Tumors of CNS v2 Gliomas Glioblastoma Part 2

Tumors of the Central Nervous System
Tumors of the Central Nervous System

Volume 2
For other titles published in this series, go to

www.springer.com/series/8812
Tumors of the Central Nervous

System
Volume 2
Tumors of the Central

Nervous System
Gliomas: Glioblastoma (Part 2)
Edited by
M.A. Hayat
Distinguished Professor
Department of Biological Sciences,
Kean University, Union, NJ, USA
123
Editor
M.A. Hayat
Department of Biological Sciences
Kean University
Union, NJ, USA
ehayat@kean.edu
ISBN 978-94-007-0617-0
e-ISBN 978-94-007-0618-7
DOI 10.1007/978-94-007-0618-7
Springer Dordrecht Heidelberg London New York
Library of Congress Control Number: 2011923069
Springer Science+Business Media B.V. 2011
No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by
any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written
permission from the Publisher, with the exception of any material supplied specifically for the purpose of
being entered and executed on a computer system, for exclusive use by the purchaser of the work.
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)
Although touched by technology, surgical pathology always has

been, and remains, an art. Surgical pathologists, like all artists, depict
in their artwork (surgical pathology reports) their interactions with
nature: emotions, observations, and knowledge are all integrated.
The resulting artwork is a poor record of complex phenomena.
Richard J. Reed MD
Preface
The primary objective of this series, Tumors of the Central Nervous System, is to
present the readers with the most up-to-date information on the initiation, progression, recurrence, metastasis, and treatment of the CNS tumors. As in volume 1,
volume 2 has discussed in detail biomarkers and diagnosis of gliomas, especially
glioblastoma. The role of a large number of biomarkers in the diagnosis of glioblastoma is included. Advantages and limitations of the use of biomarkers for diagnosis
are presented. The role of TP53 gene mutation in the initiation and progression of
glioblastoma is presented as well as germline mutations of this gene. Role of oncogenes and tumor suppressor genes is also discussed. Also, is discussed the role of
specific genes in the resistance to drug therapy.
The importance of the use of imaging modalities (e.g., PET, CT, MRI, and SPECT)
in clinical diagnosis, treatment assessment, and recurrence determination is pointed
out. It is well established that early diagnosis is the key to cancer cure. Prognosis
is highly dependent on the stage of the disease. Thus, a simple and reliable screening
method would be of tremendous advantage. Imaging techniques in clinical practice
are used for the staging of tumors, detection of tumor recurrence, monitoring of efficacy of therapy, and differentiation between malignant and benign tissues. In this
volume, use of PET in diagnosing glioma and in assessment of biological target volume in high-grade glioma patients is explained. Also is discussed the use of MRI in
glioma surgery.
Present and future therapeutic drugs for malignant gliomas are described. The efficacy of several drugs, such as cyclosporine, interferon, heparin, and cannabinoids
in treating glioblastoma is explained. Effectiveness of therapies, such as resection,
radiation, chemotherapy, and immunotherapy, against high-grade gliomas is detailed.
Therapy for recurrent high-grade glioma with bevacizumab and irinotecan is presented. Use of dendritic cell therapy and adenoviral vectors for glioblastoma is
discussed. Brainstem gliomas are also described, so is tumor-associated epilepsy.
This work consists of 37 chapters that were contributed by 101 authors representing 16 countries. The high quality of each manuscript made my work as the editor
an easy one. Strictly uniform style of manuscript writing has been accomplished.
The results are presented in the form of both black-and-white and color images and
diagrams.
I am indebted to the contributors for their promptness in accepting my suggestions,
and appreciate their dedication and hard work in sharing their knowledge and expertise with the readers. Each chapter provides unique individual, practical knowledge
based on the expertise of a large number of researches and physicians. A vast medical
vii
viii
Preface
field such as tumors of the CNS can be discussed adequately only by a large number
of experts. It is my hope that this volume will be published expediously.
I am thankful to Dr. Dawood Farahi, Dr. Kristie Reilly, and Mr. Philip Connelly
for recognizing the importance of scholarship in an institution of higher education,
and providing resources for completing this project.
Union, New Jersey
September 2010
M.A. Hayat
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
M.A. Hayat
Part I
Biomarkers and Diagnosis . . . . . . . . . . . . . . . . . . . .
2 Gliomagenesis: Advantages and Limitations of Biomarkers . . . . .

Michel Wager, Lucie Karayan-Tapon, and Christian-Jacques Larsen
11
3 Molecular Subtypes of Gliomas . . . . . . . . . . . . . . . . . . . . .

Lonneke A.M. Gravendeel and Pim J. French
25
4 Glioblastoma: Germline Mutation of TP53 . . . . . . . . . . . . . . .

Haruhiko Sugimura, Hidetaka Yamada, Shinji Kageyama,
Yasuhiro Yamamura, Naoki Yokota, Hiroki Mori,
Moriya Iwaizumi, Kazuya Shinmura, Kiyotaka Kurachi, Toshio
Nakamura, Masaru Tsuboi, Masato Maekawa, and Tomoaki Kahyo
31
5 Familial Gliomas: Role of TP53 Gene . . . . . . . . . . . . . . . . . .

Soufiane El Hallani and Ilham Ratbi
39
6 The Role of IDH1 and IDH2 Mutations in Malignant Gliomas . . . .

Yukihiko Sonoda, Ichiyo Shibahara, Ryuta Saito,
Toshihiro Kumabe, and Teiji Tominaga
47
7 Malignant Glioma: Isocitrate Dehydrogenases 1 and 2 Mutations . .

Zachary J. Reitman and Hai Yan
53
8 Metabolic Differences in Different Regions of Glioma Samples . . . .

Francisca M. Santandreu, Jordi Oliver, and Pilar Roca
63
73
Glioblastoma Patients: Role of Methylated MGMT . . . . . . . . . .

Giulio Metro and Alessandra Fabi
10 Brain Tumor Angiogenesis and Glioma Grading:

Role of Tumor Blood Volume and Permeability Estimates
Using Perfusion CT . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Rajan Jain
81
ix
Contents
11
Vasculogenic Mimicry in Glioma . . . . . . . . . . . . . . . . . . . .

Zhong-Ping Chen and Yin-Sheng Chen
12
Newly Diagnosed Glioma: Diagnosis Using Positron

Emission Tomography with Methionine and Fluorothymidine . . . .
Nobuyuki Kawai, Yoshihiro Nishiyama, and Takashi Tamiya
103
Role of Diffusion Tensor Imaging in Differentiation

of Glioblastomas from Solitary Brain Metastases . . . . . . . . . . .
Sumei Wang, Harish Poptani, Elias R. Melhem, and Sungheon Kim
113
13
93
14
131 I-TM-601
SPECT imaging of Human Glioma . . . . . . . . . . . .

Adam N. Mamelak and David Hockaday
123
15
Assessment of Biological Target Volume Using Positron

Emission Tomography in High-Grade Glioma Patients . . . . . . . .
Habib Zaidi and Srinivasan Senthamizhchelvan
131
16
Skin Metastases of Glioblastoma . . . . . . . . . . . . . . . . . . . .

Rebecca Senetta and Paola Cassoni
Part II
17
18
19
20
143
Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
151
Diffuse Low-Grade Gliomas: What Does Complete

Resection Mean? . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Johan Pallud
153
Quantitative Approach of the Natural Course of Diffuse

Low-Grade Gliomas . . . . . . . . . . . . . . . . . . . . . . . . . . .
Johan Pallud and Emmanuel Mandonnet
163
Impact of Extent of Resection on Outcomes in Patients

with High-Grade Gliomas . . . . . . . . . . . . . . . . . . . . . . . .
Debraj Mukherjee and Alfredo Quinones-Hinojosa
173
Glioma Surgery: Intraoperative Low Field Magnetic

Resonance Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Christian Senft
181
21
Low-Grade Gliomas: Intraoperative Electrical Stimulations . . . . .

Hugues Duffau
189
22
Malignant Gliomas: Present and Future Therapeutic Drugs . . . . .

