “acquired” Chiari I malformation also must be addressed. It is known that caudal herniation of the hindbrain, indistinguishable from the Chiari I deformity, may occur secondary to CSF hypotension, either iatrogenic (i.e., after the establishment of ventricular and subarachnoid shunts or multiple lumbar taps) [393–396] or spontaneous [397], as well as to raised intracranial pressure (i.e., due to intracranial tumors) [398, 399] or venous hypertension (i.e., arteriovenous malformations). Typically, there is complete reversal of tonsillar descent after surgical treatment of the predisposing condition, as is documented by follow-up MRIs [400]. We believe that these “acquired” conditions should be clearly distinguished from the congenital Chiari I deformity, and that the term “acquired Chiari I malformation” is a misnomer that should be abandoned in favor of the more correct “acquired tonsillar herniation.” Clinical Findings

Although the Chiari I malformation has been referred to as “adult-type Chiari malformation” because a significant proportion of affected patients present in adulthood, it is in fact a congenital abnormality that can be, and in fact often is, discovered in childhood. The advent of MRI has played an unique role in the increased recognition of this abnormality, which often is discovered incidentally; in fact, a significant proportion of children [14%–56% of cases] [386, 387, 401] is neurologically normal at presentation, and the diagnosis is made when an MRI is performed for another reason. Moreover, a significant degree of tonsillar ectopia (i.e., up to 12 mm) may be found in completely asymptomatic children [402]. Therefore, the clinical significance and appropriate management of these patients often remains uncertain

[402]. The presentation of Chiari I malformation is often characterized by vague and ambiguous symptoms. Many patients are misdiagnosed with conditions such as multiple sclerosis, fibromyalgia, chronic fatigue syndrome, or psychiatric disorders. Affected patients frequently experience symptoms months to years prior to diagnosis, and some may present after flexion injuries to the neck from chiropractic cervical manipulation [403, 404]. Initial symptomatology often is related to hindbrain compression and includes headache, neck pain, numbness, weakness, incoordination, nystagmus, and ataxia, among other symptoms [381, 386, 387, 402]. There is a general consensus in that the degree of tonsillar descent grossly correlates with the presence of signs and symptoms [402, 405]; however, cranial and brain measurements alone have not been found to correlate significantly with the prognosis and the incidence of complications [406]. When HSM is present, a host of additional symptoms that relate to spinal cord involvement become prominent and generally overshadow other clinical manifestations. Scoliosis is found in 28% of cases [387]. Signs of raised intracranial pressure may develop in cases of associated hydrocephalus. Imaging Findings

Measurement of the degree of downward displacement of the cerebellar tonsils is critical in order to diagnose a Chiari I malformation. The extent of this ectopia is measured on a midsagittal MR image from the tips of the cerebellar tonsils to a line drawn from the basion to the opisthion (Fig. 4.88) [383, 386, 407]. There has been a significant amount of controversy concerning the exact degree of tonsillar ectopia that should be regarded as abnormal. It is generally accepted that displacement of the cerebellar tonsils by less than 3 mm is normal. Aboulezz et al. [407] considered herniation of at least one cerebellar tonsil 5 mm or more below the foramen magnum to be pathologic, and between 3 and 5 mm to be borderline. Elster et al. [386] considered 3 mm herniations to be definitely pathologic if accompanied by other features such as syrinx, cervicomedullary kinking, or elongated fourth ventricle. Milhorat et al. [381] used 5 mm as a cut-off; however, 9% of their patients exhibited tonsillar herniation of less than 5 mm despite having symptoms that were considered to be typical of Chiari I malformation. Other authors have correlated the degree of tonsillar descent to patient age [408]; in general, the cerebellar tonsils ascend with increasing age. In the first decade of life, 6 mm is pathologic. In the second decade, this decreases to 5 mm. Further decreases occur in the following

175

176

P. Tortori-Donati, A. Rossi, and R. Biancheri

a

b Fig. 4.88a,b. Measurement of tonsillar ectopia on MRI. a Sagittal T1-weighted image in a normal child; b Sagittal T1-weighted image in a child with Chiari I malformation. A line is drawn to connect the basion (B) to the opisthion (0) on sagittal images. The extent of tonsillar ectopia is measured drawing a perpendicular line from the tip of the tonsils to the B–O line. Notice that in the normal individual (a), the tip of the tonsils is above the B–O line. Instead, this Chiari I individual (b) has marked tonsillar descent (9.3 mm below the B–O line). In our experience, we have used 3 mm below the B–O line as a cut-off

decades. In our clinical experience, in which all patients with a neuroradiological diagnosis of Chiari I malformation have received a thorough neurological and neurosurgical assessment, we have used 3 mm as a cut-off for normal individuals in the pediatric age group, and considered ectopia greater than 5 mm to be definitely pathological. Mild (i.e., 3–5 mm) ectopia was considered significant in presence of: (i) neurological signs or symptoms that can be related to hindbrain compression; (ii) peg-like deformation of the tonsils; (iii) a syrinx. Therefore, we believe that all individuals with 3–5 mm displaced tonsils should be carefully assessed neurologically. Asymmetric tonsillar herniation is common; unilateral tonsillar ectopia is considered significant when it exceeds 5 mm than the contralateral side [386]. Coronal MR images are especially useful to assess asymmetric tonsillar herniation (Fig. 4.86). Overcrowding of the posterior cranial fossa produces a number of additional features, including: (i) effacement of the CSF spaces in the posterior fossa, and especially of the vallecula and cisterna magna, which represents a useful clue on axial CT and MRI images; (ii) compression of the fourth ventricle; (iii) compression of the anterior aspect of the medulla by a retroflexed odontoid process; and (iv) downward displacement of the medulla in a significant amount of

cases, with the ponto-medullary junction approaching the basion. In a minority of cases, a true posterior cervicomedullary kinking, similar to that seen in Chiari II malformations, may be seen in association to tonsillar descent. These latter cases have been referred to as the “bulbar” or “myelencephalic” variant of the Chiari I malformation (Fig. 4.89) [385, 409]. This variant probably results from a smaller posterior fossa as compared to classical Chiari I malformation, with increased degree of neural overcrowding and downward displacement of the medulla. It is noteworthy that associated skeletal anomalies, such as platybasia and basilar invagination, also are relatively more common in patients with the bulbar variant than in classical Chiari I’s (P. Tortori-Donati and A. Rossi, unpublished observations). Clinically, patients with the bulbar Chiari I variant are less likely to be asymptomatic than patients with the classical Chiari I malformation. Because HSM is a significant contributor to the clinical picture and to the eventual outcome, it is critical that all patients with a diagnosis of Chiari I malformation undergo MR imaging of the spinal cord. The reverse is also true, in that all patients with an apparently isolated HSM should be investigated for a possible Chiari I malformation.

177

Brain Malformations

b Fig. 4.89a,b. Chiari I malformation, myelencephalic type. a Sagittal T1-weighted image; b Coronal T1-weighted image. Cerebellar tonsils are ectopic bilaterally (T, a, b). The vermis is in its normal position (V, a). Note elongation of the medulla oblongata (arrowheads, a) and cervicomedullary kinking (arrow, a)

a

Alterations in craniospinal brain and cord motion were discovered before the advent of MRI. In a myelographic study conducted in Chiari I patients [410], crowding of the neural structures with abnormal, abrupt downward systolic displacement of the brainstem and cerebellar tonsils resulted in a plugging effect with narrowing of the CSF pathways at the foramen magnum, whereas diastolic recoil of the nervous structures disimpacted the foramen magnum. Flow-sensitive phase-contrast MRI studies have been used to demonstrate noninvasively anomalies in CSF flow at the foramen magnum both in normal individuals and Chiari I patients, and to assess the benefits of decompressive surgery [388, 411–413]. In normal subjects, arrival of the systolic arterial pulse increases the cerebral blood volume, thereby displacing a corresponding amount of CSF from the basilar cisterns to the cervical subarachnoid space; then, cerebral venous outflow and recoil of the spinal dura mater produce an opposite, caudocranial diastolic CSF wave. Analyses of CSF flow in Chiari I patients have yielded controversial results, mainly due to the different techniques that were used by the various investigators. Some authors [388, 411] demonstrated selective obstruction of systolic CSF flow from the cranial cavity to the spine with normal upward diastolic CSF flow in Chiari I patients, whereas others [413] detected increased systolic caudal and diastolic cranial motion of the spinal cord, with unchanged

systolic and impaired diastolic CSF flow. Motion-sensitive MRI techniques detect restoration of normal flow dynamics after successful decompressive surgery [411]. Although of interest in the understanding of the pathogenesis of this condition, MRI flow studies have not yet been successfully incorporated into treatment decision-making. However, they could be helpful in demonstrating disturbances of CSF velocity and flow at the foramen magnum in patients with mild tonsillar ectopia, thereby facilitating the diagnosis in cases where conventional MRI findings are doubtful. As was previously stated, skeletal abnormalities, such as platybasia, basilar invagination, atlantooccipital assimilation, craniosynostosis, and fused cervical vertebrae, are a frequent finding [24% of cases] [386]. Associated brain malformations are much less common. In a series of 147 patients, Elster et al. [386] found only single cases in whom associated callosal lipoma, septo-optic dysplasia, and aqueductal stenosis were identified. A Chiari I malformation is found in 20% of patients with idiopathic growth hormone deficiency, in association with the typical picture of small anterior pituitary, absent stalk, and ectopic posterior pituitary [414], indicating a possible common embryologic derangement [414, 415]. An association with midline germ cell tumors of the suprasellar and pineal regions has also been described [416].

178

P. Tortori-Donati, A. Rossi, and R. Biancheri

4.5.2 Chiari II Malformation Background

The Chiari II malformation is a congenital abnormality of the hindbrain characterized by a smaller than normal posterior cranial fossa with downward displacement of the vermis, brainstem, and fourth ventricle into the foramen magnum and cervical spinal canal. A host of other abnormalities of the skull, brain, spine, and spinal cord may be associated, thereby generating a complex and polymorphous picture. There has been much confusion and disagreement in the literature concerning the definition of this malformation [417]. Among the numerous, often misleading denominations that were proposed in the past, the term “Arnold-Chiari malformation” gained much popularity, and is still today used by many as a synonym of “Chiari II malformation.” In our day-today practice, we have favored the latter denomination for the sake of a uniform classification following the original paper by Chiari [381]. The Chiari II malformation is found in all patients with open spinal dysraphisms (i.e., myelomeningoceles and myeloceles) [385, 409, 418]. This association is consistently present [418], and has been explained pathogenetically by McLone and Knepper on the basis of defective primary neurulation [419]. Consequently, the Chiari II malformation and open spinal dysraphisms are not merely “associated”; rather, they should be regarded as two aspects of a single disease entity [385, 409, 418]. Because the severity of the hindbrain malformation may be extremely variable, a number of patients may have a nearly normal-sized posterior fossa with minimal, or even absent caudal hindbrain displacement. Indeed, subtle, minimal Chiari II features will be present in all individuals with open spinal dysraphisms, and should therefore be actively sought [385, 409, 418]. Although all patients with open spinal dysraphisms harbor a Chiari II malformation, the reverse is not true. Indeed, although greater than 90% of patients with Chiari II malformation have open spinal defects, a minority of cases are discovered in patients with closed spinal dysraphisms. However, these involve exclusively the exceedingly rare myelocystoceles [418] (see Chap. 39). Pathogenesis

The pathogenesis of the Chiari II malformation has attracted the interest of numerous investigators, who conceived a large number of theories to explain its

kaleidoscopic manifestations [385]. Description of all these theories is beyond the scope of this presentation. At present, the “unified theory” proposed in 1989 by McLone and Knepper [419] is widely accepted as the most satisfactory explanation of the diverse manifestations of the Chiari II malformation and open spinal dysraphism. According to this theory, the medial walls of the primitive central canal of the neural tube (“neurocele”) normally appose and occlude the neurocele transiently during primary neurulation. In experimental mice, failure to occlude the neurocele is caused by the same defect involving failure of neurulation with formation of a myelomeningocele, i.e., defective biosynthesis of cell surface glycoproteins. Because the neurocele is patent, CSF flows downwards and leaks freely through the spinal defect into the amniotic cavity. This results in chronic CSF hypotension with collapse of the developing ventricular system. Consequently, the rhombencephalic vesicle (developing fourth ventricle) fails to expand, which causes lack of induction of the perineural mesenchyme of the posterior cranial fossa. Both the cerebellum and brainstem eventually are forced to develop within a smaller than normal posterior fossa. Consequently, the pontine flexure does not develop, and the cerebellum, medulla, and cervicomedullary junction herniate caudally through the foramen magnum. Moreover, the cerebellum also is displaced cephalically through the tentorial incisura. CSF hypotension in the supratentorial brain may impair neuronal migration, callosal development, and formation of the skull, producing various associated malformations of the nervous tissue and osteomeningeal coating. Clinical Features

The most common early symptom of this condition is respiratory stridor, often occurring within 1 or 2 weeks of birth. Stridor usually disappears spontaneously within a few days, or at most 3 months. On occasion, it may be associated with signs of hindbrain dysfunction, such as difficult swallowing, intermittent apnea, aspiration, cessation of breathing, and arm weakness. There appears to be no significant statistical correlation between the degree of caudal displacement of the brainstem and the clinical status [420]. Indeed, symptoms such as breathing and swallowing difficulties are probably related to the disorganization of brainstem nuclei rather than to mechanical distortion [421]. Hydrocephalus presents with signs of raised intracranial pressure, that include bradycardia, opisthotonus, hypertonic upper limbs, hyperreflexia, headache, and seizures. Because hindbrain

Brain Malformations

dysfunction is a leading cause of death, some patients are eligible for surgical decompression, a procedure that involves removing the posterior arch of the upper spine overlying the malformation. Because of the significant risks associated with this procedure, including profuse intraoperative bleeding, infection, and CSF fistula, the indications must be carefully evaluated. Brainstem evoked potentials are a powerful tool in the clinical assessment of these patients and may be helpful to better identify candidates for surgery. Imaging Findings

The Chiari II malformation is characterized by a host of pathological features that generate a complex picture on neuroradiological investigations. Although caudal hindbrain displacement is the principal abnormality, there is a wide array of anomalies that involve the supratentorial compartment, the skull and meninges, and the spine and spinal cord. As with other brain malformations, MRI is the single best neuroimaging modality. However, CT still plays a role in the depiction of bony abnormalities, other than in the follow-up of patients undergoing CSF shunting procedures. Skull and Cervical Spine

As was previously detailed, the principal osseous abnormality is represented by a smaller than normal posterior cranial fossa, resulting from abnormal development of the occipital somites due to insufficient expansion of the rhombencephalic vesicle. Other osteomeningeal abnormalities include lacunar skull, or “lückenschädel,” scalloping of the petrous bones, clivus, and odontoid process, and widening of the foramen magnum and upper cervical canal. These abnormalities represented a mainstay of the radiological diagnosis before the advent of CT and MRI. Lacunar skull, or craniolacunia, is a classic Xray sign of the Chiari II malformation. It is found in 55%–85% of cases [385], and is represented by patchy, irregular thinning of the calvarium, producing multiple defects that involve both the inner and outer tables. In the most severe cases, a thin fibrous layer is the only coverage to the underlying brain. This transient honeycomb appearance of the skull is already present during the 7th gestational month, and is clearly detected in the neonatal period; subsequently, the lacunae tend to diminish in size and progressively disappear after the 6th month of life. Because they result from deranged ossification of the membranous bone, they are unrelated to a condition of raised intracranial pressure [422], and their shape bears no resemblance to that of the underlying convolutions.

Scalloping of the bones encircling the posterior fossa is well depicted by CT. Concavity of the posterior wall of the petrous ridges is caused by the relentless pressure exerted by the cerebellum. Inconsistent in neonates, it may become prominent in older children and shorten the internal auditory canals. The foramen magnum is enlarged in approximately three fourths of cases. The posterior margin of the odontoid process is frequently scalloped, and the upper portion of the cervical canal is widened. A paradoxical widening of the subarachnoid spaces anterior to the cord may result from a combination of mesodermal dysplasia and abnormal CSF pulsation caudal to the plugged foramen magnum [385]. The occipital squama is flat, and may be tilted posteriorly. The posterior arch of the atlas is often incompletely developed; in such case, the intervening gap is bridged by a thick fibroelastic band that constricts the underlying nervous structures. Meningeal abnormalities mainly involve the tentorium, great cerebral falx, and falx cerebelli. The tentorium is verticalized and attaches to the occipital squama close to the foramen magnum, thereby contributing to reduce the overall size of the posterior fossa. The resulting low position of the torcular and transverse sinuses is a major risk factor for bleeding during decompression surgery, and must be carefully assessed by MR angiography. The great cerebral falx is consistently hypoplastic and fenestrated, resulting in multiple interdigitations of the cerebral hemispheres along the midline. The falx cerebelli is also consistently absent [423]. Posterior Cranial Fossa

Because the posterior fossa is smaller than normal, there is a lack of vital space for the nervous structures of the hindbrain, that are necessarily squeezed out of the posterior fossa during their growth. This process results in the pathological hallmark of the Chiari II malformation, i.e., a caudal displacement of the inferior vermis, medulla, cervicomedullary junction, and fourth ventricle into the foramen magnum and upper cervical canal, forming a cascade of herniations that is particularly well depicted on sagittal MR images (Fig. 4.90). As a result, there is crowding of the foramen magnum with chronic constriction exerted by bony and fibrous structures on the herniating brain. This may result in mechanically induced ischemia of the inferior vermis and cervicomedullary junction, revealed by increased signal on T2-weighted MR images (Fig. 4.91). Chronic mechanical compression may eventually result in a “vanishing cerebellum” [424], a condition that must be differentiated from

179

180

P. Tortori-Donati, A. Rossi, and R. Biancheri

b

a

Fig. 4.90a–c. Chiari II malformation in a newborn before shunting. a Sagittal T1-weighted image; b Axial T2-weighted image; c Coronal T2-weighted image. The membranous posterior fossa is small (black arrowheads, b, c). Midsagittal image shows caudal ectopia of the vermis, the so-called cerebellar peg (open white arrow, a); the fourth ventricle is reduced to a thin fissure (black arrowhead, a). On axial planes, the cerebellum has a hearts shape and tends to engulf the brainstem (thin arrows, b). Additional features include tectal beaking (white arrowhead, a), thickening of the interthalamic mass (thin white arrow, a), and dilatation of the suprapineal recess of the third ventricle (asterisk, a). Stenogyria is seen in the occipital region (open black arrows, b). Severe hydrocephalus is present (a, c)

c

Fig. 4.91a,b. Chiari II malformation in a newborn. a Sagittal T2-weighted image; b Axial T2-weighted image. Sagittal image shows similar features to those described in the previous case (compare with Fig. 4.90). Abnormal hyperintensity of the cerebellar peg is related to ischemic phenomena due to mechanic compression (arrow, a). The fourth ventricle is thin and flattened, and tectal beaking is more evident than in the previous case. Axial image shows another typical feature of Chiari II malformation, i.e., interdigitation of the mesial surface of the hemispheres due to focal fenestration of the falx, secondary to mesodermal dysplasia (arrows, a)

a

b

Brain Malformations

primary cerebellar hypoplasia associated with Chiari II (see below, Chiari IV malformation). Caudal vermian herniation. The herniating vermian lobules are usually represented by the nodulus, uvula, and pyramid. They extend caudally, stretching posterior to the medulla and spinal cord for a variable length, forming the so-called cerebellar peg. In most instances, the peg reaches the C2–3 level. Rarely, it extends further caudally to the lower cervical or upper thoracic vertebrae. Exceptional cases of spinal cerebellar ectopia have been described in patients with Chiari II malformation [425], and also as isolated cases [426]. It is important to recognize that the herniating midline cerebellar structure is represented by the vermis. This marks an important difference from the Chiari type I malformation, in which the cerebellar tonsils, and not the vermis, herniate into the foramen magnum. In Chiari II malformation, the cerebellar tonsils are located lateral to the medulla and may abut, but usually do not cross, the foramen magnum. Brainstem herniation. Downward displacement of the brainstem results in a low position of the pons and pontomedullary junction, which approaches the foramen magnum. The pons is also reduced in its anteroposterior diameter. The medulla herniates caudally through the foramen magnum, and so does the bulbomedullary junction. Also the cervical spinal cord tends to be displaced caudally, but the extent of such displacement is limited by the dentate ligaments, which attach to the lateral surface of the spinal cord and hold it in place [427]. As a consequence, the medulla buckles posterior to the spinal cord, forming the cervico-medullary kink. The pivot of the kink is represented by the inferior portion of the gracile and cuneate tubercles. The cervico-medullary kink

a

b

is located anterior to the cerebellar peg, and reaches further caudally than the peg in the vast majority of cases. Fourth ventricle. The fourth ventricle is usually elongated and stretched vertically, and often its inferior portion herniates, together with the choroid plexus, into the foramen magnum. Sometimes, a localized dilatation is found at its bottom end. The size of the fourth ventricle varies depending on the severity of the malformation, from slit-like or completely effaced in severe forms to normal or near-normal in minimal forms. However, one should note that an isolated fourth ventricle may look deceivingly normal in Chiari II patients. Indeed, such condition should suggest the search for shunt malfunction or other causes of fourth ventricular entrapment. Classification of the hindbrain deformity. The cervicomedullary deformities were categorized by Emery and MacKenzie in 1973, and further revised by Wolpert et al. in 1988 [420] (Fig. 4.92). In the type 1 deformity, the fourth ventricle and medulla do not descend through the foramen magnum, and the only deformity is a cerebellar peg through the foramen magnum. In type 2, the fourth ventricle descends vertically through the foramen magnum in front of the cerebellar peg. In type 3, the medulla is buckled below the spinal cord, forming the cervicomedullary kink behind the cord itself. This type is further categorized depending on whether the fourth ventricle is collapsed (3a) or dilated (3b). Lateral and cranial cerebellar herniation. The cerebellum is not only squeezed inferiorly but also extends laterally, thereby effacing the cerebellopontine angle cisterns. In some cases, the brainstem is completely

c

d

Fig. 4.92a–d. Chiari II malformation: classification of hindbrain deformity. Type 1 (a): The fourth ventricle and medulla do not descend through the foramen magnum, with the only deformity being an inferior vermian peg through the foramen magnum. Type 2 (b): The fourth ventricle descends vertically through the foramen magnum in front of the vermian peg; cervico-medullary kinking may be present. Type 3a (c): Cervico-medullary kinking is associated with a collapsed fourth ventricle. Type 3b (d): Cervico-medullary kinking is associated with a dilated fourth ventricle

181

182

P. Tortori-Donati, A. Rossi, and R. Biancheri

engulfed by the cerebellum, and a stripe of cerebellar tissue is visible in front of the brainstem on sagittal MR images (Fig. 4.93). Axial images will show the basilar trunk to be completely engulfed between the brainstem posteriorly and the cerebellum anterolaterally. As it wraps around the brainstem, the cerebellum stretches and displaces the cisternal segment of various cranial nerves, and especially the 7th–8th cranial nerve complex, which may result in hypoacusia and vestibular dysfunction. Often, the superior vermis is also displaced cranially through a widened tentorial incisura, resulting in the so-called towering cerebellum in coronal MR images [385]. Such feature is seen more often in children with long-standing ventriculoperitoneal shunts (Fig. 4.94). In fact, supratentorial ventricular dilatation usually prevents significant superior cerebellar herniation. After CSF diversion, the herniating vermis displaces laterally the temporal and occipital lobes, thereby enlarging both the interhemispheric fissure and supravermian carrefour [385]. As it squeezes through the tentorial hiatus, the vermis is distorted, and the orientation of the sulci may become abnormally sagittal. Supratentorial Abnormalities

In addition to their hindbrain deformity, patients with Chiari II malformation typically display various abnormalities of the supratentorial brain.

Tectal beak. The most consistent of these is represented by an abnormality of the quadrigeminal plate, best detected on sagittal images, which results from fusion and hypoplasia of the superior quadrigeminal tubercles and stretching and distortion of the inferior quadrigeminal tubercles. As a result, the quadrigeminal plate stretches posteroinferiorly, forming an elongation designated the tectal beak (Figs. 4.90, 4.94) [427]. The size of the beak is variable depending on the severity of the malformation. In case of extreme crowding, the beak lodges between the two cerebellar hemispheres, whereas only a thickening of the inferior quadrigeminal tubercles will be visible in minor cases [385]. Stenogyria. The gyral pattern is often abnormal, displaying multiple small gyri with normal-thickness, histologically normal cortex. This feature, designated stenogyria (Fig. 4.94) [428], is most prominent over the medial surface of the posterior parietal and occipital lobes, and is therefore best detected on sagittal MR images. Care should be taken not to mistake this abnormal gyral pattern with cortical malformations, such as polymicrogyria. Indeed, malformations of the cerebral cortex have been an exceptional finding in our experience, and have been mainly represented by isolated subependymal heterotopic nodules.

b

a Fig. 4.93a,b. Chiari II malformation before shunting. a Sagittal T1-weighted image; b Axial T2-weighted image. Sagittal image shows isointense tissue located anteriorly to the brainstem (H, a). As shown on axial images, this tissue is the anterior portion of one cerebellar hemisphere that engulfs the brainstem (arrows, b). Midsagittal image also shows the so-called “accessory lobe” above the vermis (asterisk, a), corresponding to the medial portion of the occipital lobes. All these features are related to crowding and stretching of nervous structures

183

Brain Malformations

a

b Fig. 4.94a,b. Chiari II malformation after ventriculoperitoneal shunting. a Coronal T1-weighted image. b Sagittal T1-weighted image. The cerebellum tends to extend rostrally through an enlarged tentorial hiatus and is located between the mesial surface of the two hemispheres (“towering cerebellum”) (arrows, a). Sagittal image shows dysgenesis of the corpus callosum (arrowheads, b), diffuse parieto-occipital stenogyria (asterisks, b), and marked tectal beaking (arrow, b). The caudal portion of the fourth ventricle is dilated (D, b) and trapped between the cerebellar peg and the caudalized medulla

Accessory lobe. The apposed medial surfaces of the temporal lobes often penetrate between the vermis and the straight sinus to create an apparently independent cerebral lobe, designated the accessory lobe [385], which is often visible on midsagittal scans (Fig. 4.93).

the third ventricle (Figs. 4.90, 4.93). The latter feature often persists after CSF shunting (Fig. 4.94), and is found in the vast majority of Chiari II patients. Dilatation of the lateral ventricles is often asymmetrical and prevails posteriorly.

Callosal dysgenesis. The corpus callosum is often abnormal, either because of true dysgenesis or secondary to mechanical distortion and hydrocephalus. Callosal distortion is often remarkable after CSF shunting and can result in bizarre configurations (Fig. 4.94), such as a brace shape [385].

As was previously mentioned, there is a spectrum of variable severity in the neuroimaging manifestations of Chiari II malformations, so that individual cases may look significantly different from one another. The principal determining factor is the size of the posterior fossa, which in turn is determined by the degree of expansion of the rhombencephalic vesicle during early gestation. In less severe cases, the size of the posterior fossa may look normal, no cerebellar herniation is present, and the fourth ventricle has a normal shape. In such condition, superficial observation could dismiss the picture as normal. Indeed, subtle Chiari II signs are invariably present; these typically include thickening of the inferior quadrigeminal tubercles, dilated suprapineal recess, presence of an accessory lobe, and low position of the pontomedullary junction (Fig. 4.95) [385].

Hydrocephalus. Supratentorial hydrocephalus is a consistent finding in neonates with Chiari II malformation. It can be present at birth (Fig. 4.90), but usually develops within 72 hours of myelomeningocele surgery. Therefore, CSF diversion is required in all these patients. Shunt malfunctions, cord tethering, trapped fourth ventricle, and hydrosyringomyelia are frequent causes of relapsing ventricular dilatation. Because these children often undergo CT scanning for evaluating ventricular size, care should be employed to limit X-ray exposure to the minimum. Apart from hydrocephalus, the shape of the third and lateral ventricles often is abnormal. Sagittal MR images typically show verticalization of the lamina terminalis and dilatation of the suprapineal recess of

Minimal Features

Spinal Cord

The typical coexistence with open spinal dysraphisms has been described earlier. These malformations are described in Chap. 39. In operated patients, hydro-

184

P. Tortori-Donati, A. Rossi, and R. Biancheri

a Fig. 4.95. Chiari II malformation: minimal form in a child with prior myelomeningocele surgery. Sagittal T1-weighted image. At first glance, no malformative features seem to be present. Posterior fossa size is grossly normal, and the cerebellum is completely located intracranially with no vermian herniation. The fourth ventricle is perhaps slightly larger than normal. However, closer inspection reveals a number of subtle Chiari II features, including dilatation of the suprapineal recess of the third ventricle (asterisk), hypertrophy of the inferior colliculi (thin arrow), and accessory lobe (open arrow). Moreover, there is dysgenesis of the corpus callosum (arrowheads)

syringomyelia often develops as a combined result of cord tethering and impacted foramen magnum. Elongation of these cavities may vary from localized to holocord. Scoliosis often develops as a result of cord tethering and may be aggravated by hydrosyringomyelia. Spinal cord imaging of treated myelomeningoceles often is very difficult due to the extreme spinal deformity, and may require 3D imaging with curvilinear reconstructions.

4.5.3 Chiari III Malformation The original paper by Chiari [381] included the description of a case of cervical spina bifida combined with multiple hindbrain anomalies. Since then, the definition of the Chiari III malformation has been expanded to include patients with herniation of the hindbrain in a low occipital and/or high cervical cephalocele in combination with pathological and imaging features of Chiari II malformation (Fig. 4.96) [20]. This condition has a high early mortality rate, and causes severe neurological deficits in survivors [20, 409].

b Fig. 4.96a,b. Chiari III malformation. a Sagittal T1-weighted image; b Sagittal T2-weighted image. There is an inferior occipital meningoencephalocele, containing both CSF (M, b) and cerebellar tissue (arrow, a, b). The latter is slightly T2 hyperintense, possibly due to either dysplastic or ischemic changes. There is concurrent dysgenesis of the corpus callosum (arrowheads, a), hypertrophy of the quadrigeminal plate (T, b), and hypoplasia of the pons (P, b)

The relative paucity of reported cases, with the largest series so far being that of Castillo et al. [20], which comprises 9 cases, still leaves some uncertainty as to the distinction from isolated occipital or occipitocervical cephaloceles and cervical myelocystoceles or meningoceles. In our opinion, the following conditions must be met in order to classify a posterior cephalocele under the Chiari III heading:

185

Brain Malformations

Location: the cephalocele consistently involves the occipital bone below the inion, with possible extension into the high cervical (C1–3) spine; purely cervical cephaloceles are exceptional [429], and as such, they must contain cerebellar tissue in order to match the Chiari III definition. Contents: the cephalocele must contain at least a part of the cerebellum, but may also contain the occipital lobe(s) and the medulla and pons. If not herniated, the

brainstem is usually at least distorted and stretched posteriorly. The fourth and lateral ventricles and basilar cisterns may also herniate into the cephalocele, and may be disproportionately dilated compared to the intracranial CSF spaces. Associated Chiari II features: the posterior fossa is disproportionately small, with at least one of the following associated elements: caudal displacement of the brainstem; tectal beak; lateral overgrowth of

a

c

b Fig. 4.97a–d. Chiari IV malformation in two different patients, both with prior history of myelomeningocele repair. Case #1: a Sagittal T1-weighted image; b Axial T2-weighted image. There is moderate cerebellar hypoplasia without crowding of the posterior fossa. The brainstem is caudalized (M, midbrain; P, pons), with the whole midbrain lying below an axial plane drawn perpendicularly to the dorsum sellae. Case #2: a Sagittal T1-weighted image; b Axial T2weighted image. There is marked cerebellar hypoplasia, with only a rudiment of the cerebellum visible (C) lateral and posterior to the midbrain (M). The pons (P) is markedly hypoplastic. Also in this case there is no significant crowding of the posterior fossa, although the latter is very small. The brainstem is significantly displaced caudally, as in the previous case

d

186

P. Tortori-Donati, A. Rossi, and R. Biancheri

the cerebellum into the cerebellopontine angle cisterns; enlarged interthalamic mass; dilated suprapineal third ventricular recess; and corpus callosum dysgenesis. Hydrocephalus may caused by associated aqueductal stenosis. As such, the Chiari III malformation may be viewed as a severe variant of occipital or occipitocervical cephalocele, in which the large extent of the extracranial herniation prevents adequate distension of the primitive rhombencephalic vesicle with subsequent development of a small posterior fossa. As with other cephaloceles, the herniated nervous tissues are believed to be nonfunctioning. Indeed, pathological examinations of resected specimens show areas of pressure necrosis, heterotopias, gliosis, meningeal inflammation, and fibrosis [20]. One should also be aware of the fact that venous drainage pathways are often aberrant, and major venous structures are often contained within the herniated sac. Therefore, MR angiography is a very useful tool to demonstrate venous anatomy prior to surgery, in order to avoid potentially life-threatening complications.

4.5.4 Chiari IV Malformation In his 1896 paper, Chiari described a fourth type of hindbrain anomaly, corresponding to severe cerebellar hypoplasia [430]. However, the term “Chiari IV malformation” was discarded by subsequent investigators. In 1996, Tortori-Donati [385] reintroduced the term to designate the association of severe cerebellar hypoplasia with Chiari II signs in patients with myelomeningocele (Fig. 4.97). Such association appears to represent a well-defined, albeit small, patient subgroup within the Chiari II malformation complex. Pathological and imaging findings of this condition include absent or severely hypoplastic cerebellum, small brainstem, and large posterior fossa CSF spaces, as opposed to caudal cerebellar displacement and collapsed posterior fossa CSF spaces seen in typical Chiari II patients [385]. These features make it questionable whether the embryogenetic theory of the Chiari II malformation can be used to explain also the occurrence of a Chiari IV malformation. One should be aware of the fact that extreme crowding of nervous structures into a small posterior fossa may result in pressure necrosis of the cerebellum, a condition that has been termed the “vanishing cerebellum” [424]. It is difficult to determine whether this clastic theory is fully tenable. Indeed, the cerebellar peg may deteriorate from long-standing compression at the foramen magnum; however,

it often is difficult to conclusively determine whether patients born with a small cerebellum have primary hypoplasia of congenital atrophy. We tend to favor a condition of primary cerebellar hypoplasia because of the rudimentary shape of the cerebellar remnant with little appreciation of intervening sulci.

References 1. Norman MG, McGillivray BC, Kalousek DK, Hill A, Poskitt KJ. Congenital Malformations of the Brain. Pathological, Embryological, Clinical, Radiological and Genetic Aspects. New York: Oxford University Press, 1995. 2. van der Knaap MS, Valk J. Classification of congenital abnormalities of the CNS. AJNR Am J Neuroradiol 1988; 9:315–326. 3. Sarnat HB. Molecular genetic classification of central nervous system malformations. J Child Neurol 2000; 15:675– 687. 4. Sarnat HB, Flores-Sarnat L. A new classification of malformations of the nervous system: an integration of morphological and molecular genetic criteria as patterns of genetic expression. Eur J Paediatr Neurol 2001; 5:57–64. 5. Patel S, Barkovich AJ. Analysis and classification of cerebellar malformations. AJNR Am J Neuroradiol 2002; 23:1074– 1087. 6. Galatioto S, Longo M, Tortori-Donati P. Embriologia dell’encefalo, teratologia e classificazione delle malformazioni del sistema nervoso centrale. In: Tortori-Donati P, Taccone A, Longo M (eds) Malformazioni cranio-encefaliche. Neuroradiologia. Turin: Minerva Medica, 1996:3–21. 7. Barkovich AJ, Kuzniecky RI, Jackson GD, Guerrini R, Dobyns WB. Classification system for malformations of cortical development – Update 2001. Neurology 2001; 57:2168–2178. 8. Tsuruda JS, Chew WM, Moseley ME, Norman D. Diffusionweighted MR imaging of the brain: value of differentiating between extraaxial cysts and epidermoid tumors. AJNR Am J Neuroradiol 1990; 11:925–931. 9. Sarnat HB. Cerebral Dysgenesis. Embryology and Clinical Expression. New York: Oxford University Press, 1992. 10. Naidich TP, Altman NR, Braffman BH, McLone DG, Zimmerman RA. Cephaloceles and related malformations. AJNR Am J Neuroradiol 1992; 13:655–690. 11. Bernardi B, Fonda C. Cefaloceli. Rivista di Neuroradiologia 1994; 7:171–186. 12. Simpson DA, David DJ, White J. Cephaloceles: treatment, outcome and antenatal diagnosis. Neurosurgery 1984; 15:14–21. 13. Menkes JH, Till K. Malformations of the central nervous system. In: Menkes JH (ed) Textbook of Child Neurology, 5th edn. Baltimore: Williams & Wilkins, 1995:240–324. 14. Van Allen MI, Kalousek DK, Chernoff GF, Juriloff D, Harris M, McGillivray BC, Yong SL, Langlois S, MacLeod PM, Chitayat D, Friedman JM, Wilson RD, McFadden D, Pantzar J, Ritchie S, Hall JG. Evidence for multi-site closure of the neural tube in humans. Am J Med Genet 1993; 47:723–743. 15. Campbell LR, Dayton DH, Sohal GS. Neural tube defects: a review of human and animal studies on the etiology of neural tube defects. Teratology 1986; 34:171–187. 16. Tortori-Donati P, Cama A, Fondelli MP, Rossi A, Taccone A.

Brain Malformations Cefaloceli. In: Tortori-Donati P, Longo M, Taccone A (eds) Malformazioni cranio-encefaliche. Neuroradiologia. Turin: Minerva Medica, 1996:107–124. 17. Friede RL. Developmental Neuropathology, 2nd edn. Berlin: Springer, 1989. 18. Rao AS, Rao VR, Mandalam KR, Gupta AK, Kumar S, Joseph S, Unni M. Corpus callosum lipoma with frontal encephalocele. Neuroradiology 1990; 32:50–52. 19. Lorber J, Schofield JK. The prognosis of occipital encephalocele. Z Kinderchir 1979; 28:347–351. 20. Castillo M, Quencer RM, Dominguez R. Chiari III malformation: imaging features. AJNR Am J Neuroradiol 1992; 13:107–113. 21. Naidich TP, Braffman BH, Altman N, Birchansky B. Congenital malformations involving the anterior cranial fossa. Rivista di Neuroradiologia 1994; 7:359–376. 22. Hyson M, Andermann F, Olivier A, Melanson D. Occult encephaloceles and temporal lobe epilepsy: developmental and acquired lesions in the middle fossa. Neurology 1984; 34:363–366. 23. Tortori-Donati P, Fondelli MP, Rossi A. Anomalie congenite. In: Simonetti G, Del Maschio A, Bartolozzi C, Passariello R (eds) Trattato Italiano di Risonanza Magnetica. Naples: Idelson-Gnocchi, 1998: 155–203. 24. Barkovich AJ, Vandermarck P, Edwards MS, Cogen PH. Congenital nasal masses: CT and MR imaging features in 16 cases. AJNR Am J Neuroradiol 1991; 12:105–116. 25. Younus M, Coode PE. Nasal glioma and encephalocele: two separate entities. Report of two cases. J Neurosurg 1986; 64:516–519. 26. Elster AD, Branch CL Jr. Transalar sphenoidal encephaloceles: clinical and radiologic findings. Radiology 1989; 170:245–247. 27. Levy RA, Wald SL, Aitken PA, Dorwart RH. Bilateral intraorbital meningoencephaloceles and associated midline craniofacial anomalies: MR and three-dimensional CT imaging. AJNR Am J Neuroradiol 1989; 10:1272–1274. 28. Seibert RW, Seibert JJ, Jimenez JF, Angtuaco EJ. Nasopharyngeal brain heterotopia – a cause of upper airway obstruction in infancy. Laryngoscope 1984; 94:818–819. 29. Rice JF, Eggers DM. Basal transsphenoidal encephalocele: MR findings. AJNR Am J Neuroradiol 1989; 10 (5 Suppl):S79. 30. Boulanger T, Mathurin P, Dooms G, Lambert M, Smellie S, Cornelis G. Sphenopharyngeal meningoencephalocele: unusual clinical and radiologic features. AJNR Am J Neuroradiol 1989; 10 (5 Suppl):S80. 31. Nager GT. Cephaloceles. Laryngoscope 1987; 97:77–84. 32. Cohen MM, Lemire RM. Syndromes with cephalocele. Teratology 1982; 25:161–172. 33. Naidich TP, Osborn RE, Bauer B, Naidich MJ. Median cleft face syndrome: MR and CT data from 11 children. J Comput Assist Tomogr 1988; 12:57–64. 34. Nevin NC, Leonard AG, Jones B. Frontonasal dysostosis in two successive generations. Am J Med Genet 1999; 87:251– 253. 35. Tambwekar S. Internasal dysplasia (craniofacial anomaly). Bombay Hospital Journal 1998; 40(4), article 19. 36. Lubinsky MS, Kahler SG, Speer IE, Hoyme HE, Kirillova IA, Lurie IW. von Voss-Cherstvoy syndrome: a variable perinatally lethal syndrome of multiple congenital anomalies. Am J Med Genet 1994; 52:272–278. 37. Bamforth JS, Lin CC. DK phocomelia phenotype (von VossCherstvoy syndrome) caused by somatic mosaicism for del(13q). Am J Med Genet 1997; 73:408–411.

38. Dobyns WB, Pagon RA, Armstrong D, Curry CJ, Greenberg F, Grix A, Holmes LB, Laxova R, Michels VV, Robinow M, et al. Diagnostic criteria for Walker-Warburg syndrome. Am J Med Genet 1989; 32:195–210. 39. Naidich TP, Braffman BH, Altman NR, Birchansky SB. Malformations of the posterior fossa and craniovertebral junction. Rivista di Neuroradiologia 1994; 7:423–439. 40. Erdincler P, Kaynar MY, Canbaz B, Kocer N, Kuday C, Ciplak N. Iniencephaly: neuroradiological and surgical features. Case report and review of the literature. J Neurosurg 1998; 89:317–320. 41. Nyberg DA, Mahomy BS, Pretorius DH. Diagnostic ultrasound of fetal anomalies -Text and atlas. Littleton (MA): Year Book Medical Publishers, Inc., 1990:146. 42. Salonen R, Norio R. The Meckel syndrome: clinicopathological findings in 67 patients. Am J Med Genet 1984; 18:671–689. 43. Nyberg DA, Hallesy D, Mahony BS, Hirsch JH, Luthy DA, Hickok D. Meckel–Gruber syndrome; importance of prenatal diagnosis. J Ultrasound Med 1990; 9:691–696. 44. Paavola P, Salonen R, Weissenbach J, Peltonen L. The locus for Meckel syndrome with multiple congenital anomalies maps to chromosome 17q21–q24. Nat Genet 1995; 11:213– 215. 45. Knobloch WH, Layer JM. Clefting syndromes associated with retinal detachment. Am J Ophthalmol 1972; 73:517– 530. 46. Seaver LH, Joffe L, Spark RP, Smith BL, Hoyme HE. Congenital scalp defects and vitreoretinal degeneration: redefining the Knobloch syndrome. Am J Med Genet 1993; 46:203– 208. 47. Handmaker SD, Campbell JA, Robinson LD, Chinwah O, Gorlin RJ. Dyssegmental dwarfism: a new syndrome of lethal dwarfism. Birth Defects Orig Artic Ser 1977; 13(3D):79–90. 48. Goldhammer Y, Smith JL. Cryptophthalmos syndrome with basal encephaloceles. Am J Ophthalmol 1975; 80:146–149. 49. Nguyen VD, Tyrrel R. Klippel-Feil syndrome: patterns of bony fusion and wasp-waist sign. Skeletal Radiol 1993; 22:519–523. 50. Ulmer JL, Elster AD, Ginsberg LE, Williams DW III. Klippel-Feil syndrome: CT and MR of acquired and congenital abnormalities of cervical spine and cord. J Comput Assist Tomogr 1993; 17:215–224. 51. Van Den Berg DJ, Francke U. Roberts syndrome: a review of 100 cases and a new rating system for severity. Am J Med Genet 1993; 47:1104–1123. 52. Aleksic S, Budzilovich G, Greco MA, Epstein F, Feigin I, Pearson J. Encephalocele (cerebellocele) in the GoldenharGorlin syndrome. Eur J Pediatr 1983; 140:137–138. 53. Gustavson EE, Chen H. Goldenhar syndrome, anterior encephalocele, and aqueductal stenosis following fetal primidone exposure. Teratology 1985; 32:13–17. 54. Moerman P, Fryns JP, Vandenberghe K, Lauweryns JM. Constrictive amniotic bands, amniotic adhesions, and limbbody wall complex: discrete disruption sequences with pathogenetic overlap. Am J Med Genet 1992; 42:470–479. 55. Yakovlev PI. Pathoarchitectonic studies of cerebral malformations. III. Arrhinencephalies (holotelencephalies). J Neuropathol Exp Neurol 1959; 18:22–55. 56. Roessler E, Muenke M. The molecular genetics of holoprosencephaly: a model of brain development for the next century. Child’s Nerv Syst 1999; 15:646–651. 57. Roach E, Demyer W, Conneally PM, Palmer C, Merritt AD. Holoprosencephaly: birth data, genetic and demographic

187

188

P. Tortori-Donati, A. Rossi, and R. Biancheri analyses of 30 families. Birth Defects Orig Artic Ser 1975; 11:294–313. 58. Matsunaga E, Shiota K. Holoprosencephaly in human embryos: epidemiologic studies of 150 cases. Teratology 1977; 16:261–272 59. Moog U, De Die-Smulders CE, Schrander-Stumpel CT, Engelen JJ, Hamers AJ, Frints S, Fryns JP. Holoprosencephaly: the Maastricht experience. Genet Couns 2001; 12:287–298 60. DeMyer WE, Zeman W. Alobar holoprosencephaly (arhinencephaly) with median cleft lip and palate: clinical, electroencephalographic, and nosologic considerations. Confin Neurol 1963; 23:1–36. 61. Simon EM, Barkovich AJ. Holoprosencephaly: new concepts. Magn Reson Imaging Clin N Am 2001; 9:149–164. 62. Tortori-Donati P, Rossi A. Congenital malformations in the neonate. In: Rutherford M (ed) MRI of the neonatal brain. London: Saunders, 2002: 225–249. 63. Barkovich AJ, Quint DJ. Middle interhemispheric fusion: an unusual variant of holoprosencephaly. AJNR Am J Neuroradiol 1993; 14:431–440. 64. Kalter H. Case reports of malformations associated with maternal diabetes: history and critique. Clin Genet 1993; 43:174–179. 65. Wallis D, Muenke M. Mutations in holoprosencephaly. Hum Mutat 2000; 16:99–108. 66. Simon EM, Hevner R, Pinter JD, Clegg NJ, Miller VS, Kinsman SL, Hahn JS, Barkovich AJ. Assessment of the deep gray nuclei in holoprosencephaly. AJNR Am J Neuroradiol 2000; 21:1955–1961. 67. Barkovich AJ. Magnetic resonance imaging: role in the understanding of cerebral malformations. Brain Dev 2002; 24:2–12. 68. Chen CP, Shih SL, Liu FF, Jan SW. Cebocephaly, alobar holoprosencephaly, spina bifida, and sirenomelia in a stillbirth. J Med Genet 1997; 34:252–255. 69. Nowaczyk MJ, Huggins MJ, Tomkins DJ, Rossi E, Ramsay JA, Woulfe J, Scherer SW, Belloni E. Holoprosencephaly, sacral anomalies, and situs ambiguus in an infant with partial monosomy 7q/trisomy 2p and SHH and HLXB9 haploinsufficiency. Clin Genet 2000; 57:388–393. 70. Le Douarin NM, Halpern ME. Discussion point. Origin and specification of the neural tube floor plate: insights from the chick and zebrafish. Curr Opin Neurobiol 2000; 10:23– 30. 71. Castillo M, Bouldin TW, Scatliff JH, Suzuki K. Radiologicpathologic correlation. Alobar holoprosencephaly. AJNR Am J Neuroradiol 1993; 14:1151–1156. 72. Liu DP, Burrowes DM, Qureshi MN. Cyclopia: craniofacial appearance on MR and three-dimensional CT. AJNR Am J Neuroradiol 1997; 18:543–546. 73. Kjaer I, Becktor KB, Lisson J, Gormsen C, Russell BG. Face, palate, and craniofacial morphology in patients with a solitary median maxillary central incisor. Eur J Orthod 2001; 23:63–73. 74. Osaka K, Tanimura T, Hirayama A, Matsumoto S. Myelomeningocele bifore birth. J Neurosurg 1978; 49:711–724. 75. Nowaczyk MJ, Farrell SA, Sirkin WL, Velsher L, Krakowiak PA, Waye JS, Porter FD. Smith-Lemli-Opitz (RHS) syndrome: holoprosencephaly and homozygous IVS8–1G–>C genotype. Am J Med Genet 2001; 103:75–80. 76. Lapunzina P, Musante G, Pedraza A, Prudent L, Gadow E. Semilobar holoprosencephaly, coronal craniosynostosis, and multiple congenital anomalies: a severe expression of

the Genoa syndrome or a newly recognized syndrome? Am J Med Genet 2001; 102:258–260. 77. Lin AE, Siebert JR, Graham JM Jr. Central nervous system malformations in the CHARGE association. Am J Med Genet 1990 ;37:304–310. 78. Oba H, Barkovich AJ. Holoprosencephaly: an analysis of callosal formation and its relation to development of the interhemispheric fissure. AJNR Am J Neuroradiol 1995; 16:453–460. 79. Longo M, Blandino A, Taccone A, Tortori-Donati P. Oloprosencefalia, displasia setto-ottica, malformazioni dell’asse ipotalamo-ipofisario. In: Tortori-Donati P, Taccone A, Longo M (eds) Malformazioni cranio-encefaliche. Neuroradiologia. Turin: Minerva Medica, 1996:137–158. 80. Barkovich AJ, Simon EM, Clegg NJ, Kinsman SL, Hahn JS. Analysis of the cerebral cortex in holoprosencephaly with attention to the sylvian fissures. AJNR Am J Neuroradiol 2002; 23:143–150. 81. Golden JA. Towards a greater understanding of the pathogenesis of holoprosencephaly. Brain Dev 1999; 21:513–521. 82. Veneselli E, Biancheri R, Di Rocco M, Fondelli MP, Perrone MV, Tortori-Donati P. Unusually prolonged survival and childhoodonset epilepsy in a case of alobar holoprosencephaly. Childs Nerv Syst 1999; 15:274–277. 83. Simon EM, Hevner RF, Pinter J, Clegg NJ, Delgado M, Kinsman SL, Hahn JS, Barkovich AJ. The dorsal cyst in holoprosencephaly and the role of the thalamus in its formation. Neuroradiology 2001; 43:787–791. 84. Kobori JA, Kaarsoo-Herrick M, Urich H. Arhinencephaly. The spectrum of associated malformations. Brain 1987; 110:237–260. 85. Sato N, Hatakeyama S, Shimizu N, Hikima A, Aoki J, Endo K. MR evaluation of the hippocampus in patients with congenital malformations of the brain. AJNR Am J Neuroradiol 2001; 22:389–393. 86. Robin NH, Ko LM, Heeger S, Muise KL, Judge N, Bangert BA. Syntelencephaly in an infant of a diabetic mother. Am J Med Genet 1996; 66:433–437. 87. Simon EM, Hevner RF, Pinter JD, Clegg NJ, Delgado M, Kinsman SL, Hahn JS, Barkovich AJ. The middle interhemispheric variant of holoprosencephaly. AJNR Am J Neuroradiol 2002; 23:151–156. 88. Fujimoto S, Togari H, Banno T, Wada Y. Syntelencephaly associated with connected transhemispheric cleft of focal cortical dysplasia. Pediatr Neurol 1999; 20:387–389 89. De Morsier G. Études sur les dysraphies cranio-éncephaliques. III. Agénésie du septum lucidum avec malformation du tractus optique. La dysplasie septo-optique. Schweiz Arch Neurol Psychiat 1956; 77:267–292. 90. Dattani MT, Martinez-Barbera JP, Thomas PQ, Brickman JM, Gupta R, Martensson IL, Toresson H, Fox M, Wales JK, Hindmarsh PC, Krauss S, Beddington RS, Robinson IC. Mutations in the homeobox gene HESX1/Hesx1 associated with septo-optic dysplasia in human and mouse. Nat Genet 1998; 19:125–133. 91. Lubinsky MS. Hypothesis: septo-optic dysplasia is a vascular disruption sequence. Am J Med Genet 1997; 69:235–236. 92. McMahon CL, Braddock SR. Septo-optic dysplasia as a manifestation of valproic acid embryopathy. Teratology 2001; 64:83–86. 93. Barkovich AJ, Fram EK, Norman D. Septo-optic dysplasia: MR imaging. Radiology 1989; 171:189–192. 94. Miller SP, Shevell MI, Patenaude Y, Poulin C, O’Gorman AM. Septo-optic dysplasia plus: a spectrum of malformations of cortical development. Neurology 2000; 54:1701–1703.

Brain Malformations 95. Sener RN. Septo-optic dysplasia (de Morsier’s syndrome) associated with total callosal absence. A new type of the anomaly. J Neuroradiol 1996; 23:79–81. 96. Sener RN. Septo-optic dysplasia associated with cerebral cortical dysplasia (cortico-septo-optic dysplasia). J Neuroradiol 1996; 23:245–247. 97. Kallmann FJ, Schoenfeld WA, Barrera SE. The gebeti aspects of primary eunuchoidism. Am J Ment Defic 1944; 48:203– 210. 98. Franco B, Guioli S, Pragliola A, Incerti B, Bardoni B, Tonlorenzi R, Carrozzo R, Maestrini E, Pieretti M, Taillon-Miller P, et al. A gene detected in Kallmann’s sindrome shares homology with neural cell adhesion and axonal path-finding molecules. Nature 1991; 353:529–536. 99. Legouis R, Hardelin JP, Levilliers J, Claverie JM, Compain S, Wunderle V, Millasseau P, Le Paslier D, Cohen D, Caterina D, et al. The candidate gene for the X-linked Kallmann syndrome encodes a protein related to adhesion molecules. Cell 1991; 86:423–435. 100. Duke VM, Winyard PJ, Thorogood P, Soothill P, Bouloux PM, Woolf AS. KAL, a gene mutated in Kallmann’s syndrome, is expressed in the first trimester of human development. Molec Cell Endocrinol 1995; 110:73–79. 101. Truwit CL, Barkovich AJ, Grumbach MM, Martini JJ. MR imaging of Kallmann syndrome, a genetic disorder of neuronal migration affecting the olfactory and genital systems. AJNR Am J Neuroradiol 1993; 14:827–838. 102. Schwanzel-Fukuda M, Bick D, Pfaff DW. Luteinizing hormone-releasing hormone (LHRH)-expressing cells do not migrate normally in an inherited hypogonadal (Kallmann) syndrome. Mod Brain Res 1989; 6:311–326. 103. Knorr JR, Ragland RL, Brown RS, Gelber N. Kallmann syndrome: MR findings. AJNR Am J Neuroradiol 1993; 14:845– 851. 104. Yousem DM, Turner WJD, Li C, Snyder PJ, Doty RL. Kallmann syndrome: MR evaluation of olfactory system. AJNR Am J Neuroradiol 1993; 14:839–843. 105. Barkovich AJ, Kuzniecky RI, Dobyns WB, Jackson GD, Becker LE, Evrard P. A classification scheme for malformations of cortical development. Neuropediatrics 1996; 27:59–63. 106. Leventer RJ, Phelan EM, Coleman LT, Kean MJ, Jackson GD, Harvey AS. Clinical and imaging features of cortical malformations in childhood. Neurology 1999; 53:715–722. 107. Montenegro MA, Guerreiro MM, Lopes-Cendes I, Guerreiro CA, Cendes F. Interrelationship of genetics and prenatal injury in the genesis of malformations of cortical development. Arch Neurol 2002; 59:1147–1153. 108. Barkovich AJ, Gressens P, Evrard P. Formation, maturation, and disorders of brain neocortex. AJNR Am J Neuroradiol 1993; 13:423–446. 109. Raymond AA, Fish DR, Sisodiya SM, Alsanjari N, Stevens JM, Shorvon SD. Abnormalities of gyration, heterotopias, tuberous sclerosis, focal cortical dysplasia, microdysgenesis, dysembryoplastic neuroepithelial tumor and dysgenesis of the archicortex in epilepsy. Clinical, EEG and neuroimaging features in 100 adult patients. Brain 1995; 118:629–660. 110. Hatten ME. Central nervous system neuronal migration. Annu Rev Neurosci 1999; 22:511–539. 111. Ross ME, Walsh CA. Human brain malformations and their lessons for neuronal migration. Annu Rev Neurosci 2001; 24:1041–1070. 112. Sarnat HB. Cerebral dysplasias as expressions of altered maturational processes. Can J Neurol Sci 1991; 18:196–204.

113. Takahashi T, Nowakowski RS, Caviness VS Jr. The mathematics of neocortical neuronogenesis. Dev Neurosci 1997; 19:17–22. 114. Chan WY, Lorke DE, Tiu SC, Yew DT. Proliferation and apoptosis in the developing human neocortex. Anat Rec 2002; 267:261–276. 115. Rakic P. Specification of cerebral cortical areas. Science 1988; 241:170–176. 116. Yung SY, Gokhan S, Jurcsak J, Molero AE, Abrajano JJ, Mehler MF. Differential modulation of BMP signaling promotes the elaboration of cerebral cortical GABAergic neurons or oligodendrocytes from a common sonic hedgehog-responsive ventral forebrain progenitor species. Proc Natl Acad Sci U S A 2002; 99:16273–16278. 117. Corbin JG, Nery S, Fishell G. Telencephalic cells take a tangent: non-radial migration in the mammalian forebrain. Nat Neurosci 2001; 4 (Suppl):1177–1182. 118. Gressens P, Evrard P. The glial fascicle: an ontogenic and phylogenic unit guiding, supplying and distributing mammalian cortical neurons. Brain Res Dev Brain Res 1993; 76:272–277. 119. Marin-Padilla M. Structural organization of the human cerebral cortex prior to the appearance of the cortical plate. Anat Embryol (Berl) 1983; 168:21–40. 120. Marin-Padilla M. [The evolution of the structure of the neocortex in mammals: a new theory of cytoarchitecture] Rev Neurol 2001; 33:843–853. 121. Cobos I, Puelles L, Martinez S. The avian telencephalic subpallium originates inhibitory neurons that invade tangentially the pallium (dorsal ventricular ridge and cortical areas). Dev Biol 2001; 239:30–45. 122. Misson JP, Austin CP, Takahashi T, Cepko CL, Caviness VS Jr. The alignment of migrating neural cells in relation to the murine neopallial radial glial fiber system. Cereb Cortex 1991; 3:221–229. 123. Lois C, Garcia-Verdugo JM, Alvarez-Buylla A. Chain migration of neuronal precursors. Science 1996; 271:978–981. 124. Goldman-Rakic PS. Topography of cognition: parallel distributed networks in primate association cortex. Annu Rev Neurosci 1988; 11:137–156. 125. Ghosh A, Shatz CJ. A role for subplate neurons in the patterning of connections from thalamus to neocortex. Development 1993; 117:1031–1047. 126. McConnell SK, Ghosh A, Shatz CJ. Subplate pioneers and the formation of descending connections from cerebral cortex. J Neurosci 1994; 14:1892–1907. 127. Yelgec S, Ozturk MH, Aydingoz U, Cila A. CT and MRI of microcephalia vera. Neuroradiology 1998; 40:332–334. 128. Abdel-Salam G, Czeizel AE. The second unrelated case with isolated microcephaly and normal intelligence (microcephalia vera). Clin Dysmorphol 2000; 9:151–152. 129. Peiffer A, Singh N, Leppert M, Dobyns WB, Carey JC. Microcephaly with simplified gyral pattern in six related children. Am J Med Genet 1999; 84:137–144. 130. van der Knaap MS, van Wezel-Meijler G, Barth PG, Barkhof F, Ader HJ, Valk J. Normal gyration and sulcation in preterm and term neonates: appearance on MR images. Radiology 1996; 200:389–396. 131. Barkovich AJ, Ferriero DM, Barr RM, Gressens P, Dobyns WB, Truwit CL, Evrard P. Microlissencephaly: a heterogeneous malformation of cortical development. Neuropediatrics 1998; 29:113–119. 132. Iannetti P, Schwartz CE, Dietz-Band J, Light E, Timmerman J, Chessa L. Norman-Roberts syndrome: clinical and molecular studies. Am J Med Genet 1993; 47:95–99.