Linda Coate and Warren Mason
207
23
Recurrent Malignant Glioma Patients: Treatment

with Conformal Radiotherapy and Systemic Therapy . . . . . . . . .
Abhirami Hallock and Lauren VanderSpek
24
Glioblastoma: Boron Neutron Capture Therapy . . . . . . . . . . . .

Tetsuya Yamamoto, Kei Nakai, and Hiroaki Kumada
25
Glioblastoma: Anti-tumor Action of Cyclosporin A

and Functionally Related Drugs . . . . . . . . . . . . . . . . . . . . .
Bozena Kaminska, Magdalena Tyburczy,
Konrad Gabrusiewicz, and Malgorzata Sielska
215
229
241
Contents
xi
26 Glioblastoma Patients: Chemotherapy with Cisplatin,

Temozolomide and Thalidomide . . . . . . . . . . . . . . . . . . . . .
Fable Zustovich and Giuseppe Lombardi
255
27 Glioblastoma: Role of Galectin-1 in Chemoresistance . . . . . . . . .

Florence Lefranc and Robert Kiss
261
28 Glioma-Initiating Cells: Interferon Treatment . . . . . . . . . . . . .

Atsushi Natsume, Masasuke Ohno, Kanako Yuki,
Kazuya Motomura, and Toshihiko Wakabayashi
269
29 Glioblastoma: Anti-tumor Action of Natural and Synthetic

Cannabinoids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Aleksandra Ellert-Miklaszewska, Iwona Ciechomska,
and Bozena Kaminska
277
30 Patients with Recurrent High-Grade Glioma: Therapy

with Combination of Bevacizumab and Irinotecan . . . . . . . . . .
Benedikte Hasselbalch, Ulrik Lassen, and Hans Skovgaard Poulsen
289
31 Monitoring Gliomas In Vivo Using Diffusion-Weighted

MRI During Gene Therapy-Induced Apoptosis . . . . . . . . . . . .
Timo Liimatainen, Olli Grhn, and Kimmo Lehtimki
301
32 High-Grade Gliomas: Dendritic Cell Therapy . . . . . . . . . . . . .

Hilko Ardon, Steven De Vleeschouwer, Frank Van Calenbergh,
and Stefaan W. Van Gool
313
33 Glioblastoma Multiforme: Use of Adenoviral Vectors . . . . . . . . .

Thomas Wirth, Haritha Samaranayake, and Seppo Yl-Herttuala
335
34 Fischer/F98 Glioma Model: Methodology . . . . . . . . . . . . . . .

David Fortin
349
35 Cellular and Molecular Characterization of Anti-VEGF

and IL-6 Therapy in Experimental Glioma . . . . . . . . . . . . . .
Sophie Javerzat and Martin Hagedorn
36 Adult Brainstem Gliomas: Diagnosis and Treatment . . . . . . . . .
Florence Laigle-Donadey and Jean-Yves Delattre
361
371
37 The Use of Low Molecular Weight Heparin in the Treatment

and Prevention of Thromboembolic Disease in Glioma Patients . . .
Bo H. Chao and H. Ian Robins
379
Part III
Prognosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
385
38 Brainstem Gliomas: An Overview . . . . . . . . . . . . . . . . . . . .

Marco Antonio Lima
387
39 Tumor-Associated Epilepsy in Patients with Glioma . . . . . . . . .

Anja Smits and Anette Storstein
397
40 Brain Tumors Arising in the Setting of Chronic Epilepsy . . . . . . .

Richard A. Prayson
407
xii
41
42
Contents
Low-Grade Gliomas: Role of Relative Cerebral Blood

Volume in Malignant Transformation . . . . . . . . . . . . . . . . . .
Neil Upadhyay and Adam D. Waldman
417
Angiocentric Glioma-Induced Seizures: Lesionectomy . . . . . . . .

Dave F. Clarke and Timothy M. George
425
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
435
Contents of Volume 1
1 Introduction
2 Molecular Classification of Gliomas
3 Glioblastoma: Endosialin Marker for Pericytes
4 Glioma Grading Using Cerebral Blood Volume Heterogeneity
5 The Role of Ectonucleotidases in Glioma Cell Proliferation
6 Gliomas: Role of Monoamine Oxidase B in Diagnosis
7 Glioma: Role of Integrin in Pathogenesis and Therapy
8 Proton Magnetic Resonance Spectroscopy in Intracranial Gliomas
9 Infiltration Zone in Glioma: Proton Magnetic Resonance
Spectroscopic Imaging
10 Malignant Gliomas: Role of E2F1 Transcription Factor
11 The Role of Glucose Transporter-1 (GLUT-1) in Malignant Gliomas
12 Malignant Gliomas: Role of Platelet-Derived Growth Factor
Receptor A (PDGFRA)
13 Molecular Methods for Detection of Tumor Markers in Glioblastomas
14 Role of MGMT in Glioblastomas
15 Glioblastomas: Role of CXCL12 Chemokine
16 Cell Death Signaling in Glioblastoma Multiforme: Role of
the Bcl2L12 Oncoprotein
17 Glioblastoma Multiforme: Role of Polycomb Group Proteins
18 Glioblastoma Multiforme: Role of Cell Cycle-Related Kinase
Protein (Method)
19 Markers of Stem Cells in Gliomas
20 Efficient Derivation and Propagation of Glioblastoma
Stem-Like Cells Under Serum-Free Conditions Using the
Cambridge Protocol
xiii
xiv
Contents of Volume 1
21
Glioma Cell Lines: Role of Cancer Stem Cells
22
Glioblastoma Cancer Stem Cells: Response to Epidermal

Growth Factor Receptor Kinase Inhibitors
23
Low- and High-Grade Gliomas: Extensive Surgical Resection
24
Brainstem Gangliogliomas: Total Resection and Close

Follow-Up
25
Glioblastoma: Temozolomide-Based Chemotherapy
26
Drug-Resistant Glioma: Treatment with Imatinib Mesylate

and Chlorimipramine
27
Glioblastoma Multiforme: Molecular Basis of Resistance to Erlotinib
28
Enhanced Glioma Chemosensitivity
29
Malignant Glioma Patients: Anti-Vascular Endothelial

Growth Factor Monoclonal Antibody, Bevacizumab
30
Aggravating Endoplasmic Reticulum Stress by Combined

Application of Bortezomib and Celecoxib as a Novel
Therapeutic Strategy for Glioblastoma
31
Targeted Therapy for Malignant Gliomas
32
Glioblastomas: HER1/EGFR-Targeted Therapeutics
33
Epidermal Growth Factor Receptor Inhibition as a

Therapeutic Strategy for Glioblastoma Multiforme
34
Role of Acyl-CoA Synthetases in Glioma Cell Survival and

Its Therapeutic Implication
35
Malignant Glioma Patients: Combined Treatment with

Radiation and Fotemustine
36
Malignant Glioma Immunotherapy: A Peptide Vaccine from

Bench to Bedside
37
Malignant Glioma: Chemovirotherapy
38
Intracranial Glioma: Delivery of an Oncolytic Adenovirus
39
Use of Magnetic Resonance Spectroscopy Imaging (MRSI)

in the Treatment Planning of Gliomas
40
Malignant Glioma Cells: Role of Trail-Induced Apoptosis
41
Long-Term Survivors of Glioblastoma
42
Glioblastoma Patients: p15 Methylation as a Prognostic Factor
Contributors
Hilko Ardon Department of Neurosurgery, Laboratory of Experimental

Immunology, Catholic University Leuven, Leuven, Belgium,
hilkoardon@hotmail.com
Paola Cassoni Department of Biomedical Sciences and Human Oncology,
University of Turin, Turin, Italy, Paola.cassoni@unito.it
Bo H. Chao Department of Medicine, University of Wisconsin School of
Medicine and Public Health, Madison, WI, USA, bchao@uwhealth.org
Yin-Sheng Chen Department of Neurosurgery, Cancer Center, Sun Yat-sen
University, Guangzhou, China, doctor.ch@163.com
Zhong-Ping Chen Department of Neurosurgery, Cancer Center, Sun Yat-sen
University, Guangzhou, China, chenzp57@mail.sysu.edu.cn
Iwona Ciechomska Department of Cell Biology, Nencki Institute of
Experimental Biology, Warsaw, Poland, jcjech@nencki.gov.pl
Dave F. Clarke Division of Neurology, Department of Pediatrics, Dell Childrens
Medical Center, Austin, TX, USA, dfclarke@seton.org
Linda Coate Department of Medical Oncology, Princess Margaret Hospital,
Toronto, ON, Canada, Linda.Coate@uhn.on.ca
Steven De Vleeschouwer Department of Neurosurgery, Laboratory of
Experimental Immunology, Catholic University Leuven, Leuven, Belgium,
steven.devleeschouwer@uzleuven.be
Jean-Yves Delattre Service de Neurologie Mazarin, Hpital de la
Piti-Salptrire APHP, 75651 Paris Cedex 13, France,
jean-yves.delattre@psl.aphp.fr
Hugues Duffau Department of Neurosurgery and INSERM U1051, Institute for
Neurosciences of Montpellier, Montpellier University Medical Center, Hpital Gui
de Chauliac, CHU Montpellier 34295, Montpellier Cedex 5, France,
h-duffau@chu-montpellier.fr
Soufiane El Hallani Cancer Imaging Unit, Integrative Oncology Department,
British Columbia Cancer Research Centre, Vancouver, BC, Canada V5Z 1L3,
Soufiane.elhallani@yahoo.ca
xv
xvi
Aleksandra Ellert-Miklaszewska Department of Cell Biology, Nencki Institute

of Experimental Biology, Warsaw, Poland, aellert@nencki.gov.pl
Alessandra Fabi Division of Medical Oncology, Regina Elena National Cancer
Center Institute, Rome, Italy, alessandra.fabi@virgilio.com
David Fortin Division of Neurosurgery, Department of Surgery, Sherbrooke
University, Sherbrooke, QC, Canada, David.fortin@usherbrooke.ca
Pim J. French Department of Neurology, Josephine Nefkens Institute, Erasmus
Medical Centre, Rotterdam, The Netherlands, p.french@erasmusmc.nl
Konrad Gabrusiewicz Laboratory of Transcription Regulation, Department of
Cell Biology, Nencki Institute of Experimental Biology, Warsaw, Poland,
k.gabrusiewicz@nenck.gov.pl
Timothy M. George Division of Neurosurgery, Department of Pediatrics, Dell
Childrens Medical Center, Austin, TX, USA, TMGeorge@seton.org
Lonneke A.M. Gravendeel Department of Neurology, Josephine Nefkens
Institute, Erasmus Medical Centre, Rotterdam, The Netherlands,
a.gravendeel@erasmusmc.nl
Olli Grhn Department of Neurobiology, University of Eastern Finland, Kuopio,
Finland, olli.grohn@uef.fi
Martin Hagedorn INSERM U1029, Universit Bordeaux 1, Talence, Cedex,
France, m.hagedorn@angio.u-bordeaux1.fr
Abhirami Hallock Department of Radiation Oncology, London Regional Cancer
Program, University of Western Ontario, E. London, ON, Canada,
Abhirami.hallock@londonhospitals.ca
Benedikte Hasselbalch Department of Radiation Biology, The Finsen Center, Sec
6321, Copenhagen University Hospital, Copenhagen, Denmark,
Benedikte.hasselbalch@rh.regionh.dk
M.A. Hayat Department of Biological Sciences, Kean University, Union, NJ,
USA, ehayat@kean.edu
David Hockaday Department of Neurosurgery, Cedars-Sinai Medical Center, Los
Angeles, CA, USA
Moriya Iwaizumi Department of Investigative Pathology I, Hamamatsu
University School of Medicine, Higashi-ward, Hamamatsu, Japan,
iwa_bonheur_mori@yahoo.co.jp
Rajan Jain Division of Neuroradiology, Department of Radiology and
Department of Neurosurgery, Henry Ford Health System, Detroit, MI, USA,
rajanj@rad.hfh.edu
Sophie Javerzat INSERM U1029, Universit Bordeaux 1, Talence, Cedex,
France, s.javerzat@angio.u-bordeaux1.fr
Shinji Kageyama Department of Investigative Pathology I, Hamamatsu
a_9040_3891_amegak@hotmail.com
Contributors
Contributors
xvii
Tomoaki Kahyo Department of Investigative Pathology I, Hamamatsu University

School of Medicine, Higashi-ward, Hamamatsu, Japan, kahyo@hama-med.ac.jp
Bozena Kaminska Laboratory of Transcription Regulation, Department of Cell
Biology, Nencki Institute of Experimental Biology, Warsaw, Poland,
bozenakk@nencki.gov.pl
Lucie Karayan-Tapon Department of Molecular Biology, Poitiers University
Hospital, University of Poitiers Medical School, Poitiers, France,
l.karayan-tapon@chu-poitiers.fr
Nobuyuki Kawai Department of Neurological Surgery, Faculty of Medicine,
Kagawa University, Kita-gun, Kagawa, Japan, nobu@kms.ac.jp
Sungheon Kim Department of Radiology, Center for Biomedical Imaging,
New York University School of Medicine, New York, NY, USA,
Sungheon.Kim@nyumc.org
Robert Kiss Laboratoire de Toxicologie, Facult de Pharmacie, Universit Libre
de Bruxelles (ULB), Brussels, Belgium, rkiss@ulb.ac.be
Toshihiro Kumabe Department of Neurosurgery, Tohoku University Graduate
School of Medicine, Aoba-ku, Sendai-shi, Miyagi, Japan,
kuma@nsg.med.tohoku.ac.jp
Hiroaki Kumada Proton Medical Research Center, University of Tsukuba,
Tsukuba City, Ibaraki, Japan, kumada@pmrc.tsukuba.ac.jp
Kiyotaka Kurachi Department of Investigative Pathology I, Hamamatsu
kurachi1@hama-med.ac.jp
Florence Laigle-Donadey Service de Neurologie Mazarin, Hpital de la
Piti-Salptrire APHP, Paris Cedex 13, France,
Florence.laigle-donadey@psl.aphp.fr
Christian-Jacques Larsen Poitiers University Hospital, University of Poitiers
Medical School, Poitiers, France, larsenc@ligue-cancer.net
Ulrik Lassen Department of Radiation Biology, The Finsen Center, Sec 6321,
Copenhagen University Hospital, Copenhagen, Denmark,
ulrik.lassen@rh.regionh.dk
Florence Lefranc Laboratoire de Toxicologie, Facult de Pharmacie, Universit
Libre de Bruxelles (ULB), Brussels, Belgium, fllefran@ulb.ac.be
Kimmo Lehtimki Department of Biotechnology and Molecular Medicine,
University of Eastern Finland, Kuopio, Finland, kimmo.lehtimaki@uef.fi
Timo Liimatainen Department of Biotechnology and Molecular Medicine,
University of Eastern Finland, Kuopio, Finland, timo.liimatainen@uef.fi
Marco Antonio Lima Department of Neurosurgery, Instituto Nacional de
Cncer-INCa, Centro, Rio de Janeiro, Brazil, mlima@inca.gov.br
Giuseppe Lombardi Oncologia Medica 1, I.O.V. IRCCS, Ospedale Busonera,
Padova, Italy, giuseppe.lombardi@ioveneto.it
xviii
Masato Maekawa Department of Investigative Pathology I, Hamamatsu