189

190

P. Tortori-Donati, A. Rossi, and R. Biancheri 133. Kroon AA, Smit BJ, Barth PG, Hennekam RC. Lissencephaly with extreme cerebral and cerebellar hypoplasia. A magnetic resonance imaging study. Neuropediatrics 1996; 27:273–276. 134. Ross ME, Swanson K, Dobyns WB. Lissencephaly with cerebellar hypoplasia (LCH): a heterogeneous group of cortical malformations. Neuropediatrics 2001; 32:256–263. 135. DeMyer W. Megalencephaly: types, clinical syndromes, and management. Pediatr Neurol 1986; 2:321–328. 136. Gooskens RH, Willemse J, Bijlsma JB, Hanlo PW. Megalencephaly: definition and classification. Brain Dev 1988; 10:1–7. 137. Sotos JF, Dodge PR, Muirhead D, Crawford JD, Talbot NB. Cerebral gigantism in childhood: a syndrome of excessively rapid growth with acromegalic features and a nonprogressive neurologic disorder. New Eng J Med 1964; 271:109– 116. 138. Maroun C, Schmerler S, Hutcheon RG. Child with Sotos phenotype and a 5:15 translocation. Am J Med Genet 1994; 50:291–293. 139. Douglas J, Hanks S, Temple IK, Davies S, Murray A, Upadhyaya M, Tomkins S, Hughes HE, Cole TR, Rahman N. NSD1 mutations are the major cause of Sotos syndrome and occur in some cases of Weaver syndrome but are rare in other overgrowth phenotypes. Am J Hum Genet 2003; 72:132–143. 140. Nance MA, Neglia JP, Talwar D, Berry SA. Neuroblastoma in a patient with Sotos’ syndrome. J Med Genet 1990; 27:130– 132. 141. Gohlich-Ratmann G, Baethmann M, Lorenz P, Gartner J, Goebel HH, Engelbrecht V, Christen HJ, Lenard HG, Voit T. Megalencephaly, mega corpus callosum, and complete lack of motor development: a previously undescribed syndrome. Am J Med Genet 1998; 79:161–167. 142. Hayashi M, Kurata K, Suzuki K, Hirasawa K, Nagata J, Morimatsu Y. Megalencephaly, hydrocephalus and cortical dysplasia in severe dwarfism mimicking leprechaunism. Acta Neuropathol (Berl) 1998; 95:431–436. 143. Tortori-Donati P, Fondelli MP, Rossi A, Andreussi L. Sclerosi tuberosa (m. di Bourneville). In: Tortori-Donati P, Taccone A, Longo M (eds) Malformazioni Cranio-Encefaliche. Neuroradiologia. Turin: Minerva Medica, 1996:375–390. 144. Colombo N, Citterio A, Munari C, Scialfa G. Epilessia. In: Dal Pozzo G (ed) Compendio di Risonanza Magnetica. Cranio e Rachide. Turin: UTET, 2001:657–704. 145. Barkovich AJ, Kjos BO. Nonlissencephalic cortical dysplasias: correlation of imaging findings with clinical deficits. AJNR Am J Neuroradiol 1992; 13:95–103. 146. Tassi L, Colombo N, Garbelli R, Francione S, Lo Russo G, Mai R, Cardinale F, Cossu M, Ferrario A, Galli C, Bramerio M, Citterio A, Spreafico R. Focal cortical dysplasia: neuropathological subtypes, EEG, neuroimaging and surgical outcome. Brain 2002; 125:1719–1732. 147. Colombo N, Tassi L, Galli C, Citterio A, Lo Russo G, Scialfa G, Spreafico R. Focal cortical dysplasias: MR imaging, histopathologic, and clinical correlations in surgically treated patients with epilepsy. AJNR Am J Neuroradiol 2003; 24:724–733. 148. Taylor DC, Falconer MA, Bruton CJ, Corsellis JA. Focal dysplasia of the cerebral cortex in epilepsy. J Neurol Neurosurg Psychiatry 1971; 34:369–387. 149. Barkovich AJ, Kuzniecky RI, Bollen AW, Grant PE. Focal transmantle dysplasia: a specific malformation of cortical development. Neurology 1997; 49:1148–1152.

150. Becker AJ, Urbach H, Scheffler B, Baden T, Normann S, Lahl R, Pannek HW, Tuxhorn I, Elger CE, Schramm J, Wiestler OD, Blumcke I. Focal cortical dysplasia of Taylor’s balloon cell type: mutational analysis of the TSC1 gene indicates a pathogenic relationship to tuberous sclerosis. Ann Neurol 2002; 52: 29–37. 151. Kuzniecky RI. MRI in focal cortical dysplasia. In: Guerrini R, Andermann F, Canapicchi R, Roger J, Zifkin BG, Pfanner P (eds) Dysplasias of Cerebral Cortex and Epilepsy. Philadelphia: Lippincott-Raven, 1996:145–150. 152. Bronen RA, Vives KP, Kim JH, Fulbright RK, Spencer SS, Spencer DD. Focal cortical dysplasia of Taylor, balloon cell subtype: MR differentiation from low-grade tumors. AJNR Am J Neuroradiol 1997; 18:1141–1151. 153. Palmini A, Gambardella A, Andermann F, Dubeau F, da Costa JC, Olivier A, Tampieri D, Robitaille Y, Paglioli E, Paglioli Neto E, et al. Operative strategies for patients with cortical dysplastic lesions and intractable epilepsy. Epilepsia 1994; 35 (Suppl 6):S57–71. 154. Palmini A, Gambardella A, Andermann F, Dubeau F, da Costa JC, Olivier A, Tampieri D, Gloor P, Quesney F, Andermann E, et al. Intrinsic epileptogenicity of human dysplastic cortex as suggested by corticography and surgical results. Ann Neurol 1995; 37:476–487. 155. Kuzniecky R, Ho SS, Martin R, Faught E, Morawetz R, Palmer C, Gilliam F. Temporal lobe developmental malformations and hippocampal sclerosis. Epilepsy surgery outcome. Neurology 1999; 52:479–484. 156. Urbach H, Scheffler B, Heinrichsmeier T, von Oertzen J, Kral T, Wellmer J, Schramm J, Wiestler OD, Blumcke I. Focal cortical dysplasia of Taylor’s balloon cell type: a clinicopathological entity with characteristic neuroimaging and histopathological features, and favorable postsurgical outcome. Epilepsia 2002; 43:33–40. 157. Montenegro MA, Li LM, Guerreiro MM, Guerreiro CA, Cendes F. Focal cortical dysplasia: improving diagnosis and localization with magnetic resonance imaging multiplanar and curvilinear reconstruction. J Neuroimaging 2002; 12:224–230. 158. Barkovich AJ, Rowley HA, Andermann F. MR in partial epilepsy: value of high-resolution volumetric techniques. AJNR Am J Neuroradiol 1995; 16:339–343. 159. Grant PE, Barkovich AJ, Wald LL, Dillon WP, Laxer KD, Vigneron DB. High-resolution surface-coil MR of cortical lesions in medically refractory epilepsy: a prospective study. AJNR Am J Neuroradiol 1997; 18:291–301. 160. Chan S, Chin SS, Nordli DR, Goodman RR, DeLaPaz RL, Pedley TA. Prospective magnetic resonance imaging identification of focal cortical dysplasia, including the non–balloon cell subtype. Ann Neurol 1998; 44:749–757. 161. Lee BC, Schmidt RE, Hatfield GA, Bourgeois B, Park TS. MRI of focal cortical dysplasia. Neuroradiology 1998; 40:675– 683. 162. Matsuda K, Mihara T, Tottori T, Otubo T, Usui N, Baba K, Matsuyama N, Yagi K. Neuroradiologic findings in focal cortical dysplasia: histologic correlation with surgically resected specimens. Epilepsia 2001; 42(Suppl 6):29–36. 163. Yagishita A, Arai N, Maehara T, Shimizu H, Tokumaru AM, Oda M. Focal cortical dysplasia: appearance on MR images. Radiology 1997; 203:553–559. 164. Adamsbaum C, Robain O, Cohen PA, Delalande O, Fohlen M, Kalifa G. Focal cortical dysplasia and hemimegalencephaly: histological and neuroimaging correlations. Pediatr Radiol 1998; 28:583–590.

Brain Malformations 165. Kim SK, Na DG, Byun HS, Kim SE, Suh YL, Choi JY, Yoon HK, Han BK. Focal cortical dysplasia: comparison of MRI and FDG-PET. J Comput Assist Tomogr 2000; 24:296–302. 166. Daumas-Duport C. Dysembryoplastic neuroepithelial tumors in epilepsy surgery. In: Guerrini R, Andermann F, Canapicchi R, Roger J, Zifkin BG, Pfanner P (eds) Dysplasias of Cerebral Cortex and Epilepsy. Philadelphia: Lippincott-Raven, 1996:71–80. 167. Li LM, Cendes F, Bastos AC, Andermann F, Dubeau F, Arnold DL. Neuronal metabolic dysfunction in patients with cortical developmental malformations: a proton magnetic resonance spectroscopic imaging study. Neurology 1998; 50:755–759. 168. Simone IL, Federico F, Tortorella C, De Blasi R, Bellomo R, Lucivero V, Carrara D, Bellacosa A, Livrea P, Carella A. Metabolic changes in neuronal migration disorders: evaluation by combined MRI and proton MR spectroscopy. Epilepsia 1999; 40:872–879. 169. Kaminaga T, Kobayashi M, Abe T. Proton magnetic resonance spectroscopy in disturbances of cortical development. Neuroradiology 2001; 43:575–580. 170. Pillai JJ, Hessler RB, Allison JD, Park YD, Lee MR, Lavin T. Advanced MR imaging of cortical dysplasia with or without neoplasm: a report of two cases. AJNR Am J Neuroradiol 2002; 23:1686–1691. 171. Flores-Sarnat L. Hemimegalencephaly: part 1. Genetic, clinical, and imaging aspects. J Child Neurol 2002; 17:373–384. 172. Vigevano F, Fusco L, Granata T, Fariello G, Di Rocco C, Cusmai R. Hemimegalencephaly: clinical and EEG characteristics. In: Guerrini R, Andermann F, Canapicchi R, Roger J, Zifkin BG, Pfanner P (eds) Dysplasias of Cerebral Cortex and Epilepsy. Philadelphia: Lippincott-Raven, 1996:285–294. 173. Di Rocco C. Surgical treatment of hemimegalencephaly. In: Guerrini R, Andermann F, Canapicchi R, Roger J, Zifkin BG, Pfanner P (eds) Dysplasias of Cerebral Cortex and Epilepsy. Philadelphia: Lippincott-Raven, 1996:295–304. 174. Sener RN. MR demonstration of cerebral hemimegalencephaly associated with cerebellar involvement (total hemimegalencephaly). Comput Med Imaging Graph 1997; 21:201–204. 175. Robain O, Gelot A. Neuropathology of hemimegalencephaly. In: Guerrini R, Andermann F, Canapicchi R, Roger J, Zifkin BG, Pfanner P (eds) Dysplasias of Cerebral Cortex and Epilepsy. Philadelphia: Lippincott-Raven, 1996:89–92. 176. De Rosa MJ, Secor DL, Barsom M, Fisher RS, Vinters HV. Neuropathologic findings in surgically treated hemimegalencephaly: immunohistochemical, morphometric, and ultrastructural study. Acta Neuropathol (Berl) 1992; 84:250–260. 177. Yagishita A, Arai N, Tamagawa K, Oda M. Hemimegalencephaly: signal changes suggesting abnormal myelination on MRI. Neuroradiology 1998; 40:734–738. 178. Barkovich AJ, Chuang SH. Unilateral megalencephaly: correlation of MR imaging and pathologic characteristics. AJNR Am J Neuroradiol 1990; 11:523–531. 179. Wolpert SM, Cohen A, Libenson MH. Hemimegalencephaly: a longitudinal MR study. AJNR Am J Neuroradiol 1994; 15:1479–1482. 180. Hanefeld F, Kruse B, Holzbach U, Christen HJ, Merboldt KD, Hanicke W, Frahm J. Hemimegalencephaly: localized proton magnetic resonance spectroscopy in vivo. Epilepsia 1995; 3:1215–1224. 181. Happle R. How many epidermal nevus syndromes exist? A clinicogenetic classification. J Am Acad Derm 1991; 25:550– 556.

182. Dodge NN, Dobyns WB. Agenesis of the corpus callosum and Dandy-Walker malformation associated with hemimegalencephaly in the sebaceous nevus syndrome. Am J Med Genet 1995; 56:147–150. 183. Kuster W, Ehrig T, Happle R. ”Hypomelanosis of Ito”: no entity, but a cutaneous sign of mosaicism. In: Nordlund JJ, Boissy R, Hearing V, King R, Ortonne JP (eds) The Pigmentary System and Its Disorders. New York: Oxford University Press; 1998:594–601. 184. Blaser S, Jay V, Krafchik B. Hypomelanosis of Ito (Incontinentia Pigmenti Achromians). Int J Neuroradiol 1996; 2:176–181. 185. Williams DW 3rd, Elster AD. Cranial CT and MR in the Klippel-Trenaunay-Weber syndrome. AJNR Am J Neuroradiol 1992; 13:291–294. 186. Berry SA, Peterson C, Mize W, Bloom K, Zachary C, Blasco P, Hunter D. Klippel-Trenaunay syndrome. Am J Med Genet 1998; 79:319–326. 187. Torregrosa A, Marti-Bonmati L, Higueras V, Poyatos C, Sanchis A. Klippel-Trenaunay syndrome: frequency of cerebral and cerebellar hemihypertrophy on MRI. Neuroradiology 2000; 42:420–423. 188. Wiedemann HR, Burgio GR, Aldenhoff P, Kunze J, Kaufmann HJ, Schirg E. The proteus syndrome. Partial gigantism of the hands and/or feet, nevi, hemihypertrophy, subcutaneous tumors, macrocephaly or other skull anomalies and possible accelerated growth and visceral affections. Eur J Pediatr 1983; 140:5–12. 189. Happle R. Cutaneous manifestation of lethal genes. Hum Genet 1986; 72:280. 190. Biesecker LG, Happle R, Mulliken JB, Weksberg R, Graham JM Jr, Viljoen DL, Cohen MM Jr. Proteus syndrome: diagnostic criteria, differential diagnosis, and patient evaluation. Am J Med Genet 1999; 84:389–395. 191. DeLone DR, Brown WD, Gentry LR. Proteus syndrome: craniofacial and cerebral MRI. Neuroradiology 1999; 41:840– 843. 192. Dietrich RB, Glidden DE, Roth GM, Martin RA, Demo DS. The Proteus syndrome: CNS manifestations. AJNR Am J Neuroradiol 1998; 19:987–990. 193. Gordon PL, Wilroy RS, Lasater OE, Cohen MM Jr. Neoplasms in Proteus syndrome. Am J Med Genet 1995; 57:74–78. 194. Ross GW, Miller JQ, Persing JA, Urich H. Hemimegalencephaly, hemifacial hypertrophy and intracranial lipoma: a variant of neurofibromatosis. Neurofibromatosis 1989; 2:69–77. 195. Cusmai R, Curatolo P, Mangano S, Cheminal R, Echenne B. Hemimegalencephaly and neurofibromatosis. Neuropediatrics 1990; 21:179–182. 196. Griffiths PD, Gardner SA, Smith M, Rittey C, Powell T. Hemimegalencephaly and focal megalencephaly in tuberous sclerosis complex. AJNR Am J Neuroradiol 1998; 19:1935– 1938. 197. Maloof J, Sledz K, Hogg JF, Bodensteiner JB, Schwartz T, Schochet SS. Unilateral megalencephaly and tuberous sclerosis: related disorders? J Child Neurol 1994; 9:443–446. 198. Galluzzi P, Cerase A, Strambi M, Buoni S, Fois A, Venturi C. Hemimegalencephaly in tuberous sclerosis complex. J Child Neurol 2002; 17:677–680. 199. Jay V, Becker LE, Otsubo H, Hwang PA, Hoffman HJ, Harwood-Nash D. Pathology of temporal lobectomy for refractory seizures in children. Review of 20 cases including some unique malformative lesions. J Neurosurg 1993; 79:53–61.

191

192

P. Tortori-Donati, A. Rossi, and R. Biancheri 200. Prayson RA, Estes ML, Morris HH. Coexistence of neoplasia and cortical dysplasia in patients presenting with seizures. Epilepsia 1993; 34:609–615. 201. Reiner O, Carrozzo R, Shen Y, Wehnert M, Faustinella F, Dobyns WB, Caskey CT, Ledbetter DH. Isolation of a MillerDieker lissencephaly gene containing G protein beta-subunit-like repeats. Nature 1993; 364:717–721. 202. Guerrini R, Carrozzo R. Epilepsy and genetic malformations of the cerebral cortex. Am J Med Genet 2001; 106:160–173. 203. Emes RD, Ponting CP. A new sequence motif linking lissencephaly, Treacher Collins and oral-facial-digital type 1 syndromes, microtubule dynamics and cell migration. Hum Mol Genet 2001; 10:2813–2820. 204. des Portes V, Pinard JM, Billuart P, Vinet MC, Koulakoff A, Carrie A, Gelot A, Dupuis E, Motte J, Berwald-Netter Y, Catala M, Kahn A, Beldjord C, Chelly J. A novel CNS gene required for neuronal migration and involved in X-linked subcortical laminar heterotopia and lissencephaly syndrome. Cell 1998; 92:51–61. 205. Gleeson JG, Allen KM, Fox JW, Lamperti ED, Berkovic S, Scheffer I, Cooper EC, Dobyns WB, Minnerath SR, Ross ME, Walsh CA. Doublecortin, a brain-specific gene mutated in human X-linked lissencephaly and double cortex syndrome, encodes a putative signaling protein. Cell 1998; 92:63–72. 206. Gleeson JG, Lin PT, Flanagan LA, Walsh CA. Doublecortin is a microtubule-associated protein and is expressed widely by migrating neurons. Neuron 1999; 23:257–271. 207. Miller JQ. Lissencephaly in 2 siblings. Neurology 1963; 13:841–850. 208. Dieker H, Edwards RH, ZuRhein G, Chou SM, Hartman HA, Opitz JM. The lissencephaly syndrome. In: Bergsma D (ed) The Clinical Delineation of Birth Defects: Malformation Syndromes. New York: National Foundation-March of Dimes, 1969:53–64. 209. Dobyns WB, van Tuinen P, Ledbetter DH. Clinical diagnostic criteria for Miller-Dieker syndrome [Abstr]. Am J Hum Genet 1988; 43:A46. 210. Pilz DT, Matsumoto N, Minnerath S, Mills P, Gleeson JG, Allen KM, Walsh CA, Barkovich AJ, Dobyns WB, Ledbetter DH, Ross ME. LIS1 and XLIS (DCX) mutations cause most classical lissencephaly, but different patterns of malformation. Hum Mol Genet 1998; 7:2029–2037. 211. Ross ME, Allen KM, Srivastava AK, Featherstone T, Gleeson JG, Hirsch B, Harding BN, Andermann E, Abdullah R, Berg M, Czapansky-Bielman D, Flanders DJ, Guerrini R, Motte J, Mira AP, Scheffer I, Berkovic S, Scaravilli F, King RA, Ledbetter DH, Schlessinger D, Dobyns WB, Walsh CA. Linkage and physical mapping of X-linked lissencephaly/SBH (XLIS): a gene causing neuronal migration defects in human brain. Hum Mol Genet 1997; 6:555–562. 212. Dobyns WB, Truwit CL, Ross ME, Matsumoto N, Pilz DT, Ledbetter DH, Gleeson JG, Walsh CA, Barkovich AJ. Differences in the gyral pattern distinguish chromosome 17–linked and X-linked lissencephaly. Neurology 1999; 53:270–277. 213. Pilz DT, Kuc J, Matsumoto N, Bodurtha J, Bernadi B, Tassinari CA, Dobyns WB, Ledbetter DH. Subcortical band heterotopia in rare affected males can be caused by missense mutations in DCX (XLIS) or LIS1. Hum Mol Genet 1999; 8:1757–1760. 214. Cardoso C, Leventer RJ, Dowling JJ, Ward HL, Chung J, Petras KS, Roseberry JA, Weiss AM, Das S, Martin CL, Pilz DT, Dobyns WB, Ledbetter DH. Clinical and molecular basis of classical lissencephaly: mutations in the LIS1 gene (PAFAH1B1). Hum Mutat 2002; 19:4–15.

215. Barkovich AJ, Guerrini R, Battaglia G, Kalifa G, N’Guyen T, Parmeggiani A, Santucci M, Giovanardi-Rossi P, Granata T, D’Incerti L. Band heterotopia: correlation of outcome with magnetic resonance imaging parameters. Ann Neurol 1994; 36:609–617. 216. Landrieu P, Husson B, Pariente D, Lacroix C. MRI-neuropathological correlations in type 1 lissencephaly. Neuroradiology 1998; 40:173–176. 217. Barkovich AJ, Jackson DE Jr, Boyer RS. Band heterotopias: a newly recognized neuronal migration anomaly. Radiology 1989; 171:455–458 218. Gallucci M, Bozzao A, Curatolo P, Splendiani A, Cifani A, Passariello R. MR imaging of incomplete band heterotopia. AJNR Am J Neuroradiol 1991; 12:701–702. 219. Berry-Kravis E, Israel J. X-linked pachygyria and agenesis of the corpus callosum: evidence for an X chromosome lissencephaly locus. Ann Neurol 1994; 36:229–233. 220. Dobyns WB, Berry-Kravis E, Havernick NJ, Holden KR, Viskochil D. X-linked lissencephaly with absent corpus callosum and ambiguous genitalia. Am J Med Genet 1999; 86:331–337. 221. Bonneau D, Toutain A, Laquerriere A, Marret S, SaugierVeber P, Barthez MA, Radi S, Biran-Mucignat V, Rodriguez D, Gelot A. X-linked lissencephaly with absent corpus callosum and ambiguous genitalia (XLAG): clinical, magnetic resonance imaging, and neuropathological findings. Ann Neurol 2002; 51:340–349. 222. Kitamura K, Yanazawa M, Sugiyama N, Miura H, IizukaKogo A, Kusaka M, Omichi K, Suzuki R, Kato-Fukui Y, Kamiirisa K, Matsuo M, Kamijo S, Kasahara M, Yoshioka H, Ogata T, Fukuda T, Kondo I, Kato M, Dobyns WB, Yokoyama M, Morohashi K. Mutation of ARX causes abnormal development of forebrain and testes in mice and X-linked lissencephaly with abnormal genitalia in humans. Nat Genet 2002; 32:359–369. 223. Hong SE, Shugart YY, Huang DT, Shahwan SA, Grant PE, Hourihane JO, Martin ND, Walsh CA. Autosomal recessive lissencephaly with cerebellar hypoplasia is associated with human RELN mutations. Nat Genet 2000; 26:93–96. 224. Kerner B, Graham JM Jr, Golden JA, Pepkowitz SH, Dobyns WB. Familial lissencephaly with cleft palate and severe cerebellar hypoplasia. Am J Med Genet 1999; 87:440–445. 225. Dobyns WB, Kirkpatrick JB, Hittner HM, Roberts RM, Kretzer FL. Syndromes with lissencephaly. II: Walker-Warburg and cerebro-oculo-muscular syndromes and a new syndrome with type II lissencephaly. Am J Med Genet 1985; 22:157–195. 226. Barkovich AJ. Imaging of the cobblestone lissencephalies. AJNR Am J Neuroradiol 1996; 17:615–618. 227. Aida N, Tamagawa K, Takada K, Yagishita A, Kobayashi N, Chikumaru K, Iwamoto H. Brain MR in Fukuyama congenital muscular dystrophy. AJNR Am J Neuroradiol 1996; 17:605–613. 228. Barkovich AJ. Neuroimaging manifestations and classification of congenital muscular dystrophies. AJNR Am J Neuroradiol 1998; 19:1389–1396. 229. Dobyns WB, Patton MA, Stratton RF, Mastrobattista JM, Blanton SH, Northrup H. Cobblestone lissencephaly with normal eyes and muscle. Neuropediatrics 1996; 27:70–75. 230. van der Knaap MS, Smit LM, Barth PG, Catsman-Berrevoets CE, Brouwer OF, Begeer JH, de Coo IF, Valk J. Magnetic resonance imaging in classification of congenital muscular dystrophies with brain abnormalities. Ann Neurol 1997; 42:50–59.

Brain Malformations 231. Beltran-Valero De Bernabe D, Currier S, Steinbrecher A, Celli J, Van Beusekom E, Van Der Zwaag B, Kayserili H, Merlini L, Chitayat D, Dobyns WB, Cormand B, Lehesjoki AE, Cruces J, Voit T, Walsh CA, Van Bokhoven H, Brunner HG. Mutations in the O-Mannosyltransferase gene POMT1 give rise to the severe neuronal migration disorder WalkerWarburg syndrome. Am J Hum Genet 2002; 71:1033–1043. 232. Rhodes RE, Hatten HP Jr, Ellington KS. Walker-Warburg syndrome. AJNR Am J Neuroradiol 1992; 13:123–126. 233. Gelot A, Billette de Villemeur T, Bordarier C, Ruchoux MM, Moraine C, Ponsot G. Developmental aspects of type II lissencephaly. Comparative study of dysplastic lesions in fetal and post-natal brains. Acta Neuropathol (Berl) 1995; 89:72–84. 234. Dobyns WB. Classification of the cerebro-oculo-muscular syndrome(s). Brain Dev 1993; 15:242–244. 235. Toda T, Segawa M, Nomura Y, Nonaka I, Masuda K, Ishihara T, Sakai M, Tomita I, Origuchi Y, Suzuki M [corrected to Sakai M], et al. Localization of a gene for Fukuyama type congenital muscular dystrophy to chromosome 9q31–33. Nat Genet 1993;5:283–286. 236. Yoshioka M, Kuroki S. Clinical spectrum and genetic studies of Fukuyama congenital muscular dystrophy. Am J Med Genet 1994; 53:245–250. 237. Fukuyama Y, Osawa M, Suzuki H. Congenital progressive muscular dystrophy of the Fukuyama type – clinical, genetic and pathological considerations. Brain Dev 1981; 3:1–29. 238. Aida N, Yagishita A, Takada K, Katsumata Y. Cerebellar MR in Fukuyama congenital muscular dystrophy: polymicrogyria with cystic lesions. AJNR Am J Neuroradiol 1994; 15:1755–1759. 239. Cormand B, Avela K, Pihko H, Santavuori P, Talim B, Topaloglu H, de la Chapelle A, Lehesjoki AE. Assignment of the muscle-eye-brain disease gene to 1p32–p34 by linkage analysis and homozygosity mapping. Am J Hum Genet 1999; 64:126–135. 240. Santavuori P, Somer H, Sainio K, Rapola J, Kruus S, Nikitin T, Ketonen L, Leisti J. Muscle-eye-brain disease (MEB). Brain Dev 1989; 11:147–153. 241. Valanne L, Pihko H, Katevuo K, Karttunen P, Somer H, Santavuori P. MRI of the brain in muscle-eye-brain (MEB) disease. Neuroradiology 1994; 36:473–476. 242. Santavuori P, Valanne L, Autti T, Haltia M, Pihko H, Sainio K. Muscle-eye-brain disease: clinical features, visual evoked potentials and brain imaging in 20 patients. Eur J Paediatr Neurol 1998; 2:41–47. 243. Barkovich AJ, Kuzniecky RI. Gray matter heterotopia. Neurology 2000; 55:1603–1608. 244. Guerrini R, Dravet C, Raybaud C, Roger J, Bureau M, Battaglia A, Livet MO, Gambarelli D, Robain O. Epilepsy and focal gyral anomalies detected by MRI: electroclinico-morphological correlations and follow-up. Dev Med Child Neurol 1992; 34:706–718. 245. Fox JW, Lamperti ED, Eksioglu YZ, Hong SE, Feng Y, Graham DA, Scheffer IE, Dobyns WB, Hirsch BA, Radtke RA, Berkovic SF, Huttenlocher PR, Walsh CA. Mutations in filamin 1 prevent migration of cerebral cortical neurons in human periventricular heterotopia. Neuron 1998; 21:1315– 1325. 246. Gorlin JB, Yamin R, Egan S, Stewart M, Stossel TP, Kwiatkowski DJ, Hartwig JH. Human endothelial actin-binding protein (ABP-280, nonmuscle filamin): a molecular leaf spring. J Cell Biol 1990; 111:1089–1105.

247. Dobyns WB, Guerrini R, Czapansky-Beilman DK, Pierpont ME, Breningstall G, Yock DH Jr, Bonanni P, Truwit CL. Bilateral periventricular nodular heterotopia with mental retardation and syndactyly in boys: a new X-linked mental retardation syndrome. Neurology 1997; 49:1042–1047. 248. Guerrini R, Dobyns WB. Bilateral periventricular nodular heterotopia with mental retardation and frontonasal malformation. Neurology 1998; 51:499–503. 249. Sheen VL, Dixon PH, Fox JW, Hong SE, Kinton L, Sisodiya SM, Duncan JS, Dubeau F, Scheffer IE, Schachter SC, Wilner A, Henchy R, Crino P, Kamuro K, DiMario F, Berg M, Kuzniecky R, Cole AJ, Bromfield E, Biber M, Schomer D, Wheless J, Silver K, Mochida GH, Berkovic SF, Andermann F, Andermann E, Dobyns WB, Wood NW, Walsh CA. Mutations in the X-linked filamin 1 gene cause periventricular nodular heterotopia in males as well as in females. Hum Mol Genet 2001; 10:1775–1783. 250. Harding B. Gray matter heterotopia. In: Guerrini R, Andermann F, Canapicchi R, Roger J, Zifkin BG, Pfanner P (eds) Dysplasias of Cerebral Cortex and Epilepsy. Philadelphia: Lippincott-Raven, 1996:81–88. 251. Barkovich AJ, Kjos BO. Gray matter heterotopias: MR characteristics and correlation with developmental and neurologic manifestations. Radiology 1992; 182:493–499. 252. Poussaint TY, Fox JW, Dobyns WB, Radtke R, Scheffer IE, Berkovic SF, Barnes PD, Huttenlocher PR, Walsh CA. Periventricular nodular heterotopia in patients with filamin1 gene mutations: neuroimaging findings. Pediatr Radiol 2000; 30:748–755. 253. Barkovich AJ. Morphologic characteristics of subcortical heterotopia: MR imaging study. AJNR Am J Neuroradiol 2000; 21:290–295. 254. Barkovich AJ. Subcortical heterotopia: a distinct clinicoradiologic entity. AJNR Am J Neuroradiol 1996; 17:1315– 1322. 255. Simonetti L, Elefante R, Tortori-Donati P. Anomalie della migrazione neuroblastica. In: Tortori-Donati P, Taccone A, Longo M (eds) Malformazioni Cranio-Encefaliche. Neuroradiologia. Turin: Minerva Medica, 1996:237–262. 256. Thompson JE, Castillo M, Thomas D, Smith MM, Mukherji SK. Radiologic-pathologic correlation – Polymicrogyria. AJNR Am J Neuroradiol 1997; 18:307–312. 257. Barkovich AJ, Hevner R, Guerrini R. Syndromes of bilateral symmetrical polymicrogyria. AJNR Am J Neuroradiol 1999; 20:1814–1821. 258. Piao X, Basel-Vanagaite L, Straussberg R, Grant PE, Pugh EW, Doheny K, Doan B, Hong SE, Shugart YY, Walsh CA. An autosomal recessive form of bilateral frontoparietal polymicrogyria maps to chromosome 16q12.2–21. Am J Hum Genet 2002; 70:1028–1033. 259. Villard L, Nguyen K, Cardoso C, Martin CL, Weiss AM, Sifry-Platt M, Grix AW, Graham JM Jr, Winter RM, Leventer RJ, Dobyns WB. A locus for bilateral perisylvian polymicrogyria maps to Xq28. Am J Hum Genet 2002; 70:1003–1008. 260. Barkovich AJ, Lindan CE. Congenital cytomegalovirus infection of the brain: imaging analysis and embryologic considerations. AJNR Am J Neuroradiol 1994; 15:703– 715. 261. Barkovich AJ, Rowley H, Bollen A. Correlation of prenatal events with the development of polymicrogyria. AJNR Am J Neuroradiol 1995; 16(4 Suppl):822–827. 262. Marques Dias MJ, Harmant-van Rijckevorsel G, Landrieu P, Lyon G. Prenatal cytomegalovirus disease and cerebral microgyria: evidence for perfusion failure, not disturbance

193

194

P. Tortori-Donati, A. Rossi, and R. Biancheri of histogenesis, as the major cause of fetal cytomegalovirus encephalopathy. Neuropediatrics 1984; 15:18–24. 263. Guerrini R, Barkovich AJ, Sztriha L, Dobyns WB. Bilateral frontal polymicrogyria: a newly recognized brain malformation syndrome. Neurology 2000; 54:909–913. 264. Harding B, Copp A. Malformations. In: Graham DI, Lantos PI (eds) Greenfield’s Neuropathology, 6th edn. London: Arnold, 1997:397–533. 265. Cellerini M, Mascalchi M, Salvi F, Muscas GC, Dal Pozzo G. MRI of congenital Foix-Chavany-Marie syndrome. Pediatr Radiol 1995; 25:316–317. 266. Pascual-Castroviejo I, Pascual-Pascual SI, Viano J, Martinez V, Palencia R. Unilateral polymicrogyria: a common cause of hemiplegia of prenatal origin. Brain Dev 2001; 23:216– 222. 267. Kuzniecky R, Andermann F, Guerrini R. Congenital bilateral perisylvian syndrome: study of 31 patients. The CBPS Multicenter Collaborative Study. Lancet 1993; 341:608–612. 268. Kuzniecky R, Andermann F, and the CBPS Study Group. The conenital bilateral perisylvian syndrome: imaging findings in a multicenter study. AJNR Am J Neuroradiol 1994; 15:139–144. 269. Gropman AL, Barkovich AJ, Vezina LG, Conry JA, Dubovsky EC, Packer RJ. Pediatric congenital bilateral perisylvian syndrome: clinical and MRI features in 12 patients. Neuropediatrics 1997; 28:198–203. 270. Van Bogaert P, David P, Gillain CA, Wikler D, Damhaut P, Scalais E, Nuttin C, Wetzburger C, Szliwowski HB, Metens T, Goldman S. Perisylvian dysgenesis. Clinical, EEG, MRI and glucose metabolism features in 10 patients. Brain 1998; 121:2229–2238. 271. Borgatti R, Triulzi F, Zucca C, Piccinelli P, Balottin U, Carrozzo R, Guerrini R. Bilateral perisylvian polymicrogyria in three generations. Neurology 1999; 52:1910–1913. 272. Guerreiro MM, Andermann E, Guerrini R, Dobyns WB, Kuzniecky R, Silver K, Van Bogaert P, Gillain C, David P, Ambrosetto G, Rosati A, Bartolomei F, Parmeggiani A, Paetau R, Salonen O, Ignatius J, Borgatti R, Zucca C, Bastos AC, Palmini A, Fernandes W, Montenegro MA, Cendes F, Andermann F. Familial perisylvian polymicrogyria: a new familial syndrome of cortical maldevelopment. Ann Neurol 2000; 48:39–48. 273. Guerrini R, Dubeau F, Dulac O, Barkovich AJ, Kuzniecky R, Fett C, Jones-Gotman M, Canapicchi R, Cross H, Fish D, Bonanni P, Jambaque I, Andermann F. Bilateral parasagittal parietooccipital polymicrogyria and epilepsy. Ann Neurol 1997; 41:65–73. 274. Montenegro MA, Guerreiro MM, Lopes-Cendes I, Cendes F. Bilateral posterior parietal polymicrogyria: a mild form of congenital bilateral perisylvian syndrome? Epilepsia 2001; 42:845–849. 275. Aicardi J. Aicardi syndrome. In: Guerrini R, Andermann F, Canapicchi R, Roger J, Zifkin BG, Pfanner P (eds) Dysplasias of Cerebral Cortex and Epilepsy. Philadelphia: Lippincott-Raven, 1996:211–216. 276. Pavone L, Rizzo R, Pavone P, Curatolo P, Dobyns WB. Diffuse polymicrogyria associated with congenital hydrocephalus, craniosynostosis, severe mental retardation, and minor facial and genital anomalies. J Child Neurol 2000; 15:493–495. 277. Duckett S. Pediatric Neuropathology. Philadelphia: Lippincott Williams & Wilkins, 1995. 278. Packard AM, Miller VS, Delgado MR. Schizencephaly: correlations of clinical and radiologic features. Neurology 1997; 48:1427–1434.

279. Barkovich AJ, Kjos BO. Schizencephaly: correlation of clinical findings with MR characteristics. AJNR Am J Neuroradiol 1992; 13:85–94. 280. Hayashi N, Tsutsumi Y, Barkovich AJ. Morphological features and associated anomalies of schizencephaly in the clinical population: detailed analysis of MR images. Neuroradiology 2002; 44:418–427. 281. Brunelli S, Faiella A, Capra V, Nigro V, Simeone A, Cama A, Boncinelli E. Germline mutation in the homeobox gene EMX2 in patients with severe schizencephaly. Nature Genet 1996; 12:94–96. 282. Hung PC, Wang HS, Yeh YS, Lui TN, Lee ST. Coexistence of schizencephaly and intracranial arteriovenous malformation in an infant. AJNR Am J Neuroradiol 1996; 17:1921– 1922. 283. Barkovich AJ, Norman D. MR imaging of schizencephaly. AJR Am J Roentgenol 1988; 150:1391–1396. 284. Raybaud C, Girard N, Levrier O, Peretti-Viton P, Manera L, Farnarier P. Schizencephaly: correlation between the lobar topography of the cleft(s) and absence of the septum pellucidum. Childs Nerv Syst 2001; 17:217–222. 285. Marciano S, Gambarelli D, Roussel M, Gire C, Raybaud C, Girard N. Challenge and resolution. Int J Neuroradiol 1999; 5:43–50. 286. Volpe JJ. Neurology of the Newborn, 4th edn. Philadelphia: Saunders, 2001. 287. Barkovich AJ, Peck WW. MR of Zellweger syndrome. AJNR Am J Neuroradiol 1997; 18:1163–1170. 288. Moroni I, Bugiani M, Bizzi A, Castelli G, Lamantea E, Uziel G. Cerebral white matter involvement in children with mitochondrial encephalopathies. Neuropediatrics 2002; 33:79–85. 289. Barkovich AJ, Peacock W. Sublobar dysplasia. A new malformation of cortical development. Neurology 1998; 50:1383– 1387. 290. Hallonet ME, Le Douarin NM. Tracing neuroepithelial cells of the mesencephalic and metencephalic alar plates during cerebellar ontogeny in quail-chick chimaeras. Eur J Neurosci 1993; 5:1145–1155. 291. Alvarez Otero R, Sotelo C, Alvarado-Mallart RM. Chick/ quail chimeras with partial cerebellar grafts: an analysis of the origin and migration of cerebellar cells. J Comp Neurol 1993; 333:597–615. 292. Yachnis AT, Rorke LB. Neuropathology of Joubert syndrome. J Child Neurol 1999; 14:655–659. 293. Yachnis AT. Rhombencephalosynapsis with massive hydrocephalus: case report and pathogenetic considerations. Acta Neuropathol (Berl) 2002; 103:301–304. 294. Blake JA. The roof and lateral recesses of the fourth ventricle, considered morphologically and embryologically. J Comp Neurol 1900; 10:79–108. 295. Altman NR, Naidich TP, Braffman BH. Posterior fossa malformations. AJNR Am J Neuroradiol 1992; 13:691–724. 296. Berry M, Bannister LH, Standring M. Nervous system. In: Williams PL (ed) Gray’s Anatomy, 38th edn. Edinburgh: Churchill Livingstone, 1999:1011–1065. 297. O’Rahilly R, Muller F. The meninges in human development. J Neuropathol Exp Neurol 1986; 45:588–608. 298. Raybaud C. [Cystic malformations of the posterior fossa. Abnormalities associated with the development of the roof of the fourth ventricle and adjacent meningeal structures.] J Neuroradiol 1982; 9:103–133. 299. Tortori-Donati P, Longo M, Fondelli MP, Cama A, Carini S, Rossi A. Malformazioni del cervelletto. In: Tortori-Donati P,

Brain Malformations Taccone A, Longo M (eds) Malformazioni cranio-encefaliche. Neuroradiologia. Turin: Minerva Medica, 1996:159–207. 300. Sidman RL, Rakic P. Development of the human nervous system. In: Haymaker W, Adams R (eds) Histology and Histopathology of the Nervous System. Springfield: Thomas, 1982. 301. Costa C, Hauw JJ. Pathology of the cerebellum, brain stem, and spinal cord. In: Duckett S (ed) Pediatric Neuropathology. Baltimore: Williams & Wilkins, 1995: 217–238. 302. Benda CE. The Dandy-Walker syndrome or the so-called atresia of the foramen of Magendie. J Neuropathol Exp Neurol 1954; 13:14–29. 303. Dandy WE, Blackfan KD. Internal hydrocephalus. An experimental, clinical and pathological study. Am J Dis Child 1914; 8:406–482. 304. Taggart TK Jr, Walker AE. Congenital atresia of the foramina of Luschka and Magendie. Arch Neurol Psychiatry 1942; 48:583–612. 305. Harwood-Nash DC, Fitz CR. Neuroradiology in Infants and Children, vol.3. St. Louis: Mosby, 1976:1014–1019. 306. Barkovich AJ, Kjos BO, Norman D, Edwards MS. Revised classification of posterior fossa cysts and cystlike malformations based on the results of multiplanar MR imaging. AJR Am J Roentgenol 1989; 153:1289–1300. 307. Kumar R, Jain MK, Chhabra DK. Dandy-Walker syndrome: different modalities of treatment and outcome in 42 cases. Childs Nerv Syst 2001; 17:348–352. 308. Miyamori T, Okabe T, Hasegawa T, Takinami K, Matsumoto T. Dandy-Walker syndrome successfully treated with cystoperitoneal shunting – case report. Neurol Med Chir (Tokyo) 1999; 39:766–768. 309. Maria BL, Zinreich SJ, Carson BC, Rosenbaum AE, Freeman JM. Dandy-Walker syndrome revisited. Pediatr Neurosci 1987; 13:45–51. 310. Morava E, Adamovich K, Czeizel AE. Dandy-Walker malformation and polydactyly: a possible expression of hydrolethalus sindrome. Clin Genet 1996; 49:211–215. 311. Leonardi ML, Pai GS, Wilkes B, Lebel RR. Ritscher-Schinzel cranio-cerebello-cardiac (3C) sindrome: report of four new cases and review. Am J Med Genet 2001; 102:237–242. 312. Berker M, Oruckaptan HH, Oge HK, Benli K. Neurocutaneous melanosis associated with Dandy-Walker malformation. Case report and review of the literature. Pediatr Neurosurg 2000; 33:270–273. 313. Barkovich AJ, Frieden IJ, Williams ML. MR of neurocutaneous melanosis. AJNR Am J Neuroradiol 1994; 15:859–867. 314. Chaloupka JC, Wolf RJ, Varma PK. Neurocutaneous melanosis with the Dandy-Walker malformation: a possible rare pathoetiologic association. Neuroradiology 1996; 38:486– 489. 315. Frieden IJ, Reese V, Cohen D. PHACE syndrome. The association of posterior fossa brain malformations, hemangiomas, arterial anomalies, coarctation of the aorta and cardiac defects, and eye abnormalities. Arch Dermatol 1996; 132:307–311. 316. Rossi A, Bava GL, Biancheri R, Tortori-Donati P. Posterior fossa and arterial abnormalities in patients with facial capillary haemangioma: presumed incomplete phenotypic expression of PHACES syndrome. Neuroradiology 2001; 43:934–940. 317. Peters V, Penzien JM, Reiter G, Korner C, Hackler R, Assmann B, Fang J, Schaefer JR, Hoffmann GF, Heidemann PH. Congenital disorder of glycosylation IId (CDG-IId) – A new entity: clinical presentation with Dandy-Walker malformation and myopathy. Neuropediatrics 2002; 33:27–32.

318. Humbertclaude VT, Coubes PA, Leboucq N, Echenne BB. Familial Dandy-Walker malformation and leukodystrophy. Pediatr Neurol 1997; 16:326–328. 319. Tortori-Donati P, Fondelli MP, Rossi A, Cama A, Carini S, Piatelli GL, Ravegnani M. Malformazioni cistiche della fossa cranica posteriore. Riv Neuroradiol 1994; 7:199–207. 320. Bentson JR, Alberti J. The fourth ventricle. In: Newton TH, Potts DG (eds) Radiology of the Skull and Brain. Ventricles and Cisterns. St. Louis : CV Mosby, 1978 :3303–3362. 321. Strand RD, Barnes PD, Poussaint TY, Estroff JA, Burrows PE. Cystic retrocerebellar malformations: unification of the Dandy-Walker complex and the Blake’s pouch cyst. Pediatr Radiol 1993;23:258–260. 322. Tortori-Donati P, Fondelli MP, Rossi A, Carini S. Cystic malformations of the posterior cranial fossa originating from a defect of the posterior membranous area. Mega cisterna magna and persisting Blake’s pouch: two separate entities. Childs Nerv Syst 1996;12:303–308. 323. Rossi A, Catala M, Di Comite R, Biancheri R, Cama A, Tortori-Donati P. Persistence of the Blake’s pouch: an embryogenetic theory for the occurrence of infraretrocerebellar “cysts” associated with normal cerebellum and tetraventricular hydrocephalus [Abstr]. Proceedings of the 41st Meeting of the American Society of Neuroradiology, 2003:135–136. 324. Calabrò F, Arcuri T, Jinkins JR. Blake’s pouch cyst: an entity within the Dandy-Walker continuum. Neuroradiology 2000; 42:290–295. 325. Gonsette R, Potvliege R, Andre-Balòisaux G, Stenuit J. La méga grande citerne: étude clinique, radiologique et anatomopathologique. Acta Neurol Belg 1968; 68:559–570. 326. Joubert M, Eisenring JJ, Andermann F. Familial dysgenesis of the vermis: a syndrome of hyperventilation, abnormal eye movements and retardation. Neurology 1968;18:302– 303. 327. Maria BL, Hoang KB, Tusa RJ, Mancuso AA, Hamed LM, Quisling RG, Hove MT, Fennell EB, Booth-Jones M, Ringdahl DM, Yachnis AT, Creel G, Frerking B. “Joubert syndrome” revisited: key ocular motor signs with magnetic resonance imaging correlation. J Child Neurol 1997; 12:423–430. 328. Chance PF, Cavalier L, Satran D, Pellegrino JE, Koenig M, Dobyns WB. Clinical nosologic and genetic aspects of Joubert and related syndromes. J Child Neurol 1999; 14:660– 666. 329. Satran D, Pierpont ME, Dobyns WB. Cerebello-oculo-renal syndromes including Arima, Senior-Löken and COACH syndromes: more than just variants of Joubert syndrome. Am J Med Genet 1999; 86:459–469. 330. Blair IP, Gibson RR, Bennett CL, Chance PF. Search for genes involved in Joubert syndrome: evidence that one or more major loci are yet to be identified and exclusion of candidate genes EN1, EN2, FGF8, and BARHL1. Am J Med Genet 2002; 107:190–196. 331. Maria BL, Quisling RG, Rosainz LC, Yachnis AT, Gitten J, Dede D, Fennell E. Molar tooth sign in Joubert syndrome: clinical, radiologic, and pathologic significance. J Child Neurol 1999; 14:368–376. 332. Friede RL, Boltshauser E. Uncommon syndromes of cerebellar vermis aplasia. I: Joubert syndrome. Dev Med Child Neurol 1978; 20:758–763. 333. Maria BL, Boltshauser E, Palmer SC, Tran TX. Clinical features and revised diagnostic criteria in Joubert syndrome. J Child Neurol 1999; 14:583–591. 334. Gitten J, Dede D, Fennell E, et al. Neurobehavioral develop-

195

196

P. Tortori-Donati, A. Rossi, and R. Biancheri ment in Joubert syndrome. J Child Neurol 1998; 13:391– 397. 335. Steinlin M, Schmid M, Landau K, Boltshauser E. Follow-up in children with Joubert syndrome. Neuropediatrics 1997; 28:204–211. 336. Quisling RG, Barkovich AJ, Maria BL. Magnetic resonance features and classification of central nervous system malformations in Joubert syndrome. J Child Neurol 1999; 14:628–635. 337. Maria BL, Bozorgmanesh A, Kimmel KN, Theriaque D, Quisling RG. Quantitative assessment of brainstem development in Joubert syndrome and Dandy-Walker syndrome. J Child Neurol 2001; 16:751–758. 338. Truwit CL, Barkovich AJ, Shanahan R, Maroldo TV. MR imaging of rhombencephalosynapsis: report of three cases and review of the literature. AJNR Am J Neuroradiol 1991; 12:957–965. 339. Romanengo M, Tortori-Donati P, Di Rocco M. Rhombencephalosynapsis with facial anomalies and probable autosomal recessive inheritance: a case report. Clin Genet 1997; 52:184–186. 340. Montull C, Mercader JM, Peri J, Martinez Ferri M, Bonaventura I. Neuroradiological and clinical findings in rhombencephalosynapsis. Neuroradiology 2000; 42:272–274. 341. Utsunomiya H, Takano K, Ogasawara T, Hashimoto T, Fukushima T, Okazaki M. Rhombencephalosynapsis: cerebellar embryogenesis. AJNR Am J Neuroradiol 1998; 19:547–549. 342. Jellinger KA. Rhombencephalosynapsis. Acta Neuropathol (Berl) 2002; 103:305–306. 343. Garfinkle WB. Aventriculy: a new entity. AJNR Am J Neuroradiol 1997; 17:1649–1650. 344. Aydingoz U, Cila A, Aktan G. Rhombencephalosynapsis associated with hand anomalies. Br J Radiol 1997; 70:764– 766. 345. Manzanares M, Trainor PA, Ariza-McNaughton L, Nonchev S, Krumlauf R. Dorsal patterning defects in the hindbrain, roof plate and skeleton in the dreher (dr(J)) mouse mutant. Mech Dev 2000; 94:147–156. 346. Millonig JH, Millen KJ, Hatten ME. The mouse Dreher gene Lmx1a controls formation of the roof plate in the vertebrate CNS. Nature 2000; 403(6771):764–769. 347. Costa C, Harding B, Copp AJ. Neuronal migration defects in the Dreher (Lmx1a) mutant mouse: role of disorders of the glial limiting membrane. Cereb Cortex 2001; 11:498–505. 348. Friede RL. Uncommon syndromes of cerebellar vermis aplasia. II: Tecto-cerebellar dysraphia with occipital encephalocele. Dev Med Child Neurol 1978; 20:764–772. 349. Gardner RJ, Coleman LT, Mitchell LA, et al. Near-total absence of the cerebellum. Neuropediatrics 2001; 32:62– 68. 350. Velioglu SK, Kuzeyli K, Zzmenoglu M. Cerebellar agenesis: a case report with clinical and MR imaging findings and a review of the literature. Eur J Neurol 1998; 5:503–506. 351. Barth PG. Pontocerebellar hypoplasias. An overview of a group of inherited neurodegenerative disorders with fetal onset. Brain Dev 1993; 15:411–422. 352. Barth PG. Pontocerebellar hypoplasia – how many types? Eur J Paediatr Neurol 2000; 4:161–162. 353. Stromme P, Maehlen J, Strom EH, Torvik A. [The carbohydrate deficient glycoprotein syndrome]. Tidsskr Nor Laegeforen 1991; 111:1236–1237. 354. de Koning TJ, de Vries LS, Groenendaal F, et al. Pontocerebellar hypoplasia associated with respiratory-chain defects. Neuropediatrics 1999; 30:93–95.

355. Squier MV. Development of the cortical dysplasia of type II lissencephaly. Neuropathol Appl Neurobiol 1993; 19:209– 213. 356. Uhl M, Pawlik H, Laubenberger J, Darge K, Baborie A, Korinthenberg R, Langer M. MR findings in pontocerebellar hypoplasia. Pediatr Radiol 1998; 28:547–551. 357. Goasdoué P, Rodriguez D, Moutard ML, Robain O, Lalande G, Adamsbaum C. Pontoneocerebellar hypoplasia: definition of MR features. Pediatr Radiol 2001; 31:613–618. 358. Boltshauser E, Steinlin M, Martin E, Deonna T. Unilateral cerebellar aplasia. Neuropediatrics 1996; 27:50–53. 359. Wang PJ, Maeda Y, Izumi T, Yajima K, Hara M, Kobayashi N, Fukuyama Y. An association of subtotal cerebellar agenesis with organoid nevus – a possible new variety of neurocutaneous syndrome. Brain Dev 1983; 5:503–508. 360. Serrano Gonzalez C, Prats Vinas JM. [Unilateral aplasia of the cerebellum in Aicardi’s syndrome] Neurologia 1998; 13:254–256. 361. Boltshauser E. Joubert syndrome: more than lower cerebellar vermis hypoplasia, less than a complex brain malformation [letter]. Am J Med Genet 2002; 109:332. 362. Soto-Ares G, Delmaire C, Deries B, Vallee L, Pruvo JP. Cerebellar cortical dysplasia: MR findings in a complex entity. AJNR Am J Neuroradiol 2000; 21:1511–1519. 363. Nowak DA, Trost HA. Lhermitte-Duclos disease (dysplastic cerebellar gangliocytoma): a malformation, hamartoma or neoplasm? Acta Neurol Scand 2002; 105:137–145. 364. Padberg GW, Schot JD, Vielvoye GJ, Bots GT, de Beer FC. Lhermitte-Duclos disease and Cowden disease: a single phakomatosis. Ann Neurol 1991; 29:517–523. 365. Robinson S, Cohen AR. Cowden disease and LhermitteDuclos disease: characterization of a new phakomatosis. Neurosurgery 2000; 46:371–383. 366. Demaerel P. Abnormalities of cerebellar foliation and fissuration: classification, neurogenetics and clinicoradiological correlations. Neuroradiology 2002; 44:639–646. 367. Sasaki M, Ehara S, Watabe T. Cerebellar polymicrogyria [letter]. AJNR Am J Neuroradiol 1997; 18:394–396. 368. Soto-Ares G, Devisme L, Jorriot S, Deries B, Pruvo JP, Ruchoux MM. Neuropathologic and MR imaging correlation in a neonatal case of cerebellar cortical dysplasia. AJNR Am J Neuroradiol 2002; 23:1101–1104. 369. Demaerel P, Wilms G, Marchal G. Rostral vermian cortical dysplasia: MRI. Neuroradiology 1999; 41:190–194. 370. Sasaki M, Oikawa H, Ehara S, Tamakawa Y, Takahashi S, Tohgi H. Disorganised unilateral cerebellar folia: a mild form of cerebellar cortical dysplasia? Neuroradiology 2001; 43:151–155. 371. Demaerel P, Lagae L, Casaer P, Baert AL. MR of cerebellar cortical dysplasia. AJNR Am J Neuroradiol 1998; 19:984– 986. 372. Pascual-Castroviejo I, Gutierrez M, Morales C, GonzalezMediero I, Martinez-Bermejo A, Pascual-Pascual SI. Primary degeneration of the granular layer of the cerebellum. A study of 14 patients and review of the literature. Neuropediatrics 1994; 25:183–190. 373. Bodensteiner JB, Schaefer GB, Keller GM, Thompson JN, Bowen MK. Macrocerebellum: neuroimaging and clinical features of a newly recognized condition. J Child Neurol 1997; 12:365–368. 374. Rossi A, Catala M, Biancheri R, Di Comite R, Tortori-Donati P. MR imaging of brain-stem hypoplasia in horizontal gaze palsy with progressive scoliosis. AJNR Am J Neuroradiol 2004; 25:1046–1048.