mmaekawa@hama-med.ac.jp
Adam N. Mamelak Department of Neurosurgery, Cedars-Sinai Medical Center,
Los Angeles, CA, USA, Adam.Mamelak@cshs.org
Emmanuel Mandonnet Laboratoire Kastler Brossel de lens, Paris Cedex 5,
France, mandonnet@mac.com
Warren Mason Department of Medical Oncology, Princess Margaret Hospital,
Toronto, ON, Canada, warren.mason@uhn.on.ca
Elias R. Melhem Division of Neuroradiology, Department of Radiology, Hospital
of the University of Pennsylvania, Philadelphia, PA, USA,
Elias.Melhem@uphs.upenn.edu
Giulio Metro Division of Medical Oncology, Regina Elena National Cancer
Center Institute, Rome, Italy, giulio.metro@yahoo.com
Hiroki Mori Department of Investigative Pathology I, Hamamatsu University
School of Medicine, Higashi-ward, Hamamatsu, Japan, mori_h@hmedc.or.jp
Kazuya Motomura Department of Neurosurgery, Nagoya University School of
Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan,
kmotomura@med.nagoya-u.ac.jp
Debraj Mukherjee Maxine Dunitz Neurosurgical Institute, Cedars-Sinai Medical
Center, Los Angeles, California 90048, debraj.mukherjee@cshs.org
Kei Nakai Department of Neurosurgery, Institute of Clinical Medicine, University
of Tsukuba, Tsukuba City, Ibaraki, Japan, knakai@md.tsukuba.ac.jp
Toshio Nakamura Department of Investigative Pathology I, Hamamatsu
toshi38@hama-med.ac.jp
Atsushi Natsume Department of Neurosurgery, Nagoya University School of
anatsume@med.nagoya-u.ac.jp
Yoshihiro Nishiyama Department of Radiology, Faculty of Medicine, Kagawa
University, Kita-gun, Kagawa, Japan, nisiyosi@kms.ac.jp
Masasuke Ohno Department of Neurosurgery, Nagoya University School of
tennjikunezumi@gmail.com
Jordi Oliver Universitat de les Illes Balears, Palma de Mallorca, Spain,
Jordi.oliver@uib.es
Johan Pallud Service de Neurochirurgie, Hpital Sainte-Anne, Paris Cedex 14,
France, johanpallud@hotmail.com
Harish Poptani Division of Neuroradiology, Department of Radiology, Hospital
of the University of Pennsylvania, Philadelphia, PA, USA,
Harish.Poptani@uphs.upenn.edu
Contributors
Contributors
xix
Hans Skovgaard Poulsen Department of Radiation Biology, The Finsen Center,

Sec 6321, Copenhagen University Hospital, Copenhagen, Denmark,
hans.skovgaard.poulsen@rh.regionh.dk
Richard A. Prayson Section of Neuropathology, Department of Anatomic
Pathology, Cleveland Clinic, Cleveland, OH, USA, PRAYSOR@ccf.org
Alfredo Quinones-Hinojosa Department of Neurosurgery, The Johns Hopkins
School of Medicine, Baltimore, MD, USA, aquinon2@jhmi.edu
Ilham Ratbi Department of Medical Genetics, Human Genomic Center, Faculty
of Medicine and Pharmacy, University Mohamed V Souissi, National Institute of
Health, Rabat, Marocco, ilhamratbi@yahoo.fr
Zachary J. Reitman The Pediatric Brain Tumor Foundation Institute, The Preston
Robert Tisch Brain Tumor Center, Duke University Medical Center, Durham, NC,
USA; The Department of Pathology, Duke University Medical Center, Durham, NC,
USA, zjr@duke.edu
H. Ian Robins Department of Medicine, Human Oncology, and Neurology,
University of Wisconsin School of Medicine and Public Health, Madison, WI, USA,
hirobins@wisc.edu
Pilar Roca Department de Biologia Fonamental i Cincies de la Salut. Ed.
Guillem Colom, Universitat de les Illes Balears, Carretera Valldemossa Km 7.5,
Palma de Mallorca, 07122, Balearic Islands, Spain, pilar.roca@uib.es
Ryuta Saito Department of Neurosurgery, Tohoku University Graduate School
of Medicine, Aoba-ku, Sendai-shi, Miyagi, Japan, saito2006@jupiter.ocn.ne.jp
Haritha Samaranayake Department of Biotechnology and Molecular Medicine,
A. I. Virtanen Institute, University of Kuopio, Kuopio, Finland,
Haritha.Samaranayake@uef.fi
Francisca M. Santandreu Universitat de les Illes Balears, Palma de Mallorca,
Spain, f.m.santandreu@hotmail.com
Rebecca Senetta Department of Biomedical Sciences and Human Oncology,
University of Turin, Turin, Italy, rebesenetta@gmail.com
Christian Senft Department of Neurosurgery, Johann Wolfgang
Goethe-University, Frankfurt, Germany, c.senft@med.uni-frankfurt.de
Srinivasan Senthamizhchelvan Division of Nuclear Medicine, The Russell
H. Morgan Department of Radiology and Radiological Science, Johns Hopkins
University School of Medicine, Baltimore, MD, USA, ssriniv9@jhmi.edu
Ichiyo Shibahara Department of Neurosurgery, Tohoku University Graduate
School of Medicine, Aoba-ku, Sendai-shi, Miyagi, Japan, Ichiyooo@aol.com
Kazuya Shinmura Department of Investigative Pathology I, Hamamatsu
kzshinmu@akiha.hama-med.ac.jp
Malgorzata Sielska Laboratory of Transcription Regulation, Department of Cell
Biology, Nencki Institute of Experimental Biology, Warsaw, Poland,
m.porycka@nencki.gov.pl
xx
Anja Smits Department of Neuroscience, Neurology, University Hospital

Uppsala, Uppsala, Sweden, Anja.smits@neuro.uu.se
Yukihiko Sonoda Department of Neurosurgery, Tohoku University Graduate
School of Medicine, Aoba-ku, Sendai-shi, Miyagi, Japan,
sono@nsg.med.tohoku.ac.jp
Anette Storstein Department of Neuroscience, Haukeland University Hospital,
N5021 Bergen, Norway, anette.storstein@helse-bergen.no
Haruhiko Sugimura Department of Investigative Pathology I, Hamamatsu
hsugimur@hama-med.ac.jp
Takashi Tamiya Department of Neurological Surgery, Faculty of Medicine,
Kagawa University, Kita-gun, Kagawa, Japan, tamiya@kms.ac.jp
Teiji Tominaga Department of Neurosurgery, Tohoku University Graduate School
of Medicine, Aoba-ku, Sendai-shi, Miyagi, Japan, tomi@nsg.med.tohoku.ac.jp
Masaru Tsuboi Department of Investigative Pathology I, Hamamatsu University
School of Medicine, Higashi-ward, Hamamatsu, Japan, m-tsuboi@anpyo.or.jp
Magdalena Tyburczy Laboratory of Transcription Regulation, Department of
Cell Biology, Nencki Institute of Experimental Biology, Warsaw, Poland,
m.tyburczy@nencki.gov.pl
Neil Upadhyay Institute of Clinical Science, Imperial College London, Charing
Cross Hospital, W68RF, London, UK, neil.upadhyay@imperial.ac.uk
Frank Van Calenbergh Department of Neurosurgery, Catholic University
Leuven, Leuven, Belgium, frank.vancalenbergh@uzleuven.be
Lauren VanderSpek Department of Radiation Oncology, London Regional
Cancer Program, University of Western Ontario, E. London, ON, Canada,
Lauren.Vanderspek@providence.org
Stefaan W. Van Gool Department of Pediatrics and Laboratory of Experimental
Immunology, Catholic University Leuven, Leuven, Belgium,
stefann_vangool@uzleuven.be
Michel Wager Poitiers University Hospital, University of Poitiers Medical
School, Poitiers, France, m.wager@chu-poitiers.fr
Toshihiko Wakabayashi Department of Neurosurgery, Nagoya University School
of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan,
wakabat@med.nagoya-u.ac.jp
Adam D. Waldman Institute of Clinical Science, Imperial College London,
Charing Cross Hospital, W68RF, London, UK, adam.waldman@csc.mrc.ac.uk
Sumei Wang Division of Neuroradiology, Department of Radiology, Hospital of
the University of Pennsylvania, Philadelphia, PA, USA,
Sumei.Wang@uphs.upenn.edu
Thomas Wirth Department of Biotechnology and Molecular Medicine, A. I.
Virtanen Institute, University of Kuopio, Kuopio, Finland, Thomas.Wirth@uef.fi
Contributors
Contributors
xxi
Hidetaka Yamada Department of Investigative Pathology I, Hamamatsu

h-yamada@hama-med.ac.jp
Tetsuya Yamamoto Department of Neurosurgery and Radiation Oncology,
Institute of Clinical Medicine, University of Tsukuba, Tsukuba City, Ibaraki, Japan,
tetsu_tsukuba@yahoo.co.jp
Yasuhiro Yamamura Department of Investigative Pathology I, Hamamatsu
yas.hero@hama-med.ac.jp
Hai Yan The Pediatric Brain Tumor Foundation Institute, The Preston Robert
Tisch Brain Tumor Center, Duke University Medical Center, Durham, NC, USA;
The Department of Pathology, Duke University Medical Center, Durham, NC, USA,
Yan0002@mc.duke.edu
Seppo Yl-Herttuala Department of Biotechnology and Molecular Medicine,
A. I. Virtanen Institute, University of Kuopio, Kuopio, Finland,
Seppo.Ylaherttuala@uef.fi
Naoki Yokota Department of Investigative Pathology I, Hamamatsu University
School of Medicine, Higashi-ward, Hamamatsu, Japan, yokota@syck.jp
Kanako Yuki Department of Neurosurgery, Nagoya University School of
yu_ki_knk@yahoo.co.jp
Habib Zaidi Division of Nuclear Medicine, Geneva University Hospital,
Geneva 4, Switzerland, Habib.zaidi@hcuge.ch
Fable Zustovich Oncologia Medica 1, I.O.V. IRCCS, Ospedale Busonera,
Padova, Italy, fable.zustovich@ioveneto.it
Chapter 1
Introduction
M.A. Hayat
In developed countries cancer is the second leading cause of death exceeded only by cardiovascular
diseases. There are more than 100 types of cancers
that can inflict any part of the body. In 2005, 7.6
million people died of cancer, which constitutes 13%
of the 58 million deaths worldwide. In the global population exceeding 6 billion in the year 2002, there
were approximately 10.9 million new cancer cases, 6.7
million cancer deaths, and 22.4 million surviving from
cancer diagnosed in the previous 5 years. In 2020, it
is expected that the worlds population will increase
to 7.5 billion, with 15 million new cancer cases and
12 million cancer deaths. Approximately, 1.4 million
new cases of cancer and 550,000 cancer deaths were
reported in the United States in 2008 (Am. Cancer
Soc.), These data amount to 1500 deaths caused
by cancer every day in the United States. In 2006
an estimated 19,000 new cases of brain tumors and
13,000 deaths were reported in the United States. This
figure accounts for 1.4% of all cancer cases and
2.3% of all cancer cases that cause death. More than
10,000 Americans die annually from glioblastoma.
Survival for this disease has not changed much in three
decades. Since 1970, the number of cancer survivors
has increased four-fold, with cancer survivors representing 3.5% of the United States population and
5-years survival rates increasing into the 60% range.
These raises invite issues related to long-term and late
effects of cancer treatment and the realization that
cancer survivors represent 16% of all new primary
cancers.
M.A. Hayat ()

Department of Biological Sciences, Kean University, Union, NJ
07083, USA
e-mail: ehayat@kean.edu
Glioblastoma
Gliomas can arise either spontaneously (primary
glioma) or can progress from a lower-grade to a highergrade (glioblastoma) of tumor. Malignant glioma is
the most common tumor of the CNS, and glioblastoma is the most malignant form. Glioblastoma is
characterized by rapid, highly invasive growth, extensive neovascularisation, and high mortality. The key
reason for the lack of successful therapy is the infiltration of single tumor cells into the surrounding brain
parenchyma cells, preventing complete glioblastoma
resection. This process is facilitated by two related processes: (1) angiogenesis, the sprouting of new blood
vessels from preexisting vasculature in response to
external chemical stimulation, and (2) vasculogenesis, the reorganization of randomly distributed cells
into a blood vessel network. Tumor cells can also
acquire blood supply through other ways to escape
conventional antiangiogenesis. In other words, blood
vessels are formed by tumor cells instead of endothelial cells. This novel concept in tumor vascularization
is termed as vasculogenic mimicry, which is the ability of aggressive tumor cells to express endotheliumassociated genes and form extracellular matrix-rich
vasculogenic-like networks in three-dimensional culture. Such networks recapitulate embryonic vasculogenesis, and have been observed in human aggressive
tumors such as glioblastoma (El Hallani et al., 2010,
and Chapter 11, in this volume).
Glioblastoma can be divided into two subtypes
based on amplication and mutation of different
genes, and characterization of molecular pathways has
opened new venues to targeted therapies based on the
individual genetic signature of the tumor (Ohgaki and
M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2,

DOI 10.1007/978-94-007-0618-7_1, Springer Science+Business Media B.V. 2011
Kleihues, 2007). Two main subtypes on the basis of

genetic differences are: (1) Primary glioblastomas typically occur in patients alder than 50 years of age, and
are characterized by epidermal growth factor receptor
(EGFR) amplification and mutations, loss of heterozygosity of chromosome log, deletion of the phosphatase
and tension homologue on chromosome 10 (PTEN),
and p16 deletion. (2) Secondary glioblastomas occur
in younger patients as low-grade or anaplastic astrocytomas that are transformed over a period of several years into glioblastoma multiforme. Secondary
glioblastomas are much less common than primary
glioblastoma, and are characterized by mutations in
the TP53 tumor suppressor gene, overexpression of
the platelet-derived growth factor receptor (PDGFR),
abnormalities in the p16 and retinoblastoma (Rb) pathways, and loss of heterozygosity of chromosome loq.
Although glioblastoma is the most common and
aggressive type of glioma and has the poorest survival,
a small percentage of patients survive longer than
the established median. Identification of genetic variants that influence long term survival of such patients
may provide insight into tumor biology and treatment. A recent study by Liu et al. (2010) indicates
that polymorphisms in the LIG4, BTBD2, HMGA2,
and RTEL1 genes are associated with the survival of
glioblastoma patients. LIG4 and BTBD2 are predictors of short-term survival, while CCDC26, HMGA2,
and RTEL1 are predictors of long-term survival. In
general, older patients (>50 years) have the worst survival (1.2 years), while younger patients may show
a median survival-term of 7.8 years. These genes are
known to be involved in the double strand break repair
pathway.
It is known that glioblastoma responds poorly
to conventional therapies. Glioblastoma cells thrive
despite an irregular blood supply and a frequently
hypoxic microenvironment. Compensatory mechanisms, including glucose uptake and glycolytic activity, are enhanced in these tumors. In other words,
glioblastoma cells, posses sufficient glycolytic capacity (more than that in normal brain tissue), supporting
their migration even without mitochondrial involvement (Beckner et al., 1990) A recent study has identified ATP citrate lyase (ACLY) as a positive regulator
of glycolysis in glioblastomas (Beckner et al., 2010).
ACLY can be targeted to suppress hypoxic cell migration and invasion, and restore glycolytic inhibition.
This inhibition of hypotoxic tumor cells potentially
M.A. Hayat
complement antiangiogenesis therapies that compromise the blood supply to tumor cells.
Treatment
The three most common treatments are resection, radiation, and chemotherapy, or a combination of these
methods. According to Sandmair et al. (2000), when
radiation is utilized following surgical resectioning
of the tumor, the median survival time may increase
from 14 to 40 weeks. Maximal resection of brain
glioma is usually the first or second treatment choice.
Radiation is the alternative treatment. The goal is to
maximize the effectiveness of resection, while minimize the operative risk. To accomplish this goal
is not easy. Active migration of glioblastoma cells
through the narrow extracellular spaces in the brain
makes them elusive targets for surgical management.
Glioma cells are self-propelled, and are able to adjust
their shape and volume rapidly as they invade the
brain parenchyma. The infiltrative nature of malignant gliomas results in poor demarcation of malignant
boundaries. Another reason is the frequent location of
supratentorial gliomas near or within eloquent areas.
Consequently, the advantage of maximal resection in
all cases is controversial. Image-guided surgery utilizing fluorescence with 5-aminolevulinic acid, neuronavigation, and intraoperative MRI has enabled more
complete resectioning of contrast-enhancing tumors
(Stummer et al., 2006; Nimsky et al., 2006)
Chemotherapy in conjunction with surgery and radiation can increase the estimated survival of patients by
10.1% at 1 year and 8.6% at 2 years, but these rates
apply to lower-grade glioma tumors. The current recommended chemotherapeutic agent is alkylating drug
temozolomide. The 2-year survival rate of patients
with newly diagnosed glioblastoma treated with radiotherapy and temozolomide is 26.5%, compared with
10.4% for radiotherapy alone (Stupp et al., 2005).
Telozolomide also exerts antitumor affects by
impairing angiogenic process. In vitro and in vivo studies have shown antiangiogenic activity by this drug
even when it is used alone (Mathieu et al., 2008).
The efficacy can be further enhanced by combining this treatment with bevacizumab; the later also
has an antiangiogenic effect, although with a different mechanism of action. Antiangiogenic compounds
Introduction
also increase the therapeutic benefits of radiotherapy (Nieder et al., 2006). A phase 2 pilot study of
bevacizumab in combination with telozolomide and
regional radiotherapy for the treatment of patients with
newly diagnosed glioblastoma recently reported that
toxicities were acceptable to continue enrollment and
a preliminary analysis of efficacy showed encouraging mean progression-free survival (Lai et al., 2008).
It is concluded that a range of side-effects, including
post-therapeutic neurological deterioration, can commonly or uncommonly are experienced by patients
undergoing chemotherapy.
To overcome some of the limitations mentioned
above, intraoperative eletrostimulation can be used
(Duffau, 2007, also, see Chapter 22, in this volume).
The purpose is to understand the interindividual
anatomical-functional variability in the case of glioma
patients. In order to tailor the resection for each patient,
it is mandatory to study the cortical functional organization, the affective connectivity, and the potentiality
for brain plasticity.
Another limitation is the innate inter individual
prognostic variability encountered among malignant
glioma patients. This limitation can be overcome by
carrying out analysis of prognostic factors, which can
predict the outcome of a therapy among diagnosed
malignant glioma patients. Such an information can
affect the design and conduct of clinical trials in the
case of patients with recurrent glioma. Phase II trials play a critical role in the assessment of novel
therapeutic approaches.
Factors associated with an increased risk of death
are old age (50 years or older), lower karnofsky performance score (<80), initial and on study histological data of glioblastoma multiforme, corticosteroid
use, shorter time from original diagnosis to recurrence, and tumor outside frontal lobe (Carson et al.,
2007). However, patients differ with respect to their
characteristics, including age, performance status, histological data, time from initial diagnosis to recurrence, exact location of tumor, whether the tumor is
resectable, number and type of prior therapies, and
use of concomitant medication (e.g., anticonvulsants).
Thus, patients with recurrent gliomas have significantly different prognoses depending on their characteristics including those mentioned above.
The importance of the role of immune-signaling
in the regulation and function of resident neural
stem cells in the CNS is beginning to be understood
(Carpentier and Palmer, 2009). The complexity of the