Brain Malformations 375. Mamourian AC, Miller G. Neonatal pontomedullary disconnection with aplasia or destruction of the lower brain stem: a case of pontoneocerebellar hypoplasia? AJNR Am J Neuroradiol 1994; 15:1483–1485. 376. Sarnat HB, Benjamin DR, Siebert JR, Kletter GB, Cheyette SR. Agenesis of the mesencephalon and metencephalon with cerebellar hypoplasia: putative mutation in the EN2 gene. Report of 2 cases in early infancy. Pediatr Dev Pathol 2002; 5:54–68. 377. Uematsu J, Ono K, Yamano T, Shimada M. Development of corticospinal tract fibers and their plasticity. II. Neonatal unilateral cortical damage and subsequent development of the corticospinal tract in mice. Brain Dev 1996; 18:173– 178. 378. Schachenmayr W, Friede RL. Dystopic myelination with hypertrophy of pyramidal tract. J Neuropathol Exp Neurol 1978; 37:34–44. 379. Milhorat TH, Chou MW, Trinidad EM, Kula RW, Mandell M, Wolpert C, Speer MC. Chiari I malformation redefined: clinical and radiological findings for 364 symptomatic patients. Neurosurgery 1999; 44:1005–1017. 380. Ehara S, Shimamura T. The semantics of terminology: distinguishing Arnold-Chiari malformations from Chiari malformations. J Bone Joint Surg 2002; 84A:321. 381. Chiari H. Über veränderungen des kleinhirns infolge von hydrocephalie des grosshirns. Dtsch Med Wochenschr 1891; 17:1172–1175. 382. Nishikawa M, Sakamoto H, Hakuba A, Nakanishi N, Inoue Y. Pathogenesis of Chiari malformation: a morphometric study of the posterior cranial fossa. J Neurosurg 1997; 86:40–47. 383. Turgut M. Chiari type I malformation in two monozygotic twins. Br J Neurosurg 2001; 15:279–280. 384. Cavender RK, Schmidt JH. Tonsillar ectopia and Chiari malformations: monozygotic triplets. J Neurosurg 1995; 82:497–500. 385. Tortori-Donati P, Cama A, Fondelli MP, Rossi A. Le malformazioni di Chiari. In: Tortori-Donati P, Taccone A, Longo M (eds) Malformazioni cranio-encefaliche. Neuroradiologia. Turin: Minerva Medica, 1996:209–236. 386. Elster AD, Chen MYM. Chiari I malformations: clinical and radiological reappraisal. Radiology 1992; 183:347–353. 387. Park JK, Gleason PL, Madsen JR, Goumnerova LC, Scott RM. Presentation and management of Chiari I malformation in children. Pediatr Neurosurg 1997; 26:190–196. 388. Pujol J, Roig C, Capdevila A, Pou A, Marti-Vilalta JL, Kulisevsky J, Escartin A, Zannoli G. Motion of the cerebellar tonsils in Chiari type I malformation studied by cine phase-contrast MRI. Neurology 1995; 45:1746– 1753. 389. Oldfield EH, Muraszko K, Shawker TH, Patronas NJ. Pathophysiology of syringomyelia associated with Chiari I malformation of the cerebellar tonsils. Implications for diagnosis and treatment. J Neurosurg 1994; 80:3–15. 390. Nakamura N, Iwasaki Y, Hida K, Abe H, Fujioka Y, Nagashima K. Dural band pathology in syringomyelia with Chiari type I malformation. Neuropathology 2000; 20:38–43. 391. Castillo M, Wilson JD. Spontaneous resolution of a Chiari I malformation: MR demonstration. AJNR Am J Neuroradiol 1995; 16:1158–1160. 392. Klekamp J, Iaconetta G, Samii M. Spontaneous resolution of Chiari I malformation and syringomyelia: case report and review of the literature. Neurosurgery 2001; 48:664–667. 393. Welch K, Shillito J, Strand R, Fischer EG, Winston KR. Chiari

I “malformations” – an acquired disorder? J Neurosurg 1981; 55:604–609. 394. Sathi S, Stieg PE. “Acquired” Chiari I malformation after multiple lumbar punctures: case report. Neurosurgery 1993; 32:306–309. 395. Chumas PD, Armstrong DC, Drake JM, Kulkarni AV, Hoffman HJ, Humphreys RP, Rutka JT, Hendrick EB. Tonsillar herniation: the rule rather than the exception after lumboperitoneal shunting in the pediatric population. J Neurosurg 1993; 78:568–573 396. Lazareff JA, Kelly J, Saito M. Herniation of cerebellar tonsils following supratentorial shunt placement. Childs Nerv Syst 1998; 14:394–397. 397. Atkinson JL, Weinshenker BG, Miller GM, Piepgras DG, Mokri B. Acquired Chiari I malformation secondary to spontaneous spinal cerebrospinal fluid leakage and chronic intracranial hypotension syndrome in seven cases. J Neurosurg 1998; 88:237–242. 398. Lee M, Rezai AR, Wisoff JH. Acquired Chiari-I malformation and hydromyelia secondary to a giant craniopharyngioma. Pediatr Neurosurg 1995; 22:251–254. 399. Sheehan JM, Jane JA Sr. Resolution of tonsillar herniation and syringomyelia after supratentorial tumor resection: case report and review of the literature. Neurosurgery 2000; 47:233–235. 400. Payner TD, Prenger E, Berger TS, Crone KR. Acquired Chiari malformations: incidence, diagnosis, and management. Neurosurgery 1994; 34:429–434. 401. Meadows J, Kraut M, Guarnieri M, Haroun RI, Carson BS. Asymptomatic Chiari Type I malformations identified on magnetic resonance imaging. J Neurosurg 2000; 92:920– 926. 402. Wu YW, Chin CT, Chan KM, Barkovich AJ, Ferriero DM. Pediatric Chiari I malformations. Do clinical and radiologic features correlate? Neurology 1999; 53:1271–1276. 403. Bunc G, Vorsic M. Presentation of a previously asymptomatic Chiari I malformation by a flexion injury to the neck. J Neurotrauma 2001; 18:645–648. 404. Leong WK, Kermode AG. Acute deterioration in Chiari type 1 malformation after chiropractic cervical manipulation. J Neurol Neurosurg Psychiatry 2001; 70:816–817. 405. Ball WS Jr, Crone KR. Chiari I malformation: from Dr Chiari to MR imaging. Radiology 1995; 195:602–604. 406. Christophe C, Dan B. magnetic resonance imaging cranial and cerebral dimensions: is there a relationship with Chiari I malformation? A preliminary report in children. Europ J Paediatr Neurol 1999; 3:15–24. 407. Aboulezz AO, Sartor K, Geyer CA, Gado MH. Position of cerebellar tonsils in the normal population and in patients with Chiari malformation: a quantitative approach with MR imaging. J Comput Assist Tomogr 1985; 9:1033–1036. 408. Mikulis DJ, Diaz O, Egglin TK, Sanchez R. Variance of the position of the cerebellar tonsils with age: preliminary report. Radiology 1992; 183:725–728. 409. Cama A, Tortori-Donati P, Piatelli GL, Fondelli MP, Andreussi L. Chiari complex in children: neuroradiological diagnosis, neurosurgical treatment and proposal of a new classification (312 cases). Eur J Pediatr Surg 1995; 5:35–38. 410. Du Boulay GH, Shah S, Currie JC, logue V. The macheanism of hydromyelia in Chiari type I malformation. Br J Radiol 1974; 47:579–587. 411. Bhadelia RA, Bogdan AR, Wolpert SM, Lev S, Appignani BA, Heilman CB. Cerebrospinal fluid flow waveforms: analysis in patients with Chiari I malformation by means of gated

197

198

P. Tortori-Donati, A. Rossi, and R. Biancheri phase-contrast MR imaging velocity measurements. Radiology 1995; 196:195–202. 412. Wolpert SM, Bhadelia RA, Bogdan AR, Cohen AR. Chiari I malformations: assessment with phase-contrast velocity MR. AJNR Am J Neuroradiol 1994; 15:1299–1308. 413. Hofmann E, Warmuth-Metz M, Bendszus M, Solymosi L. Phase-contrast MR imaging of the cervical CSF and spinal cord: volumetric motion analysis in patients with Chiari I malformation. AJNR Am J Neuroradiol 2000; 21:151–158. 414. Hamilton J, Blaser S, Daneman D. MR imaging in idiopathic growth hormone deficiency. AJNR Am J Neuroradiol 1998; 19:1609–1615. 415. Triulzi F, Scotti G, di Natale B, Pellini C, Lukezic M, Scognamiglio M, Chiumello G. Evidence of a congenital midline brain anomaly in pituitary dwarfs: a magnetic resonance imaging study in 101 patients. Pediatrics 1994; 93:409–416. 416. Ellis SJ, Crockard A, Barnard RO. Klinefelter’s syndrome, cerebral germinoma, Chiari malformation, and syrinx: a case report. Neurosurgery 1986;18:220–222. 417. Shaw CM, Alvord EC Jr. Hydrocephalus. In: Duckett S (ed) Pediatric neuropathology. Baltimore: Williams & Wilkins, 1995:168–179. 418. Tortori-Donati P, Rossi A, Cama A. Spinal dysraphism: a review of neuroradiological features with embryological correlations and proposal for a new classification. Neuroradiology 2000; 42:471–491. 419. McLone DG, Knepper PA. The cause of Chiari II malformation: a unified theory. Pediatr Neurosci 1989; 15:1–12. 420. Wolpert SM, Scott RM, Platenberg C, Runge VM. The clinical significance of hindbrain herniation and deformity as shown on MR images of patients with Chiari II malformation. AJNR Am J Neuroradiol 1988; 9:1075–1078.

421. Narayan P, Mapstone TB, Tubbs RS, Grabb PA, Frye T. Clinical significance of cervicomedullary deformity in Chiari II malformation. Pediatr Neurosurg 2001; 35:140–144. 422. McLone DG, Naidich TP. Developmental morphology of the subarachnoid space, brain vasculature, and contiguous structures, and the cause of the Chiari II malformation. AJNR Am J Neuroradiol 1992; 13:463–482. 423. Tubbs RS, Dockery SE, Salter G, Elton S, Blount JP, Grabb PA, Oakes WJ. Absence of the falx cerebelli in a Chiari II malformation.Clin Anat 2002; 15:193–195. 424. Boltshauser E, Schneider J, Kollias S, Waibel P, Weissert M. Vanishing cerebellum in myelomeningocoele. Europ J Paediatr Neurol 2002; 6:109–113. 425. Kudryk BT, Coleman JM, Murtagh FR, Arrington JA, Silbiger ML. MR imaging of an extreme case of cerebellar ectopia in a patient with Chiari II malformation. AJNR Am J Neuroradiol 1991; 12:705–706. 426. Chung CJ, Castillo M, Fordham L, Mukherji S, Boydston W, Hudgins R. Spinal intradural cerebellar ectopia. AJNR Am J Neuroradiol 1998; 19:897–899. 427. Naidich TP, McLone DG, Fulling F. The Chiari II malformation. Part IV. The hindbrain deformity. Neuroradiology 1983; 25:179–197. 428. Muller J. Congenital malformations of the brain. In: Rosenberg RN (ed) The Clinical Neurosciences. New York: Churchill, 1983. 429. Snyder WE Jr, Luerssen TG, Boaz JC, Kalsbeck JE. Chiari III malformation treated with CSF diversion and delayed surgical closure. Pediatr Neurosurg 1998; 29:117–120. 430. Chiari H. Über veränderungen des kleinhirns, des pons und der medullaoblongata infolge von congenitales hydrocephalie des grosshirns. Dtsch Akad Wiss, Wien 1896; 63–71.

Magnetic Resonance Imaging of the Brain in Preterm Infants

5

Magnetic Resonance Imaging of the Brain in Preterm Infants Luca A. Ramenghi, Fabio Mosca, Serena Counsell, and Mary A. Rutherford

CONTENTS 5.1 Normal Appearances of the Developing Brain 199 5.1.1 Introduction 199 5.1.2 Practical Issues 199 5.1.2.1 Magnet Systems 199 5.1.2.2 Coils 199 5.1.2.3 Transport 200 5.1.2.4 Monitoring Equipment 200 5.1.2.5 Immobilization 200 5.1.2.6 Temperature Maintenance 200 5.1.2.7 Sedation 200 5.1.2.8 Monitoring 200 5.1.2.9 Safety Issues 200 5.1.2.10 Pulse Sequences and Scanning Parameters 201 5.1.3 Normal Appearances of the Developing Brain 201 5.1.3.1 Cortical Folding 201 5.1.3.2 Ventricular System and Extracerebral Space 203 5.1.3.3 Germinal Matrix 204 5.1.3.4 White Matter and Myelination 204 5.1.3.5 Myelination 206 5.1.3.6 Basal Ganglia and Thalami 207 5.1.3.7 Preterm Brain at Term 208 5.1.4 Summary 209 5.2 5.2.1 5.2.2 5.2.2.1 5.2.2.2 5.2.2.3 5.2.2.4 5.2.3 5.2.3.1 5.2.3.2 5.2.3.3 5.2.3.4 5.2.3.5 5.2.3.6 5.2.4 5.2.5 5.2.6 5.2.7 5.2.8 5.2.9 5.2.10 5.3

Pathology of the Preterm Baby 210 The Brain of Preterm Babies 210 Periventricular Leukomalacia (PVL) 210 Introduction 210 Pathogenesis 210 Neuroimaging of PVL 211 Assessment of Prognosis 217 Germinal Matrix Hemorrhage-Intraventricular Hemorrhage (GMH-IVH) 217 Introduction 217 Pathogenesis 217 Timing of GMH-IVH 219 Complications of GMH-IVH 219 Neuroimaging of GMH-IVH 221 Assessment of Prognosis 222 Choroid Plexus Hemorrhage 222 Cerebellar Hemorrhage 223 Congenital Periventricular White Matter Cavitations 223 Infections 226 Multicystic Encephalomalacia and Asphyxia 226 Venous Thrombosis 226 Infarctions 230 Concluding Remarks and Application of Quantitative MRI 230 References

231

5.1 Normal Appearances of the Developing Brain 5.1.1 Introduction Magnetic resonance imaging (MRI) allows the developing brain to be studied in superb detail either in the fetus or in the infant born preterm. Serial imaging provides valuable insights into both normal maturation and the response of the developing brain to a variety of insults.

5.1.2 Practical Issues 5.1.2.1 Magnet Systems

The integration of an MR scanner into a neonatal intensive care unit (NICU) provides a unique opportunity to image preterm infants and infants in the acute phase of an insult, without compromising their intensive care [1, 2]. Images of normal babies in the first part of this chapter have been obtained on a 1 Tesla (Oxford Magnet Technology/Picker International) neonatal MRI system situated in the neonatal unit at the Hammersmith Hospital in London (UK). Preterm babies with brain lesions shown in the second part of the Chapter have been studied with a 1.5 Tesla Philips MR unit situated at the Leeds General Infirmary in Leeds (UK), and with a 1.5 Tesla (Horizon LX, EchoSpeed/General Electrics) situated at the “Istituti Clinici di Perfezionamento,” Buzzi Hospital, Milan (Italy). 5.1.2.2 Coils

A specially designed transmit/receive quadrature birdcage coil (32 cm × 23 cm) was used for all examinations performed at the Hammersmith Hospital. This coil allows infants to be studied up to a maxi-

199

200

L. A. Ramenghi, F. Mosca, S. Counsell, and M. A. Rutherford

mum weight of about 5 kg. A receive-only quadrature coil commonly used for the adult head and knee was utilized at Leeds and Milan hospitals. 5.1.2.3 Transport

We have used a specially designed nonmagnetic transport trolley to minimize handling of the neonate while being able to continue life monitoring, and ventilation where required, during the move from incubator to the imaging system [1–3]. 5.1.2.4 Monitoring Equipment

As monitoring equipment must be MR compatible, all noncompatible monitoring used in the NICU is switched prior to entering the scanning room. Oxygen lines are attached to the piped supply prior to entry into the scanning room. Infusion pumps are attached to a wall-mounted rail within the scanning room but beyond the 5 Gauss line. Ventilated infants are imaged using an MR compatible ventilator (babyPAC neonatal, pneuPac limited, UK) or using a conventional ventilator sited in an RF cupboard or secured outside the 5 Gauss line. 5.1.2.5 Immobilization

Effective immobilization of the neonate is vital in order to obtain high quality MR images. Immobilization may be achieved with a vacuum-pack or Olympic bag, containing small polystyrene balls. The air is evacuated from these bags with suction to ensure a close fit around the baby’s head. Ventilated neonates need to be imaged supine to accommodate the endoor nasotracheal tube and ventilator tubing. Neonates who are not ventilated were imaged on their side whenever possible. 5.1.2.6 Temperature Maintenance

In order to maintain the infant’s temperature, they are swaddled in blankets and the temperature of the scanning room is set at 27° C, the same as the temperature on the neonatal unit. Additional heating for extremely preterm neonates can be provided with bubble wrap, woolen hats, and “gel bags,” which retain their heat once warmed in a microwave oven. The temperature of the babies imaged at Leeds and Milan was maintained with similar protections,

although the scanning room temperature was set at 21°–23° C. 5.1.2.7 Sedation

Sedation with oral chloral hydrate (20–30mg/kg) is occasionally necessary, but the majority of preterm infants can be successfully imaged during natural sleep following a feed or while sedated for ventilation [4, 5]. 5.1.2.8 Monitoring

Oxygen saturation, heart rate, temperature and, where required, mean arterial blood pressure are monitored throughout the examination. We use a Hewlett Packard Merlin life support system situated in a radiofrequency-shielded cupboard within the scanning room. Oxygen saturation is measured by pulse oximetry (Nellcor oxiband or oxisensor D20 transducer with a Nellcor pulse oximeter, Nellcor Incorporated, Pleasanton, CA, USA), and heart rate is measured using chest ECG leads. ECG leads and electrodes must be MR compatible; we use Blue sensor (Medicotest, Olstykke, Denmark) ECG electrodes and ECG leads with current limiting resistors (NDM Division, American Hospital Supply Corporation, Ohio, USA). Radiofrequency fields can cause currents in conduction loops which may cause burns. Therefore, care must be taken to ensure that ECG leads do not form conductive loops and skin contact is kept to a minimum. The electrodes should be placed close together, but not touching, to minimize ECG interference by the magnetic field. ECG leads should be plaited in order to minimize loops across which potential differences may occur. Temperature may be measured by placing an MR-compatible temperature probe in the axilla. An experienced pediatrician remains in the scanning room, observing the monitors and the infant during scanning. 5.1.2.9 Safety Issues

The number of people in the scanning room is kept to a minimum. However, two neonatologists and a radiographer are essential to prepare a ventilated neonate. A metal check form, including specific neonatal items such as Serle arterial lines with terminal electrodes, electronic name tags, and metal poppers on clothes, is completed by the pediatrician or nurse

Magnetic Resonance Imaging of the Brain in Preterm Infants

caring for the baby and checked by the radiographer before transporting the baby to the magnet. Acoustic noise exposure is reduced by several measures incorporated into the design of the system, including lagging and gradient cable immobilization. The Olympic bag used to immobilize the infant’s head reduces the infant’s exposure to acoustic noise levels still further. In addition, individual ear plugs are made using dental putty and molded into the external ear. These are then covered with neonatal ear muffs (Natus Minimuffs, Natus Medical, Inc., San Carlos, CA, USA). 5.1.2.10 Pulse Sequences and Scanning Parameters

The neonatal brain has a higher water content (92%–95%) than the adult brain (82%–85%), and so T1 and T2 values are greater. These values decrease with increasing gestation [6, 7], and adult values are reached in early childhood [8]. This means that echo times (TE), repetition times (TR), and inversion times (TI) have to be increased. We have found the T2weighted FSE to be the optimal sequence for demonstrating myelination in the premature brain. The IRFSE sequence provides excellent gray/white matter contrast, and is also useful for assessing myelination [9]. The sequence parameters used in London are as follows: for the T1-weighted conventional spin-echo sequences, the parameters are 600/20/2 (TR/TE/excitations); a 4-mm section thickness, with nine sections obtained; and a 192 × 192 matrix. For the T2-weighted fast spin-echo (FSE) sequences, the parameters are 3500/208/2, 4; a 4-mm section thickness, with nine sections obtained; an echo train length of 16; interecho spacing of 16; and a 256 × 256 matrix. For the inversion recovery (IR) FSE sequences, the parameters are 3500/32/4; an inversion time of 950; a section thickness of 5 mm, with six sections obtained; an echo train length of 16; interecho spacing of 16; and a 256 × 256 matrix. The sequence parameters used in Leeds and Milan are as follows: for the T1-weighted spin-echo images 800/13 (TR/TE); 180-mm field of view (FOV); 4-mm section thickness with a 0.4 mm gap; and a scan time of 3 min 50 sec. The parameters for T2-weighted fast spin-echo images were 6000/200; echo train length of 13; FOV of 180 mm; section thickness of 3 mm with no gap; and a scan time of 5 min. We have been unable to use diffusion-weighted sequences on the MR system used in London, but we did use diffusion-weighted images in preterm infants imaged at Leeds [10] and Milan, similarly to those obtained in other small studies [11, 12].

5.1.3 Normal Appearances of the Developing Brain The images showed in this part of the Chapter were obtained from a cohort of over 200 preterm infants born at a gestation of less than 30 weeks at the Hammersmith Hospital. The gestational age (GA) for the infants was calculated from the date of the last menstrual period and confirmed with data from early antenatal ultrasound scans. 5.1.3.1 Cortical Folding

The most obvious change in the preterm brain between 24 weeks and term is the increase in cortical folding [13]. Cortical development occurs from approximately 8 weeks, initially with replication of neurons and glial cells in the germinal matrix. The cortical layers are formed by neuronal and glial cell migration from the germinal matrix. A layer adjacent to the ventricle, called the subventricular layer, is probably mainly responsible for later glial cell proliferation and migration [14]. In the cerebral hemispheres the layers form from the inside out, so that layer 6, which lies medially, is laid down first. The mature cortex consists of six layers. Neuronal migration to the cerebral cortex is completed by 20–24 weeks of gestation in the human brain. Between 29 weeks’ gestation and term cortical gray matter volume increases from approximately 60 ml to approximately 160 ml [12]. The subsequent increase in cortical gray matter volume is secondary to glial cell migration, neuronal differentiation, and organizational changes. During early gestation, the brain is smooth or “lissencephalic” in appearance, but as growth proceeds the typical convoluted pattern develops allowing a considerable increase in the surface area of the brain. Primary sulci begin as a shallow groove with widely separated side walls and straight ends (Figs. 5.1, 5.2). The groove then becomes deeper, and the side walls become progressively steeper, approximating and eventually meeting each other. These secondary sulci may show V-shaped or bifid ends. With continued maturation, the gyri and sulci become complex and side branches develop as seen in the adult brain. Each of the gyri and matching sulci are named according to their site within the brain. Postnatal MRI can be performed as early as 23 weeks’ gestation, and at this stage a rim of cortex is demonstrated as high signal intensity on T1-weighted images and low signal intensity on T2-weighted images when compared to the underlying white matter. There may be some variation in cortical maturity in infants born at the same gestation (Fig. 5.2).

201

202

L. A. Ramenghi, F. Mosca, S. Counsell, and M. A. Rutherford

a

b

c

Fig. 5.1a–c. Normal appearances at 25 weeks’ gestation. a–c Axial T2-weighted images. The brain is smooth with only minimal cortical folding. Rudimentary sulci are demonstrated: calcarine fissure (thick arrow, a); sylvian fissure (thick arrow, b); and central sulcus (thick arrow, c). Notice the optic radiation (thin arrow, a) and the prominent posterior horns of the lateral ventricles. The low signal intensity corpus callosum can also be clearly seen

a

d

c

b

e

f

g

Fig. 5.2a–g. Cortical folding from 25 weeks’ gestation to term equivalent age. a–g Axial T2-weighted images at the level of the centrum semiovale. a 25 weeks’ gestation; b 25 weeks’ gestation, showing some rudimentary central sulcation in this infant; c 26 weeks’ gestation; d 30 weeks’ gestation; e 32 weeks’ gestation; f 34 weeks’ gestation; and g term equivalent age

Magnetic Resonance Imaging of the Brain in Preterm Infants

Visual assessment of MR images is sufficient for identifying major abnormalities or discrepancies in maturation. Visual analysis of images of termequivalent preterm infants is not sufficient to detect a milder delay or abnormality in cortical folding. There are several visual scores that have been used [2, 15], but computer quantification techniques are necessary to identify more subtle changes in folding [16]. In order to study cortical development in more detail we have developed a computerized method for quantifying cortical folding (Fig. 5.3) [17]. We use T2-weighted fast spin echo images, as these give optimum contrast between the cortex and underlying white matter. The process involves drawing a cortical contour as a template for each slice of the brain. The program then measures the length and curvature of this cortical contour. A curvature code is produced. The cortical density is obtained and the product of this and the curvature code gives a cortical convolution index. Using this program we have shown an exponential increase in the whole cortical convolution index (measuring all slices through the brain but excluding the cerebellum) from 24 weeks until term. At term, equivalent age preterm infants show less cortical folding than infants born at term (Fig. 5.2) [17]. While this effect seems to relate to the degree of prematurity, the presence of chronic lung disease and exposure to steroids are also important factors [18, 19]. A reduction in cortical gray matter volume in preterms at term equivalent age and in relationship

100

to steroid treatment for chronic lung disease has also been demonstrated [12, 20]. The differences in cortical development compared with those infants born at term may be associated with the increase in neurocognitive and neurobehavioral disorders seen in expreterm children [21–23]. 5.1.3.2 Ventricular System and Extracerebral Space

Very preterm infants 5 mm) is a useful marker, as all these babies have cerebral palsy compared to those with smaller cysts. Location is another predictor, with posterior lesions having the worse prognosis. In addition, a symmetrical appearance also seems to influence the outcome, as parieto-occipital cavitations were related to cerebral palsy in all cases compared to only 60% of those having asymmetrical lesions [43]. Visual Impairment

Cerebral visual impairment is regarded as visual impairment due to a disturbance in the posterior pathways of the optic radiation and/or the primary visual cortex. Major contributions have been derived from studies by Cioni et al. [79] and Lanzi et al. [80], who discovered that almost half the babies with severe PVL had reduced visual field and abnormal ocular motility, whereby highlighting the importance of impairments in the optic radiation more than in the visual cortex. Cognitive Outcome

A major contribution in the understanding of the development of cognitive problems comes from Marin-Padilla [81, 82], who first described in autopsy studies the events that follow white matter damage of PVL. The overlying cortical layer tends to remain intact thanks to a different blood supply but is functionally deprived of incoming inputs (corticopetal fibers), and is also unable to make connections due to corticofugal fibers destruction. In other words, there is a functional isolation of this cortex, with a potential functional short-circuitry of corticofugal fibers as they try to repair by reattaching in the proximity of the cortex. This short-circuitry should correlate not only with cognitive impairments, but also with the development of epileptic foci.

In addition, Inder’s studies, using a quantitative three dimensional volume MRI technique, have shown reduction of cortical gray matter of those preterm babies with prior white matter lesions [78].

5.2.3 Germinal Matrix Hemorrhage-Intraventricular Hemorrhage (GMH-IVH) 5.2.3.1 Introduction

Intracerebral hemorrhage is more common in the neonatal period than at any other time of life. A progressive reduction in the number of different kinds of intracranial hemorrhages has been associated with a relative increase in cases of GMH-IVH, especially in premature babies [40]. GMH-IVH is now the most frequent and important cause of intracranial hemorrhage in neonates, mainly due to the fact that neonatal intensive medicine has changed life expectations of those premature babies who are most at risk of developing GMH-IVH [83]. A progressive reduction in the number of babies developing severe intraventricular hemorrhage has been noticed over the last few years in neonatal intensive care medicine, while the number of preterm babies showing ischemic lesions has remained pretty constant. 5.2.3.2 Pathogenesis

Schwartz [84] and Rydberg [85] were the first to recognize that intraventricular hemorrhage in preterm infants derives from the large terminal vein in the subependymal germinal matrix. The germinal matrix (Fig. 5.5) is a transient structure that, during fetal life, is present in all the subependymal areas and is the site of vigorous neuroblast mitotic activity until the 18th–20th week of life. After this time, despite the persistence of glioblastic mitotic activity, the relative size of the germinal matrix progressively decreases (Fig. 5.6) from a width of 2.5 mm at 23–24 weeks to 1.4 mm at 32 weeks, and it tends to disappear in a caudo-cranial direction, first around the posterior horns and finally, sometimes near term of gestation, around the anterior horns of the lateral ventricles. In general, germinal matrix tissue is most abundant in the caudate nucleus immediately adjacent to the lateral ventricles, where it is still prominent at 28–32 weeks of gestation, although matrix can also be seen around the frontal horns until the term of gestation [44, 86]. The site of GMH depends on the

217

218

L. A. Ramenghi, F. Mosca, S. Counsell, and M. A. Rutherford

maturity of the infant. Leech and Kohnen [87] found that 90% of subependymal hemorrhage occurs over the head of the caudate nucleus, adjacent to the foramen of Monro, and only 5% over the body of the caudate nucleus or in the occipital germinal matrix (Fig. 5.23). The facts that the germinal matrix surrounding the frontal horns does not originate bleeding and that the preferential site of GMH is the lateral portion of the caudate nucleus have highlighted the potential role of anatomical predisposing factors. Many authors have linked the location of hemorrhage to the course of the veins that drain this region of the brain, particularly the terminal vein (with

a

major blood flow changing direction sharply at that junction in a peculiar U-turn), in line with the view that GMH is essentially of venous origin, a hypothesis which unifies the consensus of most researchers [88–90]. Other authors believe that there is multifocality of GMH, although the bleeding in the caudothalamic zone accounts for most intraventricular extensions of hemorrhage (Figs. 5.24, 5.25). This multifocality of GMH appears to be in contrast with the above-mentioned hypothesis of local anatomical predisposition [83]. In accordance with both theories, to reinforce the idea of the venous origin of GMH there is evidence of blood tunneling along the perivenous space of the germinal matrix veins, leading to compression of patent veins and secondary rupture of smaller connecting tributaries [90]. This, in turn, is likely to cause venous stasis, increased venous pressure, and reduced perfusion pressure in the area where the terminal vein is the terminus of medullary, choroidal, and thalamostriate veins and empties into the internal cerebral vein [90]. This sequence of events associated to venous stasis may be more frequent in subjects with thrombophilia, a genetic predisposition to develop thrombosis more often in the venous rather than arterial vessels. Accordingly, we have found a clear trend to develop GMH-IVH in thrombophilic babies with factor II (antithrombin) abnormality [91]; Petaja et al. found the same trend in subjects with factor V Leiden abnormality [92].

b Fig. 5.23a,b. Germinal matrix hemorrhage in two different infants. a Axial T2-weighted image in a 28-week-old baby shows germinal matrix hemorrhage at the right caudo-thalamic notch (white arrow). Notice minimal dependent blood level in the left occipital horn (black arrow). b Axial T2-weighted image in a 26-week-old baby shows more atypical germinal matrix hemorrhage along the posterior horn of the lateral ventricle (white arrow)

Fig. 5.24. Mild intraventricular hemorrhage caused by germinal matrix bleed in a preterm baby born at 28 weeks of gestation. Axial T2-weighted image. There is dependent, hypointense blood in the posterior horns (black arrows) with the germinal matrix as the bleeding source (white arrows)

Magnetic Resonance Imaging of the Brain in Preterm Infants

a

b

From a clinical perspective, respiratory distress syndrome is the most consistently recognized predisposing risk factor to GMH-IVH in premature infants; in particular, hypercapnia, pneumothorax , and the fluctuating pattern of systemic blood pressure are the most significant factors. All these conditions may be major causes of increased cerebral venous pressure in the premature infant [93]. 5.2.3.3 Timing of GMH-IVH

Leech and Kohnen, in a large autopsy series, indicated that the risk of finding hemorrhage at autopsy increases steadily in the first day of life, but stabilizes thereafter [87]. The introduction of real-time ultrasound allows frequent scanning of high-risk infants to accurately time the onset of GMH-IVH. Ultrasound studies suggest that 90% of GMH-IVH occur in the first week of life, with the majority occurring within the first 3 days of life [93]. 5.2.3.4 Complications of GMH-IVH Parenchymal Hemorrhage/Venous Infarction

Approximately 15% of babies with GMH-IVH also present with unilateral parenchymal hemorrhage, better known as venous infarction. The region involved can be quite large, just dorsal and lateral to the external angle of the lateral ventricle, usually saving the cortical mantle; however, location and size can vary [83]. Less often, the lesion develops in more

Fig. 5.25a,b. Severe intraventricular hemorrhage in a preterm baby born at 26 weeks of gestation. a Axial T2weighted image obtained at 2 days of life; b Axial T2-weighted image obtained 4 weeks later (30 weeks corrected gestational age). A large intraventricular clot is visible in the left ventricle (a). Its size is significantly reduced 4 weeks later (b). Progression of maturation phenomena of the brain can be noted, especially with regards to cortical infolding; a thick periventricular band in the white matter (phenomena of cell migration) visible at presentation (a) can not be detected 4 weeks later (b)

posterior parts of the brain, in the temporal lobe or around the atrium (Fig. 5.26). In these cases, the inferior ventricular or lateral atrial veins are involved. It is still possible to have bilateral venous infarction, but this condition is extremely rare and well differentiated from PVL. At the beginning, the lesion appears as a triangular density often not touching the ventricle. Later, the lesion grows and extends to the ventricle, merging the area of increased density due to matrix hemorrhage [43]. Sometimes there is no progression to this stage, and the lesion remains as a triangular density. The hyperdensity area tends to decrease in size during the second week, and the actual infarcted area can result smaller than expected. Cystic degeneration is the most common evolution of severe cases, with a smooth-walled cavity in the parenchyma communicating with the ventricle (unlike in PVL) [43]. In the past, parenchymal hemorrhage was considered an extension of a very severe GMH-IVH. Presently it is thought to derive from venous infarction. The debate on the pathogenesis is still open, but the hypothesized mechanisms include the obstruction of the terminal vein with reduction in perfusion of the white matter drained by this vein. It has also been suggested that parenchymal ischemia and hypoperfusion, caused by vasoactive compounds generated by the GMH-IVH, can aggravate in the parenchyma the effects of vein obstruction [83]. This mechanism may take into account those venous infarctions occurring during less severe forms of GMH-IVH. Timing of this lesion is uncertain, although it seems to follow the initial intraventricular bleed from a few hours up to a few days [83].

219

220

L. A. Ramenghi, F. Mosca, S. Counsell, and M. A. Rutherford

a

Fig. 5.26a,b. Evolving venous infarct unusually located in the posterior horn of the left lateral ventricle in a 27-week-old baby with intraventricular hemorrhage at 10 days of life. a Posterior coronal ultrasound scan shows large venous infarct just lateral to the left atrium. b Axial T2-weighted image shows the posterior venous infarct in addition to intraventricular blood (arrows) and ventricular dilatation

b

Hydrocephalus

Progressive ventricular dilatation is not an uncommon sequela to GMH-IVH (Fig. 5.26). Multiple small clots may obstruct the ventricular system or the channels of reabsorption in the acute stage; they may also cause chronic arachnoiditis of the basal cisterns, involving the deposition of extracellular matrix proteins and obstruction of the foramina of the fourth ventricles or the subarachnoid space over the cerebral hemisphere [94]. The incidence of progressive ventricular dilatation is closely related to the severity of the initial hemorrhage, and can be acute (within days) or chronic (within weeks). The acute form is due to impairment of CSF absorption, whereas the chronic form is likely to originate from obliterative arachnoiditis [83]. Volpe stated that the severity of the initial IVH is the most critical determinant of not only the likelihood of progressive ventricular dilatation, but also the temporal evolution and the course [83]. It is difficult to disagree with this view. In fact, our experience, as well as published data from other authors, have shown that the majority of infants who develop hydrocephalus tend to have rapidly progressive ventricular enlargement initially or slightly later, especially if they have experienced very severe GMH-IVH. Most uncommon is a slow progressive ventricular dilatation, usually seen between 10 and 14 days after the GMH-IVH with a very high likelihood of spontaneous resolution. The odds of progressive ventricular enlargement in the weeks after hemorrhage increases by 5% in grade II, and 40%–80% in more severe forms [43, 83, 93]. It is also necessary to monitor those infants in whom dilatation has become static for a 4–6-week

period, since a maximum of 5% may develop progressive dilatation at a later stage [83, 93]. Ultrasound is the most appropriate imaging modality for the initial assessment of ventricular size [43]. The slightly rounded shape of the frontal horns can represent the initial appearance of dilatation, whereas balloon-shaped frontal horns are a sign of severe dilatation. Latero-lateral and diagonal measurements of the diameter are a well-established modality to monitor ventricular dilatation. Absence of widening of the frontal horns may be falsely reassuring, as neonates tend towards overdilatation of the occipital horns (“colpocephaly”). Many units have their own guidelines for measuring the frontal horns, but very often not for the posterior horns [43]. MRI can be useful, as detailed imaging is often required prior to shunting surgery and also after surgery to verify the functioning of the shunt, provided the proven MRI compatibility of the intraventricular device. Pseudocysts

Cavitation within the germinal matrix is called germinolysis and typically occurs at the caudo-thalamic notch, resulting in a “pseudocyst” due to the lack of a proper epithelium [95, 96]. During the postnatal period, pseudocysts occur mainly following small to moderate GMH-IVH, although a mechanism based on “pure infarction” of the germinal matrix has been hypothesized. Accordingly, we have sometimes detected pseudocysts at the caudo-thalamic notch a few weeks after birth in “normal” preterm babies, with no obvious reason for this. Pseudocysts can also be present at birth. In these cases, many different prenatal conditions, such as

Magnetic Resonance Imaging of the Brain in Preterm Infants

TORCH group infections, have to be excluded. Less often, prenatal asphyxia, feto-fetal transfusion, metabolic diseases, intrauterine growth retardation, and karyotype anomalies should be investigated [44]. Prenatal pseudocysts are thought to arise incidentally when the developing germinal matrix outgrows its blood supply [83, 93]. Pseudocysts can also be located around the frontal horns of the lateral ventricles. In this case, they are always prenatal and do not seem to correlate with GMH (see below, congenital periventricular cavitations) [97–99]. Although the germinal matrix may persist around the frontal horns even in late gestation, we have never observed hemorrhage in this area with MRI studies.

The evolution of intraventricular clots is quite predictable. They become isodense from the center outwards, showing central lysis; few fragments are often detectable in the cavity at a later stage [43]. CT

CT is very accurate for the diagnosis of any form of hemorrhage (Fig. 5.27), and may detect small areas of parenchymal hemorrhage in the acute phase in the periphery of the brain, which are not easily visualized on ultrasound examination. Nevertheless, it is less frequently used for obvious reasons and especially to avoid ionizing radiation exposure, thanks to the improved availability of MRI scans for infants.

5.2.3.5 Neuroimaging of GMH-IVH Ultrasound

The accuracy of ultrasound diagnosis was reported to approach 90% in old studies comparing ultrasound with CT [100]. Ultrasound has become the primary modality to identify all degrees of GMH-IVH, although the validity of ultrasonographic diagnosis is difficult to investigate. An increased sensitivity of ultrasound in diagnosing GMH-IVH was noticed, with an increase in the diameter of the germinal and intraventricular hemorrhage in a small number of studies where ultrasound was compared to postmortem findings. The sensitivity of ultrasound appeared more linearly related to size when more scans were performed, reaching a value of 100% for hemorrhages larger than 1 cm [43]. The abnormality marker used to diagnose GMH remains the detection of a globular area of intense, increased echogenicity, although a suspicion of hemorrhage can be proposed when irregular luminal plexus borders appear together with densities in the occipital horns. Sometimes, this precedes the typical rounded thickening of the plexus near the foramen of Monro. The abnormality should be demonstrated in two planes, although the first suspicion is often made during coronal scanning. Larger hemorrhages may rupture into the lateral ventricle, and can appear as an echogenic clot within the ventricle provided that there has been fibrin deposition [43]. After a few days, the ependyma becomes denser, probably due to reactive chemical inflammation in response to intraventricular blood, a process that may last for a few weeks. This finding, even in the absence of obvious clots, may be sufficient to diagnose prior IVH.

Fig. 5.27. Massive intraventricular hemorrhage in a 31-weekold preterm baby imaged at 2 days postnatal life. Axial CT scan shows huge intraventricular hemorrhage. The lateral ventricles are dilated. (Case courtesy P. Tortori-Donati, Genoa, Italy)

MRI

This technique provides excellent visualization of the germinal matrix, showing high signal on T1-weighted images and low signal on T2-weighted images, especially up to the 30th week of gestation (Figs. 5.5, 5.6). After this time, T2-weighted images remain the best sequence to follow physiological germinal matrix involution (see first section). GMH has similar characteristics to normal germinal matrix, but is detectable due to its irregular shape and asymmetry (Fig. 5.23). Very preterm babies may show small lesions, consistent with subependymal

221

222

L. A. Ramenghi, F. Mosca, S. Counsell, and M. A. Rutherford

hemorrhage, in different areas than the classical sites (caudo-thalamic notch), more often in the posterior horns (Fig. 5.23). These hemorrhages seem to not be visible on ultrasound [101]. IVH, more often identified in the posterior horns of the lateral ventricle, is an obvious diagnosis with MRI (Figs. 5.24, 5.25), making this technique the most accurate for GMH-IVH investigation, although MRI scanning is not practical for sick, unstable neonates during the first days of life. MRI has improved the detection of venous infarction associated to GMH-IVH (Fig. 5.26), although caution is needed as subependymal hemorrhages may appear to have white matter involvement due to partial volume effects. Nevertheless, Keeney et al. demonstrated that MRI performed between 29 and 44 weeks postconceptual age was superior to both ultrasound and CT in assessing the extent of any parenchymal injury associated with GMH-IVH [102]. MRI shows signal changes following any form of intracranial hemorrhage, including IVH, according to the timing of hemorrhage. The typical hyperintense hemorrhagic signal on T1-weighted images usually appears in the so-called subacute phase, usually between 4 days and 2 weeks after the initial bleed (Fig. 5.28), whereas the long lasting hemosiderin deposition seems to be visible only after 3–4 weeks [93]. 5.2.3.6 Assessment of Prognosis

Determination of neurological prognosis is very complex, as babies with GMH-IVH are usually the most

a

premature, and many confounding variables may play a role. Death can also be considered an outcome, and it seems that the more severe the degree of hemorrhage, the higher the mortality rate, especially when the size of the intraparenchymal involvement (venous infarct) is included [83]. In the absence of ventricular dilatation and parenchymal lesion, the outcome does not differ from infants born of the same degree of prematurity but without evidence of GMH-IVH [83]. The study by Fletcher et al. [103] well highlights the outcome of posthemorrhagic ventricular dilatation: preterm babies with shunted hydrocephalus performed significantly poorer in motor performance and visual-spatial tests, but not language tests, confirming that hydrocephalus seems to adversely affect nonverbal cognitive skills. With an equally severe ventricular dilatation, the presence of a venous infarct plays an important role. Major handicaps range from 60% to 100% in infants with this complication, depending on the extension and the location of the lesion (fronto-parietal lesions are less severe than posterior ones), especially in babies with a birth weight less than 1 kg [43, 83].

5.2.4 Choroid Plexus Hemorrhage Choroid plexus hemorrhage is very commonly represented in studies based on autopsy with varying percentages. Larroche reported minor bleeding into the choroid plexus associated with IVH in 25% of cases, whereas in full-term babies choroid plexus hemorrhage is the most frequent form of intraventricular

b

Fig. 5.28a,b. Late appearance of evolved venous infarct at 4 weeks of age in a baby born at 26 weeks of gestation who suffered an intraventricular hemorrhage. a Sagittal ultrasound scan; b sagittal T1-weighted image. The area of involved parenchyma is equally visible with ultrasound (arrow, a) and MRI (arrow, b). Studies performed at Leeds General Infirmary, Leeds, UK

Magnetic Resonance Imaging of the Brain in Preterm Infants

bleed [93, 94, 104]. In the most recent autopsy studies performed on preterm infants with IVH, the choroid plexus was only rarely the only source of intraventricular blood, whereas the germinal matrix was the principal source [105, 106]. The in vivo diagnosis is very complex and intricate. Ultrasound diagnosis should be based on sequential scans showing cavitation in the hematoma adjacent to the choroid plexus. A simple intraventricular clot adjacent to the choroid plexus is not sufficient to diagnose a primitive choroid plexus hemorrhage [43]. In our experience, the MRI diagnosis of choroid plexus hemorrhage was found to be quite difficult. We have imaged a number of scans of premature babies in their first days of life, and we could only very rarely ascertain the choroid plexus origin of the intraventricular bleed (Fig. 5.29). In minor intraventricular bleeding of premature babies, a hemorrhagic germinal matrix was almost always detectable (Fig. 5.23).

Fig. 5.29. Choroid plexus hemorrhage in a preterm baby born at 33 weeks’ gestation. Axial T2-weighted image. Bleeding originates from left choroids plexus (white arrow). The intraventricular phase of the choroid plexus hemorrhage is very mild and easily missed (black arrows)

5.2.5 Cerebellar Hemorrhage Cerebellar hemorrhage seems to be quite frequent on postmortem examinations performed in very low birth weight infants. It has been associated with traumatic birth and supratentorial hemorrhage; it was related to tightly bound ventilatory masks used in the past [107–111]. Data based on MRI studies

show that 8% of preterm infants less than 32 weeks GA suffer from cerebellar hemorrhage, although the incidence is increasing with decreasing GA. Ultrasound studies performed via the posterior fontanel, more sensitive than the conventional anterior fontanel, show a lower incidence (about 3%). Cerebellar lesions may account for significant cognitive impairment, but their role as isolated findings remains to be assessed [112].

5.2.6 Congenital Periventricular White Matter Cavitations Preterm newborn babies infrequently present small cavitations around the frontal horns of the lateral ventricles. This abnormality can be easily misdiagnosed for congenital leukomalacia, but it should be referred to as congenital germinolysis. It generally carries a good prognosis, especially when it represents an isolated finding [95–99]. On ultrasound, these cavitations are visible on coronal and parasagittal lateral scans. They are echofree areas that not rarely are subdivided into two or three smaller (pseudo)cysts separated by very thin septa. On coronal scans, they often appear to be isolated from the ventricles in more frontal sections, while they are adjacent to the frontal horns in slightly posterior cuts. They can be differentiated without difficulty from PVL lesions, which are located at the superior and external angle of the frontal horns of the lateral ventricle [93, 95]. On MRI, their diagnosis is quite obvious in axial and coronal sections; perhaps diagnosis is easier on T2-weighted images (Fig. 5.30). The frequent presence of multiple septa, indicating their germinolytic nature, is often visible together with remnants of germinal matrix tissue (Fig. 5.30). In more rare cases, these pseudocysts are detectable also in the temporal horns; in such locations, consideration should be given to congenital rubella (Fig. 5.31) and cytomegalovirus, although we have observed these abnormalities also in the absence of obvious congenital infections (Fig. 5.32). In general, we believe these pseudocysts have a very good prognosis when they are isolated findings. Caution is needed when in association with other findings, as they can derive from congenital infections or even chromosomal abnormalities.

223

224

L. A. Ramenghi, F. Mosca, S. Counsell, and M. A. Rutherford Fig. 5.30a–c. Congenital periventricular cavitations in two different patients. a Axial T1-weighted image obtained at 3 days life in a preterm baby born at 32 weeks of gestation. Cysts are visible in the frontal periventricular white matter (white arrows). These cysts do not represent congenital periventricular leukomalacia and are presumably germinolytic. The baby at 4 years of age remained asymptomatic. b, c Axial T2-weighted images in a baby born at 31 weeks of gestation. Usual appearance of congenital frontal “benign” cysts. Residual germinal matrix tissue (open arrow, b) is often represented around the cavitation, suggesting their germinolytic nature. The cysts are often multiple and divided by septa (thin arrows, b, c). The child was asymptomatic at 2 years corrected age

a

c

b

a

b

Fig. 5.31a,b. Infection-related periventricular cavitations in a baby born at 30 weeks with congenital rubella. a Parasagittal ultrasound scan; b coronal T2-weighted image. Periventricular cyst (arrow) is visible in right temporal lobe on ultrasounds (arrow, a). MR shows periventricular cysts in both temporal lobes (arrows, b)

Magnetic Resonance Imaging of the Brain in Preterm Infants

Fig. 5.32. Periventricular cysts in a preterm baby born at 30 weeks of gestation. Coronal T2-weighted image. Two cysts (arrows) are visible around the frontal and temporal horns. These cysts are likely to derive from germinolysis; no congenital infection was found despite repeated investigations. Neurological follow-up of this baby was normal at 3 years of age

a

a

Fig. 5.33a,b. Klebsiella infection in a preterm baby born at 32 weeks’ gestation. a, b Axial T1-weighted images at 3 weeks of age. There are congenital frontal cysts (arrows) and unusual areas of increased signal with linear appearance in the periventricular and subcortical white matter. The baby was investigated for congenital infection; at day 14 he presented with sepsis and a positive blood culture for Klebsiella. The baby was normal at 2-year follow-up

b

b

Fig. 5.34a,b. Congenital cytomegalovirus infection in a baby born at 33 weeks from an HIV-positive mother. The baby had cytomegalovirus isolated from saliva and urine. a Coronal ultrasound scan; b coronal T1-weighted image. Ultrasounds show hyperechogenicities (arrows, a) of the thalamo-striate arteries, known as “mineralizing vasculopathy.” Similar findings are not visible on MRI (arrow, b); however, MRI shows multiple cortical lesions and some punctuate lesions in the white matter, showing increased signal

225

226

L. A. Ramenghi, F. Mosca, S. Counsell, and M. A. Rutherford

5.2.7 Infections

5.2.9 Venous Thrombosis

There is no specific infectious agent affecting premature babies more often than term babies (Fig. 5.33). The most common agents causing congenital infections in term babies can also affect the developing brain of preterm babies, with special regard to cytomegalovirus (Fig. 5.34), which produces different lesion patterns depending on the GA at which infection occurs [114] (see Chapter 12). The importance of this concept can be generalized to other infections, and be considered more important than the nature of the infectious agent [114].

The frequency of this condition is not well known in full term neonates, and even less in preterm babies. The clinical symptoms in neonates are less obvious than in adults, and neuroimaging confirmation can not always be straightforward, especially during the very acute phase. The pathognomonic “empty delta sign” on CT is a “triangle” of decreased density due to the contrast-enhanced blood flowing around the clot. On MRI, the thrombus can appear hyperintense on T1-weighted images. These appearances are typical of the subacute phase, and from this timing, the diagnosis is difficult even with MRI. Phase contrast MR venography is a very helpful modality in this field. Venous thrombosis usually affects term babies, although we have observed quite a few mildly premature babies showing underlying venous thrombosis that appeared in the form of late and unexpected GMH-IVH, severe subcortical and periventricular leukomalacia, posterior fossa hemorrhage, and simple ultrasound “edema” with enlarging diameter of the venous vessels (Figs. 5.36–5.38) [115]. Dural sinus thrombosis may be secondary to trauma, increased hematocrit, sepsis, dehydration, cardiac failure, and, more often, inherited thrombophilia [62].

5.2.8 Multicystic Encephalomalacia and Asphyxia Diffuse ischemic damage to the brain late in gestation may result in multicystic encephalomalacia, although a greater number of premature babies can suffer such a severe insult (Fig. 5.35). Injury to the basal ganglia and thalami may present in the same way as asphyxia affects the brain of newborn babies at term, although the specific regions within the basal nuclei can be different and the cortical highlighting less intense. In premature babies, the most common target of hypoxia is thought to be the white matter.

a

b

c

Fig. 5.35a–c. Early MRI findings in asphyxia. 29-week-old twin newborn surviving after resuscitation in the delivery room, imaged at 4 days. The baby was born with very low hemoglobin level (5 g/dl) and subsequently died. a Axial T1-weighted image; b axial T2-weighted image; c axial diffusion-weighted image (DWI). Diffuse signal abnormalities are seen both on T1- and T2-weighted images (a, b), suggesting a very severe global ischemic insult. On DWI, multiple areas of restricted diffusion are visible (arrowheads, c), although the left occipital lobe already shows very decreased signal (arrow, c) corresponding to decreased ADC values compatible with cell death. In this preterm baby the brain seemed to respond to the global ischemic insult as does the brain of babies born at term of gestation

Magnetic Resonance Imaging of the Brain in Preterm Infants

b a

d c

e

f

Fig. 5.36a–f. Venous thrombosis in a preterm baby born at 33 weeks’ gestational age. a coronal ultrasound scan at 16 days of life; b–f MRI at 16 days; b axial T2-weighted image; c sagittal T1-weighted image; d axial diffusion-weighted image; e axial gradient-echo T2*-weighted image f, axial T2-weighted image. Coronal ultrasounds scan shows intraventricular hemorrhage (white arrows, a) that was unexpectedly discovered during earlier routine ultrasound scans performed on days 2, 7, and 12. MRI confirmed intraventricular bleeding (white arrow, b); bilateral abnormality in the periventricular areas was also suspected (arrowheads, b). Moreover, diffuse thrombosis affecting the superior sagittal sinus (open arrows, c), torcular (T), the straight sinus (arrowheads, c), and the vein of Galen and deep venous system (thin arrows, c) was identified. DWI showed linear appearance of abnormally restricted signal in the deep frontal lobes (arrows, d). Gradient-echo imaging highlights blood breakdown products with a linear pattern (arrowheads, e), corresponding to the anomalies detected with DWI. Blood is also present in the ventricular lumen (arrows, e). The abnormalities observed in the frontal white matter (arrows, f) are highly suspected for thrombosis of the medullary veins. This imaging pattern follows the anatomy of the medullary veins and confirms the original suspicion of thrombosis of small venous vessels. No anticoagulant therapy was administered due to the presence of intraventricular hemorrhage. Studies performed at Leeds General Infirmary, Leeds, UK

227

228

L. A. Ramenghi, F. Mosca, S. Counsell, and M. A. Rutherford

a

bb

c

Fig. 5.37a–c. Venous thrombosis in a baby born at 34 weeks’ gestational age, with sudden onset of seizures at day 4 postnatal life. a Coronal ultrasound scan; b axial T1-weighted images; c sagittal T1-weighted image. Ultrasounds show increased echogenicity in the periventricular areas suggestive of a hypoxic insult similar to the acute phase of periventricular leukomalacia, associated with intraventricular bleeding (open arrows, a). MRI confirms the periventricular abnormality, showing as increased signal, and the intraventricular blood (b). However, MRI also shows unexpected diffuse venous thrombosis affecting the major sinuses and the internal venous system (c). It is very likely that venous thrombosis caused both the intraventricular bleeding and the parenchymal lesion (compare with Fig. 5.36)

Fig. 5.38a–f. Venous thrombosis in a baby born at 35 weeks’ gestational age, presenting at birth with mild asphyxia and requiring ventilation for three days. a, b Coronal ultrasound scan at 6 h of life; c coronal ultrasound scan after 24 h; d axial T1-weighted image; and e coronal MR venogram at 7 days; f coronal MR venogram after 3 weeks. On initial ultrasound scan there is diffuse hyperechogenicity and unusually enlarged dural sinuses (arrows, a, b). Ultrasounds after 24 h (c) fail to identify the abnormal echogenicity and the enlarged venous vessels. At 7 days, thrombus is visible in the right transverse sinus (arrow, d); MR venogram clearly shows an obstructed right transverse-sigmoid sinus (e). It is intriguing to associate the early ultrasound signs (enlarged veins suggesting obstruction and diffuse parenchymal hyperechogenicity suggesting generalized edema) with obstruction due to venous thrombosis, that probably partially resolved somehow, as shown 24 h later by ultrasounds. MR venogram after 3 weeks of treatment with low molecular weight heparin (e) shows signs of recanalization of the previously thrombosed sinus, although a difference of caliber with the unaffected side is noticeable



Magnetic Resonance Imaging of the Brain in Preterm Infants

a

b

c

d

e

f

229

230

L. A. Ramenghi, F. Mosca, S. Counsell, and M. A. Rutherford

a

b

c

Fig. 5.39a–c. Neonatal infarction in a preterm baby born at 34 weeks’ gestational age, suffering from cardiac failure due to severe supraventricular tachycardic crisis. a Axial T1-weighted image; b axial T2-weighted image; and c axial T1-weighted image obtained at 4 weeks’ postnatal life. Both T1-weighted (a) and T2-weighted images (b) show outcome of large stroke in the territory of the left middle cerebral artery. Notice reduced size of homolateral cerebral peduncle (arrow, c) due to axonal degeneration; such finding is surprisingly precocious compared to the timing of a similar event in adult patients suffering from stroke

5.2.10 Infarctions Neonatal stroke is usually reported in full-term babies, whereas data on preterm babies are lacking. Infarction can be seen as an area of increased signal intensity on T2-weighted images. The largest reported series of preterm babies shows no essential differences in the appearance of neonatal stroke when compared to full-term babies [64]; we have had a similar experience. Outcome-defining criteria are comparable to those used for full-term infants, with special attention to the involvement of the PLIC (Fig. 5.39).

5.3 Concluding Remarks and Application of Quantitative MRI MRI is an ideal and safe technique for imaging the developing brain, with the extensive maturation that occurs from 23–40 weeks’ gestation. MRI well depicts known pathologies seen on ultrasound, and also those very subtle anomalies not always corresponding to a well-defined clinical outcome. The neuropathological correlates for neurodevelopmental impairments, especially for the cognitive deficits of very preterm babies, are not well defined. Quantitative MR studies have identified some abnormalities, such as increased ADC values in the central white matter and lower relative anisotropy as compared with infants born at term [73]. Preterm infants at term have a 6% decrease in whole

brain volume compared to infants born at term; they also have an 11.8% decrease in cortical gray matter volume, a 15.6% decrease in right hippocampal volume, a 12.1% decrease in left hippocampal volume, and a 42.0% increase in the size of the lateral ventricles. Therefore, individuals who were born very preterm continue to show noticeable decrements in brain volumes and striking increases in lateral ventricular volume into adolescence. The functional significance of these abnormalities deserves further investigation. Other selective areas of the brain, such as basal ganglia, corpus callosum, and cerebellum have showed reduced volumes compared to term born controls [116–119]. This differentiation is far from explaining clinical correlates, as previous studies on adolescent with brain lesions did not correlate with intellective performances. The main area for further research remains the investigation of ex utero brain development of the preterm brain, including aspects of neonatal intensive care on the developing brain.

Acknowledgements We would like to thank all the staff in the Robert Steiner MR Unit and within the Department of Paediatrics and the Hammersmith and Queen Charlottes Hospitals in London. We are also grateful to the radiological staff of Leeds General Infirmary in Leeds and to the pediatric neuroradiology team (Drs Fabio Triulzi, Andrea Righini, Cecilia Parazzini, and Elena Bianchini) of Milan ICP Hospitals. We are also grateful to the Medical Research Council for their continued support.

Magnetic Resonance Imaging of the Brain in Preterm Infants

References 1. Hall AS, Young IR, Davies FJ, Mohapatra SN. A dedicated magnetic resonance system in a neonatal intensive therapy unit. In: Bradley WG, Bydder GM (eds) Advanced MR Imaging Techniques. London: Martin Dunitz, 1997:281–289. 2. Battin MR, Maalouf EF, Counsell SJ, Herlihy AH, Rutherford MA, Azzopardi D, Edwards AD. Magnetic resonance imaging of the brain in preterm infants: visualization of the germinal matrix, early myelination and cortical folding. Pediatrics 1998; 101:957–962. 3. Maalouf E, Duggan P, Rutherford M, Counsell SJ, Fletcher AM, Battin M, Cowan F, Edward AD. Magnetic resonance imaging of the brain in a cohort of extremely preterm infants. J Pediatr 1999; 135:351–357. 4. Pennock J. Patient preparation, safety and hazards in imaging infants and children. In: Rutherford MA (ed) MRI of the Neonatal Brain. London: WB Saunders, 2002. 5. Cowan FM. Sedation for magnetic resonance scanning for infants and young children. In: Whitman JG, McCloy R (eds) Principles and Practice of Sedation. London: Blackwell Healthcare, 1997:209–213. 6. Herlihy AH, Counsell SJ, Rutherford MA, Bydder GM, Hajnal JV. T1 and T2 measurements of the preterm brain. ISMRM 1999, abstract #531. 7. Counsell S, Kennea N, Herlihy A, Allsop J, Harrison M, Cowan FM, Hajnal J, Edward B, Edwards AD, Rutherford M. T2 Relaxation values in the developing preterm brain. AJNR Am J Neuroradiol 2003; 24:1654–1660. 8. Johnson MA, Pennock JM, Bydder GM, Steiner RE, Thomas DJ, Hayward R, Bryant DR, Payne JA, Levene MI, Whitelaw A. Clinical NMR imaging of the brain in children: normal and neurologic disease. AJNR Am J Neuroradiol 1983; 4:1013–1026. 9. Counsell SJ, Maalouf EF, Fletcher AM, Duggan P, Battin M, Lewis HJ, Edwards AD, Bydder GM, Rutherford MA. Magnetic resonance imaging assessment of myelination in the very preterm brain. AJNR Am J Neurorad 2002; 23:872– 881. 10. Tanner SF, Ramenghi LA, Ridgway JP, Berry E, Saysell MA, Martinez D, Arthur RJ, Smith MA, Levene MI. Quantitative comparison of intrabrain diffusion in adults and preterm and term neonates and infants. AJR Am J Roentgenol 2000; 174:1643–1649. 11. Neil J, Miller J, Mukherjee P, Huppi PS. Diffusion tensor imaging of normal and injured developing human brain - a technical review. NMR Biomed 2002; 15:543–552. 12. Hüppi PS, Warfield S, Kikinis R, Barnes PD, Zientara GP, Jolesz FA, Tsuji MK, Volpe JJ. Quantitative magnetic resonance imaging of brain development in premature and mature newborns. Ann Neurol 1998; 43:224–235. 13. Barkovich AJ, Gressens P, Evrard P. Formation, maturation and disorders of the brain neocortex. AJNR Am J Neuroradiol 1992; 13:447–461. 14. Blakemore C. Introduction: mysteries in the making of the cerebral cortex. In: Ciba Foundation Symposium #193. Development of the Cerebral Cortex. Wiley and Sons, England 1995. 15. Van der Knaap MS, Wezel-Meijler G, Barth PG, Barkhof F, Adèr HJ, Valk J. Normal gyration and sulcation in preterm and term neonates: appearance on MR images. Radiology 1996; 200:389–396. 16. Saeed N, Ajayi-Obe M, Counsell S, Hajnal J, Rutherford MA. Convolution index computation of the cortex using

image segmentation and contour following. ISMRM 1998. Abstract #2076. 17. Ajayi-Obe M, Saeed N, Cowan FM, Rutherford MA, Edwards AD. Reduced development of cerebral cortex in extremely preterm infants. Lancet 2000; 356(9236):1162–1163. 18. Kapellou O, Ajaye-Obe M, Kennea N, Counsell S, Allsopp J, Saeed N, Duggan P, Maalouf E, Laroche S, Cowan F, Rutherford M, Edwards AD. Quantitation of brain development in preterm infants treated with corticosteroids. Pediatr Res 2003; 53 (Suppl):2746. 19. Modi N, Lewis H, Al Naqeeb N, Ajayi-Obe M, Dore CJ, Rutherford M. The effects of repeated antenatal glucocorticoid therapy on the developing brain. Pediatr Res 2001; 50:581– 585. 20. Murphy BP, Inder TE, Huppi PS, Warfield S, Zientara GP, Kikinis R, Jolesz FA, Volpe JJ. Impaired cerebral cortical gray matter growth after treatment with dexamethasone for neonatal chronic lung disease. Pediatrics 2001; 107:217– 221. 21. Cooke RW, Abernethy LJ. Cranial magnetic resonance imaging and school performance in very low birth weight infants in adolescence. Arch Dis Child Fetal Neonatal Ed 1999; 81: F116–F121. 22. Cooke RW, Foulder-Hughes L. Growth impairment in the very preterm and cognitive and motor performance at 7 years. Arch Dis Child 2003; 88:482–487. 23. Stewart AL, Rifkin L, Amess PN, Kirkbride V, Townsend JP, Miller DH, Lewis SW, Kingsley DP, Moseley IF, Foster O, Murray RM. Brain structure and neurocognitive and behavioural function in adolescents who were born very preterm. Lancet 1999; 353(9165):1653–1657. 24. Girard NJ, Raybaud CA. Ventriculomegaly and pericerebral CSF collection in the fetus: early stage of benign external hydrocephalus? Childs Nerv Syst 2001; 17:239–245. 25. Feess-Higgins A, Larroche JC. Development of the human fetal brain. Paris INSERM 1987. 26. Evans DJ, Childs AM, Ramenghi LA, Arthur RJ, Levene MI. Magnetic-resonance imaging of the brain of premature infants. Lancet 1997; 350(9076):522. 27. Childs AM, Ramenghi LA, Cornette L, Tanner SF, Arthur RJ, Martinez D, Levene MI. Cerebral maturation in premature infants: quantitative assessment using MR imaging. AJNR Am J Neuroradiol 2001; 22:1577–1582. 28. Childs AM, Ramenghi LA, Evans DJ, Ridgeway J, Saysell M, Martinez D, Arthur R, Tanner S, Levene MI. MR features of developing periventricular white matter in preterm infants: Evidence of glial cell migration. AJNR Am J Neuroradiol 1998; 19:971–976. 29. Felderhoff-Mueser U, Rutherford M, Squier W, Cox P, Maalouf E, Counsell S, Bydder G, Edwards AD. Relation between magnetic resonance images and histopathological findings of the brain in extremely sick preterm infants. AJNR Am J Neuroradiol 1999; 20:1349–1357. 30. Sie LTL, van der Knaap MS, van Wezel-Meijler, Valk J. MRI assessment of myelination of motor and sensory pathways in the brain of preterm and term-born infants. Neuropediatrics 1997; 28:97–105. 31. Chi JG, Dooling EC, Gilles FH. Gyral development of the human brain. Ann Neurol 1977; 1:86–93. 32. Counsell S, Allsop J, Harrison M, Larkman D, Kennea N, Kapellou O, Cowan F, Hajnal J, Edwards A, Rutherford M. Diffusion weighted imaging of the brain in preterm infants with focal and diffuse white matter abnormality. Pediatrics 2003; 112:1–7.

231

232

L. A. Ramenghi, F. Mosca, S. Counsell, and M. A. Rutherford 33. Leviton A, Gilles F. Ventriculomegaly, delayed myelination, white matter hypoplasia, and “periventricular” leukomalacia: how are they related? Pediatr Neurol 1996; 15:127–136. 34. Hüppi PS. Advances in postnatal neuroimaging: relevance to pathogenesis and treatment of brain injury. Clin Perinatol 2002; 29:827–856. 35. Slattery mm, Morrison J. Preterm delivery. Lancet 2002; 360:1489–1497. 36. Wood NS, Marlow N, Costeloe K, Gibson AT, Wilkinson AR. Neurologic and developmental disability after extremely preterm birth. EPICure Study Group. N Engl J Med 2000; 343:378–384. 37. Briscoe J, Gathercole SE, Marlow N. Everyday memory and cognitive ability in children born very prematurely. J Child Psychol Psychiatry 2001; 42:749–754. 38. Botting N, Powls A, Cooke RW, Marlow N. Attention deficit hyperactivity disorders and other psychiatric outcomes in very low birthweight children at 12 years. J Child Psychol Psychiatry 1997; 38:931–941. 39. Counsell SJ, Rutherford AM, Cowan FM, Edwards AD. Magnetic resonance imaging of preterm brain injury. Arch Dis Child Fetal Neonatal Ed 2003: 88:F269–F274. 40. Volpe JJ. Neurologic outcome of prematurity. Arch Neurol 1998; 55:297–300. 41. Volpe JJ. Cerebral white matter injury of the premature infant - more common than you think. Pediatrics 2003; 112:176–180. 42. Ramenghi LA, Jayashinge D, Childs AM, Tanner S, Arthur R, Martienz D, Levene M. White matter lesions of preterm infants and MRI appearance of glial cell migration: an intriguing correlation. Childs Nerv Syst 2000; 16:50a. 43. Govaert P, deVries LS. An Atlas of Neonatal Brain Sonography. Mac Keith-Cambridge University Press, 1997. 44. Paneth N, Rudelli R, Kazam E, Monte W. White matter damage, terminology, typology, pathogenesis in brain damage in the preterm infant. Mac Keith - Cambridge University Press, 1994. 45. Virchow R. Zur pathologischen Anatomie des Gehirns. 1. Congenitale Encephalitis und Myelitis. Archiv fur pathologische Anatomie und Physiologische und fur klinische Medicin 1868; 38:129–142. 46. Parrot J. Etude sur la steatose interstitielle diffuse de l’encephale chez le nouveau-né. Archives de Physiologie Normale et Pathologique 1868; 1:622–642. 47. Banker BQ, Larroche JC. Periventricular leukomalacia of infancy. A form of neonatal anoxic encephalopathy. Arch Neurol 1962; 7:386–410. 48. Okumura A, Hayakawa F, Kato T, Itomi K, Maruyama K, Ishihara N, Kubota T, Suzuki M, Sato Y, Kuno K, Watanabe K. Hypocarbia in preterm infants with periventricular leukomalacia: the relation between hypocarbia and mechanical ventilation. Pediatrics 2001; 107:469–475. 49. Wiswell TE, Graziani LJ, Kornhauser MS, Stanley C, Merton DA, McKee L, Spitzer AR. Effects of hypocarbia on the development of cystic periventricular leukomalacia in premature infants treated with high-frequency jet ventilation. Pediatrics 1996; 98:918:924. 50. Dammann O, Leviton A. Maternal intrauterine cytokines, and brain damage in the preterm newborn. Pediatr Res 1997; 42:1–8. 51. Kubota H, Ohsone Y, Oka F, Sueyoshi T, Takanashi J, Kohno Y. Significance of clinical risk factors of cystic periventricular leukomalacia in infants with different birthweights. Acta Paediatr 2001; 90:302–308.

52. Kadhim H, Tabarki B, Verellen G, De Prez C, Rona AM, Sebire G. Inflammatory cytokines in the pathogenesis of periventricular leukomalacia. Neurology 2001; 56:1278– 1284. 53. Back SA, Luo NL, Borenstein NS, Levine JM, Volpe JJ, Kinney HC. Late oligodendrocyte progenitors coincide with the development window of vulnerability for human perinatal white matter injury. J Neurosci 2001; 21:1302–1312. 54. Van den Bergh R. Centrifugal elements in the vascular pattern of the deep intracerebral blood supply. Angiology 1969; 20:88–94. 55. DeReuck J. The human periventricular arterial blood supply and the anatomy of cerebral infarctions. Eur Neurol 1971; 5:321–334. 56. Takashima S, Tanaka K. Development of cerebrovascular architecture and its relationship to periventricular leukomalacia. Arch Neurol 1978; 35:11–16. 57. Kuban KCK, Gilles FH. Human telencephalic angiogenesis. Ann Neurol 1985; 17:539–548. 58. Nelson MD, Gonzalez-Gomez I, Gilles FH. The search for human telencephalic ventriculofugal arteries. AJNR Am J Neuroradiol 1991; 12:215–222. 59. Levene MI, Wigglesworth JS, Dubowitz V. Hemorrhagic periventricular leukomalacia in the neonate: a real-time ultrasound study. Pediatrics 1983; 71:794–797. 60. Trounce JQ, Levene MI. Diagnosis and outcome of subcortical cystic leucomalacia. Arch Dis Child 1985; 60:1041– 1044. 61. de Vries LS, Eken P, Dubowitz LMS. The spectrum of leukomalacia using cranial ultrasound. Behav Brain Res 1992; 49:1–6. 62. Arthur R, Ramenghi L. Imaging the neonatal brain. In: Levene MI, Chevernack FA, Whittle M (eds) Fetal ad Neonatal Neurology and Neurosurgery. Edinburgh: Churchill Livingstone, 2001:57–85. 63. Baker LL, Stevenson DK, Enzmann DR. End stage periventricular leukomalacia: MR imaging evaluation. Radiology 1988; 168:809–815. 64. de Vries LS, Groenendaal F, Meiners L. Ischemic lesions in the preterm babies. In: Rutherford M (ed) MRI of the Neonatal Brain. London: Saunders, 2002. 65. Tartaro A, Sabatino G, Delli Pizzi C, Ramenghi L, Bonomo L. Usefulness and limitations of neuroradiologic test (CT, US, MRI) in periventricular leukomalacia. Radiol Med 1990; 5:604–608. 66. Lin Y, Okumura A, Hayakawa F, Kato K, Kuno T, Watanabe K. Quantitative evaluation of thalami and basal ganglia in infants with periventricular leukomalacia. Dev Med Child Neurol 2001; 43:481–485. 67. Roelants-van Rijn AM, Groenendaal F, Beek FJ, Eken P, van Haastert IC, de Vries LS. Parenchymal brain injury in the preterm infant: comparison of cranial ultrasound, MRI and neurodevelopmental outcome. Neuropediatrics 2001; 32:80–89. 68. Schouman-Claeys E, Henry-Feugeas MC, Roset F, Larroche JC, Hassine D, Sadik JC, Frija G, Gabilan JC. Periventricular leukomalacia: correlation between MR imaging and autopsy finding during the first 2 months of life. Radiology 1993; 189:59–64. 69. Barkovich AJ. Pediatric Neuroimaging, 2nd edn. Philadelphia: Lippincott Williams & Wilkins, 2000. 70. Childs AM, Cornette L, Ramenghi LA, Tanner SF, Arthur RJ, Martinez D, Levene MI. Magnetic resonance and cranial ultrasound characteristics of periventricular white

Magnetic Resonance Imaging of the Brain in Preterm Infants matter abnormalities in newborn infants. Clin Radiol 2001; 56:647–655. 71. Cornette LG, Tanner SF, Ramenghi LA, Miall LS, Childs AM, Arthur RJ, Martinez D, Levene MI. Magnetic resonance imaging of the infant brain: anatomical characteristics and clinical significance of punctate lesions. Arch Dis Child Fetal Neonatal Ed 2002; 86:F171–F177. 72. Leech RW, Alvord EC. Morphologic variations in periventricular leukomalacia. Am J Pathol 1974; 74:591–602. 73. Roelants-van Rijn AM, Nikkels PG, Groenendaal F, van Der Grond J, Barth PG, Snoeck I, Beek FJ, de Vries LS. Neonatal diffusion-weighted MR imaging: relation with histopathology or follow-up MR examination. Neuropediatrics 2001; 32:286–294. 74. Inder T, Huppi PS, Zientara GP, Maier SE, Jolesz FA, di Salvo D, Robertson R, Barnes PD, Volpe JJ. Early detection of periventricular leukomalacia by diffusion-weighted magnetic resonance imaging techniques. J Pediatr 1999; 134:631– 634. 75. Battin M, Rutherford MA. Magnetic resonance imaging of the brain in preterm infants: 24 weeks’ gestation to term. In: Rutherford M (ed) MRI of the Neonatal Brain. London: Saunders, 2002. 76. Maalouf EF, Duggan PJ, Counsell SJ, Rutherford MA, Cowan F, Azzopardi D, Edwards AD. Comparison of findings on cranial ultrasound and magnetic resonance imaging in preterm infants. Pediatrics 2001; 107:719–727. 77. Inder TE, Anderson NJ, Spencer C, Wells S, Volpe JJ. White matter injury in the premature infant: a comparison between serial cranial sonographic and MR findings at term. AJNR Am J Neuroradiol 2003; 24:805–809. 78. Inder TE, Huppi PS, Warfield S, Kikinis R, Zientara GP, Barnes PD, Jolesz F, Volpe JJ. Periventricular white matter injury in the premature infant is followed by reduced cerebral cortical gray matter volume at term. Ann Neurol 1999; 46:755–760. 79. Cioni G, Fazzi B, Ipata AE, Canapicchi R, van Hof-van Duin J. Correlation between cerebral visual impairment and magnetic resonance imaging in children with neonatal encephalopathy. Dev Med Child Neurol 1996; 38:120–132. 80. Lanzi G, Fazzi E, Uggetti C, Cavallini A, Danova S, Egitto MG, Ginevra OF, Salati R, Bianchi PE. Cerebral visual impairment in periventricular leukomalacia. Neuropediatrics 1998; 29:145–150. 81. Marin-Padilla M. Developmental neuropathology and impact of perinatal brain damage. II: white matter lesions of the neocortex. J Neuropathol Exp Neurol 1997; 56:219– 235. 82. Marin-Padilla M. Developmental neuropathology and impact of perinatal brain damage. III: gray matter lesions of the neocortex. J Neuropathol Exp Neurol 1999; 58:407– 429. 83. Volpe JJ. Neurology of the Newborn, 2nd edn. Philadelphia: Saunders, 2001. 84. Schwartz P. Die traumatische Gehirnerweichung des Neugeborenen Zeitschrift fur Kinderheilkunde 1922; 31:51–79. 85. Rydberg E. Cerebral injury in the new-born children consequent on birth trauma: with an inquiry into the normal and pathological anatomy of the neuroglia. Acta Pathologica et Microbiologica Scandinavica 1932; 10:1–247. 86. Panet N, Rudelli R, Kazam E, Monte W. The pathology of germinal matrix/intraventricular hemorrhage: a review. In: Panet N, Rudelli R, Kazam E, Monte W (eds) Brain Damage

in the Preterm Infant. London: Mac Keith - Cambridge University Press, 1994. 87. Leech RW, Kohnen P. Subependymal and intraventricular hemorrhages in the newborn. Am J Pahol 1974; 77:465– 475. 88. Nakamura Y, Okudera T, Fukuda S, Hashimoto T. Germinal matrix hemorrhage of venous origin in preterm neonates. Hum Pathol 1990; 21:1059–1062. 89. Moody DM, Brown WR, Challa VR, Block SM. Alkaline phosphatase histochemical staining in the study of germinal matrix hemorrhage and brain vascular morphology in a very-low-birth-weight neonate. Pediatr Res 1994; 35:424– 430. 90. Ghazi-Birry HS, Brown WR, Moody DM, Challa VR, Block SR, Reboussin DM. Human germinal matrix: venous origin of hemorrhage and vascular characteristics. AJNR Am J Neuroradiol 1997; 18:219–229. 91. Ramenghi LA, Fumagalli M, Mondello S, Stucchi I, Gatti L, Tenconi MP, Orsi A, Mosca F. GMH-IVH (Germinal-Matrix Intraventricular Hemorrhage) and thrombophilic pattern in preterm babies. Pediatr Res 2003; 53 (Suppl): 3041. 92. Petaja J, Hiltunen L, Fellman V. Increased risk of intraventricular hemorrhage in preterm infants with thrombophilia. Pediatr Res 2001; 49:643–646. 93. Levene MI, de Vries L. Neonatal intracranial hemorrhage. In: Levene MI, Chevernak FA, Whittle M (eds) Fetal and Neonatal Neurology and Neurosurgery. Edinburgh: Churchill Livingstone, 2001. 94. Larroche JC. Post-haemorrhagic hydrocephalus in infancy. Anatomical study. Biol Neonate 1972; 20:287–299. 95. Larroche JC. Sub-ependymal pseudocysts in the newborn. Biol Neonate 1972; 21:170–183. 96. Shaw CM, Alvord EC. Subependymal germinolysis. Arch Neurol 1974; 31:374–381. 97. Ramenghi LA, Domizio S, Quartulli L, Sabatino G. Atypical site of congenital pseudocysts of germinal matrix. ESPR Annual Meeting, Edinburgh 1993. 98. Rademaker KJ, de Vries L, Barth PG. Subependymal pseudocysts: ultrasound diagnosis and findings at follow-up. Acta Paediatr Scand 1993; 82:394–399. 99. Ramenghi LA, Domizio S, Quartulli L, Sabatino G. Prenatal pseudocysts of the germinal matrix in preterm infants. J Clin Ultrasound 1997; 25:169–173. 100. Dewbry KC, Bates RI. The value of transfontanellar ultrasound in infants. Br J Radiol 1981; 54:1044–1052. 101. Blankenberg FG, Norbash AM, Lane B, Stevenson DK, Bracci PM, Enzmann DR. Neonatal intracranial ischemia and hemorrhage: diagnosis with US, CT, and MR imaging. Radiology 1996; 199:253–259. 102. Keeney SE, Adcock EW, McArdle CB. Prospective observations of 100 high-risk neonates by high field (1.5 Tesla) magnetic resonance imaging of the central nervous system: 1. Intraventricular and extracerebral lesions. Pediatrics 1991; 87:421–430. 103. Fletcher JM, McCauley SR, Brandt ME, Bohan TP, Kramer LA, Francis DJ, Thorstad K, Brookshire BL. Regional brain tissue composition in children with hydrocephalus. Relationships with cognitive development. Arch Neurol 1996; 53:549–557. 104. Moriette G, Relier JP, Larroche JC. Intraventricular hemorrhages in hyaline membrane disease. Arch Fr Pediatr 1977; 34:492–504. 105. Hope PL, Gould SJ, Howard S, Hamilton PA, Costello AM, Reynolds EO. Precision of ultrasound diagnosis of patho-

233

234

L. A. Ramenghi, F. Mosca, S. Counsell, and M. A. Rutherford logically verified lesions in the brains of very preterm infants. Dev Med Child Neurol 1988; 30:457–471. 106. Darrow VC, Alvord EC Jr, Mack LA, Hodson WA. Histologic evolution of the reactions to hemorrhage in the premature human infant’s brain. A combined ultrasound and autopsy study and a comparison with the reaction in adults. Am J Pathol 1988; 130:44–58. 107. Grunnet ML, Shields WD. Cerebellar hemorrhage in the premature infant. J Pediatr 1976; 88:605–608. 108. Martin R, Roessmann U, Fanaroff A. Massive intracerebellar hemorrhage in low-birth-weight infants. J Pediatr 1976; 89:290–293. 109. Tuck S, Ment LR. A follow-up study of very low birth weight infants receiving ventilatory support by face mask. Dev Med Child Neurol 1980; 22:633–641. 110. Donat JF, Okazaki H, Kleinberg F. Cerebellar hemorrhage in newborn infants. Am J Dis Child 1979; 133:441. 111. Kennea NL, Rutherford MA, Counsell SJ, Herlihy AH, Allsop JM, Harrison MC, Cowan FM, Hajnal JV, Edwards AD . Brain injury in extremely preterm infants at birth and at term corrected age using magnetic resonance imaging. Pediatr Res 2002; 51(Suppl):2559. 112. Merrill JD, Piecuch RE, Fell SC, Barkovich AJ, Goldstein RB. A new pattern of cerebellar hemorrhages in preterm infants. Pediatrics 1998; 102:E62. 113. Scott RB, Stoodley CJ, Anslow P, Paul C, Stein JF, Sugden

EM, Mitchell CD. Lateralized cognitive deficits in children following cerebellar lesions. Dev Med Child Neurol 2001; 43:685–691. 114. Barkovich AJ, Linden CL. Congenital cytomegalovirus infection of the brain: imaging analysis and embryologic considerations. AJNR Am J Neuroradiol 1994; 15:703– 715. 115. Ramenghi LA, Gill BJ, Tanner SF, Martinez D, Arthur R, Levene MI. Cerebral venous thrombosis, intraventricular haemorrhage and white matter lesions in a preterm newborn with factor V (Leiden) mutation. Neuropediatrics 2002; 33:97–99. 116. Rutherford MA. Hemorrhagic lesions of the newborn brain. In: Rutherford M (ed) MRI of the Neonatal Brain. London: Saunders, 2002. 117. Nosarti C, Al-Asady MH, Frangou S, Stewart AL, Rifkin L, Murray RM. Adolescents who were born very preterm have decreased brain volumes. Brain 2002; 125:1616–1623. 118. Stewart AL, Rifkin L, Amess PN, Kirkbride V, Townsend JP, Miller DH, Lewis SW, Kingsley DP, Moseley IF, Foster O, Murray RM. Brain structure and neurocognitive and behavioural function in adolescents who were born very preterm. Lancet 1999;353:1653–1657. 119. Abernethy LJ, Palaniappan M, Cooke RW. Quantitative magnetic resonance imaging of the brain in survivors of very low birth weight. Arch Dis Child 2002; 87:279–283.

Neonatal Hypoxic-Ischemic Encephalopathy

6

Neonatal Hypoxic-Ischemic Encephalopathy Fabio Triulzi, Cristina Baldoli, and Andrea Righini

6.1 Introduction

CONTENTS 6.1

Introduction 235

6.2

Neuropathology and Pathogenesis 235

6.2.1 6.2.2

Selective Neuronal Necrosis 236 Parasagittal Cerebral Injury 237

6.3

Neuroradiology

6.3.1 6.3.2 6.3.3 6.3.3.1

Ultrasounds 237 Computerized Tomography 237 Magnetic Resonance Imaging 237 MR Techniques in the Evaluation of HIE 237 Magnet 237 Coils 238 Conventional Sequences 239 Advanced Imaging Techniques 240

237

6.4

Conventional MRI Features in HIE

6.4.1 6.4.2 6.4.3

Selective Neuronal Necrosis 241 Multicystic Encephalomalacia 245 Parasagittal Lesions 245

6.5

Advanced MRI Features of HIE 248

6.5.1 6.5.2

Diffusion Imaging 248 MR Spectroscopy 259

6.6

Ischemic Infarction in the Newborn 254

6.6.1

MRI Features References

254

255

240

Hypoxic-ischemic encephalopathy (HIE) is the major recognized perinatal cause of neurological morbidity both in premature and full-term newborns [1]. During the perinatal period, hypoxemia and/or ischemia usually occur as a result of asphyxia; brain biochemical changes associated with hypoxemia, ischemia, and asphyxia are extremely complex and, other than to duration and severity of the event, are strictly related to the state of brain development and maturation. Different from adults, in whom diffuse and prolonged anoxic-ischemic injury causes diffuse brain injury that predominantly involves the gray matter, neonatal HIE is generally more selective, both in premature and full term infants. The selectiveness of brain damage caused by HIE in newborns follows the rapid changes in biochemical, cellular, and anatomical constitutes of the neonatal nervous system. Relative energy requirements in various portion of the brain are related to the state of brain maturity at the time of the injury, as is well documented in vivo by PET studies [2]. To a large extent, it is possible to state that in case of mild to moderate brain injury the premature brain shows greater vulnerability in white matter, whereas full term infants more typically exhibit gray matter damage. In case of prolonged and profound HIE, gray matter involvement occurs also in the premature newborn, whereas full-term newborns exhibit a diffuse gray matter injury leading to multicystic encephalomalacia.

6.2 Neuropathology and Pathogenesis In order to understand the complexity of neuroradiologic features of HIE, a schematic review of neuropathology and pathogenetic mechanisms of brain injury is proposed. According to Volpe [1], two major varieties of injuries can be observed in full-term new-

235

236

F. Triulzi, C. Baldoli, A. Righini

born HIE, i.e., a) selective neuronal necrosis, and b) parasagittal cerebral injury.

6.2.1 Selective Neuronal Necrosis Selective neuronal necrosis (SNN) is the most common variety of injury in neonatal HIE. According to the different type of insult, four basic regional patterns of SSN have been identified: a) Diffuse pattern: severe and prolonged insult in both term and premature infants; b) Cerebral cortex–deep nuclear pattern: moderately to severe, relatively prolonged insult, primarily in term infants; c) Deep nuclear–brainstem pattern: severe, abrupt insult, primarily in term infants; d) Pontosubicular pattern: undefined temporal pattern, primarily in preterm infants Obviously, the pattern refers to the areas of predominant neuronal injury, and a considerable overlap is common. a) Diffuse pattern. In case of very severe and prolonged insult, a diffuse neuronal injury occurs in both premature and term infants. Neurons of the cerebral cortex are extremely vulnerable; the most vulnerable cortical region is the hippocampus, followed by the perirolandic and calcarine cortex. In very severe insults, a diffuse involvement of the cortex occurs and neurons in the deeper cortical layers (particularly those located in the depth of the sulci) are significantly affected. Thalamic neurons can be consistently affected, as well as hypothalamic neurons and those of the lateral geniculate bodies. Neurons of the putamen are somewhat more likely to be affected in the term infant, whereas neurons of the globus pallidus are predominantly affected in the preterm infant. The combination of putaminal-thalamic neuronal injury is one of the typical features of HIE, particularly in term infants. Neurons of the brainstem can be involved as well, and their involvement can frequently be observed in combination with basal ganglia-thalamic involvement. Finally, also cerebellar neurons seem to be extremely vulnerable to hypoxic-ischemic insults; Purkinje cells are the most vulnerable cerebellar neurons in the term infant, as opposed to internal granule cell neurons in the premature infant. b) Cerebral cortex–deep nuclear pattern. This is probably the most typical magnetic resonance imaging (MRI) pattern, and refers to an involvement

of some characteristic cortical areas (typically the perirolandic area) together with the putamen and thalamus. It could be secondary to a moderate or moderate-severe insult that evolves in a gradual manner. c) Deep nuclear–brainstem pattern. In approximately 15%–20% of infants with HIE, involvement of deep nuclear structures (basal ganglia, thalamus, and brainstem tegmentum) is the predominant lesion; however, until the advent of MRI, detection of this kind of injury in the first weeks or months of life was not frequent. This is probably due to the fact that at least part of these lesions may evolve into the so-called status marmoratus, a condition that cannot be neuropathologically detected before age 9–10 months. d) Pontosubicular pattern. In this type of SNN, neurons of the basis pons and the subiculum of hippocampus are primarily involved. The lesion is characteristic of premature infants, and is strongly associated with periventricular leukomalacia. Pathogenesis of SNN

Different factors may explain the typical selective vulnerability to asphyxia demonstrated by some neuronal groups in the neonatal brain. In recent years, there has been increasing evidence regarding the importance of regional metabolic factors and of the regional distribution of glutamate receptors. As demonstrated in vivo by PET studies [2], the neonatal gray matter exhibits strong regional differences in FDG consumption; as a general rule, areas with higher myelinogenesis and/or synaptogenesis, i.e., areas already having activated functions, show higher metabolic rate and energy utilization. These metabolic conditions may consequently render these neurons particularly vulnerable to severe, abrupt ischemic insults. On the other hand, the regional distribution of glutamate receptors, particularly of the NMDA type, now appears to be the single most important determinant of the distribution of selective neuronal injury [1]. The topography of glutamate synapses parallels that of hypoxic-ischemic neuronal death in vivo, and the particular vulnerability of certain neuronal groups in the perinatal period correlates with a transient, maturation-dependent density of glutamate receptors. Finally, factors related to the severity and temporal characteristics of the insult appear to be of particular importance in determining SNN. As previously reported, in the different patterns of SNN the severity and the duration of the injury can greatly modify the regional distribution of neonatal brain lesions.

Neonatal Hypoxic-Ischemic Encephalopathy

6.2.2 Parasagittal Cerebral Injury The term parasagittal cerebral injury (PCI) refers to lesions involving both the cerebral cortex and subcortical white matter, with a bilateral, parasagittal distribution at the level of the superomedial aspect of the cerebral convexity. Albeit usually roughly symmetrical, PCI can be more evident in one hemisphere; the parieto-occipital aspect of the hemispheres is usually more severely affected that the anterior part. PCI is characterized by cortical necrosis, usually without hemorrhagic component. Atrophic gyri or ulegyria are the chronic neuropathological features. PCI almost invariably affects full-term infants suffering from perinatal asphyxia. Pathogenesis of PCI

The pathogenesis of PCI primarily relates to cerebral perfusion imbalance. It can be considered the dominant ischemic lesion of the full-term infant [1]. The areas of necrosis involve the border zones between the end fields of the major cerebral arteries, particularly at the level of the parieto-occipital aspect at the border zone of the three major cerebral arteries. Border zones are the brain regions most susceptible to falls in cerebral perfusion pressure. To a large extent, it can be stated that diminished cerebral blood flow secondary to systemic hypotension is the major pathogenetic factor in neonatal PCI.

6.3 Neuroradiology 6.3.1 Ultrasounds Cranial ultrasounds still have a considerable value in the evaluation of term infants with HIE. The most typical findings in the acute phase are represented by poor differentiation between gray and white matter with effacement of the sulci, as well as diffuse increased parenchymal echogenicity. These features can be related primarily to diffuse cerebral edema. In the acute phase, hyperechogenic basal ganglia predict poor outcome [3] (Fig. 6.1); on the other hand, as many as 50% of ultrasound scans in neonates with HIE are normal [4]. At present, ultrasound examinations can be considered as a first step technique for each suspected HIE, allowing prompt evaluation of neonates with poor prognosis. How-

ever, an MRI study should follow if a clinical history of HIE is present.

6.3.2 Computerized Tomography Computerized tomography (CT) provides considerable diagnostic information in neonates with HIE. However, the value of CT is greater several weeks after injury. During the acute phase, CT scan demonstrates diffuse brain hypodensity with loss of cortical gray-white matter differentiation. The presence of low attenuation in the thalami on postnatal days 2–4 is predictive of poor outcome [5]. In the subacute phase (4–7 days), especially in presence of hemorrhagic necrosis, the affected cortical and deep gray matter may exhibit diffuse hyperdensity, whereas the white matter is still diffusely hypodense (Fig. 6.2). Although both acute thalamic hypodensity and subacute hyperdensity are reliable prognostic predictors, a negative CT does not have valuable prognostic significance. Therefore, the role of CT in the acute phase of HIE is presently of limited importance, as opposed to MRI studies.

6.3.3 Magnetic Resonance Imaging In recent years, MRI has become the technique of choice in evaluating the greater part of neonatal CNS diseases. MRI is the only imaging technique that can discriminate myelinated from neonatal unmyelinated white matter. It offers the highest sensitivity in detecting acute anoxic injury of neonatal brain. With proper coils and sequences, it can exquisitely depict neonatal brain anatomy and locate pathology, offering a robust and reliable tool in the prognostic assessment of neonatal CNS diseases [6]. Recent MR developments, such as the use of diffusion techniques or the increasing reliability of MR spectroscopy, have further improved the MRI capability to investigate the neonatal brain. 6.3.3.1 MR Techniques in the Evaluation of HIE a) Magnet

To obtain reasonably high signal-to-noise ratio, a high field magnet (1.5 T) should be preferred. Such a magnet has a number of advantages in the evaluation of the newborn brain. First, it allows more flex-

237

238

F. Triulzi, C. Baldoli, A. Righini

a

b

c

d

Fig. 6.1a–d. HIE in a full-term newborn: ultrasound evaluation. a Normal full-term newborn, coronal section passing through basal ganglia. b–d Full-term newborn with severe HIE. b Two days after the insult. Bilateral mildly hyperechogenic thalami (arrows). c One week after the insult. Perisylvian region is also bilaterally hyperechogenic. The lateral ventricles are slightly enlarged. d Three weeks after the insult. Diffuse hyperechogenicity involves both basal ganglia and perisylvian cortex. The subarachnoid spaces and lateral ventricles are enlarged

ible use of imaging basic parameters, such as a sufficiently thin section or a reasonably small field of view. Second, it provides a more reliable use of techniques such as echo planar imaging and spectroscopy. For modern imaging applications such as diffusion tensor imaging (DTI), a powerful gradient system should be considered with a strength greater than 30 mT/m and a slew rate greater than 100 T/m/s. The era of ultra-high field magnet (3T and greater) is still in its first days concerning neonatal applications. However, it seems to be the most promising step for using more technical demanding techniques, such as high resolution spectroscopy or high resolution DTI.

b) Coils

The small brain of a neonate is undoubtedly better evaluated by a dedicated coil. The typical transverse coverage of a standard head coil is 25–30 cm, whereas to optimize the signal-to-noise ratio for a neonatal brain a transverse coverage of 15–20 cm should be required. Even though a dedicated neonatal head coil is usually not offered by the majority of manufacturers, a variety of quadrature extremity coils are well suitable for the neonatal head. These type of coils should be used instead of standard head coils in the routine examination of the neonatal brain. For evalu-

Neonatal Hypoxic-Ischemic Encephalopathy

Fig. 6.2. Severe HIE, acute-subacute pattern (6 days after injury). Axial CT images. The affected cortical and deep gray matter (arrows) exhibit diffuse hyperdensity, whereas the white matter is still diffusely hypodense

ating brain by means of SENSE-like techniques, a new generation of coils has been designed. The SENSE head coil has a relatively small diameter and seems easily suitable for neonatal imaging.

tocol should first of all provide good quality T1- and T2-weighted images with a maximum field of view of 16–18 cm and a slice thickness of 3–4 mm. Only as a second step should other sequences be considered.

c) Conventional Sequences

T1-weighted sequences. A conventional spin-echo sequence with TR shorter than 600 ms and TE shorter than 15 ms is still the sequence of first choice in evaluating neonates with HIE. Inversion recovery sequences offer an extremely high contrast between gray and white matter and between myelinated and unmyelinated white matter. With fast inversion recovery techniques this result is now obtained in few minutes. However, in our experience conventional spin-echo T1-weighted images have exhibited greater contrast than fast inversion recovery sequences in the evaluation of hyperintense lesions secondary to

For conventional MR imaging of neonatal HIE, the basic information that still remains necessary is represented by T1- and T2-weighted images. In our experience, proton-density and FLAIR images have not been mandatory. T1- and T2-weighted images reliably depict the typical MRI features of anoxic-ischemic damage. This should be considered when a pulse sequence protocol is tailored for the neonatal brain, particularly in newborns that undergo MR examination without sedation. A neonatal MR imaging pro-

239

240

F. Triulzi, C. Baldoli, A. Righini

anoxic-ischemic insult, both in premature and fullterm newborns. T2-weighted sequences. To achieve good contrast resolution between gray and unmyelinated white matter, conventional spin-echo (CSE) sequence parameters should be different from those used for adult brain imaging [7]. This is particularly true for T2-weighted images, where a TR of at least 3,000 ms and a TE of at least 120 ms should be selected to obtain sufficiently high contrast between gray and unmyelinated white matter. The major disadvantage of CSE T2-weighted sequence with a TR of 3 sec is the long acquisition time. For this reason, the use of fast spin-echo (FSE) T2-weighted sequences is progressively growing. The shorter acquisition time of FSE T2-weighted sequences is more suitable for a neonatal study without sedation. Moreover, the accuracy of both techniques appears to be similar [8]. However, the progressive loss of T2 paramagnetic effects with the increase of echo train length must be considered, particularly in case of HIE with suspected hemorrhagic components.

neurologic/cognitive outcome have not been fully determined [9]. Due to a progressive decrease of signal-to-noise ratio, longer b values are presently not used in neonatal diffusion imaging. Values between 600 and 700 are usually reported. MR spectroscopy. Recently, proton magnetic resonance spectroscopy (1H-MRS) has allowed one to study normal metabolite concentrations in the human brain at different developmental stages, providing a better understanding of the pathophysiological mechanisms in the neonatal brain [10]. Single voxel techniques are now reliable and faster; thus, 1H-MRS has become an important tool in the evaluation of neonatal brain, particularly regarding metabolic and anoxic injuries. It was recently demonstrated that 1H-MRS is even more accurate than DWI in the acute detection of perinatal asphyxia [11].

d) Advanced Imaging Techniques

6.4 Conventional MRI Features in HIE

Diffusion-weighted imaging (DWI). Diffusion imaging has rapidly proved to be an exciting technique, thanks to the availability of very strong and rapidly acting gradients. New generation high-field MR units routinely offer DWI sequences along the three orthogonal planes, with a b value of usually 1,000, together with the mean of the three images (the so-called trace), in order to limit the difference in signal intensity related to white matter anisotropy. To obtain reliable information from DWI, the calculation of the apparent diffusion coefficient (ADC) is necessary. Parametric images of calculated ADC can easily differentiate truly modified diffusion values from artifact due to T2 contamination (i.e., T2 shine-through). More recent developments in diffusion techniques have made it possible to obtain both longer b values (up to 10,000) and a greater number of axes to apply diffusion gradients. The latter capability permits calculation of not only the ADC but also the dominant diffusion vector for each voxel and the relative anisotropy (RA) values for different tissues (diffusion tensor imaging: DTI). Both ADC and RA rapidly change with gray and white matter maturation and development, and can be variably affected by different neonatal injuries, first of all by ischemic-anoxic injury. Thus, DTI seems to have a dramatic potential in the investigation of both neonatal brain maturation and injuries. However, one should keep in mind that correlations between abnormalities on DTI and

Neuroradiological counterparts of neuropathologic lesions are still not completely elucidated. Barkovich [12] subdivided MRI findings in the term infant with HIE according to the severity of hypotension. In case of mild to moderate hypotension, typical MRI features are mainly characterized by parasagittal lesions involving vascular boundary zones, whereas profound hypotension involves primarily the lateral thalami, posterior putamina, hippocampi, and corticospinal tracts including the perirolandic gyri, but mainly sparing the remaining cortex. Independently from the severity of the hypoxiahypotension and according to recent insights on the neuropathology and pathogenesis of HIE, it seems that parasagittal lesions occurring in the vascular boundary zones are primarily, if not exclusively, related to hypotension. Instead, lesions of the basal ganglia, perirolandic area, and hippocampi are primarily related to impairment of energy substrates that selectively damages areas with higher metabolic requirements or with higher regional distribution of glutamate receptors, particularly of the NMDA type. However, it seems evident that some overlap between these two conditions is possible. To a large extent, the reason why individual infants with ischemia may primarily develop either PCI or the various patterns of SNN is not entirely clear. However, the severity and the temporal characteristics of the insult are likely to be very important factors [1].

Neonatal Hypoxic-Ischemic Encephalopathy

6.4.1 Selective Neuronal Necrosis In our personal experience, the most frequent type of injury on MRI in term newborns consists of selective involvement of areas with higher energy requirements, i.e., the lateral thalami, posterior lentiform nuclei, hippocampi, and perirolandic cortex. A. Thalami and lentiform nuclei. The thalami are involved most frequently in their ventro-lateral portions, whereas involvement of lentiform nuclei basically occurs in the posterior portion of the putamina. In the acute phase, when diffuse brain edema predominates (Fig. 6.3), the thalami and lentiform

nuclei can be relatively hypointense to isointense with respect to normal thalami and lentiform nuclei on T1weighted images and hyperintense on T2-weighted images. They progressively become hyperintense on T1-weighted images after 3–7 days from the insult, and eventually hypointense on T2-weighted images a little later (6–10 days) (Fig. 6.4) [13]. In the very acute phase, it can be difficult to detect signal abnormalities on both T1- and T2-weighted images. However, it was reported that CSE T2-weighted sequences with TR 3,000 ms and TE 60 ms may confidently show isointensity of the deep gray matter nuclei with the white matter, instead of the normal hypointensity [14]. Some difficulties can arise in differentiating pathological T1 hyperintensities from normal myelination in milder forms of HIE. As pointed out by Ruth-

Fig. 6.3a,b. Severe HIE, early acute pattern (first day) on conventional MRI. a Axial T2-weighted image; b axial T1-weighted image. There is diffuse, slight T2 hyperintensity and T1 iso-hypointensity at the level of thalami and lentiform nuclei

a

b

a

b

c

d

e

f

Fig. 6.4a–f. Different degrees of putaminal-thalamic neuronal injury during the acute-subacute phase (1 week). a, c, e Axial T1weighted images; b, d, f axial T2-weighted images. a, b Only the posterior part of the putamen seems to be involved; c, d more typical putaminal-thalamic variant; e, f more diffuse involvement of deep structures is evident. Note that, at this time, lesions are hypointense on T2-weighted images

241

242

F. Triulzi, C. Baldoli, A. Righini

erford [15], a complete loss or change in the normal signal intensity of the posterior limb of internal capsule (PLIC) may be seen following perinatal asphyxia. The posterior part of PLIC is normally myelinated in full-term newborns, whereas the signal from myelin may be diminished in case of anoxic insult. Conversely, the thalami and lentiform nuclei adjacent to PLIC show a clear-cut increase in signal intensity (Fig. 6.5). According to Rutherford et al.[16], absence of a normal PLIC in asphyxiated infants has a high positive predictive value (100% of abnormal neurodevelopmental outcome). Different hypotheses can account for the characteristic T1 shortening of the lateral thalami and

a

b

a

b

posterior putamina that becomes appreciable after 3–5 days. Among these, cellular reaction of glial cells and macrophages containing lipid droplets [17] and micromineralization of necrotic cells occurring in neurons or swollen axons near perinatal lesions [18] could be significant factors. Other locations. In order of frequency, other locations of brain involvement in full-term newborns with HIE include the perirolandic and primary auditory cortex and optic radiation (Fig. 6.6), the hippocampal formation and limbic cortex (Fig. 6.7), and the dorsal mesencephalic structures (Fig. 6.7). The evolution of the MRI signal intensity pattern in these locations is

Fig. 6.5a,b. Putaminal-thalamic neuronal injury. a Axial T1-weighted image, normal control; b axial T1-weighted image, newborn with HIE: a on normal images, the most hyperintense structure is the posterior limb of the internal capsule; b in putaminal-thalamic neuronal injury, the most hyperintense structures are the posterior part of lentiform nuclei and the medial aspects of the thalami

c

Fig. 6.6a–c. Moderate HIE one week from injury. a–c Axial T1-weighted images. Abnormal T1 hyperintensity is present at the level of primary auditory cortex (thin arrows, a, b) and motor cortex (arrow, c), as well as in the optic radiations (thick arrows, a, b)

Neonatal Hypoxic-Ischemic Encephalopathy

similar to that described for thalamic and putaminal lesions. Of note, milder HIE can show only small bilateral putaminal or lateral thalamic lesions (Fig. 6.4), whereas more severe HIE shows involvement of both deep and superficial gray matter. Moreover, as reported by Barkovich [12], thalamic and lenticular lesions, as well as mesencephalic lesions, can also be detected in premature newborns with documented severe HIE, most frequently in association with typical periventricular white matter injuries (Fig. 6.8).

a

b

The evolution of selective lesions is characterized by progressive atrophy of the basal ganglia and thalami, with possible transitory cavitation and persistent T2 hyperintensity in the damaged rolandic cortex (Figs. 6.9–11). The timing of disappearance of T1 hyperintensities is still not completely understood. However, it is a matter of fact that, after a couple of months from the anoxic insult, they are no longer detectable, except for particularly severe cases (see below), in which they tend to coalesce and progressively correspond to calcifications detectable on CT.

c

Fig. 6.7a–c. Severe HIE, 1 week from injury. a–c Axial T1-weighted images. The limbic system is involved at the level of the hippocampus (arrows, a, b) and cingulate gyrus (arrows, c). The dorsal mesencephalon and superior vermis also show abnormal hyperintense signal (a)

a

b

Fig. 6.8a, b. HIE in a 35 gestational weeks premature newborn, 2 weeks after the insult. a, b Axial T1-weighted images. There is diffuse hyperintensity of thalami and lentiform nuclei, as well as thin periventricular hyperintensities (arrows), more typical of premature injury

243

244

F. Triulzi, C. Baldoli, A. Righini

a

b

c

Fig. 6.9a–c. Same case of Fig. 6.7a–c Axial T1-weighted images, 1 month after injury. The involvement of the limbic system is confirmed by significant bilateral hippocampal atrophy (a). Cavitations ensue at the level of thalami and lentiform nuclei (b), whereas diffuse involvement of rolandic areas and superior frontal gyri is also evident (c)

a

c

b

d

Fig. 6.10a–d. Moderate HIE. a, c Axial T1-weighted images and b, d axial T2weighted images at 1 week after insult (a, b) and at three years of age (c, d). Follow-up study shows hypotrophic thalami, appearing hyperintense on T2-weighted image and slightly hypointense on T1-weighted images. Slight T2 hyperintensity is also visible at the level of the posterior part of lentiform nuclei

Neonatal Hypoxic-Ischemic Encephalopathy

a

b

c

Fig. 6.11a–c. Moderate to severe HIE, end stage (1 year follow-up). a, b Axial T2weighted images; c axial CT image. Thalamic hypotrophy is evident on both MRI and CT. The postero-lateral aspects of thalami are hyperintense on T2-weighted images and slightly hyperdense on CT. The rolandic cortex (b) is thin, and the subcortical white matter is hyperintense on T2-weighted images

Similar to premature infants with periventricular leukomalacia, midsagittal MR images will often show thinning of the corpus callosum, most commonly involving the middle or posterior body, resulting from degeneration of transcallosal fibers.

6.4.2 Multicystic Encephalomalacia When the anoxic insult is particularly severe and prolonged, a diffuse damage of the brain ensues.

Brain swelling with diffuse T1 hypointensity and T2 hyperintensity can be detected in the first 2– 3 days, followed by marked T1 hyperintensity and T2 hypointensity involving the basal ganglia and thalami. The lesions rapidly evolve in cavitation, with the final aspect of multicystic encephalomalacia (Figs. 6.12, 13). Calcifications can be detected on CT within the first 3–4 weeks after the anoxic insult (Fig. 6.13). The only preserved regions in multicystic encephalomalacia are the medulla oblongata and cerebellum.

245

246

F. Triulzi, C. Baldoli, A. Righini

a

b

c

d

Fig. 6.12a–d. Severe and prolonged HIE. a, b Axial T1-weighted images at 1 week; c axial and d coronal T1-weighted images at 3 weeks after the insult. Initial study shows diffuse gray matter hyperintensities (a, b). Follow-up study (c, d) shows diffuse cavitations and brain atrophy (multicystic encephalomalacia)

6.4.3 Parasagittal Lesions In our experience, isolated parasagittal lesions are far less common that deep ones. They are usually associated with milder forms of HIE, and typically involve the parasagittal cortico-subcortical frontoparietal, and sometimes occipital, regions. During the acute phase, the affected cortex shows increased

T2 signal intensity and decreased T1 signal intensity, related to the presence of edema. Cortical T1 hyperintensities can be found after 4–7 days from the insult (Fig. 6.14). As the lesion evolves, cortical shrinkage with marked cortical thinning becomes evident. This pattern forms a typical mushroom-shaped aspect of the affected gyri, called ulegyria by neuropathologists (Fig. 6.15) [18].

Neonatal Hypoxic-Ischemic Encephalopathy

a

c

e

b

d

f

Fig. 6.13a–f. Severe and prolonged HIE. a Axial T1-weighted and b axial T2-weighted images 1 week after the insult. c Axial T1weighted image, d axial T2-weighted image, and e, f axial CT images 1 month after the insult (e, f). The MR study in the subacute phase (a, b) shows marked gray matter hyperintensity on T1- weighted images and hypointensity on T2-weighted images. One month follow-up study shows striking diffuse cavitations and brain atrophy, with basal ganglia calcifications on CT

a

b

Fig. 6.14a, b. Parasagittal lesions 1 week after the insult. a Coronal T1-weighted image; b axial proton density-weighted image. Bilateral parasagittal cortical hyperintensities are evident on both T1- and PD-weighted images (arrows)

247

248

F. Triulzi, C. Baldoli, A. Righini

a

b

c

Fig. 6.15a–c. Parasagittal lesions. a Axial CT during subacute phase (10 days after injury); b, c axial T1-weighted images during chronic phase (9 months after injury). The lesions are particularly severe, and ulegyric cortex is detectable on follow-up MR study. The rolandic strip and basal ganglia seem almost partly preserved

6.5 Advanced MRI Features of HIE 6.5.1 Diffusion Imaging Preliminary reports suggest that DWI may be useful in the early detection of perinatal brain ischemia, even though some limitations can occur in the very early phases of injury. Robertson et al. [19] reported 12 patients with diffuse perinatal ischemia, in whom line-scan DWI was performed between 13 hours and 8 days from perinatal asphyxia. Line-scan DWI mainly showed deep gray matter and perirolandic white matter lesions earlier than conventional MR imaging, even though it may underestimate the eventual extent of injury. Delayed neuronal and oligodendroglial cell death due to apoptosis in areas with lower metabolic demand is one of the most intriguing factors that can explain DWI underestimation of final injury extent [20]. Barkovich et al. [11] reported 7 patients with HIE studied with DWI within 24 hours of injury. In that study, diffusion images showed small abnormalities in the lateral thalami or internal capsules in all seven patients, whereas T1-weighted images were normal in all; T2-weighted images were normal in three patients and showed T2 prolongation in the basal ganglia or white matter in the other four. Comparison with clinical course in all seven patients and with

follow-up MR studies in four showed that diffusion images underestimated the extent of brain injury. Very recently, McKinstry et al. [21] prospectively evaluated DTI in ten newborns with HIE during the first week of life; they confirmed that MR diffusion images obtained on the first day after injury do not necessarily show the full extent of ultimate injury in newborn infants. According to these recent studies, the current role of DWI techniques in the evaluation of acute HIE can be summarized as follows: a) Images obtained during the first 24 hours after injury can demonstrate focal abnormalities not recognized by conventional acquisition techniques; however, they do not necessarily show the full extent of ultimate injury in newborn infants (Figs. 6.16–20); b) Images obtained between the second and fourth days of life (when conventional images can still be negative or poorly informative) reliably indicate the extent of injury (Fig. 6.21); c) By 7–10 days, DWI is progressively less sensitive to perinatal brain injury, due to the transient pseudonormalization of diffusion images. During this period, conventional images can be more sensitive than diffusion (Fig. 6.22); d) Overall, even though diffusion imaging can readily demonstrate whether brain lesions followed perinatal asphyxia, it may not be suitable as a gold standard for detection of brain injury during the first day after injury in newborn infants.

Neonatal Hypoxic-Ischemic Encephalopathy

a

b

d

e

c

Fig. 6.16a–e. Moderate to severe HIE, first day study. a Axial T1-weighted image; b axial post-Gd T1-weighted image; c axial T2weighted image; d, e diffusion-weighted images (diffusion gradient through the slices). Conventional pre- and postcontrast images show diffuse brain edema without evidence of focal lesions. Diffusion-weighted images show abnormal hyperintense signal involving both thalami (arrows, d, e), as well as the white matter of the left cerebral hemisphere

6.5.2 MR Spectroscopy Markedly elevated lactate levels in infants suffering from severe perinatal asphyxia were first reported about 10 years ago [22]. In that paper, 1H-MRS data demonstrated regional differences in lactate elevation, with greater increase in Lac/NAA ratio in the basal ganglia than in the occipitoparietal cerebrum. This approximately corresponded to signal abnormalities that were observed with early DWI after term hypoxia-ischemia. However, differently from DWI, MR spectroscopy performed in the first 24 hours after birth is sensitive to the presence of hypoxic-ischemic brain injury, and seems to be suitable as a gold stan-

dard for detection of brain injury during the first day after injury in newborn infants [11]. Moreover, as recently reported both by Zarifi et al. in a study on 26 full-term infants with HIE [23] and by Cappellini et al. [24], higher lactate-choline ratios in basal ganglia and thalami of infants with perinatal asphyxia are predictive of worse clinical outcomes. Absolute ADC values in the same brain regions did not indicate a statistically significant relationship with clinical outcome. Therefore, cerebral lactate levels seem to be more useful than DWI techniques in identifying infants who would benefit from early therapeutic intervention.

249

250

F. Triulzi, C. Baldoli, A. Righini

a

b

c

d

e

f

g

h

Fig. 6.17a–h. Severe HIE, same case of Fig. 6.3 a, b axial T2-weighted images; c, d axial T1-weighted images; e, f diffusion-weighted images (trace); g, h ADC images (trace). Conventional T1- and T2-weighted images show diffuse edema. Both the basal ganglia and rolandic cortex are roughly isointense with white matter on both sequences (arrows, b, d). Both diffusion weighted and ADC sequences demonstrate abnormal diffusion within superficial and deep gray matter, although signal abnormalities are more striking at the level of the rolandic cortex (arrows, f, h)

Neonatal Hypoxic-Ischemic Encephalopathy

a

b

c

d

Fig. 6.18a–d. Diffusion-weighted and ADC images: comparison between a neonate with severe HIE (a, c) (same case of Fig. 6.17) and a normal newborn (b, d). Notice severe signal abnormality of perirolandic gray matter in the patient with HIE (a, c)

251

252

F. Triulzi, C. Baldoli, A. Righini

a

b

c

d

a

b

c

Neonatal Hypoxic-Ischemic Encephalopathy

a

Fig. 6.19a–d. Same patient of Fig. 6.17, 1-month follow-up. Sagittal T1weighted images on first day (a, b) and after 1 month (c, d). Comparison between first day exam (a, b) and 1month follow-up (c, d) demonstrates good prediction value of first day diffusion studies (see Fig. 6.17). The gray matter is diffusely involved, with eventual multicystic encephalomalacia. However, chronic evolution of gray matter injury displays some differences between gray matter areas that are more frequently involved in HIE (i.e., rolandic cortex, putamen, thalamus, etc.) and the remaining cortex. While the latter shows typical cystic or malacic pattern, the former still exhibits clear-cut T1 hyperintense signal (arrows, c, d), probably due to the different histological composition of more “mature” gray matter: more myelin, more synapses, more dendrites, etc.

Fig. 6.20a–c. Moderate HIE. a Axial T1weighted image and b diffusion-weighted images on day 1; c axial T1-weighted image after 1 week. Acute-stage diffusion-weighted image clearly shows the involvement of thalami (arrow, b) and optic radiation, whereas auditory cortex and putaminal injuries are not seen. However, both putaminal and thalamic involvement is evident on conventional imaging after 1 week

b

Fig. 6.21a,b. Mild HIE. a, b Diffusion-weighted images after 3 days. Small lesions in limbic system at the level of the hippocampal formation (arrows, a) and isthmus of cingulated gyrus (arrows, b) were only shown by diffusion images. The lesions were confirmed in follow-up study

a

b

c

d

Fig. 6.22a–d. Moderate HIE with parasagittal lesions, study at 10 days. a Axial T1weighted image; b axial T2-weighted image; c diffusion-weighted image; d ADC image. Conventional sequences better show the real extent of injury. Diffusion-weighted and ADC images show only part of the lesions (arrows, c, d)

253

254

F. Triulzi, C. Baldoli, A. Righini

6.6 Ischemic Infarction in the Newborn Typical neonatal focal infarction ensues within 2– 3 days from birth, and presents more commonly with seizures [1]. The greater part of neonatal infarctions are idiopathic. Among the recognizable forms, coagulopathies are the most common cause. However, it is a fact that during the first 2–3 days of life the newborn baby is at greater risk for focal cerebral infarction. Therefore, a correlation with the complex hemodynamic changes occurring during the first minutes and hours of extrauterine life must be considered. During the first 2 days, cranial ultrasounds are poorly sensitive in the detection of focal cerebral infarction. When visible, infarction appears as an area of slight hyperechogenicity, ill-defined and difficult to evaluate in its peripheral cortical extension due to the limited field of view of ultrasonography.

a

6.6.1 MRI Features MRI is the technique of choice in evaluating neonatal focal cerebral infarctions. DWI is the best modality to identify infarction in the acute phase [19]. DWI shows striking increase in signal intensity in the acute phase of the infarction, similarly to that demonstrated in the adult age group. A clear-cut decrease of ADC values (Fig. 6.23) can as well be easily detected in calculated images. Both DWI and ADC images can define more precisely than conventional imaging the extension of the focal infarction, even in the hyperacute phase. The signal intensity on DWI usually normalizes after 6–10 days. On both T1- and T2-weighted images, acute infarction appears as a “missed” cortical segment (Fig. 6.23). The edematous cortex shows increase of T2 signal intensity and decrease of T1 signal intensity, thus approaching the signal intensity of the unmyelinated white matter; this results in the “disappeared cortex” sign. Large infarcts, especially when they involve the basal ganglia, are frequently hemorrhagic, and paramagnetic aspects of degradation products of hemoglobin can be detectable after few days (Fig. 6.24).

b

c Fig. 6.23a–c. Acute arterial infarction, first day exam. a Axial T1-weighted image; b axial T2-weighted image; c ADC image. On both T1- and T2-weighted images acute infarction appears as a “missed” cortical segment (arrows, a, b) due to gray matter edema. In cortical infarction, the ADC technique (c) is usually more reliable in defining the final extent of injury, even in the first day of injury

Neonatal Hypoxic-Ischemic Encephalopathy

a

b

References 1. Volpe JJ. Neurology of the Newborn, 4th edn. Philadelphia: Saunders, 2001. 2. Chugani HT, Phelps ME, Mazziotta JC. Positron emission tomography study of human brain functional development. Ann Neurol 1987; 22:487–497. 3. Connolly B, Kelehan P, O’Brien N, Gorman W, Murphy JF, King M, Donoghue V. The echogenic thalamus in hypoxic ischaemic encephalopathy. Pediatr Radiol 1994; 24:268– 271. 4. Stark JE, Seibert JJ. Cerebral artery Doppler ultrasonography for prediction of outcome after perinatal asphyxia. J Ultrasound Med 1994; 13:595–600. 5. Roland EH, Poskitt K, Rodriguez E, Lupton BA, Hill A. Perinatal hypoxic-ischemic thalamic injury: clinical features and neuroimaging. Ann Neurol 1998; 44:161–166. 6. Triulzi F, Baldoli C, Parazzini C. Neonatal MR imaging. Magn Reson Imaging Clin N Am 2001; 9:57–82. 7. Nowell MA, Hackney DB, Zimmerman RA, Bilaniuk LT, Grossman RI, Goldberg HI. Immature brain: spin-echo pulse sequence parameteres for high-contrast MR imaging. Radiology 1987; 162:272–273 8. Engelbrecht V, Malms J, Kahn T, Grunewald S, Modder U. Fast spin-echo MR imaging of the pediatric brain. Pediatr Radiol 1996; 26:259–264. 9. Neil J, Miller J, Mukherjee P, Huppi PS. Diffusion tensor imaging in normal and injured developing human brain – a technical review. NMR Biomed 2002; 15:543–552. 10. Huppi P. Advances in postnatal neuroimaging: relevance to pathogenesis and treatment of brain injury. Clin Perinatol 2002; 29:827–856. 11. Barkovich AJ, Westmark KD, Bedi HS, Partridge JC, Ferriero DM, Vigneron DB. Proton spectroscopy and diffusion imaging on the first day of life after perinatal asphyxia: preliminary report. AJNR Am J Neuroradiol 2001; 22:1786– 1794. 12. Barkovich AJ. Pediatric Neuroimaging, 3rd edn. Philadelphia: Lippincott Williams & Wilkins, 2000. 13. Barkovich AJ, Westmark K, Partridge C, Sola A, Ferriero DM. Perinatal asphyxia: MR findings in the first 10 days. AJNR Am J Neuroradiol 1995; 16:427–438. 14. Barkovich AJ, Hajnal BL, Vigneron D, Sola A, Partridge JC,

Fig. 6.24a,b. Subacute arterial infarction, 1week exam. a Axial T1-weighted image; b axial T2-weighted image. The deeper part of the lesion shows some T1 hyperintensity and T2 hypointensity, probably due to degradation products of hemoglobin

Allen F, Ferriero DM. Prediction of neuromotor outcome in perinatal asphyxia: evaluation of MR scoring system. AJNR Am J Neuroradiol 1998; 19:143–149. 15. Rutherford MA. MRI of the Neonatal Brain. London: WB Saunders, 2002. 16. Rutherford MA, Pennock JM, Counsell SJ, Mercuri E, Cowan FM, Dubowitz LM, Edwards AD. Abnormal magnetic resonance signal in the internal capsule predicts poor developmental outcome in infants with hypoxic-ischemic encephalopathy. Pediatrics 1998; 102:323–328. 17. Schouman-Claeys E, Henry-Feugeas MC, Roset F, Larroche JC, Hassine D, Sadik JC, Frija G, Gabilan JC. Periventricular leukomalacia: correlation between MR imaging and autopsy findings during the first 2 months of life. Radiology 1993; 189:59–64. 18. Friede RL. Developmental Neuropathology, 2nd edn. Berlin: Springer, 1989:534–545. 19. Robertson RL, Ben-Sira L, Barnes PD, Mulkern RV, Robson CD, Maier SE, Rivkin MJ, du Plessis A. MR line-scan diffusion-weighted imaging of term neonates with perinatal brain ischemia. AJNR Am J Neuroradiol 1999; 20:1658– 1670. 20. Pulera MR, Adams LM, Liu H, Santos DG, Nishimura RN, Yang F, Cole GM, Wasterlain CG. Apoptosis in a neonatal rat model of hypoxia-ischemia. Stroke 1998; 29:2622–2630 21. McKinstry RC, Miller JH, Snyder AZ et al. A prospective, longitudinal diffusion tensor imaging study of brain injury in newborns. Neurology 2002;59:824–33 22. Groenendaal F, Veenhoven RH, van der Grond J, Jansen GH, Witkamp TD, de Vries LS. Cerebral lactate and N-acetylaspartate/choline ratios in asphyxiated full-term neonates demonstrated in vivo using proton magnetic resonance spectroscopy. Pediatr Res 1994; 35:148–151. 23. Zarifi MK, Astrakas LG, Poussaint TY, Plessis Ad A, Zurakowski D, Tzika AA. Prediction of adverse outcome with cerebral lactate level and apparent diffusion coefficient in infants with perinatal asphyxia. Radiology 2002; 225:859– 870. 24. Cappellini M, Rapisardi G, Cioni ML, Fonda C. Acute hypoxic encephalopathy in the full-term newborn: correlation between Magnetic Resonance Spectroscopy and neurological evaluation at short and long term. Radiol Med 2002; 104:332–340.