immune signaling becomes apparent considering that
some aspects of this signaling are beneficial by promoting intrinsic plasticity and replacement of injured
cells, while others inhibit the regenerative response
that might restore or replace neural networks lost in
disease. The broad implication is that tissue environment may influence the activity and fate of endogenous or transplanted neural stem cells. Immunotherapy
alone or in combination with other treatments has
also been tried with some success. For example,
bevacizumab has shown some promise for recurrent
glioblastoma, although recurrence is common following treatment with this antibody. Thus, this treatment
is transient.
Another approach is to develop novel drugs that target glioblastoma cells. Role of Na+ /K+ -ATPAse (the
sodium pump, a membrane protein) in the migration of glioblastoma cells is interesting. It has been
shown that the 1 subunit of the ubiquitous sodium
pump is over-expressed in glioblastoma (Lefranc et al.,
2008). This study demonstrates that UNBS1450 drug,
a cardionolide compound that inhibits sodium pump,
impairs glioblastoma cell migration through disorganization of the actin cytoskeleton and has potent
proautophagic effects on glioblastoma cell lines. Thus,
sodium pump may prove to be another useful target
for drug development against this disease. The importance of prior laboratory prediction of individual drug
response cannot be overemphasized. Drug sensitivity
is determined by multiple genes and accessibility to the
target tumor. The complexity of mechanisms involved
in drug sensitivity becomes apparent, considering that
gene expression profile in response to drug exposure
vary considerably among individuals even for the same
drug or regimen and tumor type.
It is known that gliomas, in their proliferative
stage, are highly vascularized tumors, and their persistence and growth depend on the pathological formation of new capillary blood vessels. Thus, angiogenesis inhibition is an efficient therapeutic strategy
for treating malignant gliomas. However, the benefit of current anti-angiogenic agents has been at best
modest. Reasons for limited effectiveness of such
agents include their short circulating half-life, doselimiting toxicity, and the noninvasive methods to monitor anti-angiogenic therapies in vivo (van Eekelen
et al., 2010). Recently, these authors have engineered
and characterized a secretable form of antiangiogenic
thrombospondin-1 (TSP-1) that targets the vascular

component of gliomas and reduces tumor blood vessel
density, resulting in the inhibition of tumor progression
and increased survival of mice bearing highly malignant human gliomas. A large number of standard and
novel treatments against malignant gliomas are discussed by other authors in this and volume 1 of this
series.
Glioblastoma tumors containing CD133-positive
cells display strong capability of resistance to
chemotherapy (temozolomide, carboplatin, VP16, and
Taxol) (Liu et al., 2006). Brain cancer stem cells
(CSCs) with such resistance also show a higher expression of drug resistance genes (e.g., BCRPI) and DNA
mismatch repair genes (e.g., MGMT) as well as antiapoptotic proteins (e.g., Bcl-2). Growing evidence
also indicates that CD133-posiotive CSCs show higher
expression of mRNA levels of inhibitors of apoptotic
proteins such as IAPs (Liu et al., 2006). Based on this
and other evidence it can be concluded that apoptotic
pathways contribute to resistance of CD133-posiotive
to chemotherapy and medical radiation. Thus, CD133positive cells can be considered as a marker for
glioblastoma CSCs that can be targeted either biochemically or immunologically without harming normal brain tissue stem cells.
One approach to treat glioblastoma is to induce
differentiation of CD133 brain tumor cells, critically
weakening their tumor-forming ability. Piccirillo et al.
(2006) have shown that bone morphogenetic proteins (BMPs) prompt the differentiation of such cells.
This study demonstrated that the BMP-treated tumor
cells engrafted into mice were more mature and less
invasive, and CD133 cells could not be recovered
from these small tumors. The implication is that a
differentiation-promoting agent (BMP) is a potential treatment for brain tumors. This experimental
approach is ripe for further testing.
As stated earlier, CD133 (prominin-1) has been
proposed as a marker for brain tumor-initiating cells.
However, a recent study shows that tumorigenic potential also exists among CD133 cells to form tumors
(Prestegarden et al., 2010). Nestin glial fibrillary acidic
protein and neuron-specific enolase markers were
expressed in CD133 cells. It seems that the ability to
form tumors maybe a general trait associated with different glioma cell phenotypes, rather than a property
limited to an exclusive subpopulation of glioma stem
cells.
M.A. Hayat
Nevertheless, targeting CSCs pathways will ultimately prove to be an effective therapeutic strategy
against malignant gliomas. In order to accomplish this
goal, it is necessary to understand and link cellular, molecular, genetic, and epigenetic mechanisms to
compare the similarities and differences between normal neural stem cells and glioblastoma-initiating stem
cells.
The relevant question is whether chemoresistance of
glioblastoma stem cells is due to reduced drug uptake
or due to drug efflux. An in vivo study indicates that
neither of these two alternatives is fully applicable to
answer this question (Eramo et al., 2006). According to
this study, drug resistance by glioblastoma stem cells
depends on the abnormalities of the cell death pathways such as overexpression of antiapoptotic factor or
silencing of key death effectors. In other words, the
altered expression of apoptosis related proteins renders
normal neural stem cells strongly resistant to death
receptor ligands and inflammatory cytokines. More
extensive studies are required to fully understand the
mechanisms of chemoresistance by glioblastoma stem
cells.
As indicated earlier, cerebral glioma shows the
highest incidence rate among malignant intracranial
tumors, and the therapeutic efficacy of surgery, radiotherapy, and chemotherapy for the former are not
satisfactory, and these tumors show a tendency to
recur after the treatments. Many drugs exert their anticancer actions only after entering the cells, but some
drug-resistant tumor cells can prevent the drugs from
penetrating and thus avoid being destroyed. Ultrasound
can increase the intracellular bioaccumulation of drugs
by increasing the membrane permeability in tumor
cells and thus reducing the thereshold values of cell
death (Deckers et al., 2008).
Currently, inducing tumor cell death is considered
the endpoint in most nonsurgical therapies, but there
is increasing evidence that different death modes of
tumor cells have different effects on the functions of
immune cells, especially macrophages. Macrophages
play important roles in the occurrence and development of tumors. Recently, Xu et al. (2009) studied
the effects of ultrasound on the cell death induced
by arsenic trioxide, and the secondary activation
of macrophages; this study was carried out using
rat glioma cell line C6. Arsenic trioxide has been
approved by the FDA for the treatment of refractory
leukemia. Kim et al. (2008) have also used arsenic
Introduction
trioxide for sensitizing human glioma cells. It seems