255

Cerebrovascular Disease in Infants and Children

7

Cerebrovascular Disease in Infants and Children Robert A. Zimmerman and Larissa T. Bilaniuk

CONTENTS 7.1

Introduction 257

7.2 7.2.1 7.2.2 7.2.3 7.2.4 7.2.5 7.2.6 7.2.7 7.2.7.1 7.2.7.2 7.2.7.3 7.2.7.4 7.2.8 7.2.9 7.2.10

Imaging Approach to Cerebrovascular Disease 257 Conventional Ultrasound 258 Transcranial Doppler Ultrasound 258 Computed Tomography 258 Computed Tomographic Angiography 259 Magnetic Resonance Imaging 259 Magnetic Resonance Arteriography and Venography 260 Magnetic Resonance Diffusion-Weighted Imaging 260 Trace 261 CSF and Fluid Spaces 261 Gray Matter 261 White Matter 261 Magnetic Resonance Perfusion 262 Magnetic Resonance Spectroscopy 262 Conventional Angiography 262

7.3 7.3.1 7.3.2 7.3.2.1 7.3.2.2 7.3.2.3 7.3.3 7.3.4 7.3.4.1 7.3.4.2 7.3.4.3 7.3.4.4 7.3.5 7.3.5.1 7.3.5.2 7.3.5.3 7.3.5.4 7.3.5.5 7.3.6 7.3.7 7.3.8 7.3.9 7.3.10

Childhood Cerebrovascular Disease 262 Sickle Cell Disease 263 Vasculitis 265 Infectious 265 Takayasu’s Arteritis 266 Other Vasculitides 266 Heart Disease 266 Hypercoagulable States with Thrombosis 268 Protein C and S Deficiency 268 Antiphospholipid Antibodies 268 Homocystinuria 268 Chemotherapeutic Agents 269 Primary and Secondary Vascular Wall Disease 269 Ehlers-Danlos Syndrome 269 Neurofibromatosis Type I (NF1) 269 Fibromuscular Dysplasia 270 Moyamoya Syndrome 270 Radiation Vasculopathy 270 Congenital Vascular Hypoplasia 271 Vascular Trauma 273 Migraine Headaches 275 Metabolic Diseases 275 Hypoxia 275

7.4 7.4.1 7.4.2 7.4.3 7.4.4 7.4.5

Nontraumatic Intracerebral Hemorrhage 277 Cerebral Aneurysms 277 Hemorrhagic Infarction 279 Coagulopathies 279 Hypertension 279 Sinovenous Occlusion 280

7.5

Conclusions 282 References

283

7.1 Introduction The nature of cerebrovascular disease affecting the pediatric patient is different from that in the adult. Atherosclerotic sources of thrombotic occlusion and thrombotic or plaque emboli are not usually an issue in infants and children. The nature of cerebrovascular injury in infants and children is dependent on the age at which the insult occurs and the mechanism of the insult. Insults in the premature infant, which consist of periventricular leukomalacia (PVL) and germinal matrix hemorrhages (GMH) (see Chapter 5), are quite different and distinct from injuries in term infants, who sustain deep gray matter infarctions, parasagittal watershed infarctions and middle cerebral artery infarctions (see Chapter 6). The term infant injuries are somewhat, but not that different from those that occur in children beyond infancy. Complicating the large number of different etiologies for the primary cerebrovascular pediatric diseases that result in injuries, such as infarctions and bleeds, there are the secondary cerebrovascular disease processes that have arisen due to various recent therapeutic methods and innovations in dealing with neoplastic diseases. For example, leukemia and brain neoplasms require chemotherapy and radiation therapy and at times also bone marrow transplantation. The spectrum of pediatric cerebrovascular disease includes thrombotic infarction secondary to chemotherapy, radiation-induced vasculopathy, and infarctions and aggressive infections related to aspergillosis in immunosuppressed patients (see Chapter 11).

7.2 Imaging Approach to Cerebrovascular Disease The imaging evaluation of the patient with cerebrovascular disease will depend on the age of the patient (fetus, infant, child, adolescent), the acuteness and stability of the illness, the availability of

257

258

R. A. Zimmerman and L. T. Bilaniuk

imaging modality, and the knowledge and skill of the people ordering and performing the examinations. The options include conventional ultrasound (US), computed tomography (CT), computed tomographic angiography (CTA), magnetic resonance imaging (MRI), magnetic resonance angiography divided into magnetic resonance arteriography (MRA) and magnetic resonance venography (MRV), magnetic resonance spectroscopy (MRS), magnetic resonance diffusion weighted imaging (DWI), magnetic resonance perfusion imaging (PI), and conventional cerebral arteriography.

7.2.1 Conventional Ultrasound US is limited to the neonatal and infant age group because of the need for an acoustic window, i.e., the anterior fontanelle. Once the fontanelle closes, the window is gone. The advantages of US are its portability to the crib-side of sick infants, its lack of need for sedation, and its relatively quick study time. Ventricular enlargement or compression, the presence of central brain bleeds, the shift of the midline structures, and the presence of gross areas of brain edema are findings that can be demonstrated with US. The disadvantages are that large areas of the brain are usually not visualized and therefore not examined, such as portions of the frontal lobes and the posterior parietal and occipital lobes. The posterior fossa is rarely studied successfully unless additional acoustic windows are utilized. The differentiation of hemorrhage within the white matter from ischemic infarction is not accurate, and the recognition of small areas of PVL or even large symmetric areas of PVL is not necessarily easy. The use of US in infants is operator dependent and, therefore, it is the skill of the technician or doctor who performs the examination that either maximizes or minimizes the information obtainable.

7.2.2 Transcranial Doppler Ultrasound Transcranial Doppler (TCD) ultrasound is also a technique that requires very skilled acquisition of data. TCD is used to measure the velocity of flow in cm/s in the major vessels at the circle of Willis. Knowledge of the vascular anatomy is critical if reproducible results are to be obtained. With good data, TCD can be used to demonstrate both reduced and increased velocity of flow in the internal carotid and middle cerebral

arteries, which reflects change in caliber of a vessel, i.e., stenosis.

7.2.3 Computed Tomography CT is the initial imaging test in most children and adolescents presenting with acute stroke symptoms, whether it is due to ischemic infarction or to a hemorrhagic bleed within the brain. Unless an arteriovenous malformation (AVM), infection, or a tumor is suspected, most CT is done without contrast enhancement. Single slice helical or spiral scanning takes only minutes to perform a brain study. Multiple slice CT (4 slice) takes on the order of a minute or less. These rapid scan times have decreased the need for sedation in emergency patients. Strokes are shown as hypodensity of the involved area of the brain with mass effect depending on the size and age of the infarction. Mass effect in larger ischemic infarcts is seen by 24 h post ictus, progresses for several days, and then resolves. Hypodensity in the mature myelinated brain of the older child and adolescent begins as a blurring of gray/white matter definition [1], followed by clear-cut lower density at the site of infarction. Loss of density can be seen as early as 6 h after ictus [2]. The presence of a thrombus or calcific embolus in the middle cerebral or basilar arteries may be seen as hyperdensity (Fig. 7.1) [3]; however, while frequent in adults, in our own experience it has been uncommon in children. Moreover, a slightly dense basilar artery is usually normal in children as long as its density is similar to that of the internal carotid arteries. Contrast enhancement occurs at the site of infarction, once the blood brain barrier (BBB) becomes open with the ingrowth of new vessels. Generally, some enhancement can be seen as early as 3–5 days post ictus and it can last for 3–5 weeks. Reabsorption of infarcted tissue produces encephalomalacia. In the basal ganglia, thalamus, and brainstem, chronic cystic changes are referred to as lacunar infarctions. Atrophy in the form of ventricular dilatation and sulcal prominence accompanies the chronic stages of larger cerebral infarctions. Demonstration of acute hemorrhage is a particularly important aspect of CT. The globin molecule of hemoglobin is dense, absorbing more of the x-ray beam than the adjacent brain. Hounsfield units (HU) for clots are in the 50–100 range, denser than brain, less dense than calcification [4]. Localization of the site of blood, parenchymal, ventricular, subarachnoid (SAH), subdural, or epidural, is relatively accurate with CT.

259

Cerebrovascular Disease in Infants and Children

Fig. 7.1. Thrombosis of the left middle cerebral artery. Unenhanced axial CT. A thrombosed hyperdense (arrow) left middle cerebral artery is seen. Subsequent MRI (not shown) demonstrated a large left middle cerebral artery infarction

SAH is best seen on CT, with fluid attenuated inversion recovery (FLAIR) MR sequence as a second option (Fig. 7.2) [4]. Blood within the brain takes approximately 3 h to undergo clot retraction, during which time serum exudes and the hematoma becomes more dense [5]. The extruded serum may be helpful in outlining the hematoma. Surrounding vasogenic edema within the adjacent brain is also not uncommon, a manifestation of adjacent cerebral injury secondary to the process producing the bleeding and the bleeding itself. With time, the globin molecule breaks down and the density of the clot is lost [6]. For example, it takes about 24 days for a 2.5 cm clot to become isodense to the brain, and still later to become hypodense [6]. Ultimately, all but a slit-like collection of hemosiderin will remain, along with adjacent changes due to the damage caused by the bleed and/or the process that produced the bleeding (i.e., AVM, aneurysm, etc.).

a

b Fig. 7.2a,b. Subarachnoid hemorrhage in a 16-month-old female. a Unenhanced axial CT; b axial FLAIR image. CT scan (a) shows hyperdense blood outlining the midbrain, part of the chiasmatic cistern, and the left cerebellar sulci. There is early hydrocephalus with dilatation of the temporal horns. MRI at the level of the odontoid through the upper cervical spinal canal shows the cervical spinal cord to be bathed in subarachnoid blood which is bright. A fluid level (arrow, b) is present superiorly

7.2.4 Computed Tomographic Angiography CTA is the rapid acquisition of data utilizing spiral or helical CT during the bolus injection of iodinated contrast material, and then 3D reconstruction of the data set. The main application of this technique has been to study adults with aneurysms, although there is no reason why this should not be used in children as well. The main limitation in children has been related to the size and rapidity of the bolus contrast material injection when venous access is limited.

7.2.5 Magnetic Resonance Imaging In older infants and children, once myelination of the brain is complete, standard T1- and T2-weighted pulse sequences can demonstrate acute, subacute, and chronic infarcts, blood products, and vascular flow voids in a manner sufficient to diagnose many of the routine vascular diseases (Fig. 7.3). FLAIR imaging

260

R. A. Zimmerman and L. T. Bilaniuk

a

b

c

Fig. 7.3a–c. Watershed (partially hemorrhagic) infarctions in parasagittal distribution in a congenital heart disease patient. a Axial T2 weighted image (6000/99); b axial FLAIR image; c axial DWI. Bilateral parasagittal watershed high signal intensity abnormalities consistent with infarcts are seen (a). Within the areas of high signal intensity are lower signal intensity zones of deoxyhemoglobin. FLAIR image (b) shows the bilateral parasagittal watershed infarcts between the anterior and middle cerebral arteries anteriorly, and between the posterior cerebral and middle cerebral arteries posteriorly. DWI (c) shows restricted motion of water with high signal intensity bilaterally at sites of infarction.

demonstrates water within tissue to a much greater degree and more exquisitely than T2 turbo spin-echo imaging. With FLAIR, the water within the ventricles and sulci is black, but the water within the infarct generates a very bright signal intensity (Fig. 7.3) [7]. Gadolinium enhancement with T1-weighted images is used to demonstrate BBB disturbances in subacute infarction. Enhancement may be seen 3 days after infarction and can last for a variable period of time, from many weeks to months. Hemorrhage can be detected and its evolution characterized by MRI. T1-weighted images show oxyhemoglobin as isointense, deoxyhemoglobin as slightly hypointense, methemoglobin as bright, and hemosiderin as dark signal intensity. T2-weighted images show oxyhemoglobin as bright, deoxyhemoglobin as dark, intracellular methemoglobin as dark, extracellular methemoglobin as bright, and hemosiderin as dark. Gradient-echo T2*-weighted imaging is a valuable adjunct to routine MRI sequences. It demonstrates with great sensitivity blood products within cerebral tissue by a loss of signal (blackness) at the affected sites. The black foci of no signal are produced by dephasing of the protons that occur secondary to the paramagnetic effects of iron in the components and by-products of blood.

7.2.6 Magnetic Resonance Arteriography and Venography Today, MRA is performed to demonstrate both arteries and veins and to help characterize the nature of blood vessel abnormalities [8]. This can be done with 3D time-of-flight (TOF), 2D TOF, 3D phase contrast (PC), and 2D PC MRA. When there is hemorrhage present at the site of vessels investigated, PC MRA is utilized to avoid incorporation of the high signal intensity of the methemoglobin of the hematoma into the image of vascular flow. 3D TOF and 2D TOF are used for arteries and veins, respectively. A saturation pulse is applied at the top of the head when venous flow needs to be removed from the image, whereas a saturation pulse is applied at the bottom of the image of the brain when arteries need to be removed, such as when an MRV is performed.

7.2.7 Magnetic Resonance Diffusion-Weighted Imaging Random motion of water molecules arises predominantly from diffusion, but also the flow of blood from the capillary bed contributes to this effect. Although these effects can be separated, the techniques for doing this are difficult, so that usually the combined effect is measured and is called the apparent diffu-

Cerebrovascular Disease in Infants and Children

sion coefficient (ADC) [9]. Although the principle of the technique has been known for many years, DWI became available only a few years ago, because very strong rapidly acting gradients are required for in vivo studies, and these have only recently become available on whole-body MRI scanners. The apparent diffusion can be measured in any desired direction (x-, y-, or zaxis) by acquiring images in the presence of very large field gradients applied in that direction. Acquisition of a series of images with different gradients allows the ADC to be accurately quantitated. The effect of the motion of water molecules decreases the signal intensity from that expected as a result of the usual contrast mechanisms (T1, T2, T2*); regions of tissue with higher ADC will be less intense on the images. 7.2.7.1 Trace

The “trace” is the average of the ADC in three perpendicular directions. This value is relatively easy to measure, and like anisotropy, is indicative of the physiological state of the tissue [10]. In normal adult brain, the ADC values fall broadly into three different ranges that are characteristic of the three categories of material in the brain. 7.2.7.2 CSF and Fluid Spaces

The ADC values are those expected from unrestricted diffusion. The trace ADC values of CSF are about 3.0 × 103 mm2/s, and are the same in all directions (anisotropy) [11]. 7.2.7.3 Gray Matter

Tissue restricts the diffusion of water, so that the trace ADC values for gray matter are lower than that for CSF, lying in the range of 0.8–1.2 × 103 mm2/s. However, the restriction is the same in all directions, so there is no anisotropy [11]. 7.2.7.4 White Matter

The axon sheath acts as a significant barrier to the diffusion in a direction perpendicular to the fiber axis, and thus the ADC of white matter in the brain is isotropic, variable with the direction of measurement of the diffusion. The trace value is somewhat less than that of gray matter, and falls in the range 0.4–0.6 × 10 -3 mm2/s [11].

Following birth, the ADC changes in a fashion commensurate with the expected postnatal myelination [12]. The ADC of white matter decreases and becomes more anisotropic during the first 6 months. While there are many factors that affect the ADC, it is possible to explain the main features of the changes that occur in edema in terms of the redistribution of water between two compartments. In the intracellular compartment, the diffusion of water is highly restricted so that the ADC is low, whereas [13] in the extracellular compartment, the diffusion of water is less than that in free water (or CSF), but nonetheless, is significantly higher than in the intracellular compartment. The measured ADC is essentially an average of these two values, weighted according to the relative volumes of water in the two compartments. The observed changes in the different type of edema are: (1) cytotoxic edema: the movement of extracellular water into the cell increases the fraction of water in the compartment with the lower ADC, so that ADC value falls; (2) vasogenic and interstitial edema: here the water is predominantly extracellular, so that the component with the larger ADC dominates and the ADC is larger than normal [14]. Brain edema following ischemia has been the focus of many studies using DWI. Within the first hour, when acute cytotoxic edema develops, the ADC decreases, so that the lesion appears brighter on DWI. Changes are visible on DWI before any change can be appreciated on conventional CT or T1- and T2-weighted images. With time, as the vasogenic edema develops, the lesion will darken on DWI because the ADC increases. The regions of abnormal brain indicated on DWI and on the T2-weighted images are similar, but not identical: the region in which the diffusion is changed is, in general, greater than the region in which the T2 is altered [15]. DWI has revolutionized our ability to see abnormalities in tissue due to change in the restriction of movement of water molecules, such as acute infarctions (Fig. 7.3) The ability to process these images rapidly to determine accurately numerical values, in the form of apparent diffusion coefficients, may have meaning as to how severe the changes are. Can diffusion imaging differentiate infarction from ischemia? To answer this question, more studies and research are needed. As DWI finds wider application to the evaluation of the pediatric patient presenting with acute neurological picture, it is anticipated that the incidence and promptness with which cerebrovascular disease is diagnosed in this population will increase.

261

262

R. A. Zimmerman and L. T. Bilaniuk

7.2.8 Magnetic Resonance Perfusion

7.2.10 Conventional Angiography

There are two techniques for looking at vascular perfusion in the brain. These are the tracer techniques, one involving bolus gadolinium injection with T1weighted imaging [16] and the other the EPI star technique, without gadolinium injection, in which the incoming blood is altered by an RF pulse placed over the carotid artery in the neck and then imaged in the brain [17]. Quantification of blood flow by either technique is difficult, but relative perfusion can be seen. Bolus techniques measurement of time-to-peak enhancement and production of maps of the relative cerebral blood volumes throughout the brain have proven useful.

Conventional arteriography remains the gold standard for the demonstration of vasculitis, collateral blood flow, emboli within vessels, feeding vessels of an AVM, and the anatomy of an aneurysm [8]. The process requires experience and skill in that vascular access is not as easy in children and infants as in adolescents and adults, and vascular spasm of the catheterized vessel is more common. The catheters utilized are smaller in younger patients, and the volumes of contrast that can be used are affected (restricted) by body weight. Much of the former need for arteriography in the pre-MRI era has been circumvented by the development of MRI with MRA, MRV, and DWI.

7.2.9 Magnetic Resonance Spectroscopy

7.3 Childhood Cerebrovascular Disease

Acute infarction in the brain alters the measurable levels of cerebral metabolites seen with MRS. N-acetylaspartate (NAA) decreases as neurons are destroyed, while lactate, a by-product of anaerobic metabolism not normally seen, becomes elevated (Fig. 7.4) [18]. With complete tissue necrosis, MRS shows complete absence of cerebral metabolites.

Estimates of the incidence of stroke in children in the United States and Canada, both ischemic and hemorrhagic, has been given as 2.5 per 100,000/year versus an incidence of 100–300 per 100,000/year in adults [19]. This series [19] had the hemorrhagic strokes

Fig. 7.4. Proton spectroscopy of acute infarction. A single voxel spectrum was acquired from cortex and some white matter in predominantly the right posterior temporal/occipital region. T2weighted MRI showed bilateral watershed infarctions. Proton spectroscopy was performed with a TE of 10 ms and a TR of 1,600 ms. The study shows decreased N-acetyl aspartate (NAA) and markedly elevated lactate. The lactate appears at 1.3 ppm as a doublet, while NAA is at 2 ppm

Infarction

Lac

Cr

4

ml

3.5

Cho Cr

3

Glu/ NAA Gln

2.5

2

Chemical Shift (ppm)

1.5

1

0.5

Cerebrovascular Disease in Infants and Children

at 1.89/100,000/year, at a higher rate than the ischemic strokes, which were 0.63/100,000/year. With the advent of MRI, the incidence of both will have increased significantly, but with the further advent of DWI, the incidence of ischemic stroke will have clearly exceeded that of hemorrhagic by a major factor. There is a need for current MRI and MR diffusion-based statistics to reveal the real incidence of stroke in children, which appears to be much higher than previous estimates. A later study of pediatric stroke from 1995, performed in Canada, gave the incidence of ischemic and sinovenous stroke as 1.2/100,000/year. In the Canadian study, infants less than 1 year of age represented 38.5% of strokes, between 1 and 5 years 24%, while from 5–18 years 37.6%. Large vessel arterial disease was present in 42.7%, small vessel arterial disease in 36.5%, and sinovenous disease in 20.8%. The etiology of the stroke was cardiac disease in 20.8%, infection or inflammation in 16.7%, coagulopathy in 7.3%, large vessel dissection in 6.3%, and metabolic disease in 3.1%. The etiology was not determined in 19.8% in this series, whereas in older series the etiology was undetermined in approximately one third [20]. Younger children are more often affected by complications of congenital heart disease, producing either hypoperfusion watershed infarction or embolic stroke within a vascular territory [21]. The older child and adolescent fall victim to traumatic vascular injury, such as dissections of the carotid and middle cerebral artery, leading to thrombosis or thromboembolism [22, 23].

7.3.1 Sickle Cell Disease Within both the pediatric and adult stroke population there are patients with sickle cell disease. These patients have both small vessel hypoperfusion disease causing lesions within the white matter of the centrum semiovale, and stenoses and/or occlusions of the more proximal major vessels, leading to infarctions within vascular territories [21]. Systematic evaluation of sickle cell stroke data has given us some insight into the incidence of stroke in this population [22, 23]. The incidence of clinically apparent infarctions in children with sickle cell disease is 21/416 (Fig. 7.5) [24]. These infarcts are all due to large vessel stenoses or occlusions, and involve the cortex ± white matter and/or the basal ganglia. In a cohort of sickle cell children followed from age 6 for 8 years, an incidence of silent infarctions of 17% was found at the start of the study,

and an incidence of 22% at the end of the study [25]. These patients were clinically asymptomatic, but when tested neuropsychologically were found to have lower scores than controls on arithmetic, vocabulary, visual motor speed, and coordination evaluation [26]. The patients with clinical infarctions performed significantly worse on most neuropsychological tests [26]. In follow-up MR studies of the cohort, patients with prior infarction had a 12.1% incidence of a new or worse infarction, while patients with an initial normal MRI had only a 1.9% risk of subsequent infarction [25]. Altogether, the sickle cell stroke population is thought to constitute 6% of United States children with the acute onset of hemiplegia. The annual risk of initial stroke in sickle cell disease is estimated to be 0.7%/year. Subsequent infarction, after the first stroke, occurs in 50%–75% of untreated patients (not transfused), with 80% having a second infarction within 36 months. A second trial regarding sickle cell disease and infarction is also of interest. This study deals with TCD evaluations, and is called the STOP 1 Trial. A TCD velocity of > 200 cm/s in the distal internal carotid or middle cerebral artery was classified as elevated. Of 1,934 sickle cell children evaluated, 9.4% were at or above this level. The patients also had MRIs, and of the patients with elevated TCDs 36.9% had silent infarcts (Fig. 7.5) [27]. The patients with elevated TCDs were randomized into two groups. The first received standard treatment, and the second were put on chronic transfusion therapy. In the first group there were 13 clinical infarctions and 1 bleed in 71 patients (19.7%), versus only 1 infarction in 56 (1.8%) in the other group. As a result, the study was closed after only 16 months [27].

7.3.2 Vasculitis 7.3.2.1 Infectious

Bacterial infection of the wall of both arteries and veins can occur with meningitis, resulting in stenoses or thrombosis of the vessel, which leads to infarction. The incidence of infarctions in meningitis in the preMRI literature varies from a low 4.7% in hemophilus type B meningitis to a high 27% [28, 29] in a variety of other organisms. Based on our own experience, MRI with diffusion detects small infarctions of perforators that otherwise would not be found by any other method and thus missed, indicating that the true incidence of infarction is underestimated in the

263

264

R. A. Zimmerman and L. T. Bilaniuk

a

b

d

e

c

Fig. 7.5a–f. Sickle cell disease (different patients).a Axial FLAIR image; b, c, e axial DWI; d axial T2WI (6000/99); f 3D TOF MRA. MRI (a) shows deep white matter, right-sided hyperintensity consistent with infarction. In the same patient, DWI shows restricted motion of water in the deep white f matter sites of infarction as well as two small cortical areas, one in the right parietal and one in the left frontal (arrows, b). In a symptomatic 7-year-old female, DWI (c) shows acute infarction on the right in both the middle and posterior cerebral artery vascular territories. In a different patient (a 2-year-old male, found to have abnormal TCD on the left, in distal internal carotid and middle cerebral artery), MRI done the next day shows a slight increased signal (arrows, d) in the head of the caudate and anterior putamen. In the same patient, DWI shows restricted diffusion in the head of the caudate and anterior putamen (arrow, e). In this case, 3D TOF MRA shows that the distal internal carotid artery on the left is markedly narrowed, as is the proximal middle cerebral artery (arrow, f). The proximal anterior cerebral artery A1 segment is not seen on the left

Cerebrovascular Disease in Infants and Children

literature. Most of our patients are studied because of seizures, new onset of weakness, decreased mental status, or prolonged fevers. Most of the patients with infarction complicating meningitis are infants in the first 2 years of life, and the organisms tend to be the more virulent ones, such as streptococcus. If MRA demonstrates large vessel vasculitis, in our experience the outcome is very poor (Fig. 7.6). Fortunately, the infarctions occur more often in the territories of the perforators and tend to involve the frontal white matter, with outcomes that are not poor. Syphilis was once one of the more frequent causes of cerebral vascular disease in adults but not children, as it occurs as a vasculitis in the tertiary stage [30, 31]. Congenital syphilis remains a possible cause of infarction in children, especially when it has not been recog-

nized clinically. Some viral infections are also able to produce a vasculitis, leading to infarction [20]. Herpes simplex virus type I and II are the best known, but varicella may be a more frequent cause [32, 33]. HIV infection has a recognized vascular infection component, occurring late in the disease and involving large vessels in one of two ways: either as a inflammatory fibrosis of the wall producing narrowing that leads to infarction, or as an inflammatory disruption of the internal elastic membrane with resultant aneurysmal dilatation of the involved vessels (Fig. 7.6) [34]. Among fungal infections, aspergillosis and mucormycosis are the recognized causes of large and small vessel disease where vessel wall invasion and thrombosis lead to infarction. Aspergillosis is found in the setting of the immunosuppressed patient (i.e., bone marrow trans-

Fig. 7.6a–d. Infectious vasculitis (different patients). a MRA; b. axial ADC map; c, d. Contrast-enhanced axial CT scan. In a 1-year-old female with streptococcal meningitis and vasculitis, MRA shows that the proximal right middle cerebral artery is markedly narrowed (arrow, a). The left middle cerebral artery is occluded. In this case, ADC map derived from diffusion images shows infarction of a large portion of the left hemisphere, seen as hypointensity (arrows, b). There is also involvement of both thalami and of the right internal capsule region. In a different hemophiliac patient (c) in whom HIV vasculitis has caused aneurysmal dilatation of the middle and anterior cerebral arteries, enhanced CT demonstrates aneurysmal dilatation of some of the vessels involving the circle of Willis. In a different patient (d), HIV infection has produced vascular stenosis and infarction, and enhanced CT shows contrast enhancing infarcts in the left basal ganglionic region. In addition, there is atrophy of the brain manifested as dilated sulci and ventricles

a

b

c

d

265

266

R. A. Zimmerman and L. T. Bilaniuk

plant, chemotherapy) [35], and mucormycosis in the setting of diabetes and ketoacidosis [36]. The infectious causes of vasculitis are only a small component of the overall list of etiologies (Table 7.1) [20]. However, they are among the more treatable causes if diagnosed early and treated effectively. In case of other types of vasculitis, often there is a more wide ranging spectrum of involvement of other organ systems, and this complicates management. In many of these, CNS imaging is used to identify the presence of cerebral involvement when the underlying disease is known. In some cases, findings on brain MRI stimulate a search for the underlying disorder. Table 7.1. Cerebral Vasculitis Infections Necrotizing Collagen vascular Systemic diseases Giant cell arteritides Hypersensitivity Primary CNS Miscellaneous

from hypertension with renal artery or aortic stenosis, in a lesser number of patients [38]. Vascular imaging, whether by MRA or, better yet, by conventional arteriography (the gold standard for diagnosis of vascular stenosis), delineates the distribution of vascular involvement. MRI is utilized to demonstrate cerebral ischemic strokes. 7.3.2.3 Other Vasculitides

In most patients with vasculitis, MRI will be the most successful test for showing infarction, whether acute, subacute, or chronic. However, recognizing the involvement of the artery by vasculitis is frequently beyond the current capability of MRA. Severe forms may be obvious on MRA, but most will require conventional arteriography in order to be detected. Some may not be seen even on these studies, and only be identified by biopsy of the vascular wall or on the basis of laboratory studies.

7.3.3 Heart Disease 7.3.2.2 Takayasu’s Arteritis

This disease affects the aorta and its branches, primarily in young females. Involvement of the subclavian artery is present in 85% of patients [37], of the carotid arteries in 50% [37], and of the vertebral arteries in 20% [37]. Cerebrovascular disease is due either to ischemic stroke from vascular stenosis, in 5%–10% of patients (Fig. 7.7), or to intracerebral hemorrhage

Infarction in the brain of pediatric patients with heart disease has a multitude of possible causes (Table 7.2). Among all etiologies of ischemic stroke in young patients past the neonatal period, between one fifth to one third are presumed cardiac in origin, often embolic. Congenital heart disease is the most common cardiac disorder that causes stroke. Patent foramen ovale and atrial septal defect, relatively mild forms of congenital heart disease, have both been

Fig. 7.7. Takayasu’s disease involving multiple vessels in a 18-year-old female with prior ischemic infarction in basal ganglia on the left. 3D TOF MRA, coronal projection. MRA shows involvement of anterior and middle cerebral artery (arrows), as well as the distal internal carotid artery on the left

Cerebrovascular Disease in Infants and Children Table 7.2. Cardiac causes of infarction Congenital heart disease Valvular heart disease Cardiac arrhythmias Other cardiac Myocarditis Left ventricular aneurysm Cardiac tumor Cardiac surgery Kawasaki disease

correlated with an increased risk of embolic stroke (Fig. 7.8) [39, 40]. In children with more severe congenital heart disease, such as hypoplastic left heart syndrome (HLHS), ischemic brain lesions have been found in utero, at delivery, and in the immediate postnatal period when the ductus arteriosus closes. In one series reported, 18 of 40 HLHS patients had hypoxic-ischemic or hemorrhagic lesions [41]. PVL, ischemic infarction, and

a

brainstem necrosis were the major hypoxic-ischemic lesions. A significant relationship has been found between cardiac surgery and brain damage [41], especially when cardiopulmonary bypass with hypothermic total circulatory arrest is used. Our study (Tavani et al., unpublished data) found 5 of 24 (20.8 %) patients to have such abnormalities prior to surgery, and 15 of 24 (62.5 %) of patients to have such findings following surgery. PVL (Fig. 7.8) and some of the ischemic strokes seen in this patient population are not embolic, but represent a hypoxic-hypoperfusion watershed injury. In pediatric patients with cardiomyopathy, congestive heart failure or other end-stage cardiac disease, especially when complicated by pulmonary problems, episodes of bradycardia, hypotension, or hypoxia, lead to watershed infarction in the parasagittal distribution between anterior and middle and posterior and middle cerebral arteries. Acutely, these are best demonstrated by DWI (Fig. 7.3).

b

c

Fig. 7.8a–c. Congenital heart disease in different patients. a Axial CT scan; b axial DWI; c sagittal T1-weighted image (600/14). In a 17-year-old male with congenital heart disease and acute onset of right hemiparesis, CT scan shows a vague hypodensity involving the cortex in the left hemisphere in the precentral gyrus (arrow, a). In the same patient, DWI done the same day shows restricted motion of water as high signal intensity in the perirolandic region. The infarction was thought to be embolic based on clinical findings with echo. In a different term patient, several days old, who underwent cardiac repair for congenital heart disease, MRI (c) shows abnormal increased signal intensity in the periventricular white matter

267

268

R. A. Zimmerman and L. T. Bilaniuk

7.3.4 Hypercoagulable States with Thrombosis There are a large number of etiologies of hypercoagulable states, both primary and secondary (Table 7.3), many quite rare (i.e., homocystinuria) and some thought previously to be rare (i.e., protein C deficiency), but more recently recognized with increasing frequency as a cause of stroke in pediatric patients. Imaging findings in most of the ischemic infarcts due to hypercoagulability do not reveal the underlying disorder, and are thus nonspecific. However, their presence necessitates that hypercoagulable states be excluded as part of the clinical diagnostic workup of pediatric stroke when more obvious etiologies (i.e., presence of a dissection) are not present.

Primary Protein C deficiency Protein S deficiency Antithrombin III deficiency Fibrinogen disorders Anticardiolipin antibodies Lupus anticoagulant Secondary Polycythemia vera Essential thrombocythemia Homocystinuria Hyperviscosity Sickle cell disease Chemotherapeutic agents

lipids shared by platelets, coagulation factors, and endothelial cells, producing thrombosis in young patients [45, 46]. Imaging findings are nonspecific ischemic infarctions.

7.3.4.1 Protein C and S Deficiency

Protein C and S are vitamin-dependent plasma proteins; C, an anticoagulant, and S, a cofactor for C. Protein C is formed in the liver. Acquired and inherited forms of deficiency occur resulting in stroke, typically an ischemic infarct due to spontaneous thrombosis of the proximal vessel. Imaging findings are nonspecific (Fig. 7.9). Recognition of this etiology of infarction in young patients is increasing in frequency [42–44]. 7.3.4.2 Antiphospholipid Antibodies

IgG and IgM antibodies, including lupus anticoagulant and anticardiolipin antibodies, target phospho-

a

Table 7.3. Hypercoagulable diseases

7.3.4.3 Homocystinuria

Homocystinuria is an autosomal recessive genetic defect in which methionine metabolism is affected. Skeletal and cutaneous abnormalities, dislocated lens, and light skin with 50% incidence of mental retardation and a 15% incidence of seizures are found clinically [47]. Spontaneous thrombosis of both arteries and veins occurs, resulting in infarctions, which, by themselves, are nonspecific [48]. Vascular catheterization for arteriography should be avoided when the diagnosis is known or suspected. The physical appear-

b

Fig. 7.9a,b. Hypercoagulable infarction. a Axial FLAIR image; b 3D TOF MRA. MRI (a) shows an acute infarction as increased signal intensity within the right basal ganglia and thalamus. There is mass effect on the body and frontal horn of the lateral ventricle. In the same patient, MRA shows focal area of narrowing (arrow, b) in the horizontal portion of the right middle cerebral artery

Cerebrovascular Disease in Infants and Children

ance of the patient with the history of stroke should lead to the appropriate clinical laboratory workup.

7.3.5.1 Ehlers-Danlos Syndrome

7.3.4.4 Chemotherapeutic Agents

This is a genetic disease of collagen formation with multiple subtypes based on inheritance and clinical manifestation, in which the skin is excessively elastic, the joints hyperextensible, and the vascular bed potentially abnormal. Type IV, which represents less than 20% of Ehlers-Danlos cases, arises from a defect on chromosome 2 [50], and is the variety associated with cerebrovascular complications [51]. Impairment in the synthesis of type III collagen, the major form of vascular collagen, results in vascular weakening with aneurysmal dilatation of internal carotid and vertebral arteries. There is also an increased incidence of intracranial aneurysms. Spontaneous development of carotid cavernous fistulae has also been reported. Catheterization for conventional arteriography is not without risk, as dissections are possible.

L-asparaginase has been implicated in producing a hypercoagulable state by suppressing liver production of clotting factors after several weeks of therapy [49]. Decreased levels of plasminogen, antithrombin, fibrinogen, and factors IX and XII are noted [49]. Both arterial and veno-sinus occlusions as well as intracerebral hemorrhage occur (see Chapter 11 for details).

7.3.5 Primary and Secondary Vascular Wall Disease Both primary and secondary diseases occur that affect the walls of the major feeding arteries in the neck, skull base, or proximal intracranial circulation. In general, the infarctions that result are more often ischemic than hemorrhagic, whether due to luminal narrowing or emboli that are formed and launched from the sites of involvement. The infarctions are nonspecific as to etiology in most, but not all cases as, for example, vertebral dissection pattern of infarction. The location of the infarction indicates the need for evaluation of the vascular circulation supplying the involved distributions, usually by conventional arteriography, although MRI and MRA have proven to be very successful at demonstrating vascular dissections.

7.3.5.2 Neurofibromatosis Type I (NF1)

NF1, a genetic disease localized to chromosome 17 (see Chapter 16), can produce cerebrovascular disease through two mechanisms: (1) involvement of renal arteries can result in hypertension which can lead to intracerebral hemorrhage [52], and (2) involvement of cerebral vessels can produce stenosis and moyamoyalike picture [53] (Fig. 7.10), or infarction in the distal distribution of the vessel.

Fig. 7.10. Neurofibromatosis type I with moyamoya in a 14-year-old male. Coronal 3D TOF MRA. There is marked narrowing of the distal left internal carotid artery (arrow), irregularity and narrowing of the distal right internal carotid artery, and enlargement of the basilar artery with extensive collaterals supplying the anterior circulation

269

270

R. A. Zimmerman and L. T. Bilaniuk

7.3.5.3 Fibromuscular Dysplasia

Fibromuscular dysplasia (FMD) is a nonatherosclerotic angiopathy of young women and occasionally boys [54] that is inherited as an autosomal dominant in some. Most patients are asymptomatic and never identified. Transient ischemic attacks and ischemic infarcts arise from emboli and/or dissection at the site of involvement. Identification of FMD depends on conventional angiography as MRA is unreliable. The “string of beads” appearance of the vessel, a series of constrictions and dilatations (Fig. 7.11), is due to both hyperplasia and thinning of the media of the vessel.

Fig. 7.11. Fibromuscular dysplasia. Common carotid arteriogram. The distal internal carotid artery is shown in its cervical and intrapetrosal portions to be irregularly beaded (arrows). This patient also had renal artery involvement

degree in the Japanese population, where it was first described [56]. Clinical symptoms vary, from none to headaches, transient ischemic attacks, and hemiplegia. CT is poor at identifying the disease other than in showing an acute infarction or hemorrhage (from rupture of a collateral). On CT, contrast enhancement of numerous collaterals at the base of the brain may mimic subarachnoid infiltrative process. MRI and MRA are the noninvasive diagnostic procedures that are most successful in demonstrating the disease. Attenuation of the flow void of the distal internal carotid artery, decrease in size of the flow void of the middle cerebral arteries, and the presence of abnormally numerous small collateral blood vessel flow voids within the basal ganglia, hypothalamus, and thalamus are the T2 findings plus or minus new or old infarcts [57] (Figs. 7.12, 13). MRA shows the internal carotid artery stenosis or occlusion and the myriad of small collateral blood vessels [57]. This condition is an example where the postgadolinium-enhanced 3D TOF MRA, in addition to the pregadolinium MRA, is useful. Gadolinium is also useful in demonstrating subacute infarctions by the disturbance in their blood brain barrier (BBB) producing contrast enhancement of the infarction (Figs. 7.12, 13). Conventional arteriography is the ultimate gold standard in the evaluation of patients suspected to have moyamoya syndrome. Much of the collateral blood supply arises from the distal posterior circulation (basilar artery, posterior communicating arteries, posterior cerebral arteries), necessitating a vertebral artery injection in addition to study of both internal carotid arteries (Fig. 7.12). In general, MRI and MRA are diagnostic (Figs. 7.12, 13), and conventional arteriography is reserved for presurgical evaluation prior to vascular anastomosis. Surgery is done by using the vasculature of the scalp placed via a burr hole, to help develop additional collateral circulation (encephaloduroarteriosynangiosis) [58]. Preoperative arteriography should include both external carotid arteries.

7.3.5.4 Moyamoya Syndrome

7.3.5.5 Radiation Vasculopathy

Moyamoya (in Japanese meaning puff of smoke) describes the arteriographic finding of collateral blood vessels and their angiographic blush, which is due to stenosis and/or occlusions of the distal internal carotid arteries [55]. Many diseases can produce this picture, i.e., NF1, sickle cell disease, and radiation vasculopathy; however, the disease is most often idiopathic and, while worldwide, is found to a greater

One of the effects of irradiation, as used in therapy of pediatric brain tumors, is damage to the endothelium of blood vessels. This occurs not only in the vascular bed of the tumor, but also can occur in the normal vasculature of the cerebrum. Opening up of the BBB can begin in the weeks to months following completion of radiation therapy, and is seen on MRI as the development of new vasogenic edema along with contrast enhancement postgadolinium injection. Often

Cerebrovascular Disease in Infants and Children

a

c

b

Fig. 7.12a–c. Moyamoya disease (different patients). a Axial T2WI (6000/99); b Gdenhanced axial T1-weighted image; c Conventional arteriogram. In a 4 year old female with known moyamoya, MRI (a) shows multiple areas of prior infarction on the right, seen as areas of high signal intensity. There are numerous small hypointense collateral flow voids in both the basal ganglia and in the region of the splenium of the corpus callosum. The arrows (a) point to only a few of the many. In a different patient aged 9 years with moyamoya disease and recent acute onset of symptoms, postcontrast MRI shows contrast enhancing infarcts in the right hemisphere within the caudate head and lenticular nucleus (arrows, b). In another case (a 7-year old-male with known moyamoya), conventional arteriogram with common carotid artery injection on the left shows occlusion of the origin of the middle cerebral artery from the internal carotid (arrow, c). The distal middle cerebral artery is reconstituted from extensive collateral moyamoya vessels

small vessels are affected and there may or may not be a permanent residual injury. In children who are longterm survivors of brain tumor, i.e., hypothalamic and chiasmatic gliomas and medulloblastomas, large and small vessel injury may be seen [59]. Stenosis of the distal internal carotid arteries (Fig. 7.14) with development of collaterals can produce a moyamoya-like picture. Infarctions within the affected vascular territory, with or without moyamoya pattern of collaterals, are the primary concern. Clinically, these can present as transient ischemic attacks, but their progression to frank infarction occurs at an unpredictable rate. MRI and MRA remain the best ways to evaluate postradiotherapy patients noninvasively. Radiation vasculopathy can also present as small areas of hemorrhage (see Chapter 11). Initially, this was thought to be radiation-induced formation of telangiectasia or small cavernomas, which bled. Subsequently, this has been called into question and the

hemorrhages may be manifestations of small vessel injury [60]. These small bleeds are best demonstrated by MRI and are seen as foci of signal intensity changes which depend on the chemical state of the blood products (Fig. 7.14). Gradient-echo T2*-weighted images are the most sensitive for all types of blood products, especially those that are minute.

7.3.6 Congenital Vascular Hypoplasia Developmental hypoplasia or absence of a vessel at the circle of Willis is not an uncommon finding on conventional arteriography, MRA, or at autopsy (50%) [61]. Most often this variation is not of clinical significance, until an acquired occlusion of a major vessel occurs in which the absence of a vessel does not permit the establishment of adequate collateral

271

272

R. A. Zimmerman and L. T. Bilaniuk

a

c

b

d

Fig. 7.13a–d. Moyamoya disease in a 10-month-old girl with Fanconi syndrome. a Axial T2-weighted image; b Coronal T2-weighted image; c Gd-enhanced axial T1-weighted image; d 3D TOF MRA, coronal projection. T2-weighted images (a, b) show diffuse reduction of the vessels of the circle of Willis and an excess amount of tiny flow voids in the basal cisterns, consistent with collateral circulation (rete mirabilis). An infarction with blood-brain-barrier damage is also visible in the right rolandic region (c). MRA shows stenosis of the carotid siphons (with a mouse-tail appearance) (arrows, d), diffuse stenosis of the vessels of the circle of Willis, and the rete mirabilis in the region of the perforating arteries. (Courtesy Dr. P. Tortori-Donati, Genoa, Italy)

circulation in order to prevent the development of an infarction. Under these circumstances, the variations in the circle of Willis take on a much greater significance. Hypoplasia or absence of major brachiocephalic vessels supplying the brain is much less common than at the circle of Willis. True agenesis of the carotid arteries is associated with absence of the carotid canals in the skull base [62]. Carotid agenesis or hypoplasia have also been found in association with congenital facial hemangiomas [63] and in the setting of PHACE syndrome, a vascular phakomatosis [64] (see Chap-

ters 17 and 35), and other orbital region developmental anomalies, such as microphthalmos. In a young adult or child, a hypoplastic internal carotid artery is a setup for potential infarction with any hypotension, arrhythmia, or other insult. MRA and conventional arteriography are the most specific diagnostic methods, although MRI, by demonstrating the size of the arterial flow void, may be the first test that leads to recognition of the problem. Vertebral arteries are commonly of different sizes, the left more often larger than the right. Marked discrepancies in size do occur with the hypoplastic

Cerebrovascular Disease in Infants and Children

b

Fig. 7.14a,b. Radiation vasculopathy (different patients). a. Conventional left common carotid arteriogram; b axial FLAIR image. In a teenage female who presents with transient ischemic attack, postradiation therapy for hypothalamic astrocytoma, conventional arteriogram shows marked narrowing of the distal internal carotid artery (arrow, a). In a different teenage male, long-term survivor of posterior fossa ependymoma treated with surgery, chemotherapy and irradiation, MRI shows a hypointense rim of hemosiderin outlining a cavernoma in the left thalamus (arrow, b). There are extensive high signal intensity radiation changes at the site of the boost portal in the parietal lobes

a

vertebral artery ending in only the posterior inferior cerebellar artery. This is a setup for potential problems. Transient occlusion of the vertebral artery is thought to be possible with marked rotation of the head and cervical spine. Forced rotation, as with chiropractic manipulation, has been found to result in occlusions and dissections of the vertebral artery, leading to infarction in the cerebellum. The lack of effective collateral sources of blood supply in the case of hypoplasia of the vertebral artery may predispose these patients to such events, whether the turning has been forced, occurs accidentally (i.e., car accident), or is self induced.

7.3.7 Vascular Trauma The effect of both penetrating and blunt vascular trauma, if exsanguination does not result, is on the distal vascular bed supplied by the involved vessel, leading to ischemia and infarction. When the brachiocephalic, carotid, or vertebral vessels are damaged, cerebral infarction is not rare. With regard to

brain injury, vascular trauma is not limited to the chest and neck region, as injury to the internal carotid artery can occur within the skull base, at the intracavernous or supraclinoid segment, or at its bifurcation. Penetrating injuries can occur anywhere intracranially. Blunt injury can produce intimal tear, spasm, thrombosis, intramural dissection, dissection aneurysm, pseudoaneurysm, and complete transection [65]. Intimal tear is a site for thrombus formation, which can be a source of emboli to distal sites within the brain. The intimal tear also permits blood access to the wall of the carotid artery, producing an intramural dissecting hematoma. The subintimal blood folds the intima in an irregular fashion, producing the “classic” arteriographic picture (Fig. 7.15) seen in the carotid, middle cerebral, and vertebral arteries with dissections. Conventional arteriography is the gold standard, while MRA and MRI (T1-weighted images with fat suppression to look for intramural clot) are often diagnostic. Distal emboli produce infarction. In the vertebral circulation, involvement is most often of the thalami, cerebellum, occipital lobes, and occasionally the brain stem (Fig. 7.15). The history of

273

274

R. A. Zimmerman and L. T. Bilaniuk

a

b

c

d Fig. 7.15a–d. Vascular trauma (different patients). a Internal carotid arteriogram; b left vertebral arteriogram; c axial T2-weighted image (3000/120); d selective internal carotid arteriogram. In a teenage male with blunt trauma and internal carotid artery dissection, marked narrowing and irregularity of the internal carotid artery in the neck is seen (arrows, a). In a child with vertebral artery dissection, left vertebral arteriogram with reflux into the right vertebral artery shows irregular narrowing on the right (arrow, b). In a different patient (a 4-year-old male with vertebral artery dissection from trauma), MRI shows multiple areas of high signal intensity infarctions involving the left cerebellum, cerebellar peduncle, and brainstem (arrows, c). In another young child with pseudoaneurysm secondary to puncture wound of carotid artery, selective internal carotid arteriogram shows pseudoaneurysm (arrow, d) arising from the anterior wall of the carotid in the neck

blunt trauma may not be obvious, therefore requiring careful questioning about the mode of trauma, such as shooting a rifle, football practice, jumping on a trampoline, or a visit to the chiropractor. In carotid dissections, the new onset of a Horner’s syndrome is an important sign as it points to disruption of the sympathetic fibers in the carotid wall and indicates a need for imaging evaluation. Pseudoaneurysms develop as a result of disruption of the vascular wall with formation of a periarterial

hematoma contained by fascia. Once the clotted hematoma cavitates, the hole communicates with the arterial lumen. The risk to the cerebrum is from emboli formed within the pseudoaneurysm. MRI, MRA, and conventional arteriography demonstrate pseudoaneurysms in the neck and intracranially (Fig. 7.15). Penetrating injuries in adults are due to missiles (bullets) and stab wounds (knives), but in children are more often related to intraoral trauma, when the child falls on a stick or a pencil held in the mouth.

Cerebrovascular Disease in Infants and Children

The object is forced either superiorly through the skull base toward the anterior cerebral arteries, the frontal lobes (transcribriform), or, more commonly, posteriorly into the carotid artery in the neck, where a dissection or pseudoaneurysm occurs, a source of intracranial emboli and infarction.

7.3.8 Migraine Headaches Complicated migraine have the occurrence of transient neurological deficit or altered mentation in association with the headache. Permanent neurological deficits may occur, but are uncommon in children (1 in 132) [66]. The exact incidence of stroke from vasospasm is unknown both in the adult and pediatric population. MRI FLAIR imaging is used to identify lesions in the white matter which are consistent with past episodes of ischemia. These appear as scattered bright areas. In addition, we have now seen a number of teenagers with and without neurological deficits, but with watershed infarctions (Fig. 7.16), both acute and chronic, in which the history has been migraine headache. We have also seen several adolescents with acute thalamic infarctions identified on DWI during the course of their acute complicated migraine attack. In all of these patients it is necessary to exclude an underlying source for infarction, such as protein S deficiency.

7.3.9 Metabolic Diseases Mitochondrial dysfunctions are one of the genetic causes of stroke in children (Table 7.4). These include mitochondrial encephalomyopathy, lactic acidosis and stroke-like episodes (MELAS) [67]; myoclonic epilepsy with ragged red-fibers (MERRF); MERFF/ MELAS overlap syndrome; and Kearns-Sayre syndrome. Mitochondrial encephalopathies are suspected when lactic acidosis (diagnosed by serum, CSF analysis, or MRS) occurs in association with seizures, recurrent strokes, and respiratory failure. DNA classification of these disorders is now possible in many instances [68]. The proposed mechanism of stroke is regional failure to produce sufficient energy Table 7.4. Genetic causes of stroke Hereditary connective tissue disorders Ehlers-Danlos syndrome (type IV) Marfan syndrome Homocystinuria Menkes syndrome Organic acidemias Methylmalonic Propionic Glutaric type III Mitochondrial MELAS MERRF Kearns-Sayre syndrome Leigh disease Neurocutaneous syndromes Neurofibromatosis type I Hereditary dyslipoproteinemia Tangier disease Progeria

to maintain cell function. Sites of normal high energy metabolism appear to be at greatest risk when attacks occur. MRI, DWI, and MRS provide the imaging information for diagnostic workup (Fig. 7.17). MRA is useful to exclude other vascular diseases that might mimic mitochondrial disease, such as basilar artery stenosis.

7.3.10 Hypoxia

Fig. 7.16. Migraine in a 17-year-old male with visual problems and headaches. Axial FLAIR image. There is abnormal increased signal in the left parietal watershed, consistent with infarction

Hypoxic injury to the brain does occur in older infants, children, and adolescents, but does not represent the same magnitude of problem as it does in the preterm and term infant. Often, the hypoxic injury is a complication of another disease process, such as trauma, when the patient has been thrown from the car and

275

276

R. A. Zimmerman and L. T. Bilaniuk

Fig. 7.17. Mitochondrial disease in a 4-year old-male with known mitochondrial disease presenting with acute language problems. Axial DWI. There is restricted motion of water as high signal intensity in posterior left temporal lobe, left thalamus, and left frontal cortex

a

b

d

e

the medullary respiratory center has failed, or in case of child abuse with whiplash impact injury when the same occurs. In these situations, both the primary brain injury and the secondary hypoxia contribute to the ultimate outcome. Resuscitated cardiorespiratory arrests, seizures with poor airway, and near drowning are among other potential etiologies of more specific hypoxic injury. In older infants, the central gray matter and frontal, temporal, and occipital cortex are involved, often with relative sparing of the rolandic cortex and white matter. Such findings are seen in near sudden infant death syndrome (SIDS), transiently during imaging before death. Such a picture, while compatible with SIDS, is also consistent with suffocation as a part of child abuse. In the older child and adolescent, when near drowning occurs, the high metabolic rate areas of the brain, such as the lenticular nuclei, are first to be affected. MRI including DWI is the procedure of choice in demonstrating the acute imaging findings as high signal intensity (Fig. 7.18). CT is often much less revealing.

c

Fig. 7.18a–e

Cerebrovascular Disease in Infants and Children

7.4 Nontraumatic Intracerebral Hemorrhage Cerebrovascular disease presenting as hemorrhage (subarachnoid, intraparenchymal, and/or intraventricular), is relatively frequent even when premature neonates with germinal matrix are excluded. The etiology of the bleeds (Table 7.5) is varied, but vascular malformations and bleeding diathesis are most common. Most often, patients suffering from intracranial hemorrhage have sudden onset of symptoms and require emergency care and rapid and precise evaluation by imaging. Emergency noncontrast CT is the initial step in identifying the presence of blood, its location and size, the degree of mass effect, and any accompanying hydrocephalus (Fig. 7.19). The workup of the etiology of the hemorrhage generally requires MRI, MRA, perhaps MRV (i.e., to diagnose sinus thrombosis), and conventional arteriography as the definitive gold standard in treatment planning of AVMs and aneurysms. Arteriovenous malformations and occult vascular malformations are described in Chapters 8 and 9, respectively. Table 7.5. Etiologies of nontraumatic intracerebral hemorrhage Vascular malformation Aneurysm Hypertension Bleeding diathesis Drug Brain tumor Sinovenous occlusive disease

7.4.1 Cerebral Aneurysms Aneurysms of cerebral blood vessels are uncommon in children, despite the concept that these are usually congenital in origin, indicating that only the weakness in the vessel’s wall is congenital, but not the aneurysm. When an intracerebral hematoma (ICH) is found in an infant or child, without history of trauma, an AVM is usually the first consideration as the eti-

ology. However, aneurysms do occur, although the total number in pediatrics is only 5% of the total number of aneurysms in the general population [69]. Pediatric aneurysms are of interest because there is a distinct male dominance, a greater incidence of large and giant aneurysms, and a lower incidence of multiple aneurysm than in the adult population [70, 71]. Twenty-four to fifty-four percent of pediatric aneurysms are located in the distal internal carotid artery circulation, while more than 15% are in the posterior circulation and 4% are found on the posterior communicating artery [71]. The types of aneurysms in pediatrics are as follows: saccular 50%–70%, giant 20%–40%, mycotic (due to infectious process) 5%– 15%, traumatic 5%–15%, and multiple 2% [71]. The relatively high incidence of traumatic and mycotic aneurysms means that attention has to be paid to the possibility of peripherally located aneurysms which, in addition to trauma and infection, include other etiologies such as tumor emboli (atrial myxoma, rhabdomyoma) and arteritis [72]. Aneurysms also occur with AVMs, moyamoya syndrome, and sickle cell vascular disease [72]. Symptoms related to pediatric aneurysms are due to mass effect of the aneurysm in up to 45% of patients over age 5 years [72]. Both seizures and signs of infarction are present in less than 10% of pediatric patients [73]. In pediatric patients there is a greater concern for conditions which predispose to aneurysm formation. These include aortic coarctation, polycystic kidney disease, fibromuscular dysplasia, Marfan’s Syndrome and Ehlers-Danlos syndrome type IV [72]. CT is most often the initial diagnostic study performed in evaluation of patients suspected to have an aneurysm. The imaging findings are either SAH, SAH and ICH, or a hyperdense mass, the aneurysm itself (Fig. 7.19). MRI and MRA are the next step, although if the clinical situation dictates, conventional arteriography before surgery and/or interventional treatment may be indicated (Fig. 7.19). On T2-and sometimes on T1-weighted images, MRI shows the hypointense flow void of the lumen of the nonclotted aneurysm (Fig. 7.19). Blood products, due to rupture, may be found adherent to the aneurysm, a good sign of which aneurysm has bled when more than one is present.

Fig. 7.18a–e. Hypoxic brain injury in two different patients. a axial FLAIR image; b ADC map; c, d axial FLAIR images; e coronal T2-weighted image. In a near-drowning child, axial FLAIR image (a) shows high signal intensity abnormalities in the lenticular nuclei bilaterally and in the thalamus. In the same patient, ADC map (b) shows that the abnormalities are more extensive, involving the corpus callosum and temporal occipital cortex and subadjacent white matter. In a different patient (a 10-year-old boy with prolonged cardiac arrest) axial FLAIR images (c, d) show hyperintense signal in the posterior putamen and thalamus bilaterally (thin arrows c, d), left calcarine region (open arrow, c), and rolandic cortex bilaterally (arrowheads d). Basal ganglia abnormalities give high signal also on T2-weighted images (arrows, e)

277

278

R. A. Zimmerman and L. T. Bilaniuk

a

b

d

e

c

Fig. 7.19a–f. Cerebral aneurysms (different patients). a, b. Axial CT scan; c. selective left vertebral arteriogram; d. coronal T2-weighted image (6000/99); e, f 3D TOF MRA. In a 6-month-old male with ruptured anterior cerebral artery aneurysm presenting with acute subarachnoid hemorrhage and intracerebral hematoma, CT scan shows blood in the right frontal lobe (arrow, a) and within the subarachnoid space between two frontal lobes extending into the chiasmatic cistern and around the brainstem. There is early hydrocephalus with dilatation of the temporal horns. In a different 4-year-old male with giant basilar artery aneurysm presenting with f subarachnoid hemorrhage and hydrocephalus, CT scan shows a calcified mass (arrow, b) in the interpeduncular chiasmatic cistern region. There is increased density in the interhemispheric fissure and in the fissures of the middle cerebral arteries, consistent with subarachnoid blood. The temporal horns are markedly enlarged from hydrocephalus. In the same patient, selective left vertebral arteriogram (c) shows giant basilar artery aneurysm. In another infant with peripheral giant mycotic aneurysm and presenting with the CT scan picture of a right frontal intercerebral hematoma suspected to be child abuse, MRI shows a hypointense flow void representing the patent lumen of the aneurysm (white arrow, d). There is vasogenic edema in the surrounding brain (black arrowhead, d). Between the vasogenic edema and the patent lumen of the aneurysm there is laminated clot in the wall of the giant aneurysm. In the same patient, MRA shows the patent lumen of the aneurysm on the right (white arrow, e). In a different patient (a 12-year-old male presenting with acute severe headache), coronally reconstructed 3D TOF MRA shows a 3 mm aneurysm (arrow, f) arising from the A1 segment of the anterior cerebral artery

Cerebrovascular Disease in Infants and Children

Blood products may be found within the aneurysm when some degree of clotting has occurred (Fig. 7.19). CT is more sensitive to acute SAH than MRI, while FLAIR MRI is more sensitive to subacute SAH than CT [4]. MRI (T2 and susceptibility sequence) is equal to CT in sensitivity for acute ICH and IVH, but more sensitive than CT for subacute and chronic blood products. MRA is performed with 3D TOF sequence for the demonstration of aneurysm, unless methemoglobin is shown on T1-weighted images (Fig. 7.19). If methemoglobin is present, then 3D PC MRA is the procedure of choice in order to avoid confusion of clotted and flowing blood. Aneurysms greater than 3 mm should be visible with good segmentation and/or multiplanar reconstruction (MPR) technique (Fig. 7.19) [8]. Contrast-enhanced MRA can be used to increase the conspicuity of the lumen of the aneurysm. An advantage of MRI and MRA over conventional arteriography is the ability to demonstrate the wall of the aneurysm (thickness, clot) and the condition of the surrounding brain (i.e., infarction, mass effect, hydrocephalus). MRA can also show whether spasm of the intracranial vessels is present or not.

7.4.2 Hemorrhagic Infarction In adults, autopsy studies have found that 50%–70% of embolic strokes are hemorrhagic, compared to 2%–20% of nonembolic ones. Most hemorrhagic strokes found at autopsy were not hemorrhagic on CT within the first 48 h after ictus. The pathogenesis of hemorrhage in an embolic infarct has one of two mechanisms: (1) the embolus moves on or lyses, flow is re-established, and blood breaks through the wall of the damaged vessel; (2) there is good collateral blood flow which, by retrograde flow, reperfuses the damaged vascular bed, producing the same insult. Cardiac disease is the most common etiology of pediatric embolic disease, and congenital heart disease is the most common. Numerous other disease processes can be associated with embolism (i.e., fibromuscular dysplasia, traumatic vascular injury, cerebral aneurysms). Venous sinus occlusive disease, especially with cortical vein thrombosis, is another cause of hemorrhagic infarction. The mechanism of infarction and bleeding is different than in embolic stroke. The venous outlet becomes obstructed so that the blood coming in via the arteries is impeded, builds up pressure by increasing the local cerebral blood volume,

produces ischemia, and eventually ruptures into the surrounding damaged tissue. On CT, the high-density hemorrhagic infarction is within a vascular territory when embolic in nature, and not when due to venous occlusive disease. Transformation of a nonhemorrhagic ischemic infarction to a hemorrhagic one can be shown both by CT and MRI over a course of time. MRI shows the blood products according to their chemical state, and often shows nonhemorrhagic components as part of the infarction. Susceptibility scanning and DWI are important techniques that should be used in addition to routine T1- and T2-weighted images.

7.4.3 Coagulopathies Bleeding diathesis can be found in association with certain hematologic malignancies, such as acute leukemia, where thrombocytopenia below 20,000 platelets/mm3 is problematic [20]. Hemophilia has a reported incidence of 2.2%–14% of intracranial hemorrhage, and it is this bleeding that is the leading cause of death [74]. Anticoagulation for surgical purposes (as in operative treatment of congenital heart disease) or as therapy for a hypercoagulable or thrombotic disorder are other potential etiologies for ICH. Vitamin K deficiency in neonates results in decreased factors II, VII, IX, and X [75]. Today, 1 mg of vitamin K is given at birth; however, previously, when it was not given, intracranial hemorrhage was a risk in breast fed babies.

7.4.4 Hypertension High blood pressure in pediatric patients is relatively rare, less than 1% of children [76]. Causes include coarctation of aorta, renal disease, and endocrine diseases. Most of the hypertensive events that affect the brain in the pediatric population either result in edema or ischemia in the posterior watershed between the posterior cerebral and middle cerebral arteries. Damage to the BBB by increased pressure can lead to vasogenic edema, best seen on FLAIR MRI (Fig. 7.20) but not positive for acute infarction on DWI. However, vasospasm can also occur, resulting in infarction, which can be diagnosed very promptly when acute with DWI. Bleeds in the basal ganglia, brainstem, and cerebellum or cerebral hemispheres do occur in children (Fig. 7.20), but are less common than in adults with hypertension.

279

280

R. A. Zimmerman and L. T. Bilaniuk

a

b

Fig. 7.20a,b. Hypertension (different patients). a coronal FLAIR image; b gradient-echo susceptibility scan. In a 10-year-old child with acute lymphocytic leukemia and chemotherapeutically induced hypertension with disturbance in the blood brain barrier in the biparietal watershed, MRI shows hyperintensity of the cortex (arrows, a). Diffusion imaging was negative, and the changes resolved with treatment. In a different case (a 12-year-old male with severe chronic hypertension secondary to renal disease), MRI (c) shows multiple bilateral hypointense sites of prior bleeding within the basal ganglia

7.4.5 Sinovenous Occlusion Two types of venous occlusive disease present as neurologic compromise with acute onset of symptoms [77]. These are: (1) superficial dural sinus thrombosis with or without cortical vein thrombosis; and (2) deep venous thrombosis. The setting for deep venous thrombosis is the very young age of patients, often newborns or young infants, and significant dehydration. Superficial dural sinus thrombosis tends to occur in older patients (not exclusively) and to have a wide variety of etiologies, including dehydration, sepsis, and hypercoagulable states (Table 7.6). Superior sagittal sinus thrombosis may be relatively asymptomatic if only that structure is involved; however, with the frequent involvement of transverse sinuses and cortical veins, cerebral infarction, often hemorrhagic, becomes more frequent and acutely symptomatic. Superior sagittal sinus thrombosis may present with chronic symptoms, such as papilledema with pseudotumor cerebri, and given that clinical diagnosis, imaging evaluation is required. Obstruction of the deep venous system affects the thalami and basal ganglia and, at times, the periventricular white matter, and produces infarction, often hemorrhagic in part.

Table 7.6. Sinovenous thrombosis: etiologies Infections of head and neck Dehydration Chemotherapeutic agents Inflammatory disorders, i.e., bowel disease Collagen vascular disease Hypercoagulable states Flow-related, i.e., AVM Iatrogenic

Noncontrast CT is often the initial diagnostic study in the patient with sinovenous occlusion. The most specific CT finding is the presence of hyperdense clot in the sinus or deep veins (Fig. 7.21). Unfortunately, infants tend to have normally relatively large dense dural venous sinuses that can be misdiagnosed as thrombosed by the less experienced. The presence of hemorrhage and edema in the brain and/or thrombosed cortical veins make the diagnosis easier (Fig. 7.21). The delta sign is a sign of subacute superior sagittal sinus thrombosis described on contrastenhanced CT, wherein the central less dense thrombus in the sinus is outlined by denser enhancing dural coverings of the sinus. MRI and MRV are now the gold standards in diagnosis of sinovenous thrombosis [8]. Conventional

Cerebrovascular Disease in Infants and Children

a

d

b

c

e Fig. 7.21a–i. Venous thrombosis (different patients). a–c axial CT scan; d, e sagittal T1-weighted image; f coronal 2D TOF MRV with saturation of arterial input; g, i axial T2-weighted images; h axial DWI. In a 1-month-old male with dehydration and deep venous thrombosis of the internal cerebral vein and tributary veins, the two adjacent CT slices show clot in the internal cerebral veins (arrow, a) as well as subependymal veins and choroid plexus. There is hypodensity in the left thalamus consistent with infarction. In a teenage female with cortical vein thrombosis, CT scan (c) shows a large parietal area of infarction with mass effect on the lateral ventricle and subfalcial shift of the midline structures. The infarct is both hypodense and peripherally hyperdense. The peripheral hyperdensity (arrow, c) is an area of hemorrhagic infarction. In a different patient with sagittal sinus thrombosis, MRI shows a hyperintense clot within the sagittal sinus. A slightly hyperintense clot is present in the straight sinus (arrow, d). A thrombosed cortical vein can also be seen (open arrow, d). In a 1-month-old dehydrated infant with deep venous thrombosis, MRI shows a hyperintense clot in the internal cerebral vein (arrow, e) as well as within the vein of Galen and straight sinus. The thalamus, beneath the internal cerebral vein, is swollen and decreased in signal intensity.

f

continued on next page >

281

282

R. A. Zimmerman and L. T. Bilaniuk

g

h

i

Fig. 7.21a–i. (continued) ... In another case with sagittal sinus and transverse sinus thrombosis with collateral venous drainage, MRV with saturation of arterial input (f) shows segments of flow at the site of the superior sagittal sinus interrupted by gaps due to clots (arrows, f). There are multiple small collateral veins over both hemispheres. There is absence of both transverse venous sinuses. In a 9-year-old female with cortical vein thrombosis and infarction, MRI shows cortical hyperintensity in the right parietal lobe at the site of infarction. A large hypointense thrombosed cortical vein can be seen (arrow g). In the same patient, DWI (h) shows the area of restricted motion of water due to infarction to be larger than seen on the T2WI. In a one-month-old dehydrated male with medullary vein thromboses due to deep venous thrombosis, MRI (i) shows bilateral multiple white matter areas of hypointensity, indicating thrombosis within deep medullary veins

arteriography is almost never used. On MRI, absence of the flow void in the sinus is the initial finding in the hyperacute-acute thrombosis. This is best seen on T1weighted images. Over several days, the clot becomes T1 hyperintense (methemoglobin) (Fig. 7.21). 2D TOF MRV is used before methemoglobin is present, and demonstrates the thrombosis as a lack of flow-related enhancement in the affected sinus (Fig. 7.21). Once methemoglobin has formed, the procedure for MRV is 2D phase contrast, wherein the image is insensitive to the T1 hyperintensity of methemoglobin and is made up by only the moving protons. In the absence of blood movement, the sinus lacks signal from flow. Brain tissue affected by the lack of venous drainage is visualized on T2-weighted images, FLAIR, susceptibility, and DWI (Fig. 7.21). Vasogenic edema can occur from venous obstruction, and can be best seen on FLAIR. Acute infarction requires DWI, but can be seen on T2-weighted images. Hemorrhagic components, while seen on T2-weighted images (Fig. 7.21) and perhaps on T1-weighted images, are best shown by T2 susceptibility sequences.

7.5 Conclusions The imaging diagnostic workup of the pediatric patient with cerebrovascular disease depends to a large extent on the neuroradiologists and their ability to perform and interpret correctly the appropriate studies. There is now a wide array of studies with varying specificity and some pitfalls. However, it has never been easier than now to obtain the important information. CT remains the initial screening test for most acute events occurring beyond early infancy, but while it is good for SAH, ICH, and IVH, it is often less informative regarding etiology of bleeds and in the demonstration of acute ischemic infarction and hypoxia. MRI with T1, T2, FLAIR, DWI, susceptibility scanning, MRA, and MRV have revolutionized the capability of the neuroradiologist to provide diagnostic information to the clinician. Despite the imaging advances, arriving at the etiology of ischemic stroke is still mostly dependent on the history and laboratory findings. Conventional arteriography has been relegated to a limited, but important, role in providing definitive information in specific circumstances (aneurysm, AVM, vasculitis, etc.).

Cerebrovascular Disease in Infants and Children

References 1. Tomura N, Uemura K, Inugami A, Fujita H, Higano S, Shishido F. Early CT finding in cerebral infarction: Obscuration of the lentiform nucleus. Radiology 1988; 168:464467. 2. Moulin T, Cattin F, Crepin-Leblond T, Tatu L, Chavot D, Piotin M, Viel JF, Rumbach L, Bonneville JF. Early CT signs in acute middle cerebral artery infarction; predictive value for subsequent infarct locations and outcome. Neurology 1996; 47:366–375. 3. Bastianello S, Pierallini A, Colonnese C, Brughitta G, Angeloni U, Antonelli M, Fantozzi LM, Fieschi C, Bozzao L. Hyperdense middle cerebral artery CT sign: comparison with angiography in the acute phase of ischemic supratentorial infarction. Neuroradiology 1991; 33:207–211. 4. Noguchi K, Ogawa T, Seto H, Inugami A, Hadeishi H, Fujita H, Hatazawa J, Shimosegawa E, Okudera T, Uemura K. Subacute and chronic subarachnoid hemorrhage; diagnosis with fluid-attenuated inversion-recovery imaging. Radiology 1997; 203:257–262. 5. New PF, Aronow S. Attenuation measurements of whole blood and fractions in computed tomography. Radiology 1976;121:635–640. 6. Dolinskas C, Bilaniuk LT, Zimmerman RA, Kuhl DE. Transmission CT observations on hematoma resolution. Part 2. AJR Am J of Roentgenology 1977;129:689–691. 7. Noguchi K, Ogawa T, Inugami A, Fujita H, Hatazawa J, Shimosegawa E, Okudera T, Uemura K, Seto H. MRI of acute cerebral infarction: a comparison of FLAIR and T2 weighted fast spin-echo imaging. Neuroradiology 1997, 39:406–410. 8. Zimmerman RA, Naidich TP. Magnetic resonance angiography. In: Salcman MN (ed) Current Techniques in Neurosurgery. Philadelphia: Current Medicine, 1993. 9. Sorensen AG, Buonanno FS, Gonzalez RG, Schwamm LH, Lev MH, Huang-Hellinger FR, Reese TG, Weisskoff RM, Davis TL, Suwanwela N, Can U, Moreira JA, Copen WA, Look RB, Finklestein SP, Rosen BR, Koroshetz WJ. Hyperacute stroke: evaluation with combined multi-section diffusion-weighted and hemodynamically weighted echoplanar MRI. Radiology 1996; 199:391–401. 10. Brunberg JA, Chenevett TL, McKeever PE, Ross DA, Junck LR, Muraszko KM, Dauser R, Pipe JG, Betley AT. In vivo MR determination of water diffusion coefficients and diffusion anisotropy: correlation with structural alteration in gliomas of the cerebral hemispheres. AJNR Am J Neuroradiol 1996; 16:361–371. 11. Le Bihan D, Turner R, Douek P, Patronas N. Diffusion MR imaging: clinical applications. AJR Am J Roentgenol 1992; 159:591–599. 12. Wimberger DM, Roberts TP, Barkovich AJ, Prayer LM, Moseley ME, Kucharczyk J. Identification of ‘premyelination’ by diffusion-weighted MRI. J Comput Assist Tomogr 1995; 19:28–33. 13. Bandettini PA, Wong EC, Hinks RS, Tikofsky RS, Hyde JS. Time course EPI of human brain function during task activation. Magn Reson Med 1992; 25:390–397. 14. Matsumoto K, Lo EH, Pierce AR, Wei H, Garrido L, Kowall NW. Role of vasogenic edema and tissue cavitation in ischemic evolution on diffusion-weighted imaging: comparison with multiparameter MR and immunohistochemistry. AJNR Am J Neuroradiol 1995; 16:1107–1115. 15. Moseley ME, Kucharczyk J, Minotorovitch J, Cohen Y, Kurhanewicz J, Derugin N, Asgari H, Norman D. Diffusion-weighted MRI of acute stroke. Correlation with T2-

weighted and magnetic-susceptibility-enhanced MRI. AJNR Am J Neuroradiol 1990; 11:423–429. 16. Kwong KK, Chesler DA, Weisskoff RM, Donahue KM, Davis TL, Ostergaard L, Campbell TA, Rosen BR. MR perfusion studies with T1-weighted echo planar imaging. Magn Reson Med 1995; 34:878–887. 17. Edelman RR, Siewert B, Darby DG, Thangaraj V, Nobre AC, Mesulam MM, Warach S. Qualitative mapping of cerebral blood flow and functional localization with echo-planar MR imaging and signal targeting with alternating radio frequency. Radiology 1994; 192:513–520. 18. Higuchi T, Fernandez EJ, Maudsley AA, Shimizu H, Weiner MW, Weinstein PR. Mapping of lactate and N-acetyl Laspartate predicts infarction during focal ischemia: In vivo 1H magnetic resonance spectroscopy in rats. Neurosurgery 1996; 38:121–130. 19. Schoenberg BS, Mellinger JF, Schoenberg DG. Cerebrovascular disease in infants and children: A study of the incidence, clinical features and survival. Neurology 1978; 28:763–768. 20. Biller J. Strokes in the young. In: Toole JF (ed) Cerebrovascular Disorders. Philadelphia: Lippincott Williams & Wilkins, 1999:283. 21. Kandeel AY, Zimmerman RA, Ohene-Frempong K. Comparison of magnetic resonance angiography and conventional angiography in sickle cell disease; clinical significance and reliability. Neuroradiology 1996; 38:409–416. 22. Zimmerman RA, Bilaniuk LT, Packer RJ, Goldberg HI, Grossman RI. Computed tomographic-arteriographic correlates in acute basal ganglionic infarction of childhood. Neuroradiology 1983; 24:241–248. 23. Solomon GE, Hilal SK, Gold AP, Carter S. Natural history of acute hemiplegia of childhood. Brain 1970; 93:107–120. 24. Kinney TR, Sleeper LA, Wang WC, Zimmerman RA, Pegelow CH, Ohene-Frempong K, Wethers DL, Bello JA, Vichinsky EP, Moser FG, Gallagher DM, DeBaun MR, Platt OS, Miller ST. Silent cerebral infarcts in sickle cell anemia: a risk factor analysis. The Cooperative Study of Sickle Cell Disease. Pediatrics 1999; 103:640–645. 25. Pegelow CH, Macklin EA, Moser FG, Wang WC, Bello JA, Miller ST, Vichinsky EP, DeBaun MR, Guarini L, Zimmerman RA, Younkin DP, Gallagher DM, Kinney TR. Longitudinal changes in brain MRI findings in children with sickle cell disease. Blood 2002; 99:3014–3018. 26. Armstrong FD, Thompson RJ, Wang W, Zimmerman R, Pegelow CH, Miller S, Moser F, Bello J, Hurtig A, Vass K. Cognitive functioning and brain magnetic resonance imaging in children with sickle cell disease. Neuropsychology Committee of the Cooperative Study of Sickle Cell Disease. Pediatrics 1996; 97:864–870. 27. Wang WC, Gallagher DM, Pegelow CH, Wright EC, Vichinsky EP, Abboud MR, Moser FG, Adams RJ. Multicenter comparison of magnetic resonance imaging and transcranial Doppler ultrasonography in the evaluation of the central nervous system in children with sickle cell disease. J Pediatr Hematol Oncol 2000; 22:335–339. 28. Syrjanen J. Infection as a risk factor for cerebral infarction. Eur Heart J 1993; 14:17–19. 29. Friede RL. Cerebral infarcts complicating neonatal leptomeningitis. Acta Neuropath 1973; 23:245–252. 30. Simon RP. Neurosyphilis. Arch Neurol 1985; 42:606–613. 31. Marcus JC. Congenital neurosyphilis: a re-appraisal. Neuropediatrics 1982; 13:195–199. 32. Kamholz J, Tremblay G. Chickenpox with delayed contra-

283

284

R. A. Zimmerman and L. T. Bilaniuk lateral hemiparesis caused by cerebral angiitis. Ann Neurol 1985; 18:358–360. 33. Gray F, Belec L, Lescs MC, Chretien F, Ciardi A, Hassine D, Flament-Saillour M, de Truchis P, Clair B, Scaravilli F. Varicella-zoster virus infection of the central nervous system in the acquired immune deficiency syndrome. Brain 1994; 117 (Pt 5): 987–999. 34. Shah S, Zimmerman RA. Cerebrovascular complications of pediatric CNS HIV disease. AJNR Am J Neuroradiol 1996; 17:1913–1917. 35. Walsh TJ, Hier DB, Caplan RL. Aspergillosis of the central nervous system: Clinicopathological analysis of 17 patients. Ann Neurol 1985; 18:574–582. 36. Press GA, Weinding SM, Hesselink JR, Ochi JW, Harris JP. Rhinocerebral mucormycosis: MR manifestations. J Comput Assist Tomogr 1988; 12:744–749. 37. Lupi-Herrera E, Sanchez-Torres G, Marcushamer J, Mispireta J, Horwitz S, Vela JE. Takayasu’s arteritis. Clinical study of 107 cases. Am Heart J 1977; 93:94–103. 38. Kohrman MH, Huttenlocher PR. Takayasu arteritis: A treatable cause of stroke in infancy. Pediatr Neurol 1986; 2:154–158. 39. Cabanes L, Mas JL, Cohen A, Amarenco P, Cabanes PA, Oubary P, Chedru F, Guerin F, Bousser MG, de Recondo J. Atrial septal aneurysm and patent foramen ovale as risk factors for cryptogenic stroke in patients less than 55 years of age; a study using transesophageal echocardiography. Stroke 1993; 24:1865–1873. 40. Lechat PH, Pas JL, Lascault G, Loron P, Theard M, Klimczac M, Drobinski G, Thomas D, Grosgogeat Y. Prevalence of patent foramen ovale in patients with stroke. N Engl J Med 1988; 318:1148–1152. 41. Glauser TA, Rorke LB, Weinberg PM, Clancy RR. Acquired neuropathologic lesions associated with the hypoplastic left heart syndrome. Pediatrics 1990; 85:991–1000. 42. Whitlock JA, Janco RL, Phillips JA. Inherited hypercoagulable states in children. Am J Pediatr Hematol Oncol 1989; 11:170–173. 43. Mannucci PM, Tripodi A, Bertina RM. Protein S deficiency associated with juvenile arterial and venous thromboses. Thromb Haemost 1986; 55:440. 44. Israels SJ, Seshia SS. Childhood stroke associated with protein C or S deficiency. J Pediatr 1987; 111:562–564. 45. Brey RL, Hart RG, Sherman DG, Tegeler CH. Antiphospholipid antibodies and cerebral ischemia in young people. Neurology 1990; 40:1190–1196. 46. Ginsburg KS, Liang MH, Newcomer L, Goldhaber SZ, Schur PH, Hennekens CH, Stampfer MJ. Anticardiolipin antibodies and the risk for ischemic stroke and venous thrombosis. Ann Intern Med 1992; 117:997–1002. 47. Schoonderwaldt HC, Boers GHJ, Cruysberg JRM, Schulte BP, Slooff JL, Thijssen HO. Neurologic manifestations of homocystinuria. Clin Neurol Neurosurg 1981; 83:153– 162. 48. Zimmerman RA, Sherry RG, Elwood JC, Gardner LI, Streeten BW, Levinsohn EM, Markarian B. Vascular thrombosis and homocystinuria. Radiologic-pathologic correlation conference. AJR Am J Roentgenol 1987; 148:953–957. 49. Priest JR, Ramsay NKC, Steinherz PG, Tubergen DG, Cairo MS, Sitarz AL, Bishop AJ, White L, Trigg ME, Levitt CJ, Cich JA, Coccia PF. A syndrome of thrombosis and hemorrhage complicating L-asparaginase therapy for childhood acute lymphoblastic leukemia. J Pediatr 1982; 100:984–989.

50. Nicholls AC, De Paepe A, Narcisi P, Dalgleish R, De Keyser F, Matton M, Pope FM. Linkage of a polymorphic marker for the type III collagen gene (COL3A1) to atypical autosomal dominant Ehlers-Danlos syndrome type IV in a large Belgian pedigree. Hum Genet 1988; 78:276–281. 51. Roach ES, Zimmerman CF. Ehlers-Danlos syndrome. In: Caplan LR, Bogouslavsky J (eds) Cerebrovascular Syndromes. London: Oxford University Press, 1995. 52. Pellock JM, Kleinman PK, McDonald BM, Wixson D. Childhood hypertensive stroke with neurofibromatosis. Neurology 1980, 30:656–659. 53. Rizzo JF, Lessell S. Cerebrovascular abnormalities in neurofibromostis type 1. Neurology 1994; 44: 1000–1002. 54. Emparanza JI, Aldamiz-Echevarria L, Perez-Yarza E, Hernandez J, Pena B, Gaztanaga R. Ischemic stroke due to fibromuscular dysplasia. Neuropediatrics 1989; 20:181– 182. 55. Suzuki J, Takaku A. Cerebrovascular “moyamoya” disease. Arch Neurol 1969; 20:288–289. 56. Kudo T. Spontaneous occlusion of the circle of Willis. Neurology 1968; 18:485–496. 57. Yamada I, Matsushima Y, Suzuki S. Moya-moya disease: diagnosis with three-dimensional time-of-flight angiography. Radiology 1992; 184:773–778. 58. Rooney CM, Kaye EM, Scott M, Klucznik RP, Rosman NP. Modified encephaloduroarteriosynangiosis as a surgical treatment of childhood moyamoya disease: Report of five cases. J Child Neurol 1991; 6:24–31. 59. Mitchell WG, Fishman LS, Miller JH, Nelson M, Zeltzer PM, Soni D, Siegel SM. Stroke as a late sequela of cranial irradiation for childhood brain tumors. J Child Neurol 1991; 6:128–133. 60. Allen JC, Miller DC, Budzilovich GN, Epstein FJ. Brain and spinal cord hemorrhage in long-term survivors of malignant pediatric brain tumors: a possible late effect of therapy. Neurology 1991; 41:148–150. 61. Menkes JH, Sarnat HB. Cerebrovascular Disorders. In: Menkes JH, Sarnat HB (eds) Child Neurology. Lippincott Williams & Wilkins, 2000:885. 62. Quint DJ, Silbergleit R, Young WC. Absence of the carotid canals at skull base. Radiology 1992; 182:477–481. 63. Pascual-Castroviejo, Viano J, Pascual-Pascual SI, Martinez V. Facial haemangioma, agenesis of the internal carotid artery and dysplasia of cerebral cortex: case report. Neuroradiology 1995; 37:692–695. 64. Rossi A, Bava GL, Biancheri R, Tortori-Donati P. Posterior fossa and arterial abnormalities in patients with facial capillary haemangioma: presumed incomplete phenotypic expression of PHACES syndrome. Neuroradiology 2001; 43:934–940. 65. Zimmerman RA. Vascular injuries of the head and neck. Neuroimaging Clin N Am 1991; 1:443–459. 66. Hockaday JM. Basilar migraine in childhood. Dev Med Child Neurol 1979; 21:455–463. 67. Kim IO, Kim JH, Kim WS, Hwang YS, Yeon KM, Han MC. Mitochondrial myopathy-encephalopathy-lactic acidosisand strokelike episodes (MELAS) syndrome: CT and MR findings in seven children. AJR Am J Roentgenol 1996; 166:641–645. 68. DiMauro S, Moraes CT, Schon EA.Mitochondrial encephalopathies: problems of classification. In: Sato T, DiMauro S (eds) Mitochondrial Encephalopathies. New York: Raven Press, 1991: 113–127. 69. Locksley HB. Report on the cooperation study of intracra-

Cerebrovascular Disease in Infants and Children nial aneurysms and subarachnoid hemorrhage. J Neurosurg 1966; 25:219–239. 70. Zimmerman RA, Atlas S, Bilaniuk LT, Hackney DB, Goldberg HI, Grossman RI. Magnetic resonance imaging of cerebral aneurysm. Acta Radiol Suppl 1986; 369:107– 109. 71. Choux M, Lena G, Genitori L. Intracranial aneurysms in children. In: Raimondi A, Choux M, Di Rocco C (eds) Cerebrovascular disease in children. Vienna: Springer, 1992:123–131. 72. Lasjaunias P. Intracranial aneurysms in children. In: Lasjaunias P (ed) Vascular Diseases in Neonates, Infants and Children: Interventional Neuroradiology Management. Berlin: Springer, 1997.

73. Sedzimir CB, Robinson J. Intracranial hemorrhage in children and adolescents. J Neurosurg 1973; 38:269–281. 74. Kerr CB. Intracranial haemorrhage in hemophilia. J Neurol Neurosurg Psychiatry 1964; 27:166–173. 75. Visudhiphan P, Bhanchet P, Lakanapichanchat C, Chiemchanya S. Intracranial hemorrhage in infants due to acquired prothrombin complex deficiency. J Neurosurg 1974; 41:14– 19. 76. DeSwiet M. The epidemiology of hypertension in children. Br Med Bull 1986; 42:172–175. 77. Barron TF, Gusnard DA, Zimmerman RA, Clancy RR. Cerebral venous thrombosis in neonates and children. Pediatr Neurol 1992; 8:112–116.

285

Arteriovenous Malformations: Diagnosis and Endovascular Treatment

8

Arteriovenous Malformations: Diagnosis and Endovascular Treatment Laecio Batista, Augustin Ozanne, Marcos Barbosa, Hortensia Alvarez, and Pierre Lasjaunias

CONTENTS 8.1 8.1.1 8.1.2

Introduction 287 Role of Imaging in the Pediatric Age Group 288 Arterial and Venous System in Children 290

8.2 8.2.1 8.2.2 8.2.3 8.2.3.1 8.2.3.2 8.2.3.3 8.2.3.4 8.2.3.5

Classification of Cerebral Vascular Malformations 290 Pial, Dural, or Choroidal Lesions 291 Timing of Genesis 292 Angioarchitecture and Morphology 292 Angioarchitecture 292 Size 294 Location 295 Multiplicity 295 Complex Cranio-Facial Vascular Lesions 295

8.3 8.3.1 8.3.2 8.3.3 8.3.4

Specificities per Type 296 Pial Lesions 296 Dural Lesions 297 Vein of Galen Aneurysmal Malformation 302 Craniofacial Metameric Lesions 309

8.4 8.4.1 8.4.2

Treatment 310 Therapeutic Objectives and Follow-Up Imaging of Embolic Agents 313 References

310

316

8.1 Introduction Cerebral arteriovenous malformations (CAVMs) or shunts have different characteristics in children when compared with adults with respect to multifocal lesions, induced remote arteriovenous shunts [1, 2], venous thrombosis, systemic manifestations, large venous ectasias, high flow lesions, and rapid cerebral atrophy [3–5]. Conversely, high-flow angiopathic changes are seldom seen with the same frequency as in adults; flow-related arterial aneurysms are absent [6], whereas proximal occlusive arteriopathic changes are more frequent. For this reason, management protocols derived from experience in adults cannot be applied to the pediatric population. In particular, adult-based AVM grading is especially inappropriate for children, because (1) cerebral elo-

quence is difficult to assess, particularly in the first years of life; (2) most lesions are fistulas or multifocal; (3) their drainage usually involves the entire venous system; and (4) the possibility for recovery in children is different from adults. In addition, a high grade AVM that would be dangerous to operate in an adult may not necessarily be dangerous for the patient if left unoperated, or at least more dangerous than a lowgrade AVM. The decision-making process in children includes, in addition to the conventional objectives, various specific factors, such as anatomic analysis of veins and the brain myelination process. Thereafter, progressive deficits related to congested cerebral veins, poorly controlled seizures, hemorrhages with or without specific arterial or venous changes upstream and downstream the AVMs, or headaches without hydrocephalus or macrocrania can become immediate objectives for staged or partial targeted treatment. In our experience, all children that were considered completely cured at follow-up did not show recurrent arteriovenous shunting or new symptoms at later clinical or angiographic follow-up. Neurocognitive evaluation is an important followup criterion in neonates and infants, even without deficit, hemorrhage or seizure; it helps to assess treatment quality and success. Failure to obtain a normal maturation process can constitute a therapeutic failure if the optimal moment for intervention has been missed. This points to the difference between early irreversible damage and damage by persistent disorder and secondary destruction. Brain calcifications, which can be observed in both instances, indicate failure to obtain normal hemo- and hydrodynamic conditions for the maturing brain. When discussing CAVM or vascular diseases in children, one wonders whether it represents an artificial grouping. AVMs in children are primarily characterized by specific diagnostic and therapeutic challenges in which they develop. Some rare lesions are exclusively encountered in children, mainly in neonates and infants. The anatomic characteristics of neonatal and infant brain and the immaturity of its system flexibility create a specific group of symptoms and thera-

287

288

L. Batista, A. Ozanne, M. Barbosa, H. Alvarez, P. Lasjaunias

peutic objectives. This vulnerability means that the lesion rapidly becomes lethal or creates a disabling state, whereas a similar lesion in adult would produce fewer symptoms and damage.

8.1.1 Role of Imaging in the Pediatric Age Group Children have higher risk of complications related to femoral puncture than adults, due to the small caliber of the femoral artery and its vulnerability to vasospasm. Contrast restriction exists in this group (6cc/ kg). Due to small body surface and organic immaturity of neonates and infants, in cases of secondary cardiac overload due to the CAVM, fluid volume should be cautiously monitored. In our experience, the procedure should not exceed 2 h. Thus, catheter angiography is reserved in most of the cases: (1) if endovas-

cular treatment is foreseen in the same session, and (2) in the presence of persistent unclear diagnosis by other methods, i.e., clinical, computerized tomography (CT), and magnetic resonance imaging (MRI). Numerous useful morphological and functional information, such as feeders, location, mass effect, brain damage, and veno-dural patency, can nowadays be anticipated from noninvasive examinations. Ducreux [7] demonstrated that functional MRI in proliferative angiopathy provides data for a better understanding of symptoms (Fig. 8.1). Antenatal ultrasounds and/or MRI can demonstrate vein of Galen aneurysmal malformations (VGAMs) (Fig. 8.2), dural sinus malformations (DSMs) (Fig. 8.3), or exceptionally cerebral arteriovenous malformations (CAVMs). Presence of fetal cardiac failure or encephalomalacia are suggestive of brain pathology. In other situations, the antenatal diagnosis will guide the delivery of the child in the presence of a qualified pediatric unit (Fig. 8.4).

Fig. 8.1. Functional-MRI in proliferative angiopathy (Ducreux 2002). MR perfusion imaging in a vein of Galen aneurysmal malformation with proliferative activity: nidal area (1), at the level of the centra semiovale (2), and at the level of the central sulci (3). From left to right, T2-weighted (a), TTP (b), CBV (c) and CBF (d) parametric perfusion images. The color scale shows high values in red, mean values in green, and low values in blue. Note the high CBV and CBF values in the nidal zone (white arrow). Note increased perfusion in the left corona radiata (yellow arrow), and the bilaterally increased TTP values remote from the nidus in the white matter (white arrowheads). Truncation of the images in the right posterior region is due to artifact from the shunt valve

289

Arteriovenous Malformations: Diagnosis and Endovascular Treatment

a

Fig. 8.2. Fetal MRI showing a vein of Galen aneurysmal malformation

b

Fig. 8.3. Fetal MRI: diagnosis of dural sinus malformation (DSM)

Fig. 8.4a–c. Antenatal diagnosis of cortical arteriovenous fistula (CAVF) in a female baby with severe neonatal cardiac failure. Two CAVFs at angiography. Embolization of both CAVFs was performed at age 17 days with immediate clinical improvement. Normal cardiac function at 2 months

c

290

L. Batista, A. Ozanne, M. Barbosa, H. Alvarez, P. Lasjaunias

8.1.2 Arterial and Venous System in Children The arterial tree in neonates is likely to have an immature pericytic coverage, which suggests higher fragility and risk of rupture during vascular procedures (surgical and endovascular) in this age group. Gross anatomy does not differ from adults [8], with few exceptions: (a) the existence of a more competent and complete circle of Willis, (b) usual normal kinking of the cervical internal carotid arteries, and (c) widely patent anastomoses between the external and internal carotid arteries. Children have higher cardiac pulse than adults, and one often can see veins during arterial phase examination in TOF-sequences (2DTOF/3D-TOF), even in the absence of arteriovenous shunts. The normal postnatal maturation demands a vascular environment, especially venous, to be free of venous hypertension or congestion. The venous system in the pediatric age group presents numerous peculiarities. The hydric dural resorption system of neonates is still immature, which results in the medullary venous system being responsible for both the cerebral venous drainage and the intrinsic and extrinsic cerebral water dynamics. The granulations are not yet functional before 2 years of age; their maturation is delayed if the hemodynamics of the dural sinuses is abnormal. The maturation of the jugular bulb is simultaneous with that of the jugular foramen and also occurs after birth, allowing free flow through the sigmoid sinus towards the internal jugular vein. Simultaneously, the occipital and marginal sinuses regress [9]. At birth, the torcular drains the entire venous flow towards the posterior outlets, as the sylvian veins have not yet been “captured” by the parasellar venous plexus. This alternate pathway is usually present at 6 months. Some uneven ballooning of the transverse sinuses is observed in the roentgenograms of the fetuses from the second half of the fourth to the seventh month. After 20 weeks, the inner caliber of the transverse sinuses gradually becomes even. From birth to age 1 year, the inner diameter of the transverse sinus decreases somewhat. After age 1 year, the sinus has an adult appearance. In our experience, the inner diameters of the sigmoid sinuses remained relatively constant and small during fetal life, ranging in size from 1–2 mm. Up to the sixth fetal month, the course of the sigmoid sinuses differs from that of adults in that they follow a gentle convex curve medially in Towne’s view. After the 6th fetal month, the sigmoid sinuses follow a gentle convex curve laterally. At this stage, the course of the sinuses approaches that of adults.

The inner diameter of the jugular sinus increases by only 1 mm from the 3rd to the 7th fetal month, and is extremely small (1–2 mm on average). After birth, it rapidly increases, forming a bulb-like enlargement at 2 years of age. This is the formation of the (high) jugular bulb. During the period when the jugular sinus is small and poorly developed, the amount of stagnant venous blood drains from (mostly the convexity of) the cerebrum and cerebellum. The formation of the (high) jugular bulbs takes place after birth (clearly visible on angiograms, 2 years after birth) and is likely related to hemodynamic factors, resulting from a change from the fetal lying-down position to the postnatal erect posture [9]. In the presence of intracranial arteriovenous shunts, dysmaturation of basal sinus drainage may occur. Thus, embryonic dural sinus patterns can persist, accompanying the underlying disease as seen on venous phase magnetic resonance angiography (MRA) (Fig. 8.5). Under high sinusal pressure, hydric imbalance initially produces macrocrania, ventriculomegaly (if the cranial sutures are open), and later hydrocephalus with transependymal resorption. Tonsillar prolapse associated to severe degree of brain atrophy (melting syndrome) and pial venous congestion occur later, when sutures are already closed with occluded jugular bulbs. Hence, normal head circumference is misleading.

8.2 Classification of Cerebral Vascular Malformations Arteriovenous shunts represent an heterogeneous group of vascular lesions, traditionally classified according to morphologic criteria or technical accessibility [10–12]. Clinical experience shows that apparently identical lesions can produce different symptoms. Conversely, quite morphologically different vascular lesions can produce identical clinical manifestations. Nowadays, endovascular techniques allow distal approach to the territories and access to lesions located in the immediate vicinity of eloquent areas. Classification of vascular cerebral malformations aims to reflect their natural history with regard to degree of vascular adaptation (system response) [13] to the conditions created by the CAVM.

291

Arteriovenous Malformations: Diagnosis and Endovascular Treatment

a

b

c

d Fig. 8.5a–d. Neonatal cerebral drainage. a, b Schematic representation. At this age, venous drainage converges into superior and posterior sinuses. Posterior fossa drainage is via the latero-mesencephalic vein and the superior petrosal sinus to the cavernous sinus and caudally to the jugular bulb. There is no cortical venous drainage to the cavernous sinus yet (hatched). c, d Late phase of a carotid angiogram in a baby with vein of Galen aneurysmal malformation

8.2.1 Pial, Dural, or Choroidal Lesions According to the meningeal space from which arteriovenous (AV) lesions primarily develop, one can classify AV lesions into pial, dural, or choroidal (subarachnoid). Cerebral pial AVMs are located in the subpial space [15–17], and are supplied by pial cerebral arteries (Fig. 8.6). Pial AV shunts include cerebral AVM, cor-

tical AVF, proliferative angiopathy, and hemorrhagic angiopathy. Dural arteries arising from the external and internal (ICA siphon) carotid or vertebral arteries may contribute to the supply of dural lesions. In the pediatric age group, the middle meningeal artery is almost always involved, and often develops flow-related aneurysms. This group includes dural sinus malformations (DSM) and juvenile dural AV shunts.

292

L. Batista, A. Ozanne, M. Barbosa, H. Alvarez, P. Lasjaunias

Choroidal plexus AV shunts occur in the choroidal fissure; the vein of Galen aneurysmal malformation (VGAM) is the embryonic entity of this group. The choroidal branches arise from the basilar tip, posterior cerebral, anterior choroidal, and anterior cerebral (pericallosal or anterior communicating) arteries.

sidering all the harmful effects of high-flow lesions on the growing and maturing brain, it is difficult to believe that most CAVMs discovered in adults were present as such at birth [16]. The word “congenital” means during the time of gestation, based on the postulate that the fetus is fully developed at birth. However, construction of a vascular tree is not a static process, but rather results from complex biological factors and events starting in the embryo and continuing in the fetus, neonate, and infant. The vascular tree is maintained, repaired, and modified according to metabolic demands, resulting with time in a renewal of the entire structure [17], which is genetically programmed and controlled. Alterations in the required program or cellular logistics will result in a different construction, and the vascular tree may eventually be abnormal. An abnormal vascular architecture may therefore be due to a construction failure (embryonic, fetal, or perinatal) or a failure in the renewal process (postnatal). Thus, CAVMs can result from a congenital event, and albeit morphologically unexpressed at birth they will become detectable later. The impact of this congenital event is primarily cellular, somehow affecting the vascular modeling and remodeling chain [18]. For this quiescent dysfunction to be revealed as a morphological abnormality, secondary “revealing” triggers or “mutations” are needed. They are not yet identified, but are likely to include mechanical, hormonal, pharmaceutical, hemodynamic, thermal, radiation, viral, infective, and metabolic factors. The different nature and timing of the revealing trigger (or triggers) could result in the variety of AVMs that are known today [16] (Table 8.1). VGAMs (Fig. 8.2), and DSMs (Fig. 8.3) are often diagnosed in utero. The VGAM defect obviously occurs during the embryonic period and becomes manifest during fetal life. The DSM defect occurs during the early fetal period and also develops at the end of this period. All other CAVMs occur and develop during the perinatal period at the earliest, but most likely after infancy and during adulthood. In our experience, only 1% of nongalenic CAVFs were diagnosed in utero over the past 20 years (Fig. 8.4).

8.2.2 Timing of Genesis

8.2.3 Angioarchitecture and Morphology

CAVMs are considered to be congenital in nature; it is implicit that a “malformation” is present at birth, even though it may be discovered much later in life. However, there is no evidence that AV shunts diagnosed in adults are present as such in children. Con-

8.2.3.1 Angioarchitecture

a

b Fig. 8.6a,b. Pial malformation is placed in the subpial space. 1, venous drainage; 2, venous pouch; 3, venous reflux; 4, dural opening; 5, cortical reflux; 6, subpial reflux; 7, medullary reflux; 8, subcortical reflux; 9, secondary pial reflux; 10, dural sinus flow

Appropriate analysis of angioarchitecture allows understanding of past events and anticipation of

Arteriovenous Malformations: Diagnosis and Endovascular Treatment

Cerebral vasculature (capillary-venous junction)

Cellullar substrate = target Window of exposure = timing (system in activity, ...)

Mutation or primary trigger (humoral, immune, infectious, trauma ...) Remaining vasculature Dormant Defect (transmissible)

Reversible focal change or irreversible dormant focal change (dna repair systems, p53, ...) (failure to repair) Mutation or revealing trigger (humoral, immune, infectious, trauma ...) Arterio venous shunt Induced stresses (shear stresses, ...) Brain “AVM” nidus

Reversible or permanent regional changes (high flow angiopathy)

future ones; it helps to identify potential causes of symptoms, recognize underlying anatomical risk factors, and guide our therapeutic attitude. For instance, epilepsy or neurological deficit are more frequent in temporal pial PAVMs with venous ectasia and cortical or veno-sinusal high pressure, especially if a long subpial course of a cortical vein is present. Assessment of both arteries and veins is an important element at this point. Nidus versus fistula. A nidus consists of a group of small AV shunts within a vascular meshwork. The term fistula describes a direct opening of one or several arteries into an enlarged draining vein (Fig. 8.7). Arterial vascular changes in cerebral AV shunts. Flowrelated aneurysms are due to shear forces established on the arterial wall due to AV shunt; these are absent in children. If distal or intranidal false aneurysms (normally in older children) can be seen in the acute stage of a hemorrhagic episode, endovascular embolization may become a priority, as early recurrence is more frequent in children than in adults. Headache is often associated with arterial stenosis and, in this population, with a raised intracranial pressure. A moyamoya type of high flow angiopathic process [1]

Normal vasculature

Table 8.1. Cerebral vasculature (capillary-venous junction)

expresses a certain degree of angiogenesis, as confirmed by the presence of transdural supply several years after establishment of the diagnosis. Such a situation should be differentiated from proliferative angiopathy. Venous angiopathy. Abnormalities of the venous system are more frequent than arterial ones. It is likely that most symptoms before the age of 3 years are venous-related. Venous pouches are often noted in children, since thrombosis and high flow are characteristic features in this age group. Paradoxically, huge pouches in children can be diagnosed with almost no mass-related symptoms, illustrating the ability of infant’s head to enlarge. However, seizures are frequently associated with spontaneous (partial) thrombosis of the pouch. Increasing size of the pouches can be observed over time on serial CT/ MRI, subsequent to the development of restriction of the downstream outlets [13]. Venous “ischemia” is a generic denomination that includes various types of symptomatic venous congestion/restriction, from focal to melting-brain syndrome. The latter is specific to infants. It consists of rapid destruction of the brain, usually involving the subcortical white matter with ventriculomegaly. This phenomenon is associated with severe neurological manifestations and no

293

294

L. Batista, A. Ozanne, M. Barbosa, H. Alvarez, P. Lasjaunias

a a

b b Fig. 8.7a,b. a Arteriovenous fistula with enlarged draining vein. b Presence of nidus in an arteriovenous malformation

signs of increased intracranial pressure, although these are usually present before the syndrome takes place. It may be diffuse and symmetrical, or focal in pial CAVMs. These findings are never encountered in adults, illustrating the role played by subpial and medullary veins in the maintenance and development of the white matter in the pediatric age group. The melting brain syndrome can be observed in CAVMs and DSMs. In the latter, the acute onset of venous ischemia often leads to hemorrhagic venous infarction, rather than to subacute atrophic changes (Fig. 8.8).

Fig. 8.8a,b. Axial (a) and sagittal (b) T1-weighted images show focal melting brain syndrome in the posterior fossa due to the early venous ischemia produced by the cerebellar arteriovenous malformation. Note ventricular enlargement

Flow velocity: AV shunts can show very slow flow (as in DSM), mild flow (as in micro-AVMs), or very high flow (as in cases of VGAM and pial CAVF) (Fig. 8.4). 8.2.3.2 Size

The shunting zone can be variably sized, and is therefore named micro, macro, or giant. Thus, one has micro-AVM, micro-AVF, macro-AVM, and macroAVF according to angioarchitecture. Cerebral AVMs

Arteriovenous Malformations: Diagnosis and Endovascular Treatment

do not grow, although they can enlarge by angiogenesis or an angioectatic process. If an intrauterine defect occurs, the impact area, size, and severity will be related to the timing of the causative event in relation to migration and the transformation of the “stem” to “committed.” The earlier the causative event, the larger the area of impact and the higher the chances of multifocality. The later the event, the more focal the defect and the smaller the lesion. Growth of an AVM as such will not occur. A large nidus will not result from the growth of a small one, but express a defect shared by a large group of cells fired by the revealing trigger on a clone of cells carrying a dormant defect [16]. Most of so-called AVM growths are due to high flow angiopathy changes and angiogenesis in response to clot, ischemia, and operative procedures (embolization, surgery or radiotherapy), or to proliferative angiopathy. 8.2.3.3 Location

In contrast to some diseases such as moyamoya, CAVM can occur anywhere in the brain. Whereas fistula are always superficial and cortical, nidustype PAVMs can be superficial, cortico-subcortical, deep, paraventricular or mixed, supra- or infratentorial. The nidus itself does not entrap parenchymal or neural structures. On MRI sequences, CAVMs are typically cortico-ventricular and wedge-shaped. Deep diencephalic lesions can be isolated or included in a cerebro-facial metameric syndrome (CAMS). Dural sinus malformations (DSMs) and infantile types of dural AV shunt (IDAVS) occur at the level of the torcular-lateral sinus complex and will be discussed later. The adult type of dural shunts seen

in children (ADAVS) occurs in the cavernous sinus. VGAMs lie on the choroid fissure from the pineal gland area to the interventricular foramina, more or less seated on the midline. 8.2.3.4 Multiplicity

Cerebral AVMs can be single or multiple. In our experience, multifocality in children is twice that of adults (18% versus 9%). They are usually spread bilaterally or supra- and infratentorially. Mixed shunt characteristics rarely occur in multifocality (i.e., association of nidus and fistulous types). More often, one finds the same type of angioarchitecture in all sites in the same individual, i.e., multiple fistulae or multiple niduses. Such multiple CAVF are highly suggestive of hereditary hemorrhagic telangiectasia (HHT-1). 8.2.3.5 Complex Cranio-Facial Vascular Lesions

Complex cranio-facial vascular diseases are rare lesions that include two recently recognized forms of presentation, both carrying an underlying metameric linkage between the intracranial and facial vascular territories: (1) cerebrofacial arteriovenous metameric syndrome (CAMS) and (2) cerebrofacial venous metameric syndrome (CVMS). CAMS corresponds to Wyburn-Mason or Bonnet-DechaumeBlanc syndromes, whereas CVMS to Sturge-Weber syndrome (SWS) [19, 20]. According to the involved metamere (hypothalamo-nasal, prosencephaloorbito-maxillar, or rhombencephalo-mandibular), these lesions are named CAMS-1, CAMS-2, CAMS3 and CVMS-1, CVMS-2 or CVMS-3, respectively (Fig. 8.9).

Fig. 8.9. Cerebrofacial topographic correspondence in arteriovenous and venous metameric syndromes (CAMS and CVMS). (From [20], with permission)

295

296

L. Batista, A. Ozanne, M. Barbosa, H. Alvarez, P. Lasjaunias

8.3 Specificities per Type 8.3.1 Pial Lesions Pial cerebral AV shunts include CAVM, CAVF, proliferative angiopathy, and hemorrhagic angiopathy. The common denominator is a subpial location and its postnatal diagnosis. CAVM can be cortical, deep, or subependymal (paraventricular) in location, or buried in the white matter; however, different location does not correspond to pathological diversity. Cortico-ventricular CAVMs typically show a pyramidal shape on CT and/or MRI. The lesion itself can appear as a fistula, a nidus-type, or a mixed type. The CAVF type always lies superficial on the cortical surface, supra- or infratentorially. In this situation, the dilated arterial feeders (usually one to three) more commonly originate from pial arteries away from the midline (basilar artery or circle of Willis) and drain into very dilated veins, often associated with venous pouch. CAVMs drain into the pial venous system and interfere early with the local hydrovenous function. However, in some cases the rapid passage into the subarachnoid spaces probably prevents some of these CAVF from damaging the brain, as seen normally in similar spinal cord lesions. In symptomatic patients, nidus-type CAVMs produce functional symptoms, such as seizures (resulting from venous congestion or bleeding), hemorrhage, and progressive neurological deficit including

developmental delay related to early interference between AVM drainage and cerebral venous maturation (Fig. 8.10). Single CAVF and multifocal lesions in children are highly suggestive of HHT1, also called Rendu-OslerWeber syndrome (ROW). CAVMs of the central nervous system are present in approximately 8% of HHT1 patients [21]; Willinsky [5] reported that 28% of patients with multiple CAVMs had HHT1. HHT1 is an autosomal (9q) disorder of strong penetrance and variable expression; in adults, it is characterized by multiple mucocutaneous and visceral telangiectasias, which produce recurrent hemorrhagic complications. Epistaxis is common in adults with HHT1 but rare in children [22, 23] (Fig. 8.11). The disease involves failure of the remodeling process due to an abnormal binding protein (endoglin) to TGFβ1, which compromises locally the reconstitution of the venous capillary junction [24]. Merland [25] emphasized the role played by pulmonary fistulas in the neurological symptoms. They can produce CNS manifestations as the result of emboli (septic or not). Among these, formation of brain abscesses may be a prominent manifestation of this disease, and may rarely represent its initial presentation. When CNS manifestations are seen in patients with HHT1, the search for pulmonary AVF is mandatory, and fistulous disconnection should be performed. If CAVM is associated to pulmonary AVF, the latter should be treated first. Multiple CAVM can be expressed in a sporadic form, without familial history or manifestations of HHT1. Batista [26] reported a rare association of two

a

b Fig. 8.10a,b. Cortical arteriovenous fistula in a young infant with focal, yet significant, melting brain syndrome

Arteriovenous Malformations: Diagnosis and Endovascular Treatment

to recurrent hemorrhages within the first 6 months. These lesions are also placed in the CAVM group, yet we do not consider these as a true CAVM or a proliferative angiopathy, as their clinical and angioarchitectural characteristics are specific: small scattered “nidus-like” lesions in subcortical/white matter areas. The arterial feeders are normal in size, and the area drains into normal veins. Due to the type of nidus, which is intermingled with the normal brain, embolization or surgical treatment is not indicated. These lesions respond dramatically to radiotherapy, and we observed excellent results in such cases (Fig. 8.14). Imaging Findings

Fig. 8.11. Cortical arteriovenous fistula in a child with hereditary hemorrhagic telangiectasia 1 (HHT1)

pial high-flow fistulas with nonvascular intracranial malformations (lipoma, arachnoid cysts, and cortical dysplasia) (Fig. 8.12). In neonates with CAVF, high flow interfering with normal brain homeostasis can induce, in few weeks, focal melting syndrome secondary to venous ischemia and/or hemorrhage, justifying rapid treatment. In other cases, the long subpial course of cortical vein draining the fistula may produce epileptic foci or neurocognitive delay. Proliferative angiopathy represents a type of pial AV shunt characterized by angiogenetic and angioectatic activity. Angiographic studies show discrepancy between the apparent size of the “nidus-like” network of vessels (involving one complete lobe or hemisphere, uni- or bilaterally) and the draining veins, which are often normal or slightly enlarged. Secondary proximal arterial stenosis is part of this pathology. Transdural supply in various locations demonstrates the diffuse character of the angiogenetic activity of this disease, that is often confused with PAVM and named diffuse or holo-hemispheric nidus (Fig. 8.13). Proper recognition and classification is important, as it implies anticipation of normal brain tissue intermingled with the vascular spaces on MRI images [7]. Seizures are the most common clinical symptoms at presentation, although headache and progressive deficit are also possible. Single hemorrhage is an exception; however, recurrent ones are very frequent. Hemorrhagic angiopathy. Revealed by intracerebral hemorrhage, this lesion demonstrates high tendency

MRI sequences give important information. T1weighted sequences should be obtained at least on axial and sagittal planes, and allow detailed analysis of the nidus. The brain status (edema, atrophy, mass effect, encephalomalacia, and transependymal resorption) is much better appreciated on FLAIR and T2-weighted sequences. Calcification and hemorrhage require gradient-echo images. In symptomatic intranidal partial venous thrombosis, the hypersignal in T1- and T2-weighted sequences indicates slow flow, rather than hemorrhage. Mass effect exerted by venous pouches on the surrounding brain can result from partial thrombosis and extravascular reaction. On CT scan, one has to be cautious not to misinterpret intravascular blood (into giant venous pouch) as hemorrhage (parenchymal hematoma); contrastenhanced sections will confirm the intraluminal nature of blood. CT scan can also demonstrate thin parenchymal calcifications, indicating brain insult and chronic venous ischemia. In the presence of multiple pial lesions, HHT1 should be suspected, and lung screening by CT is recommended. In multiple AVMs, correlation of MRI with clinical history and angiographic findings helps to determine which among the various lesions is the one responsible for symptoms; thereafter, proper treatment can be initiated.

8.3.2 Dural Lesions Pediatric dural arteriovenous shunts (DAVSs) are rare, with only few reports dealing with these lesions in the literature [1, 27, 28]. In contrast to adults, dural arteriovenous fistulae (DAVF) in children are likely to be congenital [29], despite reported cases in literature as part of adult series [30–32].