that ultrasound synergisticulary enhances the cell death
effect by promoting arsenic trioxide entry into the cell
line C6, and macrophages are activated by the killed
C6 cells (Xu et al., 2009). This study provides a theoretical basis for the clinical use of this drug and
ultrasound in glioma treatment.
Because conventional treatments are not very effective against glioblastoma, there is an urgent need to
develop novel, effective therapeutic strategies against
this disease. One such therapy involves the role played
by brain tumor-derived stem cells (BTSC) and neural
stem cells (NSC) in cancer initiation and progression
(Germano et al., 2010). BTSC are a small subpopulation of cancer cells having similarities, such as selfrenewal, multipotency, and relative quiescence. BTSC
originate from a population of endogenous NSC. It is
thought that BTSC play a crucial role in the recurrence of primary brain tumors and treatment resistance.
Therapies targeting BTSC might be more effective in
treating this disease.
One of the new treatments to improve patient survival is targeting cancer stem cells that contain higher
NOTCH activity. The NOTCH signaling pathway regulates stem cells in the brain. It has been shown that
the NOTCH pathway blockade depletes stem-like cells
in the glioblastoma, suggesting the usefulness of using
NOTCH inhibitors (e.g., -secretase) as chemotherapeutic reagents to target cancer stem cells (Fan et al.,
2010). Another critical role of NOTCH signaling is
that it promotes radio resistance of glioma stem cells
(Wang et al., 2009). Thus, inhibition of this signaling holds promise to improve the efficiency of current
radiotherapy in glioma therapy.
Glioblastoma Multiforme
Glioblastoma multiforme is the most common primary
intrinsic brain tumor of adulthood, and the most malignant glioma subtype. Although significant advances
have taken place during the last 25 years in the basic
understanding of tumor pathogenesis, the median survival of patients has increased only 3.3 months (from
11.3 to 14.6 months). This poor prognosis is due to
the near inevitability of the recurrence of the tumor
in spite of the initial use of maximal safe surgical
resection, radiotherapy, and chemotherapy. Under the
circumstances, additional systemic and local therapies, as well as repeat surgery, are considered; these
therapies have potential benefits and risks, especially
resulting from surgery.
Surgery does provide benefits, such as immediate
decrease in tumor burden and improvement of tumorrelated neurologic symptoms and deficits. A potential
risk is the exacerbation or new onset of the same, as
well as a temporary or permanent exclusion from other
therapies. In addition, potential injury to the M1 and
M2 segments of the middle cerebral artery can result
in damage to the eloquent brain regions they supply
(Park et al., 2010).
In the light of risks and uncertainity involved in
the outcome of the reoperation, these patients deserve
to know their treatment options. In the past, studies have been carried out assessing the outcome of
reoperation of patients with recurrent glioblastoma
multiforme (Ammirati et al., 1987). However, such
studies did not establish guidelines for providing preoperative advice to patients considering reoperation.
Recently, Park et al. (2010) have devised a preoperative
scale that predicts survival after resurgery for recurrent
glioblastoma multiforme. This scale identifies patients
likely to have poor, intermediate, or good relative outcomes after surgical resection of the recurrent tumor.
On the basis of this scale, survival benefit and the attendant risk can be evaluated, on case-by-case basis, from
surgery. Because surgery is not curative, it is important to identify potentially effective treatments with
minimal risk.
Medical Imaging
A large of imaging modalities, including computed tomography (CT), position emission tomography (PET), magnetic resonance imaging (MRI), functional magnetic resonance imaging (fMRI), and single
photon emission computed tomography (SPECT), are
being routinely used for diagnosis, treatment, and
assessment of treatments (Hayat, 2008). Combined
PET/CT is a well-established approach that has been
extensively validated in routine clinical practice; CT
provides anatomical information, while PET contributes functional information. The introduction of
PET/CT was a revolutionary milestone in clinical
imaging. However, because the effective radiation
dose is the sum of the doses from PET and CT,

this combined examination (particularly diagnostic one
requiring high resolution CT) results in increased
patient exposure and a higher theoretical radiationinduced cancer risk than that with PET or CT alone.
Understandably, but unfortunately, the acceptance of a
treatment by patients is much higher when the information primarily addresses the benefits than the risk.
A more recently emerging therapeutic modality for
glioblastoma consists of boron neutron capture protocol (Yamamoto et al., 2008). This technique theoretically allows tumor-selective destruction, while sparing
normal tissue, even when the cells have microscopically spread to the surrounding normal brain. The
tumor cell selective irradiation depends on the nuclear
reaction between the stable isotope of boron (10 B) and
thermal neutrons, which release and 7 Li particles
within a limited path length through the boron neutron capture reaction, 10 B (n,) 7 Li. The selectivity
depends on the dose from the boron neutron capture
reaction, i.e., the accumulation of boron-10 in tumor
cells. Prospective randomized clinical trials are needed
to confirm the efficacy of this modality.
Risk of Medical Radiation

At the outset it is pointed out that although radiation
therapy has been the standard treatment of glioblastoma for four decades, it is only transiently effective,
and offers no lasting cure. Medical radiation fails in
the long run because it cannot kill the subpopulation of CD133 tumor-initiating cells. Although many
nuclear medicine and radiological procedures are safe,
there is still the need to justify these procedures on
the basis of a favorable risk: benefit ratio. Hazards
associated with medical radiation exposure do exist.
Radiation, for example, may induce necrosis at doses
of 60 Gy and above. The recent dramatic increase
in the number of diagnostic medical procedures and
nuclear medicine procedures has resulted in a very significant increase in cumulative exposure to radiation
and related risk of carcinogenesis. In the United States
in 2006, 377 million diagnostic and interventional
radiological examinations and 18.6 million nuclear
medicine examinations were performed (Mettler et al.,
2009). Among the diagnostic nuclear medicine procedures, cardiac studies followed by bone scanning
M.A. Hayat
are the most frequent nuclear medicine procedures.