297

298

L. Batista, A. Ozanne, M. Barbosa, H. Alvarez, P. Lasjaunias

a b

c

d

e f

g

Fig. 8.12a–g. Cortical arteriovenous fistula (CAVF). a Carotid injection shows an ipsilateral cortical arteriovenous fistula on the temporal cortex. b Vertebral angiogram, lateral view shows a pial arteriovenous fistula supplied by a short circumferential artery from the basilar tip and draining into an ectatic tegmental vein. c Angio-3D shows the exact fistulous point. d The “basket” created into the venous ectasia and occluding the AVF by two small coils. e Vertebral angiogram, lateral view: immediate control shows complete occlusion of the basilar fistula. f, g Angiographic control at left internal carotid artery and vertebral artery, lateral view: complete occlusion of both CAVFs

Arteriovenous Malformations: Diagnosis and Endovascular Treatment

a

b Fig. 8.13a,b. Proliferative angiopathy involving the entire left frontal lobe: MRI (a) and angiographic (b) features

a

b

c

Fig. 8.14a–c. Hemorrhagic angiopathy: before treatment (a), when it produced a cerebellar hematoma (b), and 1 year after stereotaxic radiotherapy (c)

299

300

L. Batista, A. Ozanne, M. Barbosa, H. Alvarez, P. Lasjaunias

In general, authors tend to apply to children the experience gained in the adult population, ignoring the specificities of the pediatric population such as cerebral and skull maturation, dural changes, venous vasculature, and CSF pathways. In neonates, clinical features include cardiac failure (usually mild) and coagulation disorders (consumption), whereas in infants there is increased intracranial pressure (with irritability, macrocrania, neurocognitive delay, and seizures). By analyzing the clinical history and anatomical features obtained with catheter angiography, CT, and MRI, three different types of DAVS can be described in this age group: (1) dural sinus malformation (DSM); (2) infantile or juvenile type DAVS (IDAVS); and (3) adult-type DAVS (ADAVS) [28]. DSM: the AV shunts are a secondary event to the dural sinus malformation. Most often, DSM are revealed clinically in the first months of life; 20% of cases are diagnosed in utero by ultrasounds. Two types can be distinguished: 1. DSM involving the torcular and adjacent sinuses, with giant pouches and mural AV shunts (Fig. 8.15). The malformation is usually seen as a giant dural sinus lake, with partial thrombosis already apparent. The dural sinus lake communicates with the other sinuses; normal cerebral veins open into the malformed sinus. There often are restricted outlets, due to dysmaturation of one jugular bulb. The degree of lateralization (lateral, sigmoid sinus, jugular bulb) or the midline location (torcular-superior sagittal sinus) and also the magnitude of the dural lake and thrombosis are predictive factors, with a more favorable evolution for DSM away from the torcular. a

Midline-located DSM, associated with multiple slow-flow AV shunts within the sinus wall, contribute to the venous congestion of the brain, even in absence of pial venous reflux. Spontaneous thrombosis of the outlets further compromises the cerebral venous drainage, leading to venous infarction and parenchymal hemorrhage (Fig. 8.16). Postnatal growth of these lesions has been seen, with appearance of cavernomas or association with facial or calvarial venous malformations. In the presence of patent venous outlets, the clinical evolution remains subacute with macrocrania and mental retardation, but cerebral damage can be expected. Treatment must preserve the lake, preventing a radical venous approach. The aim is to preserve appropriate venous drainage to the brain and to occlude the mural AV shunts, until the cavernous capture allows the exclusion of the malformation. In some instances, progressive separation of both circulations can be achieved to avoid retrograde venous congestion. In lateralized dispositions that spare the torcular, the capability of the normal side to provide brain drainage is the most important prognostic sign. A similarly favorable evolution is seen in distal superior sagittal sinus locations away from the torcular. 2. DSM of the jugular bulb with otherwise normal sinuses produce an apparent sigmoid sinus-jugular bulb “diaphragm” and are associated with a petromastoid-sigmoid sinus high-flow AVF, usually single-holed. Such disposition is very different from the normal variations seen in the region where the sigmoid sinus is absent [33]. The malformation corresponds to a dysmaturation of the jugular bulb, with occlusion of the distal sigmoid sinus and normal postb

Fig. 8.15a,b. Partially thrombosed dural sinus malformation (DSM) involving the torcular and adjacent sinuses. Catheter angiogram (a) and schematic (b)

Arteriovenous Malformations: Diagnosis and Endovascular Treatment

a

b

Fig. 8.16a–c. Superior sagittal sinus DSM. MRI and CT performed at 7 and 8 months of age, respectively, demonstrate the superior sagittal sinus DSM and the spontaneous thrombosis (c) of the sinuses with rapid venous infarction, hemorrhage, and ventricular rupture (b)

c

natal occlusion of the marginal sinus. Although the AV shunt is usually of the high-flow type (the cerebral venous drainage is preserved), symptoms are mild in most cases. Prognosis is excellent with occlusion of the AVF, eventually with thrombosis of the sigmoid sinus distal to the superior petrosal sinus opening. Infantile DAVS: IDAVS are lesions with high flow and low pressure, with multifocal dural AV shunts. The sinuses are much larger than normal, but no lakes exist. They remain patent for a long time. Several unusual high flow angiopathic changes are noted, in particular on dural arteries with arterial aneurysms (Fig. 8.17). During evolution the venous drainage is craniofugal, usually unilateral, without any contralateral dural sinus reflux despite the high flow. The evolution includes: (1) Persistent high flow, with

patent outlets: the sinus sump effect creates remote cortical pial AVSs. The latter are usually asymptomatic, and may regress after occlusion of several dural AVSs. (2) Uni- or bilateral jugular bulb stenosis and occlusion, with increased hydrodynamic manifestations and bilateral pial vein congestion and reflux. In this case, tonsillar prolapse and even syringohydromyelia may take place [34]. Clinical onset often occurs at preadolescence, often a few years after well-tolerated, minor craniofacial traumatic or bleeding events. Cranial nerve deficits are often the initial symptom. If the first symptom is neglected, macrocrania and mental retardation develop. Progressive neurological symptoms (involving both the cerebrum and posterior fossa) occur at a later stage, usually after 10 years of age. Rerouting of the venous drainage towards the cavernous sinus may

301

302

L. Batista, A. Ozanne, M. Barbosa, H. Alvarez, P. Lasjaunias

a

b Fig. 8.17a,b. Juvenile or infantile dural arteriovenous shunt with multifocal fistulas in a 12-year-old child. Typical MRI (a) and angiographic (b) features

also produce proptosis and facial vein enlargement. In cases with rerouting into cerebral pial veins, seizures, neurological deficits, mental retardation, and brain calcifications appear. Treatment by arterial embolization is always partial. Multifocality leads to recurrence, with new shunts opening near the previously embolized ones. The long term prognosis is poor with neurological deterioration in early adulthood. Treatment in these cases is usually disappointing, as the cerebral impairment is only partially reversible. Hemorrhagic events can occur at this final phase. Adult-type DAVS: Anatomically, almost all ADAVS are located at the cavernous venous plexus. Sigmoid ADAVS are very rare, and multifocality has not been described. Extra-sinusal ADAVS (duro-subdural and osteodural) have not been encountered. Clinically, ADAVS are encountered in very young children, and several cases of neonatal cavernous fistulas have been described [35–37]. Many of these lesions become spontaneously thrombosed. It is noteworthy that neonatal ones never drain in the pial veins. If treatment is contemplated, cases with ophthalmic venous drainage alone will benefit from manual compression at the medial cantus; other agents or venous approach to the parasellar region have been used. Imaging Findings

MR images confirm the type of lesion (DSM, IDAVS, or ADAVS) as well as the anatomical distribution and extent of the dural involvement. In DSM, the initial

examination is important to establish clinical prognosis, as shunts away from midline have a much better evolution than midline ones (with torcular involvement). Tonsillar prolapse can be assessed on sagittal views. Postcontrast T1-weighted images differentiate AV shunts from cystic posterior fossa malformations, such as the Dandy-Walker malformation or mega cisterna magna. In ADAVS, thickening of the cavernous sinus can be observed. Enlargement of the ophthalmic veins correlates well with conjunctival congestion. In rare cases of DSM, complex venous anomalies may exist, and MRI will demonstrate developmental venous anomalies (DVA) associated to cavernomas; the latter can produce hemorrhage which must be differentiated from hemorrhage from deep venous infarction [38].

8.3.3 Vein of Galen Aneurysmal Malformation VGAM is a nonhereditary disease with a 3:1 male predominance. The first description of a possible VGAM occurred in 1895 [39], in which a false VAGM was represented: it was in fact a cerebral AVM draining into a dilated vein of Galen. It is convenient to distinguish true and false VGAM (i.e., vein of Galen aneurysmal malformation from vein of Galen aneurysmal dilatation, VGAD) (Fig. 8.18), as they represent different diseases. Anatomically, VGAM represents an AVM of the median vein of prosencephalon, the embryonic precursor of the vein of Galen itself, as was first recognized by Charles Raybaud [39] (Fig. 8.19).

303

Arteriovenous Malformations: Diagnosis and Endovascular Treatment

a

b

Fig. 8.18a–c. Vein of Galen aneurysmal dilatation (false vein of Galen aneurysmal malformation). 10-year-old boy with untreated macrocrania. a, b vertebral and c late carotid angiograms. Note the deep venous reflux, testifying the presence of a matured vein of Galen confluent

c

Fig. 8.19. Vein of Galen aneurysmal malformation. Choroidal arterial feeders of the arteriovenous shunt. Dilatation of the vein of Galen. Venous reflux in longitudinal sinus and cortical veins. Persistence of the falcine and occipital sinuses

304

L. Batista, A. Ozanne, M. Barbosa, H. Alvarez, P. Lasjaunias

The arterial disposition (persistent limbic arch) (Fig. 8.20) and persistent venous embryonic routes, such as dorsal diencephalic vein (the so-called epsilon) (Fig. 8.21) and the falcotentorial sinus (Fig. 8.22), support the embryonic nature of this AV shunt [40]. The arterial supply in VGAM usually involves all the choroidal arteries from the posterior cerebral, anterior cerebral, anterior communicating, supraclinoid carotid, and tip of the basilar arteries (Fig. 8.19); subependymal arteries from the circle of Willis can also feed the lesion. The shunt itself lies into the subarachnoid space of the choroidal fissure. The nidus of the lesion is located on the midline, and receives bilateral and often symmetrical supply. When one feeder side is more prominent, the venous pouch, due

to jet force effect, will be shifted to the opposite side. According to angioarchitecture, two forms of VGAM exist, i.e., choroidal and mural. The choroidal type corresponds to a very primitive condition, with a contribution of several choroidal arteries and an interposed network before the opening into the large venous pouch. This condition is encountered in most neonates with poor clinical scores (Fig. 8.23). The mural type is frequently encountered in infants, who tend to have a better tolerance to the disease and better clinical scores (Fig. 8.24). Intermediate forms exist, but their importance is only from a technical/management point of view. So far, we have not seen a true VGAM associated with another type of CAVM.

Fig. 8.20. Persistent limbic arch between the posterior and anterior cerebral arteries in an embolized vein of Galen aneurysmal malformation

a

b Fig. 8.21a,b. Vein of Galen aneurysmal malformation. Lateral view and 3D angiography during venous phase show the typical epsilon shape draining the internal cerebral vein

Arteriovenous Malformations: Diagnosis and Endovascular Treatment

a

b Fig. 8.22a,b. Persistent embryonic patterns of dural sinuses in presence of vein of Galen aneurysmal malformation: occipital sinus (arrow, b) and marginal sinus. Absence of straight sinus associated with presence of falcine dural channel (arrowhead, b) draining the venous pouch towards the posterior third of the superior sagittal sinus

a

b Fig. 8.23a,b. Vein of Galen aneurysmal malformation, choroidal type. Vertebral angiogram, anteroposterior (a) and lateral (b) views

The venous drainage in both forms is towards the dilated median vein of the prosencephalon; no communication exists with the deep venous system of the brain. The thalamo-striate veins (diencephalic primitive veins) join the subtemporal ones, demonstrating a typical “epsilon” shape on lateral view angiograms (Fig. 8.21). Analysis of dural sinuses shows absence of the straight sinus, with a falcine (falcotentorial) dural channel draining the venous pouch towards the pos-

terior third of the superior sagittal sinus (Fig. 8.21). Occipital and marginal embryonic sinuses persist in neonates and young infants. Clinically, congestive cardiac failure is present during the neonatal period, and is mainly encountered in choroidal forms. After VGAM is suspected by clinical examination, a pretherapeutic evaluation should be obtained, including: (1) clinical history since birth (i.e., convulsion and hemorrhage do not occur in VGAM at neonatal age, unless brain damage

305

306

L. Batista, A. Ozanne, M. Barbosa, H. Alvarez, P. Lasjaunias

a

b Fig. 8.24a,b. Vein of Galen aneurysmal malformation, mural type. a,b Carotid angiograms

Infant

Necknate

has already taken place); (2) evaluation of renal and liver function; (3) transfontanellar ultrasounds for possible encephalomalacia; (4) echocardiogram to assess cardiac tolerance and associated anomalies (i.e., cardiac defects); (v) MRI to provide morphological information (the diagnosis of pial high-flow AVF in neonates would have completely different therapeutic consequences and approach) and evaluate the status of myelination; and (vi) electrocardiogram. Since the neurological prognosis is difficult in neonates, we have established a neonatal score based on nonneurological manifestations and gross neurological status (Tables 8.2, 3). A score of less than 8/21 results in a decision not to treat, as the prognosis is independent from what could be achieved on the VGAM itself, the damage being irreversible even if not yet visible on imaging. A score between 8 and 12/21 requires emergency endovascular intervention. A score greater than 12/21 leads to medical treatment until the child is aged 5 months, providing there is no failure to thrive; at 5 months, embolization is performed regardless of the symptoms. In our experience, angiography and treatment at 5 months has shown the best efficacy of embolization against the risk of cerebral maturation delay. Unnecessary earlier management does not improve the neurological status and carries a higher risk of failure and technical morbidity. Therefore, pediatric follow-up criteria include monthly evaluation of head circumference and weight, and an MRI examination at 3 months. Obviously, alteration in any of these parameters will prompt endovascular management. Decrease in head circumference indicates loss of brain substance and early suture fusion.

Child 5 years

Congestive cardiac failure Macrocrania hypercephalies

Multiorgan failure encephalomalacia

hydrodynamic disorders Dural venous thrombosis (sigmoid s., jugular bulb)

Fall off in head circumference

Dural venous tcongestion and supratentorial pial reflux bone hyperthrophybulb)

Infratentorial pial reflux and congestion

Tonsillar prolapse

Facial venosis collateral circulation epistatis

Optimal therapeutic window Hydromyelia or syringomyelia

Neurocognitive delay s/ependymal atrophy (pseudy ventriculomegaly) calcifications (chronic venous ischemia)

Conclusions neurological deficitis cerebro-meningeal haemorrhages

Epilepsy neurological deficitis

Table 8.2. Natural history of vein of Galen aneurysmal malformations

307

Arteriovenous Malformations: Diagnosis and Endovascular Treatment Table 8.3. Neonatal evaluation score Hepatic function 3 No hepatomegaly. Normal function 2 Hepatomegaly. Normal function 1 Moderate or transient hepatic insufficiency 0 Coagulation disorder. Elevated enzymes. Renal function 3 Normal 2 Transitory anuria 1 Instable diuresis under treatment 0 Anuria Respiratory function 5 Normal 4 Polypnea. Bottle finished 3 Polypnea. Bottle not finished 2 Assisted ventilation. Nl saturation-FI021 cm in diameter; - associated with solid components; - associated with unusual clinical signs (precocious puberty, Parinaud syndrome, hydrocephalus)

10.3 Extra-axial Tumors 10.3.1 Choroid Plexus Tumors Choroid plexuses are composed of an ependymal layer and an underlying mesenchymal stroma with fibroblasts, vascular elements, and arachnoidal cellular nests [196]. Choroid plexus neoplasms may originate from the ependymal layer or the mesenchymal stroma [196]. Neoplastic degeneration of the epithelium may produce benign (papillomas) or malignant tumors (carcinomas); uncommon mesenchymal masses include meningiomas, hemangiomas, and lipomas [196]. 10.3.1.1 Choroid Plexus Papilloma Epidemiology and Clinical Picture

Choroid plexus papillomas (CPPs) are rare tumors of neuroectodermal origin, accounting for 0.5%–3% of intracranial tumors in children according to various reports [124, 197–201]. In the great majority of cases, these tumors are discovered within the first two years of life [38, 197, 198, 200, 202]. In 12.5%–42% of cases, as suggested by its large size, the tumor is congenital [199, 201], i.e., the mass develops during intrauterine life and becomes clinically evident within the first

two months of life [203]. As many as 42% of tumors presenting in infants within 60 days of birth are CPPs [201]. CPP is the third most common neoplasm in children younger than 2 years, after astrocytoma and MB [199]. Male predilection is widely reported [38, 124, 197, 200, 201]. In females, a high incidence of CPP is found in the Aicardi syndrome [124, 196], an X-linked dominant entity characterized by infantile spasms, callosal hypo-agenesis, and chorioretinopathy [204]. The morphology of the corpus callosum should therefore be carefully evaluated in all girls with CPP. CPPs may be located wherever a choroid plexus is found within the ventricles [197, 201]; however, in children they typically are supratentorial, and involve the lateral ventricles (80% of cases) [196, 197, 201, 205, 206]. Generally, these neoplasms originate from the epithelium of the choroid glomus. Therefore, they are located in the ventricular atria [205]. Their size may be huge (up to 7 cm in diameter) [198]. Only very rarely are they located in the third ventricle (4% of pediatric cases) [197, 199, 201], whereas the fourth ventricle (16% of pediatric cases) [197, 201] and the cerebellopontine angle cisterns are typical adult locations [38, 124, 197, 199]. In the first year of life, plasticity of the incompletely ossified skull and unmyelinated nervous tissue allows congenital neoplasms, and particularly CPPs, to attain a considerable size before becoming manifest clinically [197, 199, 200, 207, 208]. Therefore, infants generally present with macrocrania that often is asymmetric [197]; mental delay, fever, vomiting, lethargy, irritability, sensorimotor deficit, seizures, widened cranial sutures, and papilledema may follow [197, 200, 206]. Intraventricular hemorrhage due to tumor bleeding is an exceptional presentation [206]. With the exception of such cases, CSF biochemistry usually is normal [197]. Biological Behavior and Neuropathology

Macroscopically, CPP is a dark pink or gray-reddish cauliflower-like mass [38, 124, 206], in some cases more or less markedly calcified, but generally friable [197, 198]. The mass is markedly vascular [197, 201] and, when supratentorial, its blood supply is provided by the anterior, posterolateral, and posteromedial choroid arteries [205], whose tumoral branches are constantly hypertrophied, tortuous, and elongated [199]. Small hemorrhagic foci commonly are found within the central portions of the mass [124]. The involved ventricle is locally expanded, but the adjacent nervous tissue always is spared [124]. Vasogenic edema involves the adjacent nervous tissue in up to one-third of cases [209]. When located in a lateral

Brain Tumors

ventricle, the tumor originates as a pedunculated mass, generally in the inferior portion of the trigone at the level of the glomus [38, 197, 205]. The microscopic appearance is similar to that of the normal choroid plexus [197, 199], showing papillae with a single layer of columnar or cuboidal epithelium and an underlying vascularized connectival stroma [38, 197, 198, 206]. Occasionally, CPPs may show local aggressive features, such as loss of papillary architecture, invasion of surrounding nervous tissue [202], cellular anaplasia with giant nuclei, and increased mitoses [38, 206]; these are the so-called anaplastic papillomas [124], which may recur and metastasize [206] and have been considered precursors of choroid plexus carcinomas [197, 210]. Tumor cells retain the property of CSF production, which is typical of the normal choroid plexus; however, experimental studies have demonstrated that CSF production by CPP is four- to fivefold higher than normal [199, 211]. Overproductive hydrocephalus, a condition widely described both in the neuropathological [38, 198, 206] and neuroradiological literature [124, 196, 197, 199–201, 207, 212, 213], is the result of this uncontrolled CSF production. Hydrocephalus typically recedes after surgical excision of the mass [124, 199, 207]. The large size of the tumor also may obstruct the adjacent CSF pathways [197], causing entrapment of a temporal horn [199, 205] or occlusion of the foramina of Monro [199, 201].

to the normal structure of choroid plexus, generally iso-to hypointense on T1-weighted images and iso-to hyperintense to gray matter on T2-weighted images (Fig. 10.84) [196, 197, 200, 214]. Calcification and/or hemorrhage may locally modify the signal behavior of the tumor [124, 197, 200]. Enlarged feeding arteries may sometimes be identified by MRA (Fig. 10.84). Intense and homogeneous CE is due to rich vascularity (Fig. 10.84, 10.85) [124]. The main differential diagnosis issue is to recognize the benign or malignant nature of the tumor; in the former case, the neoplasm has a homogeneous appearance and is entirely located within the involved ventricle, whereas the adjacent nervous tissue is preserved. Oppositely, carcinomas show an inhomogeneous appearance and an invasive behavior; the surrounding nervous tissue is infiltrated and a variable amount of perilesional edema and mass effect is present, whereas anaplastic CPPs may not show a similar behavior. In exceedingly rare cases [210], benign forms can also seed the CSF (Fig. 10.86), making differentiation from anaplastic forms difficult. Other neoplasms that should be ruled out include ependymomas, PNETs, astrocytomas and, less commonly, meningiomas, metastases, and colloid cysts [197, 201]. 10.3.1.2 Choroid Plexus Carcinoma

Imaging Studies

Baseline CT generally demonstrates a smooth or lobulated, well-marginated mass, sometimes resembling a bunch of grapes, which is homogeneously isodense (Fig. 10.84) [206] or hyperdense to the surrounding nervous tissue [124, 196, 197, 200, 201, 214]. This behavior is consistent with the vascular nature of the tumor and with its location within CSF spaces [205]. Hemorrhagic foci may be seen within the mass [124, 196, 201]. Calcifications are not common in pediatric CPP [197, 200, 201]; however, because the physiological calcification of the choroid plexuses generally appears in the second decade of life, choroid plexus calcification in a small child is highly suspect, particularly when hydrocephalus is present [197]. Mixed density or heterogeneous enhancement suggest anaplastic degeneration [197, 200], which may be accompanied by ependymal and periventricular white-matter infiltration and perilesional edema [124]. However, a certain amount of periventricular white-matter edema may be seen in up to one-third of cases of benign CPP [209]. On MRI (Fig. 10.84, 10.85), CPP is a smooth or lobulated intraventricular mass, sometimes very similar

Epidemiology and Clinical Picture

The epidemiological features of choroid plexus carcinomas (CPCs) have not been completely assessed yet [198], but they would account for 20% [200] to 40 % [124, 199] of all choroid plexus neoplasms. The incidence is greater below age 5 years [124]; however, unlike CPPs, CPCs are not typical congenital or neonatal tumors. The prevailing location is in the trigones of the lateral ventricles [200], and the incidence is higher in males [124]. These malignant (grade IV) neoplasms usually arise de novo, although they may represent malignant degeneration of a preexisting CPP (anaplastic papilloma) [197, 210]. Clinical presentation is with headaches, signs of raised intracranial pressure, and focal neurological signs. Biological Behavior and Neuropathology

Histopathologically, CPC differs from CPP in that the adjacent nervous tissue usually is invaded [197, 213]. Macroscopically, the tumor tends to lose the papillary cytoarchitecture typical of both the normal choroid plexus and CPP [38, 198]. The mass is inhomogeneous due to necrotic, cystic, and hemorrhagic changes, but

407

408

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama

a

b

c

d

e

Fig. 10.84a–e Choroid plexus papilloma in a 3-month-old boy. a Axial CT scan. b Axial T1-weighted image. c Axial T2-weighted image. d Gdenhanced sagittal T1-weighted image. e 3D TOF MR angiography, axial MIP. CT shows an isodense mass within the right ventricular atrium (a). The lesion is isointense with gray matter on T1-weighted images (b), slightly hyperintense on T2-weighted images (c), and shows marked enhancement (d). A rounded, unenhancing portion (arrowhead, a–d) may represent trapped CSF among the grape-like tumor strands. Notice that hydrocephalus cannot be explained by CSF pathway obstruction; instead, it results from CSF overproduction by the tumor. Also notice hypertrophy of the anterior choroidal artery feeding the tumor, already visible in the baseline images (arrows, c,d) and confirmed by MRA (arrows, e)

Brain Tumors

Fig. 10.85 Choroid plexus papilloma in a 4-month-old boy. Gdenhanced sagittal T1-weighted image. Large, markedly enhancing, multilobulated mass fills the third ventricle and plugs the aqueduct (arrow). Hydrocephalus can be explained by CSF pathway obstruction in this case

retains marked vascular features [197]. Marked perilesional edema is present, and the mass effect upon the surrounding structures usually is pronounced [197, 213]. Individual cells show variable shape and size, and markedly resemble adenocarcinomatous cells [196]. Nuclei are variably sized, and mitoses are numerous [38]. Cellularity is dense and organized in a pseudostratified columnar epithelium [197]. Foci of completely disorganized growth are common [198]. Infiltration, focal necrosis, loss of demarcation between stroma and parenchyma, and proliferation of vascular structures are found [197], although the tumor may rarely be avascular [205]. The propensity to seed the CSF spaces, producing ventricular or leptomeningeal metastases, is marked [38, 196, 200]. Unlike CPPs, CPCs are infrequently associated with hydrocephalus. When present, hydrocephalus is caused by CSF pathway obstruction due to either the tumor itself or ependymal and leptomeningeal metastases [197].

a b

c

Fig. 10.86a–c Histologically benign choroid plexus papilloma with CSF seeding in a 6-year-old boy with macrocrania. a Gd-enhanced axial T1weighted image. b Gd-enhanced sagittal T1-weighted image. c Axial FLAIR image. Mass involving the choroid glomus of the left atrium (a) and showing a papillary pattern, similar to the normal choroid plexus. Tetraventricular hydrocephalus probably was initially caused by CSF overproduction, but is now worsened by obstruction of the distal portion of the aqueduct by an unenhancing metastatic nodule (black arrow, b). Neoplastic septa are recognizable at the level of the anterior third ventricle (arrowheads, b). In the posterior fossa, another intraparenchymal location involves the left cerebellar hemisphere (arrows, c). Enhancement of the interpeduncular fossa also is likely neoplastic (white arrow, b).

409

410

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama

Imaging Studies

On CT (Fig. 10.87), the mass is inhomogeneously isoto hyperdense to the adjacent nervous structures [197, 200, 205], which are more or less extensively invaded. CE is variable, but often is marked and heterogeneous [196, 200, 205]. The mass shows focal hypodense areas due to necrosis [215] or old hemorrhage and, in some cases, hyperdense portions due to calcifications (Fig. 10.87) or recent bleeding. MRI (Fig. 10.87) shows a heterogeneous mass, whose solid portions generally are iso-to hypointense on both T1- and T2-weighted images; central necrotic areas will appear hypointense on T1-weighted images and hyperintense on T2-weighted images [197, 214]. CE usually is marked and heterogeneous (Fig. 10.87). Surrounding edema, infiltration of adjacent nervous

a

b

c

d

tissue and ependyma, and leptomeningeal spread are exquisitely depicted by MRI [196]. Entrapment of a horn of the involved ventricle is detected in up to 80% of cases [124]. Straightforward intraventricular CPCs usually are rather easily diagnosed based on the typical neuroradiological features. Conversely, when the mass is large and its intraventricular location is not easily detectable, other tumors enter the differential diagnosis scope. PNETs must be considered, because macroscopic features, such as hemorrhage and calcifications, are similar. Other neoplasms, such as ependymomas and GBMs, must also be included in the differential diagnosis. Owing to the common neuroepithelial nature, differentiation from ependymomas can be difficult also on microscopic examination [38].

Fig. 10.87a-d Choroid plexus carcinoma in a 10-month-old girl. a Axial CT scan. b Axial T1-weighted image. c Axial T2weighted image. d Gd-enhanced axial T1-weighted image. There is a huge mass in the left cerebral hemisphere that abridges the midline and abuts the contralateral foramen of Monro, causing dilatation of the contralateral ventricle. Notice that the shape of the lesion suggests it is located within, and markedly expands, the left lateral ventricle; also notice that only the frontal horn of the same ventricle is visible, reinforcing this opinion. The lesion is hyperdense on CT (a), isointense with gray matter both on T1-weighted (b) and T2-weighted images (c), and enhances markedly (d). There is marked perifocal edema. Calcifications (arrows, a), hemorrhage (open arrow, b), and neoformed vessels (arrowheads, b) can be recognized within the mass. Notice asymmetric macrocrania with splaying of the homolateral lambdoid suture (thick arrow, a)

Brain Tumors

10.3.2 Meningioma Epidemiology and Clinical Picture

Intracranial meningiomas are common tumors in adults; however, they are extremely rare in children, accounting for 1%–4% of all pediatric intracranial tumors [216], and fewer than 2% of all meningiomas [217]. The majority occurs in the second decade, whereas they distinctly are uncommon in children aged less than 4 years. However, a minority may be congenital. In the pediatric population, males are more frequently affected than females [216, 218, 219]. Clinically, afflicted children usually present with signs of increased intracranial pressure, focal neurological signs, and seizures [219, 220]. Hydrocephalus is common in cases of intraventricular meningiomas. Cranial nerve deficits may occur. Overall, symptoms mainly are related to tumor location. Pediatric meningiomas often occur within particular clinical settings, such as in patients with neurofibromatosis type 2 (NF2) [219, 221]. In this case, they may be multiple and associated with multiple schwannomas and ependymomas. The frequency of the association with NF2 varies from 19% to 41% according to various reports [219, 222]. Multiple meningiomas may also be found in patients without NF2, in which case the term “meningiomatosis” is applied; this is, however, rare in children (2.5% of cases) [219]. They may also occur as a complication of radiation therapy to the head [217]. However, this event is rare because of the very long latent period for the presentation of these tumors, which usually is 15 to 20 years [219], and is inversely related to the total radiation dose [217]. Biological Behavior and Neuropathology

Pathologically, pediatric meningiomas differ from those of adults in many ways, as follows:  Size Pediatric meningiomas usually are larger than adult ones: they often measure more than 5 cm in their largest dimension [222].  Structure A significant proportion of pediatric meningiomas is partially cystic [223].  Location Pediatric meningiomas tend to occur in somewhat atypical locations, such as the ventricular system, the optic chiasm, the cistern of the lamina terminalis (Fig. 10.88), the posterior fossa, or within the brain parenchyma without a dural attachment; however, “classical” locations along the convex-

ity and the falx cerebri are also possible [219, 224] (Figs. 10.89, 10.90). Infratentorial meningiomas are significantly frequent (19% of cases) [222]. Intraventricular meningiomas are observed more commonly in children (17%–24% of cases) [225, 226] than in adults (0.5%–4.5% of cases); they are thought to arise from meningothelial rests within the tela choroidea or the choroid plexus [219]. However, those located in the third ventricle are particularly rare [222, 226], and the recognition of their intraventricular location is crucial for a correct surgical approach.  Pathology It has long been believed that pediatric meningiomas are more aggressive and prone to malignant degeneration than their adult counterparts [218, 222]. However, recent evidence does not seem to support such theory. The incidence of malignant meningiomas is 2%–5%, a figure that approaches that found in adults [219, 220]. A possible reason is that meningeal sarcomas were incorrectly classified as malignant meningiomas in the past. The WHO classification comprises as many as 15 subtypes (Table 10.2); of these, the meningothelial type is the most frequent in children, followed by the fibroblastic and transitional types [224]. No significant differences occur compared to adults. Rhabdoid meningioma is a rare tumor containing patches or extensive sheets of rhabdoid cells. Most rhabdoid meningiomas have high proliferative indices, display aggressive clinical course, and correspond to WHO grade III [19]. Imaging Studies

MRI depicts an usually huge mass, sometimes separated from the adjacent brain by a rim composed of CSF and pial veins [219]. Intratumoral signal voids represent vessels or calcification. The tumor is usually isointense on T1-weighted images and shows a variable intensity on T2-weighted images (Figs. 10.88, 10.89). Enhancement generally is marked (Fig. 10.88) and better delineates the cystic components, which do not enhance (Fig. 10.89). A “dural tail sign” may be found; however, it commonly is absent in pediatric meningiomas, as is the lack of a dural attachment [219]. Surrounding edema is variable but may be marked (Fig. 10.90). CT scan easily identifies calcification (Fig. 10.90) and reactive hyperostosis in the adjacent bone. Meningiomas are isodense or slightly hyperdense to the adjacent parenchyma. We have observed a case of rhabdoid meningioma (Fig. 10.91). It was an extensively hemorrhagic mass abridging the tentorium, with a few necrotic components.

411

412

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama

b

a

d

c Fig. 10.88a–d Meningioma in a 4-year-old girl. a Sagittal T1-weighted image. b Axial CT scan. c Coronal T2-weighted image. d Gdenhanced axial T1-weighted image. Rounded mass located in the interhemispheric fissure between the genu of corpus callosum superiorly and the optic chiasm inferiorly. The anterior commissure is displaced posteriorly (arrows, a), indicating that the lesion is anterior to the third ventricle. The mass is slightly hyperdense on CT scan (b), hypointense to gray matter on T2-weighted images (c), and causes hydrocephalus by obstructing the foramina of Monro. Enhancement is marked and homogeneous (d)

Brain Tumors

a

b

Fig. 10.89a–c Meningioma. a Sagittal T2-weighted image. b Axial T1-weighted image. c Gd-enhanced coronal T1-weighted image. Huge lesion of the right rolandic-parietal convexity, surrounded by perifocal edema (a). The lesion shows mixed signal behavior on both T1-weighted (b) and T2-weighted images (a), and enhances inhomogeneously (c). Notice that the adjacent convolutions are compressed and displaced (open arrows, a), a typical feature of extra-axial tumors

a

c

b Fig. 10.90a,b Meningioma in an 8-year-old girl. a Axial CT scan. b Contrast-enhanced axial CT scan. The mass inserts onto the anterior part of the falx cerebri with contralateral midline displacement. Calcifications (arrows, a) and diffuse perifocal edema are seen. Enhancement is moderate (b)

413

414

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama

a

b

c

d Fig. 10.91a–d Rhabdoid meningioma in a 13-year-old girl. a Axial CT scan. b Axial T1-weighted image. c Axial T2-weighted image. d Gd-enhanced coronal T1-weighted image. CT shows a hyperdense mass (arrow, a) in close relationship with the right cerebellar hemisphere, which is markedly edematous (asterisk, a) with compression of the fourth ventricle (arrowhead, a). Some peripheral hypodense components are visible (open arrows, a). There is an incidental arachnoid cyst of the right temporal pole (AC, a). On MRI, the solid nodule is T1 isointense (arrow, b) and T2 hypointense (arrow, c) which, associated with the CT appearance, is strongly suggestive of acute hemorrhage. Peripheral necrosis (open arrows, b,c), cerebellar edema (asterisk, b,c), and fourth ventricular displacement (arrowheads, b,c) are well depicted, as is the arachnoid cyst (AC, b,c). Coronal sections clarify the location of the lesion, abridging the tentorium with extra-axial extension both caudally (arrowheads, d) and cranially (thick arrow, d). There is a very thin “dural tail” along the tentorium medially (thin arrows, d). Necrotic portions are persistently hypointense after Gd administration (asterisk, d)

10.3.3 Hemangiopericytoma

recently, whereas in the past they were considered an angioblastic variant of meningiomas [229].

Hemangiopericytoma (HP) is a rare tumor arising from vascular pericytes that predominates in adults and is found in only 10% of cases in children [227, 228]. Because of the histological origin, HP may arise wherever capillaries are found throughout the body. Intracranial HPs are rare tumors, representing less than 1% of CNS neoplasms in the general population [228]. They have gained an independent status only

Biological Behavior and Neuropathology

Although an infantile variant occurring within the first year of life and showing a more favorable outcome has been described, no significant histological differences can be demonstrated from the malignant HP that is found in older age groups [227]. In both cases, increased cellularity with pleomorphism, necrosis, hemorrhage, and frequent mitoses are

Brain Tumors

found. Therefore, distinction between the two forms is based on the clinical behavior rather than on histology [227]. Macroscopically, these huge (usually larger than 4 cm in their greatest dimension), sometimes multilobulated masses are well-demarcated, attached to the dura, and associated with profuse bleeding on resection [229]. They are aggressive lesions, with a high frequency of recurrence and extracranial metastasis. Imaging Studies

Neuroradiological examinations display a large, extra-axial mass that usually is located on the con-

vexity or parasagittally near the falx, similar to classical meningiomas (Fig. 10.92). The dural base of these tumors may be narrower than in meningiomas [229]. On CT, the lesion is slightly hyperdense, but usually lacks calcification. Contrary to meningiomas, erosion of adjacent skull is common, whereas hyperostosis is not [229]. On MRI, these well-demarcated masses show heterogeneous signal, partly due to numerous signal flow voids that represent prominent tumor vascularity (Fig. 10.92). The lesion generally is isointense on T1-weighted images and iso-to hyperintense on T2-weighted images, and shows marked CE [228].

a

b

c

d Fig. 10.92a–d Hemangiopericytoma in an adolescent. a Gd-enhanced sagittal T1-weighted image. b Coronal T1-weighted image. c Axial T2-weighted image. d Gd-enhanced axial T1-weighted image. Huge mass of the right frontal convexity, extending contralaterally underneath the falx (arrows, b). Surrounding edema is not pronounced (thick arrow, c). Note the “dural tails” (arrowheads, d). The corpus callosum is deformed and compressed (a,b). Multiple intralesional flow voids reflect the hypervascular nature of the mass

415

416

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama

10.3.4 Meningeal Sarcoma

10.3.5 Schwannomas

Meningeal sarcomas account for 0.7%–4.3% of intracranial tumors in childhood [230, 231]. These tumors are markedly aggressive, and can invade the adjacent brain, bone, sinuses, and subcutaneous tissues (Fig. 10.93). They have a poor outcome, with only 50% of affected children surviving 1 year after surgery [231]. These lesions may arise as discrete, poorly marginated masses that show a connection to the leptomeninges, or as a diffuse meningeal sarcomatosis, similar to secondary leptomeningeal carcinomatosis [231]. Poorly defined or fringed margins on the cerebral surface are suggestive of brain invasion, and are believed to be significantly more common in meningeal sarcomas than in benign meningiomas [221]. Cystic change within the mass is another sign that, while not uncommon in benign pediatric meningiomas, should nevertheless be regarded with suspicion for malignancy [221].

Intracranial schwannomas seldom occur in children, and almost exclusively involve the 8th cranial nerve; other cranial nerves, such as the trigeminal and oculomotor nerve, are involved only exceptionally. In children and young adults, vestibular schwannomas almost exclusively are found in, and are considered to be diagnostic of, NF2; in these patients there also is a high incidence of associated meningiomas and ependymomas. In NF2, schwannomas typically are bilateral, a feature that is considered to be diagnostic of this condition. Isolated tumors (Fig. 10.94) are exceptional, and have been found in children between 7 and 16 years of age [232]. The clinical presentation of these tumors is not different from that typical of adults, i.e., progressive sensorineural hearing loss and signs of a mass in the cerebellopontine angle [232]. The size of the

a

c

b

Fig. 10.93a–c Meningeal sarcoma. a Axial T1-weighted image. b Gd-enhanced axial T1-weighted image. c Gd-enhanced coronal T1-weighted image. Huge lateral mass in the posterior fossa, deforming and compressing the cerebellar structures and displacing the fourth ventricle contralaterally. The lesion is surrounded by a hypointense rim composed of dura and CSF (arrowheads, a), that indicates an extra-axial location. Note transverse sinus invasion (thick arrow, b) and extension into the subgaleal soft tissues (asterisk, c)

417

Brain Tumors

a

b

Fig. 10.94a–c Isolated schwannoma in a 10-year-old girl. a Axial T1-weighted image. b Axial T2-weighted image. c Gd-enhanced axial T1-weighted image. Huge extra-axial mass originating from the right internal acoustic canal (arrows, a–c). The mass shows necrotic changes (arrowheads, a–c), and enhances markedly (c). The lesion is hypointense on T1-weighted images (a) and grossly isointense with gray matter on T2-weighted image (b). The brainstem and the fourth ventricle are deformed and compressed. There is a thin CSF rim surrounding the mass (open arrows, a,b) that reveals its extra-axial location

c

mass may be variable, but children with vestibular schwannomas tend to have larger tumors than adults. Pathologically, schwannomas are solid or cystic, capsulated tumors that tend to displace, rather than infiltrate, the parent nerve. Cystic degeneration, calcification, and fatty tissue (xanthomatous variety) may be found within the tumor. A characteristic feature of pediatric schwannomas is their marked vascularity, contrary to the typical hypovascularity of their adult counterparts [232]. Therefore, there is greater risk for massive hemorrhage at surgery. Anaplastic degeneration is rare, whereas the rare malignant tumors generally are primitive and show predilection for children or young adults. In such cases, peripheral nerves are involved more frequently than nerve roots, and the tumor shows cystic degeneration and infiltrates the surrounding tissues. The differentiation from malignant or anaplastic neurofibromas may not be feasible. Vestibular nerve schwannomas typically erode the surrounding bone, resulting in an increased width of the inner acoustic meatus that may be an indirect telltale sign on CT. On MRI, schwannomas are

hypointense on T1-weighted images, hyperintense on T2-weighted images, and enhance moderately to markedly. Small tumors may be completely nested within the inner acoustic meatus or even be confined to the vestibule, whereas large tumors predominately grow in the cerebellopontine angle cisterns, displacing and compressing the brainstem; however, they will consistently extend into the inner acoustic meatus, producing a so-called “ice cream cone” appearance that excludes meningiomas or other extra-axial cerebellopontine angle lesions (Fig. 10.94) [233]. Recognition of small tumors completely nested within the inner acoustic meatus, as well as of postsurgical remnants, is obviously easier by MRI, and is greatly facilitated by contrast medium administration. The main differential diagnosis in children is intracranial capillary hemangiomas, which have a striking predilection for the inner acoustic meatus and cerebellopontine angle cistern. However, these lesions are easily excluded, because they consistently are associated with a homolateral facial hemangioma [234].

418

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama

10.3.6 Dysontogenetic Masses Epidermoid and dermoid cysts are rare, slowly growing congenital lesions, thought to arise from inclusion of epithelial elements within the neural groove at the time of neural tube closure as a result of incomplete disjunction of the neuroectoderm from the cutaneous ectoderm [88, 124, 235]. Epidermoid cysts (ECs) arise from the ectoderm-derived epidermis only, whereas dermoid cysts (DCs) also contain dermal appendages, such as hair, sebaceous, and sweat glands [124]. Dermoid Cysts

In our experience, intracranial DCs have been more common than ECs in the pediatric population [88, 124]. The most common intracranial location is midline infratentorial (i.e., adjacent to the vermis or fourth ventricle) [124, 236] (Fig. 10.95). Intracranial supratentorial locations also are possible. Other common locations include the bregmatic fontanel and the orbit, especially at level of the inner and outer orbital canthi. Clinical symptoms occur earlier than with ECs, perhaps explaining the greater frequency of DCs that are discovered during childhood as opposed to ECs. Symptoms may result from obstruction of CSF pathways, chemical meningitis related to rupture of the cyst into the ventricles or subarachnoid space, and abscess formation due to ascending infection along an associated dermal sinus [124, 237]. DCs are well-demarcated, round or oval, unilocular cystic masses, lined by stratified squamous epithelium mounted on collagenous connective tissue that, unlike ECs, contains dermal appendages, such as hair follicles, sebaceous, and sweat glands [88]. The cyst contains thick, yellowish material resulting from an admixture of sebaceous secretion and desquamated epithelium [88]. These lesions show a variable neuroimaging behavior depending on the relative proportion of the various intracystic components. On CT scan, they may be isodense with CSF (Fig. 10.95), fat (Fig. 10.96), or brain (Fig. 10.97). On MRI, the signal is highly also variable on both T1- and T2-weighted images (Fig. 10.95−10.97). T1-weighted images may show hyperintense components related to the presence of fat (Fig. 10.96), and individual lesions may be extensively bright (Fig. 10.97); however, the fact that dermoids are invariably hyperintense on T1-weighted images is a myth. Following gadolinium administration, there usually is no enhance-

ment. In some cases, portions of the cyst wall may enhance slightly. Enhancement also may involve the adjacent leptomeninges due to reactive phenomena. Abscessed dermoids show intense, ring-like enhancement (Fig. 10.98). Epidermoid Cysts

Intracranial ECs are less common than DCs in children, accounting for 0.5%–1% of all intracranial masses. Intracranial ECs usually involve the subarachnoid spaces, thus being confined to the extra-axial space; however, intra-axial locations are possible [237]. The most common location is the cerebellopontine angle, followed by the pineal region, supra- and parasellar regions, and middle cranial fossa [124]. ECs also may involve the cisterna magna, tela choroidea (usually in the temporal horn of the lateral ventricle) [237], cerebral hemispheres, and brainstem [88, 235]. Although ECs are congenital lesions, symptoms rarely occur before the third or fourth decade of life [237]. Clinical presentation depends on location. Involvement of the facial nerve and, subsequently, of the acoustic nerve with unilateral hearing loss is typical in cases of cerebellopontine angle ECs [88, 124]. Hydrocephalus may ensue in cases of suprasellar and pineal locations. Chemical meningitis secondary to leakage of tumor contents into the subarachnoid space is more common in cases of middle cranial fossa locations [124]. ECs are encapsulated lesions (the mother-of-pearl sheen capsule explains the term “pearly tumors” that sometimes is used for these cysts), filled with soft, creamy material [88]. The internal layer is composed of stratified squamous epithelium mounted on collagenous connective tissue [88]. They usually are multilocular [202]. Progressive cyst enlargement is related to epithelial desquamation with accumulation of keratin, cholesterol, and other debris within the cavity [237]. On CT scan, ECs appear as irregular, lobulated masses that are isodense with CSF and lack CE. Focal calcifications along the cyst wall have been described, but are not the rule [237]. Small ECs embedded in the subarachnoid space may remain undetected [202]. On MRI, ECs usually are isointense to CSF on both T1-weighted and T2-weighted images, and most commonly lack appreciable CE [124]. However, since signal intensity may vary significantly depending on the cyst contents, three different subtypes have been described [238]: (1) The majority of ECs (80% of cases) are T1 hypointense and T2 hyperintense due to high liquid content and low lipids and keratin; (2) about

419

Brain Tumors

a

b

c

e

d Fig. 10.95a–e Dermoid in a 13-year-old boy. a Axial CT scan. b Sagittal T1-weighted image. c Sagittal T2-weighted image. d Axial FLAIR image. e Axial diffusion-weighted image. CT shows an apparent enlargement of the vallecula with questionable hypodensity with respect to isodense CSF. Notice that the margins are irregular (arrowheads, a). MRI shows a irregularly marginated lesion that compresses the inferior aspect of the vermis (thick arrow, b,c). Notice that signal intensity is slightly higher than that of CSF on both T1-weighted (b) and T2-weighted images (c). There is a bright spot on T1-weighted images along the anterior aspect of the lesion (thin arrow, b) that is consistent with a fat deposit. On FLAIR images there is clear-cut hyperintensity with respect to CSF (arrows, d). Diffusion-weighted images (e) show restricted proton diffusion within the lesion, consistent with a dysontogenetic mass. (e, Courtesy of Dr. L. Manfrè, Catania, Italy)

420

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama

a

b

c

d Fig. 10.96a–d Dermoid in a 7-year-old boy. a Axial CT scan. b Axial T2-weighted image. c Axial T1-weighted image. d Axial fatsaturated T1-weighted image. Hypodense lesion in the right frontotemporal region shows a more markedly hypodense component posteriorly, resembling fat (arrow, a). T2-weighted images reveal a homogeneously hyperintense mass that displaces the middle cerebral artery (arrowheads, b). T1-weighted images reveal a small, hyperintense component in the posterior portion of the mass (arrow, c), corresponding to the markedly hypodense spot on CT. This hyperintense signal is suppressed on fat-saturated sequences, confirming its adipose nature (d)

20% are T1 hyperintense and T2 isointense due to elevated triglycerides and fatty acids and absent cholesterol; and (3) a small minority are T1 hyperintense and T2 hypointense due to elevated iron content. The main differential diagnosis of EC is with arachnoid cysts. Because the vast majority of ECs is isodense with CSF on CT scan and isointense to CSF on both T1- and T2-weighted MR images, this differentiation has traditionally been regarded as difficult. On CT scan, the differentiation has traditionally been based on the typically lobulated margins of ECs, becoming more evident after intrathecal administration of iodized contrast material [124], as opposed to

the smooth external surface of arachnoid cysts [237]. MRI allows an easier differentiation between EC and arachnoid cysts if a combination of MRI techniques is employed, as follows: (1) on proton-density-weighted images, ECs are almost always slightly hyperintense to CSF, as opposed to the isointensity shown by arachnoid cysts [237]; (2) on FLAIR images, the necrotic and keratinaceous-proteinaceous content of ECs gives high signal, as opposed to the CSF-like low signal given by arachnoid cysts; (3) on DWI, arachnoid cysts show unrestricted diffusion similar to CSF, whereas ECs behave like normal brain tissue [239]; and (4) on magnetization transfer imaging, ECs show significant

421

Brain Tumors

a

b

c

d Fig. 10.97a–d Dermoid in an 8-year-old girl. a Axial CT scan. b Axial T1-weighted image. c Axial T2-weighted image. d 3D TOF MR angiography, coronal MIP. CT shows an isodense lesion located anterolaterally to the medulla oblongata (arrows, a). The mass is spontaneously hyperintense on T1-weighted images (b) and isointense with gray matter on T2-weighted images (c). A hypointense spot within the lesion in T2-weighted images (arrowhead, c) corresponds to a stretched, encased right vertebral artery (arrowheads, d). Notice that the mass is faintly visible in the MIP image due to its short T1 (thick arrows, d)

transfer of magnetization from the solid matrix of the tumor to adjacent free water, whereas arachnoid cysts show no magnetization transfer [240]. The rare ECs showing hyperintensity on T1weighted images may be difficult to differentiate from dermoids or lipomas. However, the application of fat saturation pulses suppresses the high signal from both dermoids and lipomas, whereas the signal from ECs will remain unaffected [124]. Lipomas

In addition to those associated with dysgenesis of the corpus callosum (see Chap. 3), lipomas may essentially occur in two main sites: the quadrigeminal plate (Fig. 10.99) and the cerebellopontine angle cistern (Fig. 10.100). Although these lesions are more correctly referred to as malformations, it should be pointed out that their size may increase in parallel to the physiological increase of adipose tissue as the child grows. Therefore, lipomas located in criti-

cal areas such as the quadrigeminal plate or foramen magnum may generate symptoms due to obstructive hydrocephalus or cervicomedullary junction compression, respectively.

10.3.7 Chordoma Epidemiology and Clinical Picture

Chordomas are rare tumors, accounting for less than 1% of all intracranial tumors. They are slowly growing, locally infiltrating tumors, arising from intraosseous remnants of the notochord, a primitive cell line around which the base of the skull and the vertebral column develop [241] (Fig. 10.101). The nuclei pulposi of the intervertebral disks are the only normal notochordal elements that remain after birth. However, ectopic notochordal remnants may persist along the neural axis, and especially at its top and

422

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama

a

b Fig. 10.98a,b Dermoid in a 3-year-old boy. a Contrast enhanced axial CT scan. b Gd-enhanced sagittal T1-weighted image. Retrovermian dermoid (arrowhead, a,b) associated with a large abscess involving the vermis and the left cerebellar hemisphere (asterisk, a,b). There is compression of the fourth ventricle with supratentorial hydrocephalus. A dermal sinus involving the subgaleal tissues also is detected (arrow, b)

a

b Fig. 10.99a,b Lipoma in a 2-year-old boy. a Axial T1-weighted image. b Sagittal T1-weighted image. Small lipoma involving the left inferior quadrigeminal tubercle (arrow, a) and superior vermian sulci (arrowheads, a,b). There is concurrent downward ectopia of the cerebellar tonsils (T, b), consistent with Chiari I malformation, causing hydrocephalus

bottom levels; chordomas arise from these ectopias. There is no known gender predilection. In children, chordomas are very unusual; only exceptionally have these tumors been described in children aged less than 5 years, and only a single congenital tumor has been reported [242].

In the general population, intracranial locations account for 35% of all chordomas, and usually involve the clivus (Fig. 10.102) in the region of the sphenooccipital synchondrosis or, more rarely, the sellar or parasellar regions; the remainder arise in the spine, mostly in the sacrococcygeal region and rarely from a

423

Brain Tumors

a

b Fig. 10.100a,b Lipoma of the right cerebellopontine angle cistern in an 8-year-old boy. a Axial T1-weighted image. b Axial T2weighted image. There is a hyperintense lesion in the right cerebellopontine angle cistern (thick arrows, a,b). T2-weighted image clearly shows that encasement of the seventh and eighth cranial nerves (thin arrows, b). Notice chemical shift artifact (arrowheads, a,b) along the fat-brain interface

to brainstem compression; increased intracranial pressure and extension of tumor through the foramen magnum are other prominent features [241, 242]. Biological Behavior and Neuropathology

Fig. 10.101 Midsagittal section depicting the course of the notochord. (Modified from Braun IF, Nadel L. The central skull base. In: Som PL, Bergeron (eds) Head and Neck imaging, 2nd edn. St. Louis: Mosby, 1991)

vertebral body, especially in the upper cervical spine [243]. However, the sphenooccipital region is the predominant location in children [241]. Exceptionally, clival chordomas may arise as totally intradural lesions, i.e., without surgical or radiographic evidence of bone erosion [244, 245]; these intradural lesions are believed to originate from the so-called “ecchondrosis physaliphora”, intradural deposits of notochordal tissue that are found anterior to the pons in about 2% of all autopsy specimens. Children with chordomas tend to present with long tract signs and lower cranial nerve involvement due

Albeit histologically benign, chordomas typically demonstrate a potentially malignant course; they are locally invasive, destroy bone, and infiltrate the soft tissues, with a 10%–20% incidence of metastases [242]. The prevalence of atypical histological findings with corresponding aggressive behavior is greater in patients younger than 5 years, as is the incidence of metastases in the same age group [241]. Chordomas are prominently calcified and often cystic, resulting in a gelatinous semiliquid appearance on gross examination. Histologically, two distinct variants have been identified. The “typical” chordoma is composed of strands of so-called physaliphorous cells embedded in a gelatinous matrix that may be more or less represented, resulting in more compact or looser architecture. The “chondroid” chordoma has interspersed cartilaginous foci, whose amount may vary from minimal to predominately cartilaginous masses [243]. Chondroid chordomas probably represent a majority of chordomas in older children, and are reported to have a significantly better prognosis [243]. Imaging Studies

Both CT and MRI (Figs. 10.102, 10.103) should be employed in the diagnostic workup of chordomas. CT provides better demonstration of bone erosion and tumor calcification, whereas MRI is superior in delineating the full extent of the lesion [243]. The tumor

424

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama

a

b

c

d Fig. 10.102a–d Chondroid chordoma in a 13-year-old girl. a Sagittal T1-weighted image. b Axial T1-weighted image. c Axial CT scan. d Contrast enhanced axial CT scan. Huge mass arising from the spheno-occipital synchondrosis and bulging in the rhinopharynx anteriorly (asterisk, a,b). Posteriorly, the mass bulges into the anterior medullary cistern, causing deformation and compression of the medulla and pons (arrow, a). Scattered calcifications are recognizable in the extracranial portion of the mass (arrowheads, c). Enhancement is inhomogeneous (d)

is midline in most cases, and grows by infiltration of the surrounding bone, spheno-ethmoidal sinuses, and anterior nasopharynx (Fig. 10.102). The brainstem usually is displaced posteriorly (Fig. 10.102). MRI also is useful in demonstrating transdural transgression of tumor, a finding of surgical relevance [243]. Adjacent nervous structures may be invaded along the planum sphenoidale, the sella turcica, and the clivus. There also is significant displacement or encasement of the major vasculature, and especially of either the carotid or vertebrobasilar arteries. Systemic or CSF drop metastases

may be found in some cases, especially with transdural extension of the primary tumor [243]. On CT, clival chordomas are more or less heterogeneous masses, often containing calcified structures (Fig. 10.102) which may be ascribed to bony sequestration or calcified deposits of recent formation. Typical and chondroid chordomas have been reported to display a different behavior on MRI. Typical tumors are isointense on T1-weighted images in 75% of cases, and hypointense in the remainder [243]. The tumor is invariably hyperintense on T2-weighted

425

Brain Tumors

b

a Fig. 10.103a,b Chordoma in a 13-year-old boy. a Sagittal T1-weighted image. b Contrast enhanced axial CT scan. Huge lesion arising from the basiocciput (arrow, a) and extending both anteriorly to the rhinopharynx (thick arrows, a) and posterolaterally, invading the posterior fossa and foramen magnum to extend to the suboccipital soft tissues (asterisk, a,b). Notice erosion of the odontoid process of C2 and posterior arch of C1 (b). The lesion also causes deformation and compression of the cervico-medullary junction

images, although hypointense septations may separate lobulated areas of higher signal intensity [243]. In some cases, the signal behavior may be heterogeneous due to structural anomalies and calcifications. CE is variable. Chondroid chordomas have shorter T1 and T2 relaxation times, perhaps because the gelatinous matrix is in part replaced by cartilaginous foci [243].

10.3.8 Capillary Hemangioma Capillary hemangioma (CH) is the most common tumor of infancy, occurring in up to 12% of infants of Western descent in their first year of life [246]. It usually involves the head and neck and prevails in females, with a 3:1 to 5:1 preponderance [247]. It usually is located in the skin, scalp, and orbit, where it involves the palpebrae with frequent extension to the rectus muscles [248]. Fewer than one-third of CHs are present at birth [247], whereas most appear by the first month of life [249]. They typically demonstrate rapid growth up until 6–8 months, followed by a phase of plateau between 8 and 12 months [247]. High blood flow and pressure within the mass lead to vascular dilatation and perivascular collagenization [246], which accounts for their well-known tendency

to spontaneous involution [246–248, 250], beginning by 12 months and continuing until 5–10 years of age [247]. CHs may be isolated or associated with other abnormalities within the setting of a vascular phakomatosis known as PHACES syndrome, i.e., the non-random association of: (1) extracranial CH; (2) posterior fossa malformations including the DandyWalker malformation, cerebellar hemispheric hypoplasia, and cerebellar cortical dysgenesis; (3) arterial anomalies, such as the persistent trigeminal artery and hypo-aplasia of the internal carotid, external carotid, and vertebral arteries; (4) coarctation of the aorta; (5) congenital ocular abnormalities; and (6) sternal defects [251, 252] (see Chap. 17). The intracranial compartment is a distinctly uncommon location for CHs [234]. Intracranial CHs are found exclusively in patients who also harbor cutaneous lesions; they may be nodular or diffuse, and preferentially lie homolateral to the cutaneous CH. Their natural history parallels that of their extracranial counterparts, i.e., they can either regress spontaneously along with the involution of the extracranial CH, or remain unchanged if the extracranial CH does not regress [234]. Intracranial CHs may be both extra-axial and intra-axial in location. Extra-axial CHs typically involve the inner acoustic meatus and cerebello-

426

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama

a

d

e b

c

f

Brain Tumors

pontine angle cistern (Figs. 10.104, 10.105), and may mimic acoustic schwannomas; however, the differentiation is easy due to the coexistent extracranial CH. Other meningeal-based lesions may be found in the unco-hippocampal region (Fig. 10.105), or exceptionally appear as a diffuse leptomeningeal enhancement [234]. Intra-axial CHs almost exclusively are confined to the hypothalamus, where they appear as an enhancing thickening of infundibulum, tuber cinereum, and median eminence [234] (Fig. 10.104). On MRI, intracranial CHs are well-circumscribed lesions which are isointense on T1-weighted images and hyperintense on T2-weighted images, reflecting their content in unclotted blood [246]. Typically, CHs enhance markedly with contrast material administration both on CT and MRI, due to their highly vascular structure. MRA may depict a conglomerate of fine vessels, similar to the typical “blush” that is seen with catheter angiography [234] (Fig. 10.105). One important clinical consideration is warranted: as CHs are known to regress spontaneously or with systemic steroid and interferon therapy, surgery should be avoided unless (1) the lesion impairs primary functions (such as huge orbital or airway CH); (2) follow-up studies demonstrate lesion growth; or (3) the clinical picture deteriorates. Finally, it could be questioned whether all patients with CH of the head and neck should undergo screening MR studies of the brain with contrast material administration regardless of their neurological status in order to detect silent intracranial lesions. In fact, the prevalence of isolated extracranial CHs is very high compared to that of “syndromic” CHs. We believe that all infants with large, prevailingly unilateral, plaque-like or bulky CH of the head and neck should be screened for associated PHACES manifestations. This screening should include contrast-enhanced brain MRI, MRA of the epiaortic, cervical, and intracranial arteries, echocardiography, ophthalmologic examination, and arterial pressure measurements in both upper and lower limbs [252].

10.3.9 Extra-axial Locations of Typically Intra-axial Tumors Although rarely, typical intra-axial tumors may arise in the extra-axial spaces. This occurs particularly with posterior fossa tumors, such as medulloblastomas (Fig. 10.7), ependymomas, and pilocytic astrocytomas (Fig. 10.106). In these instances, it usually is very difficult to advance a diagnosis preoperatively due to the rarity of these lesions.

10.3.10 Metastases The vast majority of intracranial metastases are due to leptomeningeal spread of primary CNS tumors; highgrade malignancies, such as MBs, PNETs, and malignant astrocytomas, most commonly are involved. However, low-grade astrocytomas, particularly of the pilomyxoid type, can also show leptomeningeal spread at presentation [74]. The appearance of these lesions on contrast-enhanced CT and MRI may vary from discrete nodular masses to a diffuse leptomeningeal carcinomatosis that has also been defined as “neoplastic leptomeningitis”. Parenchymal metastases are exceedingly rare in the pediatric population. Brain tumors are involved only exceptionally; when two large, independent masses are detected at presentation the question of multicentricity should be raised [56]. Cerebral metastases from systemic, nonleukemic-lymphomatous tumors during childhood are rare (4.5% of cases) and carry a dismal prognosis. The primary disease is more often neuroblastoma, rhabdomyosarcoma, Ewing sarcoma, Wilms’ tumor, or osteogenic sarcoma [253], and the brain is secondarily involved by hematogenous spread. CT and MRI findings are not specific; variably sized, often multiple enhancing masses with extensive surrounding edema usually are found. Knowledge of the primary tumor is essential for a correct diagnosis.

Fig. 10.104a–f Capillary hemangioma. a Photograph of the girl at age 11 months. b Gd-enhanced sagittal and c Gd-enhanced axial T1-weighted images at age 11 months. d Photograph of the girl at age 3 years. e Gd-enhanced sagittal and f Gd-enhanced axial T1-weighted images at age 3 years. There is a capillary hemangioma of the left upper eyelid (a). MRI shows two additional intracranial enhancing lesions at the level of the hypothalamus (arrowheads, b) and left internal acoustic meatus (arrow,c). Notice how the hypothalamic lesion recalls the normal anatomy of the hypothalamus from ventral (median eminence) to dorsal (tuber cinereum). There is incidental downward ectopia of the cerebellar tonsils (T, b,e) consistent with Chiari I malformation. Two years later and without any treatment, there is a clear-cut regression of the external hemangioma (d), paralleled by a marked reduction of both intracranial locations (arrowheads, e and arrow, f). (Permission to publish the photographs of the child was obtained from the parents.)

427

428

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama

Fig. 10.105a–e Capillary hemangioma. a Photograph of the girl at age 1 month. b Gd-enhanced axial T1-weighted image and c Digital angiography, anteroposterior projection at age 1 month. d Photograph of the girl at age 6 months. e Gd-enhanced axial T1-weighted image at age 6 months. This child has a large facial hemangioma that involves the right side of the face and the chin. There also is an enhancing, meningeal-based lesion of the right temporal uncus bulging into the opto-chiasmatic cistern (arrow, b). Digital angiography demonstrates typical hypervascularity of a capillary hemangioma fed by the lenticulo-striate arteries (arrows, c). Five months later and without any treatment, there is a marked regression of both the facial hemangioma and the intracranial location (arrow, e). (Reproduced from Tortori-Donati et al. (ref. 231), with permission from Springer Verlag; permission to publish the photographs of the child was obtained from the parents.)

a

b

c

d

e

429

Brain Tumors

a

b

c d Fig. 10.106a–d Extra-axial pilocytic astrocytoma in a 3-year-old boy. a Axial T1-weighted image. b Axial T2-weighted image. c Gd-enhanced axial T1-weighted image. d Sagittal T1-weighted image. Huge, totally extra-axial mass at the level of the right pontocerebellar angle that deforms and shifts the brainstem (arrows, a–c) and tends to engulf the basilar artery. Gross cystic portions are recognizable along the medial and inferior portions of the lesion (d)

References 1. Colosimo C, Cerase A, Moschini M, Di Lella GM, Tartaglione T. Tumori in età pediatrica. In: Simonetti G et al (eds) Trattato Italiano di Risonanza Magnetica. Napoli: IdelsonGnocchi, 1998:459–489. 2. Walker AE, Robins M, Weinfeld FD. Epidemiology of brain tumors: the national survey of intracranial neoplasms. Neurology 1985; 35:219–226. 3. Tortori-Donati P, Rossi A. Patologia neoplastica cranio-spi-

nale. In: Dal Pozzo G (ed) Compendio di risonanza magnetica. Cranio e rachide. Torino: UTET, 2001:1223–1277. 4. Rutherford GS, Hewlett RH, Truter R. Contrast enhanced imaging is critical to glioma nosology and grading. Int J Neuroradiol 1995; 1:28–38. 5. Ruggieri PM. Practical MR spectroscopy in pediatric neuroradiology. ASNR 2000 Proceedings, pp 5–53. 6. Okamoto K, Ito J, Ishikawa K, Sakai K, Tokiguchi S. Dif-

430

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama fusion-weighted echo-planar MR imaging in differential diagnosis of brain tumors and tumor-like conditions. Eur Radiol 2000; 10:1342–1350. 7. Krabbe K, Gideon P, Wagn P, Hansen U, Thomsen C, Madsen F. MR diffusion imaging of human intracranial tumors. Neuroradiology 1997; 39:483–489. 8. Tsuruda JS, Chew WM, Moseley ME, Norman D. Diffusionweighted MR imaging of the brain: value of differentiating between extraaxial cysts and epidermoid tumors. AJNR Am J Neuroradiol 1990; 11:925–931. 9. Kim YJ, Chang KH, Song IC, Kim HD, Seong SO, Kim YH, Han MH. Brain abscess and necrotic or cystic brain tumor: discrimination with signal intensity on diffusion-weighted MR imaging. AJR Am J Roentgenol 1998; 171:1487–1490. 10. Hartmann M, Jansen O, Heiland S, Sommer C, Munkel K, Sartor K. Restricted diffusion within ring enhancement is not pathognomonic for brain abscess. AJNR Am J Neuroradiol 2001; 22:1738–1742. 11. Holtås S, Geijer B, Stromblad LG, Maly-Sundgren P, Burtscher IM. A ring-enhancing metastasis with central high signal on diffusion-weighted imaging and low apparent diffusion coefficients. Neuroradiology 2000; 42:824–827. 12. Tien RD, Felsberg GJ, Friedman H, Brown M, MacFall J. MR imaging of high-grade cerebral gliomas: value of diffusionweighted echoplanar pulse sequences. AJR Am J Roentgenol 1994; 162:671–677. 13. Holland SK, Ball WS Jr. MR perfusion imaging in pediatric cerebrovascular disease. ASNR 2000 Proceedings, pp 115– 123. 14. Principi M, Italiani M, Muti M, Feranti R, Guiducci A, Micheli C, Aprile I, Baroni M, Ottaviano P. Studio di alcuni tipi di neoplasie cerebrali intra-assiali con tecnica RM-perfusionale e correlazioni istopatologiche. Rivista di Neuroradiologia 2001; 14:25–44. 15. Naidich TP, Yousry TA. Anatomy and function: MR imaging of motor function and sensation. ASNR 2000 Proceedings, pp 63–74. 16. Biegel JA. Genetics of pediatric central nervous system tumors. J Pediatr Hematol Oncol 1997; 19:492–501. 17. Tortori-Donati P, Fondelli MP, Rossi A, Cama A, Caputo L, Andreussi L, Garré ML. Medulloblastoma in children: CT and MRI findings. Neuroradiology 1996; 38:352–359. 18. Tortori-Donati P, Fondelli MP, Carini S, Cama A, Garré ML, Salomone G. Tumori neuroectodermici primitivi (PNETs) e medulloblastoma (PNET-fossa posteriore). Rivista di Neuroradiologia 1992; 5 (Suppl 4):15−31. 19. Kleihues P, Cavenee WK. Pathology and genetics, tumors of the nervous system. Lyon, France: IARC Press, 1997:207– 214. 20. Levy RA, Blaivas M, Muraszko K, Robertson PL. Desmoplastic medulloblastoma: MR findings. AJNR Am J Neuroradiol 1997; 18:1364–1366. 21. Ur-Rahman N, Jamjoom A, Al-Rayess M, Jamjoom ZA. Cerebellopontine angle medulloblastoma. Br J Neurosurg 2000; 14:262–263. 22. Giangaspero F, Perilongo G, Fondelli MP, Brisigotti M, Carollo C, Burnelli R, Burger PC, Garré ML. Medulloblastoma with extensive nodularity: a variant with favorable prognosis. J Neurosurg 1999; 91:971–977. 23. Brandis A, Heyer R, Hori A, Walter GF. Cerebellar neurocytoma in an infant: an important differential diagnosis from cerebellar neuroblastoma and medulloblastoma? Neuropediatrics 1997; 28:235–238. 24. Bergmann M, Pietsch T, Herms J, Janus J, Spaar HJ, Terwey

B. Medullomyoblastoma: a histological, immunohistochemical, ultrastructural and molecular genetic study. Acta Neuropathol (Berl) 1998; 95:205–212. 25. Tortori-Donati P, Fondelli MP, Cama A, Garré ML, Rossi A, Andreussi L. Ependymomas of the posterior cranial fossa: CT and MRI findings. Neuroradiology 1995; 37:238–243. 26. Wang Z, Sutton LN, Cnaan A, Haselgrove JC, Rorke LB, Zhao H, Bilaniuk LT, Zimmerman RA. Proton MR spectroscopy of pediatric cerebellar tumors. AJNR Am J Neuroradiol 1995; 16:1821–1833. 27. Kulkarni AV, Becker LE, Jay V, Armstrong DC, Drake JM. Primary cerebellar glioblastomas multiforme in children. J Neurosurg 1999; 90:546–550. 28. Pencalet P, Maixner W, Sainte-Rose C, Lellouch-Tubiana A, Cinalli G, Zerah M, Pierre-Kahn A, Hoppe-Hirsch E, Bourgeois M, Renier D. Benign cerebellar astrocytomas in children. J Neurosurg 1999; 90:265–273. 29. Lee YY, van Tassel P, Bruner JM, Moser RP, Share JC. Juvenile pilocytic astrocytomas: CT and MR characteristics. AJR Am J Roentgenol 1989; 152:1263–1270. 30. Smirniotopoulos JG. The new Who classification of brain tumors: imaging correlations (published online). Available http://radlinux1.usuf1.usuhs.mil/rad/home/newwho/ at who2.html 31. Canapicchi R, Valleriani AM, Bianchi MC, Montanaro D, Abbruzzese A, Tosetti M. Tumori cerebellari e del quarto ventricolo in età pediatrica. Rivista di Neuroradiologia 1996; 9:707–719. 32. Chiechi MV, Smirniotopoulos JG, Jones RV. Intracranial subependymomas: CT and MR imaging features in 24 cases. AJR Am J Roentgenol 1995; 165:1245–1250. 33. Ikezaki K, Matsushima T, Inoue T, Yokoyama N, Kaneko Y, Fukui M. Correlation of microanatomical localization with postoperative survival in posterior fossa ependymomas. Neurosurgery 1993; 32:38–44. 34. Nagib MG, O’Fallon MT. Posterior fossa lateral ependymoma in childhood. Pediatr Neurosurg 1996; 24:299–305. 35. Figarella-Branger D, Civatte M, Bouvier-Labit C, Gouvernet J, Gambarelli D, Gentet JC, Lena G, Choux M, Pellissier JF. Prognostic factors in intracranial ependymomas in children. J Neurosurg 2000; 93:605–613. 36. Courville CB, Broussalian SL. Plastic ependymomas of the lateral recess. Report of eight verified cases. J Neurosurg 1961; 18:792–796. 37. Spoto GP, Press GA, Hesselink JR, Solomon M. Intracranial ependymoma and subependymoma: MR manifestations. AJNR Am J Neuroradiol 1990; 11:83–91. 38. Russell DS, Rubinstein LJ. Pathology of Tumours of the Central Nervous System, 5th edn. London: Edward Arnold, 1989. 39. Nazar GB, Hoffman HJ, Becker LE, Jenkin D, Humphreys RP, Hendrick EB. Infratentorial ependymomas in childhood: prognostic factors and treatment. J Neurosurg 1990; 72:408–417. 40. Fischbein NJ, Prados MD, Wara W, Russo C, Edwards MS, Barkovich AJ. Radiologic classification of brain stem tumors: correlation of magnetic resonance imaging appearance with clinical outcome. Pediatr Neurosurg 1996; 24:9–23. 41. Colosimo C, Tartaglione T, Di Lella GM, Cerase A. Neuroradiology of brain stem and cervicomedullary tumours. Rivista di Neuroradiologia 1997; 10:41–54. 42. Kane AG, Robles HA, Smirniotopoulos JG, Hieronimus JD, Fish MH. Radiologic-pathologic correlation. Diffuse pontine astrocytoma. AJNR Am J Neuroradiol 1993; 14:941– 945.

Brain Tumors 43. Albright AL, Guthkelch AN, Packer RJ, Price RA, Rourke LB. Prognostic factors in pediatric brain-stem gliomas. J Neurosurg 1986; 65:751–755. 44. Burger PC. Pathology of brainstem astrocytomas. Pediatr Neurosurg 1996; 24:35–40. 45. Packer RJ, Boyett JM, Zimmerman RA, Albright AL, Kaplan AM, Rorke LB, Selch MT, Cherlow JM, Finlay JL, Wara WM. Outcome of children with brain stem gliomas after treatment with 7,800 cGy of hyperfractioned radiotherapy. Cancer 1994; 74:1827–1834. 46. Epstein FJ. A staging system for brainstem gliomas. Cancer 1989; 56:1804–1806. 47. Fisher PG, Breiter SN, Carson BS, Wharam MD, Williams JA, Weingart JD, Foer DR, Goldthwaite PT, Tihan T, Burger PC. A clinicopathologic reappraisal of brain stem tumor classification. Cancer 2000; 89:1569–1576. 48. Molloy PT, Bilaniuk LT, Vaughan SN, Needle MN, Liu GT, Zackai EH, Phillips PC. Brainstem tumors in patients with neurofibromatosis type 1: a distinct clinical entity. Neurology 1995; 45:1897–1902. 49. Barkovich AJ, Krischer J, Kun LE, Packer R, Zimmerman RA, Freeman CR, Wara WM, Albright L, Allen JC, Hoffman HJ. Brain stem gliomas: a classification system based on magnetic resonance imaging. Pediatr Neurosci 1990–1991; 16:73–83. 50. Zimmerman RA. Neuroimaging of primary brainstem gliomas: diagnosis and course. Pediatr Neurosurg 1996; 25:45–53. 51. Kaplan AM, Albright AL, Zimmerman RA, Rorke LB, Li H, Boyett JM, Finlay JL, Wara WM, Packer RJ. Brainstem gliomas in children: a Children’s Cancer Group review of 119 cases. Pediatr Neurosurg 1996; 24:185–192. 52. Bognar L, Turjman F, Villanyi E, Mottolese C, Guyotat J, Fischer C, Jouvet A, Lapras C. Tectal plate gliomas part II: CT scans and MR imaging of tectal gliomas. Acta Neurochir (Wien) 1994; 127:48–54. 53. Robertson PL, Muraszko KM, Brunberg JA, Axtell RA, Dauser RC, Turrisi AT. Pediatric midbrain subgroup of brainstem gliomas. Pediatr Neurosurg 1995; 22:65–73. 54. May PL, Blaser SI, Hoffman HJ, Humphreys RP, HarwoodNash DC. Benign intrinsic tectal “tumors” in children. J Neurosurg 1991; 74:867–871. 55. Rorke LB, Packer R, Biegel J. Central nervous system atypical teratoid/rhabdoid tumors of infancy and childhood. J Neurooncol 1995; 24:21–28. 56. Tortori-Donati P, Fondelli MP, Rossi A, Colosimo C, Garré ML, Bellagamba O, Brisigotti M, Piatelli G, Andreussi L. Atypical teratoid rhabdoid tumors of the central nervous system in infancy. Neuroradiologic findings. Int J Neuroradiol 1997; 3:327–338. 57. Burger PC, Tihan T, Friedman HS, Strother DR, Kepner JL, Duffner PK, Kun LE, Perlman EJ. Atypical teratoid/rhabdoid tumor of the central nervous system: a highly malignant tumor of infancy and childhood frequently mistaken for medulloblastoma: a Pediatric Oncology Group study. Am J Surg Pathol 1998; 22:1083–1092. 58. Hanna SL, Langston JW, Parham DM, Douglass EC. Primary malignant rhabdoid tumor of the brain: clinical, imaging, and pathologic findings. AJNR Am J Neuroradiol 1993; 14:107–115. 59. Muñoz A, Carrasco A, Muñoz MJ, Esparza J. Cranial rhabdoid tumor with marginal tumor cystic component and extraaxial extension. AJNR Am J Neuroradiol 1995; 16:1727–1728.

60. Biegel J, Rorke LB, Packer RJ, Emanuel BS. Monosomy 22 in rhabdoid or atypical tumors of the brain. J Neurosurg 1990; 73:710–714. 61. Booz MM, Koelma I, Unnikrishnan M. Case report: primary malignant rhabdoid tumor of the brain. Clin Radiol 1996; 51:65–66. 62. Caldemeyer KS, Smith RR, Azzarelli B, Boaz JC. Primary central nervous system malignant rhabdoid tumor: CT and MR appearance simulates a primitive neuroectodermal tumor. Pediatr Neurosurg 1994; 21:232–236. 63. Ho PSP, Lee WH, Chen CY, Chou TY, Wang YC, Sun MJ, Liou WY. Primary malignant rhabdoid tumor of the brain: CT characteristics. J Comput Assist Tomogr 1990; 14:461–463. 64. Perilongo G, Sutton L, Czaykowski D, Gusnard D, Biegel J. Rhabdoid tumor of the central nervous system. Med Pediatr Oncol 1991; 19:310–317. 65. Agranovich AL, Griebel RW, Lowry N, Tchang SP. Malignant rhabdoid tumor of the central nervous system with subarachnoid dissemination. Surg Neurol 1992; 37:410–414. 66. Sutton LN, Laisner T, Hunter J, Rorke LB, Sanford RA. Thirteen-year-old female with hemangioblastoma. Pediatr Neurosurg 1997; 27:50–55. 67. Eskridge JM, McAuliffe W, Harris B, Kim DK, Scott J, Winn HR. Preoperative endovascular embolization of craniospinal hemangioblastomas. AJNR Am J Neuroradiol 1996; 17:525–531. 68. Gilles FH, Brown WD, Leviton A, Tavare CJ, Adelman L, Rorke LB, Davis RL, Hedley-Whyte TE. Limitations of the World Health Organization classification of childhood supratentorial astrocytic tumors. Children Brain Tumor Consortium. Cancer 2000; 88:1477–1483. 69. Osborn AG. Diagnostic Neuroradiology. St. Louis: Mosby, 1994. 70. Kuroiwa T, Numaguchi Y, Rothman MI, Zoarski GH, Morikawa M, Zagardo MT, Kristt DA. Posterior fossa glioblastoma multiforme: MR findings. AJNR Am J Neuroradiol 1995; 16:583–589. 71. Klein R, Molenkamp G, Sorensen N, Roggendorf W. Favorable outcome of giant cell glioblastoma in a child. Report of an 11year survival period. Child’s Nerv Syst 1998; 14:288–291. 72. Meyer-Puttlitz B, Hayashi Y, Waha A, Rollbrocker B, Bostrom J, Wiestler OD, Louis DN, Reifenberger G, von Deimling A. Molecular genetic analysis of giant cell glioblastomas. Am J Pathol 1997; 151:853–857. 73. Peraud A, Watanabe K, Schwechheimer K, Yonekawa Y, Kleihues P, Ohgaki H. Genetic profile of the giant cell glioblastoma. Lab Invest 1999; 79:123–129. 74. Tihan T, Fisher PG, Kepner JL, Godfraind C, McComb RD, Goldthwaite PT, Burger PC. Pediatric astrocytomas with monomorphous pilomyxoid features and a less favorable outcome. J Neuropathol Experimental Neurol 1999; 58:1061–1068. 75. Parazzini C, Triulzi F, Bianchini E, Agnetti V, Conti M, Zanolini C, Maninetti MM, Rossi LN, Scotti G. Spontaneous involution of optic pathway lesions in neurofibromatosis type I: serial contrast MR evaluation. AJNR Am J Neuroradiol 1995; 16:1711–1718. 76. Schmandt SM, Packer RJ, Vezina LG, Jane J. Spontaneous regression of low-grade astrocytomas in childhood. Pediatr Neurosurg 2000; 32:132–136. 77. Hwang JH, Egnaczyk GF, Ballard E, Dunn RS, Holland SK, Ball WS Jr. Proton MR spectroscopy characteristics of pediatric pilocytic astrocytomas. AJNR Am J Neuroradiol 1998; 19:535–540.

431

432

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama 78. Haga S, Morioka T, Nishio S, Fukui M. Multicentric pleomorphic xanthoastrocytoma: case report. Neurosurgery 1996; 38:1242–1245. 79. Petropoulou K, Whiteman MLH, Altman NR, Buce J, Morrison G. CT and MRI of pleomorphic xanthoastrocytoma: unusual biologic behavior. J Comput Assist Tomogr 1995; 19:860–865. 80. Kepes JJ, Rubinstein LJ, Eng LF. Pleomorphic xanthoastrocytoma: a distinctive meningocerebral glioma of young subjects with relatively favorable prognosis. A study of 12 cases. Cancer 1979; 44:1839–1852. 81. Tien RD, Cardenas CA, Rajagopalan S. Pleomorphic xanthoastrocytoma of the brain: MR findings in six patients. AJR Am J Roengenol 1992; 159:1287–1290. 82. Russo CG, Frederickson K, Smoker WRK, Ghatak NR, Naidich TP, Rubin D, Gonzales-Arias SM. Pleomorphic xanthoastrocytoma. Report of two cases and review of the literature. Int J Neuroradiol 1996; 2:570–578. 83. Lipper MH, Eberhard DA, Phillips CD, Vezina LG, Cail WS. Pleomorphic xanthoastrocytoma, a distinctive astroglial tumor: neuroradiologic and pathologic features. AJNR Am J Neuroradiol 1993; 14:1397–1404. 84. Tonn JC, Paulus W, Warmuth-Metz M, Schachenmayr W, Sörensen N, Roosen K. Pleomorphic xanthoastrocytoma: report of six cases with special consideration of diagnostic and therapeutic pitfalls. Surg Neurol 1997; 47:162–169. 85. Bayindir C, Balak N, Karasu A, Kasaroglu D. Anaplastic pleomorphic xanthoastrocytoma. Child’s Nerv Syst 1997; 13:50–56. 86. Davies KG, Maxwell RE, Seljeskog E, Sung JH. Pleomorphic xanthoastrocytoma: report of four cases with MRI scan appearances and literature review. Br J Neurosurg 1994; 8:681–689. 87. Giannini C, Scheithauer BW, Burger PC, Brat DJ, Wollan PC, Lach B, O’Neill BP. Pleomorphic xanthoastrocytoma. What do we really know about it? Cancer 1999; 85:2033–2045. 88. Lantos PL, Vandenberg SR, Kleihues P.Tumours of the nervous sytem. In: Graham DI, Lantos PL (eds) Greenfield’s Neuropathology, 6th edn. London: Arnold, 1997: 583–879 (vol.II). 89. Tice H, Barnes PD, Goumnerova L, Scott RM, Tarbell NJ. Pediatric and adolescent oligodendroglioma. AJNR Am J Neuroradiol 1993; 14:1293–1300. 90. Narita T, Kurotaki H, Hashimoto T, Ogawa Y. Congenital oligodendroglioma: a case report of a 34th-gestational week fetus with immunohistochemical study and review of the literature. Hum Pathol 1997; 28:1213–1217. 91. Rizk T, Mottolèse C, Bouffet E, Jouvet A, Guyotat J, Bret P, Lapras C. Cerebral oligodendrogliomas in children: an analysis of 15 cases. Childs Nerv Syst 1996; 12:527–529. 92. Burnette WC, Nesbit GM, Hall FW. Radiologic-pathologic correlation in oligodendroglioma. Int J Neuroradiol 1997; 3:503–510. 93. Natelson SE, Dyer ML, Harp DL. Delayed CSF seeding of benign oligodendroglioma. Southern Med J 1992; 85:1011– 1012. 94. Lee YY, van Tassel P. Intracranial oligodendroglioma: imaging findings in 35 untreated cases. AJNR Am J Neuroradiol 1989; 10:119–127. 95. Armao DM, Stone J, Castillo M, Mitchell KM, Bouldin TW, Suzuki K. Diffuse leptomeningeal oligodendrogliomatosis: radiologic/pathologic correlation. AJNR Am J Neuroradiol 2000; 21:1122–1126. 96. Krouwer HGJ, van Duinen SG, Kamphorst W, van der Valk

P, Algra A. Oligoastrocytomas: a clinicopathological study of 52 cases. J Neurooncol 1997; 33:223–238. 97. Furie DM, Provenzale JM. Supratentorial ependymomas and subependymomas: CT and MR appearance. J Comput Assist Tomogr 1995; 19:518–526. 98. Pizer BL, Moss T, Oakhill A, Webb D, Coakham HB. Congenital astroblastoma: an immunohistochemical study. J Neurosurg 1995; 83:550–555. 99. Cabello A, Madero S, Castresana A, Diaz-Lobato R. Astroblastoma: electron microscopy and immunohistochemical findings: case report. Surg Neurol 1991; 35:116–121. 100. Baka JJ, Patel SC, Roebuck JR, Hearshen DO. Predominantly extraaxial astroblastoma: imaging and proton MR spectroscopy features. AJNR Am J Neuroradiol 1993; 14: 946–950. 101. Bonnin JM, Rubinstein LJ. Astroblastomas: a pathological study of 23 tumors, with a postoperative follow-up in 13 patients. Neurosurgery 1989; 25:6–13. 102. Port JD, Brat DJ, Burger PC, Pomper MG. Astroblastoma: radiologic-pathologic correlation and distinction from ependymoma. AJNR Am J Neuroradiol 2002; 23:243–247. 103. Del Carpio-O’Donovan R, Korah I, Salazar A, Melançon D. Gliomatosis cerebri. Radiology 1996; 198:831–835. 104. Önal Ç, Bayindir Ç, S¸ iraneci R, I˙zgi N, Yalçin I, Altinel Z, Barlas O. A serial CT scan and MRI verification of diffuse cerebrospinal gliomatosis: a case report with stereotactic diagnosis and radiological confirmation. Pediatr Neurosurg 1996; 25:94–99. 105. Pyhtinen J, Pääkkö E. A difficult diagnosis of gliomatosis cerebri. Neuroradiology 1996; 38:444–448. 106. Felsberg GJ, Silver SA, Brown MT, Tien RD. Radiologicpathologic correlation. Gliomatosis cerebri. AJNR Am J Neuroradiol 1994; 15:1745–1753. 107. Hart ML, Earle KM. Primitive neuroectodermal tumors of the brain in children. Cancer 1973; 32:890–897. 108. Jacacki RI. Pineal and nonpineal supratentorial primitive neuroectodermal tumors. Childs Nerv Syst 1999; 15:586– 591. 109. Tortori-Donati P, Fondelli MP, Rossi A, Ravegnani M, Andreussi L. La neuroradiologia dei PNETs sopratentoriali. Rivista di Neuroradiologia 1994; 7 (Suppl 1):87–92. 110. Katayama Y, Limura S, Watanabe TY, Yoshino A, Koshinaga M. Peripheral-type primitive neuroectodermal tumor arising in the tentorium. J Neurosurg 1999; 90:141–144. 111. Mikaeloff Y, Raquin MA, Lellouch-Tubiana A, TerrierLacombe MJ, Zerah M, Bulteau C, Habrand JL, Kalifa C. Primitive cerebral neuroectodermal tumors excluding medulloblastomas: a retrospective study of 30 cases. Pediatr Neurosurg 1998; 29:170–177. 112. Figueroa RE, el Gammal T, Brooks BS, Holgate R, Miller W. MR findings on primitive neuroectodermal tumors. J Comput Assist Tomogr 1989; 13:773–778. 113. Klisch J, Husstedt H, Hennings S, van Velthoven V, Pagenstecher A, Schumacher M. Supratentorial primitive neuroectodermal tumours: diffusion-weighted MRI. Neuroradiology 2000; 42:393–398. 114. Wiegel B, Harris TM, Edwards MK, Smith RR, Azzarelli B. MR of intracranial neuroblastoma with dural sinus invasion and distant metastases. AJNR Am J Neuroradiol 1991; 12:1198–1200. 115. Kosnik EJ, Boesel CP, Bay J, Sayers MP. Primitive neuroectodermal tumors of the central nervous system in children. J Neurosurg 1978; 48:741–746. 116. Shields JA, Shields CL, Schwartz RL. Malignant teratoid medulloepithelioma of the ciliary body simulating per-

Brain Tumors sistent hyperplastic primitive vitreous. Am J Ophthalmol 1989; 107:296–298. 117. Molloy PT, Yachnis AT, Rorke LB, Dattilo JJ, Needle MN, Millar WS, Goldwein JW, Sutton LN, Phillips PC. Central nervous system medulloepithelioma: a series of eight cases including two arising in the pons. J Neurosurg 1996; 84:430– 436. 118. Dorsay TA, Rovira MJ, Ho VB, Kelley J. Ependymoblastoma: MR presentation. A case report and review of the literature. Pediatr Radiol 1995; 25:433–435. 119. Mork SJ, Rubinstein LJ. Ependymoblastoma: a reappraisal of a rare embryonal tumor. Cancer 1985; 55:1536–1542. 120. Barkovich AJ, Kuzniecky RI, Dobyns WB, Jackson GD, Becker LE, Evrard P.A classification scheme for malformations of cortical development. Neuropediatrics 1996; 27:59– 63. 121. Furie DM, Felsberg GJ, Friedman HS, Fuchs H, McLendon R. MRI of gangliocytoma of cerebellum and spinal cord. J Comput Assist Tomogr 1993; 17:488–491. 122. Duchowny M, Altman N, Bruce J. Dysplastic gangliocytoma of the cerebral hemispheres. In: Guerrini R, Andermann F, Canapicchi R, Roger J, Zifkin B, Pfanner P (eds) Dysplasia of cerebral cortex and epilepsy. Philadelphia: LippincottRaven, 1996:93–100. 123. Altman NR. MR and CT characteristics of gangliocytoma: a rare cause of epilepsy in children. AJNR Am J Neuroradiol 1988; 9:917–921. 124. Barkovich AJ. Pediatric Neuroimaging, 3nd edn. Philadelphia: Lippincott Williams & Wilkins, 2000. 125. Zentner J, Wolf HK, Ostertun B, Hufnagel A, Campos MG, Solymosi L, Schramm J. Gangliogliomas: cinical, radiological, and histopathological findings in 51 patients. J Neurol Neurosurg Psychiatry 1994; 57:1497–1502. 126. Castillo M, Davis PC, Takei Y, Hoffman JC. Intracranial ganglioglioma: MR, CT, and clinical findings in 18 patients. AJNR Am J Neuroradiol 1990; 11:109–114. 127. Sartoretti-Schefer S, Aguzzi A, Wichmann W, Valavanis A. Gangliogliomas. Correlation between appearance on magnetic resonance imaging and tumor histology. Int J Neuroradiol 1998; 4:357–365. 128. Blatt GL, Ahuja A, Miller LL, Ostrow PT, Soloniuk DS. Cerebellomedullary ganglioglioma: CT and MR findings. AJNR Am J Neuroradiol 1995; 16:790–792. 129. Castillo M. Gangliogliomas: ubiquitous or not? AJNR Am J Neuroradiol 1998; 19:807–809. 130. Matsumoto K, Tamiya T, Ono Y, Furuta T, Asari S, Ohmoto T. Cerebral gangliogliomas: clinical characteristics, CT and MRI. Acta Neurochir (Wien) 1999; 141:135–141. 131. Chintagumpala MM, Armstrong D, Miki S, Nelson T, Cheek W, Laurent J, Woo SY, Mahoney DH Jr. Mixed neuronal-glial tumors (gangliogliomas) in children. Pediatr Neurosurg 1996; 24:306–313. 132. Nakajima M, Kidooka M, Nakasu S. Anaplastic ganglioglioma with dissemination to the spinal cord: a case report. Surg Neurol 1998; 49:445–448. 133. Dorne HL, O’Gorman AM, Melanson D. Computed tomography of intracranial gangliogliomas. AJNR Am J Neuroradiol 1986; 7:281–285. 134. Provenzale JM, Ali U, Barboriak DP, Kallmes DF, Delong DM, McLendon RE. Comparison of patient age with MR imaging features of gangliogliomas. AJR Am J Roentgenol 2000; 174:859–862. 135. Tampieri D, Moumdjian R, Melanson D, Ethier D. Intracerebral gangliogliomas in patients with complex partial sei-

zures: CT and MR imaging findings. AJR Am J Roentgenol 1991; 157:843–849. 136. Hassoun J, Gambarelli D, Grisoli F, Pellet W, Salamon G, Pellissier JF, Toga M. Central neurocytoma. An electronmicroscopic study of two cases. Acta Neuropathol (Berl) 1982; 56:151–156. 137. Funato H, Inoshita N, Okeda R, Yamamoto S, Aoyagi M. Cystic ganglioneurocytoma outside the ventricular region. Acta Neuropathol (Berl) 1997; 94:95–98. 138. Goergen SK, Gonzales MF, McLean CA. Intraventricular neurocytoma: radiologic features and review of the literature. Radiology 1992; 182:787–792. 139. Tomura N, Hirano H, Watanabe O, Watarai J, Itoh Y, Mineura K, Kowada M. Central neurocytoma with clinically malignant behavior. AJNR Am J Neuroradiol 1997; 18:1175– 1178. 140. Warmuth-Metz M, Klein R, Sörensen N, Solymosi L. Central neurocytoma of the fourth ventricle. Case report. J Neurosurg 91:506–509. 141. Ashkan K, Casey ATH, D’Arrigo C, Harkness WF, Thomas DGT. Benign central neurocytoma. A double misnomer? Cancer 2000; 89:1111–1120. 142. Nishio S, Takeshita I, Kaneko Y, Fukui M. Cerebral neurocytoma. A new subset of benign neuronal tumors of the cerebrum. Cancer 1992; 70:529–537. 143. Giangaspero F, Cenacchi G, Losi L, Cerasoli S, Bisceglia M, Burger PC. Extraventricular neoplasms with neurocytoma features. A clinicopathological study of 11 cases. Am J Surg Pathol 1997; 21:206–212. 144. Taylor CL, Cohen ML, Cohen AR. Neurocytoma presenting with intraparenchymal cerebral hemorrhage. Pediatr Neurosurg 1998; 29:92–95. 145. Tortori-Donati P, Fondelli MP, Rossi A, Cama A, Brisigotti M, Pellicanò G. Extraventricular neurocytoma with ganglionic differentiation associated with complex partial seizures. AJNR Am J Neuroradiol 1999; 20:724–727. 146. Yamamoto T, Komori T, Shibata N, Toyoda C, Kobayashi M. Multifocal neurocytoma/gangliocytoma with extensive leptomeningeal dissemination in the brain and spinal cord. Am J Surg Pathol 1996; 20:363–370. 147. Kim DH, Suh YL. Pseudopapillary neurocytoma of temporal lobe with glial differentiation. Acta Neuropathol (Berl) 1997; 94:187–191. 148. Ellison DW, Zygmint SC, Weller RO. Neurocytoma/lipoma (neurolipocytoma) of the cerebellum. Neuropathol Appl Neurobiol 1993; 19:95–98. 149. Mena H, Morrison AL, Jones RV, Gyure KA. Central neurocytomas express photoreceptor differentiation. Cancer 2001; 91:136–143. 150. Daumas-Duport C, Scheithauer BW, Chodkiewicz JP, Laws ER Jr, Vedrenne C. Dysembryoplastic neuroepithelial tumor: a surgically curable tumor of young patients with intractable partial seizures. Neurosurgery 1988; 23:545–556. 151. Ostertun B, Wolf HK, Campos MG, Matus C, Solymosi L, Elger CE, Schramm J, Schild HH. Dysembryoplastic neuroepithelial tumors: MR and CT evaluation. AJNR Am J Neuroradiol 1996; 17:419–430. 152. Koeller KK, Dillon WP. Dysembryoplastic neuroepithelial tumors: MR appearance. AJNR Am J Neuroradiol 1992; 13:1319–1325. 153. Kuroiwa T, Bergey GK, Rothman MI, Zoarski GH, Wolf A, Zagardo MT, Kristt DA, Hudson LP, Krumholz A, Barry E, Numaguchi Y. Radiologic appearance of the dysembryoplastic neuroepithelial tumor. Radiology 1995; 197:233–238.

433

434

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama 154. Weissman Z, Michowitz S, Shuper A, Kornreich L, Amir J. Dysembryoplastic neuroepithelial tumor: a curable cause of seizures. Pediatr Hematol Oncol 1996; 13:463–468. 155. Daumas-Duport C. Dysembryoplastic neuroepithelial tumors in epilepsy surgery. In: Guerrini R, Andermann F, Canapicchi R, Roger J, Zifkin B, Pfanner P (eds) Dysplasia of cerebral cortex and epilepsy. Philadelphia: LippincottRaven, 1996:71–80. 156. Rypens F, Esteban MJ, Lellouch-Tubiana A, Bastien-Fagnou C, Renier D, Baraton J, Brunelle F. Desmoplastic supratentorial neuroepithelial tumours of childhood: imaging in 5 patients. Neuroradiology 1996; 38:S165–S168. 157. Tenreiro-Picon OR, Kamath SV, Knorr JR, Ragland RL, Smith TW, Lau KY. Desmoplastic infantile ganglioglioma: CT and MRI features. Pediatr Radiol 1995; 25:540–543. 158. Lopez JI, Ereño C, Oleaga L, Areitio E. Meningioangiomatosis and oligodendroglioma in a 15-year-old boy. Arch Pathol Lab Med 1996; 120:587–590. 159. Wiebe S, Munoz DG, Smith S, Lee DH. Meningioangiomatosis. A comprehensive analysis of clinical and laboratory features. Brain 1999; 122:709–726. 160. Gomez-Anson B, Munoz A, Blasco A, Madero S, Esparza J, Cordobes F, Orejon G, Mateos F. Meningioangiomatosis: advanced imaging and pathological study of two cases. Neuroradiology 1995; 37:120–123. 161. Garen PD, Powers JM, King JS, Perot PL. Intracranial fibroosseous lesion. J Neurosurg 1989; 70:475–477. 162. Magaudda A, Dalla Bernardina B, De Marco P, Sfaello Z, Longo M, Colamaria V, Daniele O, Tortorella G, Tata MA, Di Perri R, et al. Bilateral occipital calcification, epilepsy and coeliac disease: clinical and neuroimaging features of a new syndrome. J Neurol Neurosurg Psychiatry 1993; 56:885–889. 163. Sugama S, Yoshimura H, Ashimine K, Eto Y, Maekawa K. Enhanced magnetic resonance imaging of leptomeningeal angiomatosis. Pediatr Neurol 1997; 17:262–265. 164. Smirniotopoulos JG, Rushing EJ, Mena H. Pineal region masses: differential diagnosis. RadioGraphics 1992; 12:577– 596. 165. Tien RD, Barkovich AJ, Edwards MS. MR imaging of pineal tumors. AJR Am J Roentgenol 1990; 155:143–151. 166. Hoffman HJ, Yoshida M, Becker LE, Hendrick EB, Humphreys RP. Pineal region tumors in childhood. Pediatr Neurosurg 1994; 21:91–104. 167. Engel U, Gottschalk S, Niehaus L, Lehmann R, May C, Vogel S, Jänisch W. Cystic lesions of the pineal region – MRI and pathology. Neuroradiology 2000; 42:309–402. 168. Packer RJ, Cohen BH, Coney K. Intracranial germ cell tumors. Oncologist 2000; 5:312–320. 169. Fujimaki T, Matsutani M, Funada N, Kirino T, Takakura K, Nakamura O, Tamura A, Sano K. CT and MRI features of intracranial germ cell tumors. J Neurooncol 1994; 19:217–226. 170. Zimmerman RA, Bilaniuk LT. Age-related incidence of pineal calcification detected by computed tomography. Radiology 1982; 142:659–662. 171. Sumida M, Uozumi T, Kiya K, Mukada K, Arita K, Kurisu K, Sugiyama K, Onda J, Satoh H, Ikawa F, Migita K. MRI of intracranial germ cell tumours. Neuroradiology 1995; 37:32–37. 172. Allen JC, Nisselbaum J, Epstein F, Rosen G, Schwartz MK. Alphafetoprotein and human chorionic gonadotropin determination in cerebrospinal fluid: an aid to the diagnosis and management of intracranial germ-cell tumors. J Neurosurg 1979; 51:368–374. 173. Drummond KJ, Rosenfeld JV. Pineal region tumours in

childhood. A 30-year experience. Child’s Nerv Syst 1999; 15:119–127. 174. Nakajima H, Iwai Y, Yamanaka K, Yasui T, Kishi H. Primary intracranial germinoma in the medulla oblongata. Surg Neurol 2000; 53:448–451. 175. Shinoda J, Yamada H, Sakai N, Ando T, Hirata T, Miwa Y. Placental alkaline phosphatase as a tumor marker for primary intracranial germinoma. J Neurosurg. 1988; 68:710–720. 176. Shinoda J, Miwa Y, Sakai N, Yamada H, Shima H, Kato K, Takahashi M, Shimokawa K. Immunohistochemical study of placental alkaline phosphatase in primary intracranial germ-cell tumors. J Neurosurg 1985; 63:733–739. 177. Moon WK, Chang KH, Han MH, Kim IO. Intracranial germinomas: correlation of imaging findings with tumor response to radiation therapy. AJR Am J Roentgenol 1999; 172:713–716. 178. Ganti SR, Hilal SK, Stein BM, Silver AJ, Mawad M, Sane P. CT of pineal region tumors. AJR Am J Roentgenol 1986; 146:451–458. 179. Hoffman HJ, Otsubo H, Hendrick EB, Humphreys RP, Drake JM, Becker LE, Greenberg M, Jenkin D. Intracranial germcell tumors in children. J Neurosurg 1991; 74:545–551. 180. Higano S, Takahashi S, Ishii K, Matsumoto K, Ikeda H, Sakamoto K. Germinoma originating in the basal ganglia and thalamus: MR and CT evaluation. AJNR Am J Neuroradiol 1994; 15:1435–1441. 181. Satoh H, Uozumi T, Kiya K, Kurisu K, Arita K, Sumida M, Ikawa F. MRI of pineal region tumours: relationship between tumours and adjacent structures. Neuroradiology 1995; 37:624–630. 182. Bavbek M, Altinörs N, Caner H, Ag˘ ildere M, Cinemre O, Güz T, Erekul S. Giant posterior fossa teratoma. Childs Nerv Syst 1999; 15:359–361. 183. Srinivasan US, Joseph T. Intra-fourth-ventricle teratoma. Pediatr Neurosurg 1999; 30:200–202. 184. Storr U, Rupprecht T, Bornemann A, Ries M, Beinder E, Böwing B, Harms D. Congenital intracerebral teratoma: a rare differential diagnosis in newborn hydrocephalus. Pediatr Radiol 1997; 27:262–264. 185. Borit A, Blackwood W, Mair WCP. The separation of pineocytoma from pineoblastoma. Cancer 1980; 45:1408–1418. 186. Chiechi MV, Smirniotopoulos JG, Mena H. Pineal parenchymal tumors: CT and MR features. J Comput Assist Tomogr 1995; 19:509–517. 187. Nakamura M, Saeki N, Iwadate Y, Sunami K, Osato K, Yamaura A. Neuroradiological characteristics of pineocytoma and pineoblastoma. Neuroradiology 2000; 42:509–514. 188. Al Omari A, Mustafa MM, Hessler R, Rejjal A, Kattan A. Retinoblastoma as a congenital primary intracranial tumor. J Pediatr Hematol Oncol 1999; 21:296–298. 189. Bader J, Miller R, Meadow A, Zimmerman L, Champion L, Voute P. Trilateral retinoblastoma. Lancet II 1980; 582–583. 190. Bagley LJ, Hurst RW, Zimmerman RA, Shields JA, Shields CL, De Potter P. Imaging in the trilateral retinoblastoma syndrome. Neuroradiology 1996; 38:166–170. 191. MacKay CI, Baeesa SS, Ventureyra ECG. Epidermoid cysts of the pineal region.Child’s Nerv Syst 1999; 15:170–178. 192. Kuzma BB, Goodman JM. Epidermoid or arachnoid cyst? Surg Neurol 1997; 47:395–396. 193. Mamourian AC, Yarnell T. Enhancement of pineal cysts on MR images. AJNR Am J Neuroradiol 1991; 12:773–774. 194. Golzarian J, Balèriaux D, Bank WO, Matos C, FlamentDurand J. Pineal cyst: normal or pathological? Neuroradiology 1993; 35:251–253.

Brain Tumors 195. Hayman LA, Hinck VC. Clinical brain imaging: normal structure and functional anatomy. St.Louis: Mosby, 1992. 196. Wolpert SM, Barnes PD. MRI in Pediatric Neuroradiology. St.Louis: Mosby, 1992. 197. Coates TL, Hinshaw DB, Peckman N, Thompson JR, Hasso AN, Holshouser BA, Knierim DS. Pediatric choroid plexus neoplasms: MR, CT and pathologic correlation. Radiology 1989; 173:81–88. 198. Zulch KJ. Brain Tumors, 5th edn. Berlin: Springer, 1986. 199. Tomita T, McLone DG. Brain tumors during the first twentyfour months of life. Neurosurgery 1985; 17:913–919. 200. McConachie NS, Worthington BS, Cornford EJ, Balsitis M, Kerslake RW, Jaspan T. Computed tomography and magnetic resonance in the diagnosis of intraventricular cerebral masses. Br J Radiol 1994; 67:223–243. 201. Anderson DR, Falcone S, Bruce JH, Mejidas AA, Post MJD. Radiologic-pathologic correlation. Congenital chororid plexus papillomas. AJNR Am J Neuroradiol 1995; 16:2072– 2076. 202. Zimmerman RA. Pediatric brain tumors. In: Lee SH, Rao KCVG, Zimmerman RA (eds) Cranial MRI and CT, 3rd edn. New York: McGraw-Hill, 1992:381–416. 203. Radkowski MA, Naidich TP, Tomita T, Byrd SE, McLone DG. Neonatal brain tumors: CT and MR findings. J Comput Assist Tomogr 1988; 12:10–20. 204. Tortori-Donati P, Fondelli MP, Rossi A, Tomà P. Disgenesie e lipomi del corpo calloso. In: Tortori-Donati P, Taccone A, Longo M (eds) Malformazioni Cranio-Encefaliche. Neuroradiologia. Torino: Minerva Medica, 1996:125–136, 1996. 205. Fitz CR, Rao KCVG. Cranial Computed Tomography, 2nd edn. New York: McGraw-Hill, 1992. 206. Sanford RA, Laurent JP. Intraventricular tumors of childhood. Cancer 1985; 56:1795–1799. 207. Cama A, Tortori-Donati P, Carini S, Andreussi L. Idrocefalo congenito e neonatale. In: Tortori-Donati P, Taccone A, Longo M (eds) Malformazioni Cranio-Encefaliche. Neuroradiologia. Torino: Minerva Medica, 1996:313–339. 208. Asai A, Hoffman HJ, Hendrick EB, Humphreys RP, Becker LE. Primary intracranial neoplasms in the first year of life. Childs Nerv Syst 1989; 5:230–233. 209. Kendall B, Reider-Grosswasser I, Valentine A. Diagnosis of masses presenting within the ventricles on computed tomography. Neuroradiology 1983; 25:11–22. 210. Chow E, Jenkins JJ, Burger PC, Reardon DA, Langston JW, Sanford RA, Heideman RL, Kun LE, Merchant TE. Malignant evolution of choroid plexus papilloma. Pediatr Neurosurg 1999; 31:127–130. 211. Milhorat TH, Hammock MK, Davis DA, Fenstermacher JD. Choroid plexus papilloma. I. Proof of cerebrospinal fluid overproduction. Childs Brain 1976; 2:273–289. 212. Gajno TM, Locatelli D, Bonfanti N, Spanu G, Pezzotta S. Tumori dei plessi corioidei. Osservazioni di diagnostica neuroradiologica. Rivista di Neuroradiologia 1989; 2:285–289. 213. Morrison G, Sobel DF, Kelley WM, Norman D. Intraventricular mass lesions. Radiology 1984; 153:435–442. 214. Tortori-Donati P, Fondelli MP, Rossi A, Piatelli G, Balzarini C. Intraventricular supratentorial tumors in childhood. Rays 1996; 21:26–49. 215. Silver AJ, Ganti SR, Hilal SK. Computed tomography of tumors involving the atria of the lateral ventricles. Radiology 1982; 145:71–78. 216. Ferrante L, Acqui M, Artico M, Mastronardi L, Rocchi G, Fortuna A. Cerebral meningiomas in children. Childs Nerv Syst 1989; 5:83–86.

217. Starshak RJ. Radiation-induced meningioma in children: report of two cases and review of the literature. Pediatr Radiol 1996; 26:537–541. 218. Yoon HK, Kim SS, Kim IO, Na DG, Byun HS, Shin HJ, Han BK. MRI of primary meningeal tumours in children. Neuroradiology 1999; 41:512–516. 219. Darling CF, Byrd SE, Reyes-Mugica M, Tomita T, Osborn RE, Radkowski MA, Allen ED.MR of pediatric intracranial meningioma. AJNR Am J Neuroradiol 1994; 15:435–444. 220. Drake JM, Hendrick EB, Becker LE, Chuang SH, Hoffman HJ, Humphreys RP. Intracranial meningiomas in children. Pediatr Neurosci 1985–86; 12:134–139. 221. Hope JKA, Armstrong DA, Babyn PS, Humphreys RR, Harwood-Nash DC, Chuang SH, Marks PV. Primary meningeal tumors in children: correlation of clinical and CT findings with histologic type and prognosis. AJNR Am J Neuroradiol 1992; 13;1353–1364. 222. Erdinçler P, Lena G, Sarioglu AC, Kuday C, Choux M. Intracranial meningiomas in children: review of 29 cases. Surg Neurol 1998; 49:136–141. 223. Katayama Y, Tsubokawa T, Yoshida K. Cystic meningiomas in infancy. Surg Neurol 1986; 25:43–48. 224. Sheikh BY, Siqueira E, Dayel F. Meningioma in children: a report of nine cases and review of the literature. Surg Neurol 1996; 45:328–335. 225. Kloc W, Imielin´ ski BL, Wasilewski W, Stempniewicz M, Jende P, Karwacki Z. Meningiomas of the lateral ventricles of the brain in children. Childs Nerv Syst 1998; 14:350–353. 226. Huang PP, Doyle WK, Abbott IR. Atypical meningioma of the third ventricle in a 6-year-old boy. Neurosurgery 1993; 33:312–316. 227. Herzog CE, Leeds NE, Bruner JM, Baumgartner JE. Intracranial hemangiopericytomas in children. Pediatr Neurosurg 1995; 22:274–279. 228. Cole JC, Naul LG. Intracranial infantile hemangiopericytoma. Pediatr Radiol 2000; 30:271–273. 229. Chiechi MV, Smirniotopoulos JG, Mena H. Intracranial hemangiopericytomas: MR and CT features. AJNR Am J Neuroradiol 1996; 17:1365–1371. 230. Paulus W, Slowik F, Jellinger K. Primary intracranial sarcomas: histopathological features of 19 cases. Histopathology 1991; 18:395–402. 231. Pfluger T, Weil S, Weis S, Bise K, Egger J, Hadorn HB, Hahn K. MRI of primary meningeal sarcomas in two children: differential diagnostic considerations. Neuroradiology 1997; 39:225–228. 232. Allcutt DA, Hoffman HJ, Isla A, Becker LE, Humphreys RP. Acoustic schwannomas in children. Neurosurgery 1991; 29:14–19. 233. Tortori-Donati P, Fondelli MP, Rossi A, Andreussi L. I tumori extra-assiali della fossa posteriore in età pediatrica. Rivista di Neuroradiologia 1996; 9:721–730. 234. Tortori-Donati P, Fondelli MP, Rossi A, Bava GL. Intracranial contrast-enhancing masses in infants with capillary haemangioma of the head and neck: intracranial capillary haemangioma? Neuroradiology 1999; 41:369–375. 235. Caruso G, Germano A, Caffo M, Belvedere M, La Torre D, Tomasello F. Supratentorial dorsal cistern epidermoid cyst in childhood. Pediatr Neurosurg 1998; 29:203–207. 236. Gelabert-Gonzalez M. Intracranial epidermoid and dermoid cysts. Rev Neurol 1998; 27:777–782. 237. Atlas SW. Magnetic Resonance Imaging of the Brain and Spine, 2nd edn. Philadelphia: Lippincott-Raven, 1996. 238. Gandon Y, Hamon D, Carsin M, Guegan Y, Brassier G, Guy

435

436

P. Tortori-Donati, A. Rossi, R. Biancheri, M. L. Garrè, and A. Cama G, Husain A, Simon J. Radiological features of intradural epidermoid cysts. Contribution of MRI to the diagnosis. J Neuroradiol 1988; 15:335–351. 239. Tsuruda JS, Chew WM, Moseley ME, Norman D. Diffusionweighted MR imaging of the brain: value of differentiating between extraaxial cysts and epidermoid tumors. AJNR Am J Neuroradiol 1990; 11:925–931. 240. Chew WM, Rowley HA, Barkovich AJ. Magnetization transfer contrast imaging in pediatric patients. Radiology 1992; 185 (P):281. 241. Borba LAB, Al-Mefty O, Mrak RE, Suen J. Cranial chordomas in children and adolescents. J Neurosurg 1996; 84:584–591. 242. Probst EN, Zanella FE, Vortmeyer AO. Congenital clivus chordoma. AJNR Am J Neuroradiol 1993; 14:537–539. 243. Sze G, Uichanco LS, Brant-Zawadzki MN, Davis RL, Gutin PH, Wilson CB, Norman D, Newton TH. Chordomas: MR imaging. Radiology 1988; 166:187–191. 244. Niida H, Tanaka R, Tamura T, Takeuchi S. Clival chordoma in early childhood without bone involvement. Childs Nerv Syst 1994; 10:533–535. 245. Tashiro T, Fukuda T, Inoue Y, Nemoto Y, Shakudo M, Katsuyama J, Hakuba A, Onoyama Y. Intradural chordoma: case report and review of the literature. Neuroradiology 1994; 36:313–315. 246. Willing SJ, Faye-Petersen O, Aronin P, Faith S. Radiologic-

pathologic correlation. Capillary hemangioma of the meninges. AJNR Am J Neuroradiol 1993; 14:529–536. 247. Lasjaunias P. Vascular Diseases in Neonates, Infants and Children. Berlin: Springer, 1997:565–591. 248. Dillon WP, Som PM, Roseneau W. Hemangioma of the nasal vault: MR and CT features. Radiology 1991; 180:761–765. 249. Burrows PE, Mulliken JB, Fellows KE, Strand RD. Childhood hemangiomas and vascular malformations: angiographic differentiation. AJR Am J Roentgenol 1983; 141:483–488. 250. Mafee MF, Schatz CJ The orbit. In: Som PM, Bergeron RT (eds) Head and Neck Imaging, 2nd edn. St Louis: Mosby Year Book, 1991:799. 251. Frieden IJ, Reese V, Cohen D. PHACE syndrome. The association of posterior fossa brain malformations, hemangiomas, arterial anomalies, coarctation of the aorta and cardiac defects, and eye abnormalities. Arch Dermatol 1996; 132:307–311. 252. Rossi A, Bava GL, Biancheri R, Tortori-Donati P. Posterior fossa and arterial abnormalities in patients with facial capillary haemangioma: presumed incomplete phenotypic expression of PHACES syndrome. Neuroradiology 2001; 43: 934–940. 253. Tasdemiroglu E, Patchell RA. Cerebral metastases in childhood malignancies. Acta Neurochir (Wien) 1997; 139:182– 187.

Hemolymphoproliferative Diseases and Treatment-Related Disorders

11 Hemolymphoproliferative Diseases and Treatment-Related Disorders Paolo Tortori-Donati, Andrea Rossi, and Roberta Biancheri

Introduction

CONTENTS Introduction 11.1 11.1.1 11.1.1.1 11.1.1.2 11.1.1.3 11.1.2 11.1.2.1 11.1.2.2 11.1.2.3 11.1.3 11.1.3.1 11.1.3.2 11.1.3.3 11.1.4 11.2 11.2.1 11.2.1.1 11.2.1.2 11.2.1.3 11.2.2 11.2.2.1 11.2.2.2

437

Direct CNS Involvement from Primary Disease 437 Leukemia 437 Meningeal Disease 438 Cranial Nerve Involvement 440 Chloromas and Other Leukemic Masses Lymphoma 443 Cerebral Lymphomas 443 Head and Neck Lymphomas 444 Spinal Lymphomas 445 Langerhans Cell Histiocytosis 445 Parenchymal Involvement 446 Skull Involvement 447 Spinal Involvement 447 Familial Erythrophagocytic Lymphohistiocytosis 449 CNS Disease Related to Underlying and Secondary Effects of HLD 450 Hematological and Cerebrovascular Complications 450 Hemorrhage 450 Cerebral Ischemia/Infarction 450 Dural Sinus/Venous Thrombosis 451 Infection 451 Sinusitis 451 Intracranial Infection 452

11.3 11.3.1 11.3.1.1 11.3.1.2 11.3.1.3 11.3.1.4 11.3.1.5 11.3.1.6 11.3.1.7 11.3.2 11.3.2.1 11.3.2.2 11.3.2.3 11.3.2.4 11.3.3 11.3.3.1

Treatment-related Complications 453 Radiotherapy 454 Focal Radiation Necrosis 454 Radiation Leukoencephalopathy 454 Mineralizing Microangiopathy 456 Cavernomatous Malformations 456 Neuroendocrine Pathology 457 Neuropsychological Delay 458 Second Neoplasms 459 Chemotherapy 459 Methotrexate Neurotoxicity 460 L-asparaginase Neurotoxicity 460 Cytarabine Neurotoxicity 461 Steroids 461 Bone Marrow Transplantation 461 Posterior Reversible Encephalopathy Syndrome (PRES) 462 11.3.3.2 GVHD Vasculitis 464 11.4

Conclusions 466 References

466

440

Pediatric hemolymphoproliferative diseases (HLDs) are a constellation of disorders that prominently include leukemia, lymphoma, and histiocytoses. These conditions are invariably severe, and are burdened by elevated morbidity and mortality in the absence of proper treatment. HLDs are systemic disorders, i.e., they typically affect multiple organs and systems in the human body. While, as a whole, leukemia and lymphomas account for about 40% of all malignancies in children [1], the majority of histiocytoses are not malignant. Intensive treatment regimens have resulted in a significant increase in the number of survivors, but have also disclosed new phases of the natural history of these disorders, among which is involvement of the central nervous system (CNS). CNS disease is not consistent, but when present, it is commonly associated with worsened prognosis. Moreover, current treatment modalities, including systemic therapy (i.e., combination chemotherapy) and specific CNS prophylaxis (i.e., intrathecal chemotherapy with or without cranial irradiation), are potentially neurotoxic. Schematically, a rational approach to neuroimaging of CNS manifestations of HLD basically involves three perspectives [2]: (1) direct involvement of the CNS from primary disease, (2) other CNS disease related to underlying and secondary effects of disease, and (3) treatment-related complications.

11.1 Direct CNS Involvement from Primary Disease 11.1.1 Leukemia Leukemia is the most common neoplasm in the pediatric age group, accounting for 2.7% of new cancer diagnoses and 3.7% of cancer deaths in the United States [3]. Over 25 subtypes exist according to the

437

438

P. Tortori-Donati, A. Rossi, and R. Biancheri

involved stem cell line; these are further classified into acute and chronic leukemia [2, 4]. Acute lymphoblastic leukemia (ALL) is, by far, the most common of these subtypes in the pediatric age group, accounting for 78% of all pediatric leukemia and 40 new cases per million per year in children younger than 15 years [5]. Acute myeloid leukemia (AML) accounts for 16% of pediatric leukemia, whereas the remaining 6% is represented by chronic myeloid leukemia and other specified and unspecified leukemia [4]. Owing to steadfast improvements in treatment modalities, survival for leukemic children has markedly increased over the past 40 years. Although this is true for leukemia as a whole, it is especially valid for ALL. In 1960, 5-year disease-free ALL survivors were 3%, whereas they are about 80% at present [4]. CNS complications were rare in earlier years because of rapid fatality of the disease [6], whereas at present they are seen in 25%–50% of all patients with leukemia [2]. Direct involvement of the CNS is only uncommonly the initial manifestation of disease, and occurs in only 5% of patients at the onset. More often, the CNS is involved in patients with known leukemia or as a relapse after initial remission [2]. Arrival of leukemic cells to the CNS occurs either through hematogenous mechanisms or by direct spread from adjacent infiltrated bone marrow [7]. The meninges are involved much more commonly than the nervous tissues are.

a

11.1.1.1 Meningeal Disease

Meningeal infiltration can be dural, leptomeningeal, or both. Clinical presentation is similar to other forms of meningitis, with signs of raised intracranial pressure including headache, nausea and vomiting, irritability, lethargy, and papilledema [2]. Dural infiltration is rare. It can be focal or diffuse, and appears on MRI as an abnormal dural thickening that enhances strongly and homogeneously with gadolinium administration (Fig. 11.1). Leptomeningeal infiltration is much more common. As with other forms of meningeal disease (either inflammatory or neoplastic), unenhanced MRI is generally inconclusive, and contrast material injection is mandatory for identifying pathology. In most cases, subarachnoid nodules, diffuse leptomeningeal carcinomatosis (Fig. 11.2), or both, are detected on post-contrast MR images. Neoplastic sedimentation is a third, uncommon form of leptomeningeal disease characterized by accumulation of neoplastic cells at the bottom of the thecal sac due to gravity phenomena. In this case, the normal signal intensity of CSF is replaced by an abnormal intermediate signal intensity in all

b Fig. 11.1a,b Localized dural infiltration in a 13-year-old boy with acute lymphoblastic leukemia. a Gd-enhanced axial T1-weighted image. b Gd-enhanced coronal T1-weighted image. Following gadolinium administration, MRI shows a diffusely enhancing dura, with marked focal thickening over the left inferior frontal convexity (arrow, a,b). The dura is also abnormally thickened over the vertex (arrowheads, b)

MR sequences (Fig. 11.3). Typically, neoplastic sedimentations are discovered incide