Approximately, 12% of all radiological procedures and
50% of nuclear medicine procedures carried out worldwide are performed in the United States (Salvatori and
Lucignani, 2010). These startling findings underline
the pressing need to optimize and justify all exposures
of patients to medical ionizing radiation.
Starting from last decade, there has been a substantial increase in the use of PET/CT in children. In all
patients, especially in children, it is crucial to keep
radiation doses as low as is clinically feasible. It is
known that some children who undergo medical radiation, are prone to develop cancer or heart disease later
in life. Medical radiation protection has implication
for a range of persons: patients, medical, technical,
and nursing staff, friends, relatives, caregivers, and
members of the public.
References
Ammirati M, Galicich JH, Arbit E, Liao Y (1987) Reoperation
in the treatment of recurrent intracranial malignant gliomas.
Neurosurgery 21:607614
Beckner ME, Fellows-Mayle W, Zhang Z, Agostino NR, Kant
JA, Day BW, Pollack IF (2010) Identification of ATP citrate lyase as a positive regulator of glycolytic function in
glioblastomas. Nat J Cancer 126:22822295
Beckner ME, Stracke ML, Liotta LA, Schiffmann E (1990)
Glycolysis as primary energy source in tumor cell chemotaxis. J Nat Cancer Inst 82:18361840
Carpentier PA, Palmer TD (2009) Immune influence on adult
neural stem cell regulation and function. Neuron 64:7992
Carson KA, Grossman SA, Fisher JD, Shaw EG (2007)
Prognostic factors for survival in adult patients with recurrent glioma enrolled onto the new approaches to brain tumor
therapy CNS consortium phases I and II clinical trials. J Clin
Oncol 25:26012606
Deckers R, Rome C, Moonen CT (2008) The role of ultrasound
and magnetic resonance in local drug delivery. J Magn Reson
Imag 27:400409
Duffau H (2007) Contributation of cortical and subcortical electrostimulation in brain glioma surgery: methodological and
functional considerations. Clin Neurophysiol 37:373382
El Hallani S, Boisselier B, Peglion F, Rousseau A, Colin C,
Idbaih A, Marie Y, Mokhtari K, Thomas JL, Eichmann A,
Delattre JY, Maniotis AJ, Sanson M (2010) A new alternative mechanism in glioblastoma vascularization: tubular
vasculogenic mimicry. Brain 133:973982
Eramo A, Ricci-Vitiani L, Zeuner A, Pallini R, Lotti F, Sette
G, Pilozzi E, Larocca LM, Peschle C, De Maria R (2006)
Chemotherapy resistance of glioblastoma stem cells. Cell
Death Diff 13:12381241
Fan X, Khaki L, Zhu TS, Soules ME, Talsma CE, Gul N, Koh C,
Li JY-M, Maciaczyk J, Nikkhah G, DiMeco F, Piccirillo S,
Introduction
Vescpvi AL, Eberhart C (2010) NOTCH pathway blockade depletes CD133-positive glioblastoma cells and inhibits
growth of tumor neurospheres and xenografts. Stem Cells
28:516
Germano I, Swiss V, Casaccia P (2010) Primary brain tumors,
neural stem cell, and brain tumor cancer cell: where is the
link?. Neuropharmacology 58:903910
Hayat MA (ed) (2008) Cancer imaging, volumes 1 and 2.
Elsevier, San Diego, CA
Kim EH, Yoon MJ, Kim SU, Kwon TK, Sohn S, Choi
KS (2008) Arsenic trioxide sensitizes human glioma cells,
but not normal astrocytes, to TRAIL-induced apoptosis
via CCAT/enhancer-binding protein homologous proteindependent DR5 upregulation. Cancer Res 68:266275
Lai A, Filka E, McGibbon B, Leia P et al (2008) Phase II
pilot study of bevacizumab in combination with temozolomide and regional radiation therapy for up-front treatment
of patients with newly diagnosed glioblastoma multiforme:
interim analysis of safety and tolerability. Int J Radiat Oncol
Biol Phys 71:13721380
Lefranc F, Mijatovic T, Kondo Y, Sauvage S et al (2008)
Targeting the 1 subunit of the sodium pump to combat
glioblastoma cells. Neurosurgery 62:211222
Liu Y, Shete S, Erzel CJ, Scheurer M, Alexiou G, Armstrong G,
Tsavachidis S, Liang F-W, Gilbert M, Aldape K, Armstrong
T, Houlson R, Hosking F, Robertson L, Xiao Y, Wiencke
J, Wrensch M, Andersson U, Melin BS, Bondy M (2010)
Polymorphisms of LIG4, BTBD2, HMGA2, and RTEL1
genes involved in the double-strand break repair pathway
predict glioblastoma survival. J Clin Oncol 28:24672474
Liu G, Yuan X, Zeng Z, Tunici P, Ng H, Abdulkadir IR, Lu L,
Irvin D, Black KL, Yu JS (2006) Analysis of gene expression and chemoresistance of CD133 cancer stem cells in
glioblastoma. Mol Cancer 5:6768
Mathieu V, De Neve N, Le Mercier M, Dewelle J, Gaussin
J-F, Dehoux M, Kiss R, Lefranc F (2008) Combining bevacizumab with temozolomide increases the antitumor efficacy of temozolomide in a human glioblastoma orthotopic
xenograft model. Neoplasia 10:13831392
Mettler Jr,FA, Bhargavan M, Faulkner K, Gilley DB, Gray JE,
Ibbott GS et al (2009) Radiologic and nuclear medicine studies in the United States and worldwide: frequency, radiation
dose, and vomparison with other radiation sources-1950
2007. Radiology 253:520531
Nieder C, Wiedenmann N, Andratschke N, Molls M (2006)
Current status of angiogenesis inhibitors combined with
radiation therapy. Cancer Treat Rev 32:348364
Nimsky C, Ganslandt O, von Keller B, Fahlbusch R (2006)
Intraoperative visulization for resection of gliomas: the role
7
of functional neuronavigation and intraoperaive 1.57 MRI.
Neurol Res 28:482487
Ohgaki H, Kleihues P (2007) Genetic pathways to primary and
secondary glioblastomas. Am J Pathol 170:14451453
Park JK, Hodges T, Arko L, Shen M et al (2010) Scale to predict
survival after surgery for recurrent glioblastoma multiforme.
J Clin Oncol 28:38383843
Piccirillo SGM, Reynolds BA, Zanetti N, Lamorte G, Binda
E, Broggi G, Brem H, Olivi A, Dimeco F, Vescovi AL
(2006) Bone morphogenetic proteins inhibit the tumorigenic
potential of human brain tumour-initiating cells. Nature
444:761765
Prestegarden L, Svendsen A, Wang J, Sleire L, Skaftnesmo KO,
Bjerkvig R, Yan T, Askland L, Persson A, Sakariassen PQ,
Enger PQ (2010) Glioma cell populations grouped by different cell type markers drive brain tumor growth. Cancer Res
70:42744279
Salvatori M, Lucignani G (2010) Radiation exposure, protection,
and risk from nuclear medicine procedures. Eur J Nucl Med
Mol Imaging 37:122512231
Sandmair AM, Vapalahti M, Yla-Herttuala S (2000)
Adenovirus-mediated herpes simplex thymidine kinase gene
therapy for brain tumors. Adv Exp Med Biol 465:163170
Stummer W, Pichlmeier U, Meinel T, Wiestler OD, Zanella
F, Reulen H-J (2006) Fluorescence-guided surgery with 5aminolevulinic acid forresection of malignant glioma: a randomised controlled multicentre phase III trial. Lancet Oncol
7:392401
Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher
B, Taphoorn MJB, Belanger K, Brandes AA, Marosi C,
Bogdahn U, Curschmann J, Janzer RC, Ludwin SK, Gorlia
T, Allgeier A, Locombe D, Cairncross G, Eisenhauer E,
Mirimanoff RO (2005) Radiotherapy plus concomitant and
adjuvant temozolomide for glioblastoma. New Engl J Med
352:987996
Van Eekelen M, Sasportas LS, Kasmieh R, Yip S, Figueiredo JL, Louis DN, Weissleder R, Shah K (2010) Human stem cells
expressing novel TSP-1 variant have anti-angiogenic effect
on brain tumors. Oncogene 29:31853195
Wang J, Wakeman TP, Lathia JD, Hjelmeland AB, Wang X-F,
White RR, Rich N, Sullenger BA (2009) Notch promotes
radioresistance of glioma stem cells. Stem Cells 28:1728
Xu Y, Zhang Y, Liu X, Wang Z, Ma J, Wang J, Yue W (2009) The
effects of ultrasound and arsenic trioxide on neurogliocytoma cells and secondary activation of macrophages. Tumori
95:780788
Yamamoto T, Nakai K, Matsumura A (2008) Boron neutron
capture therapy for glioblastoma. Cancer Lett 262:143152
Part I
Biomarkers and Diagnosis
Chapter 2
Gliomagenesis: Advantages and Limitations of Biomarkers

Michel Wager, Lucie Karayan-Tapon, and Christian-Jacques Larsen
Abstract During the last years, spectacular progress

in the field of molecular biology raised the hope that
it might replace or at least improve significantly the
World Health Organization classification of tumors
of the central nervous system. Although up to now,
this has not been the case, a few biomarkers have
found their place as prognostic factors for certain
tumor types on the one hand, and as predictive
factors of response to treatments on the other hand.
The present chapter will deal with the following
tumor types: astrocytic tumors (including Diffuse
astrocytoma, Anaplastic astrocytoma, Glioblastoma,
and Pilocytic astrocytoma), Oligodendroglial tumours
(including Oligodendroglioma and Anaplastic
oligodendroglioma), and mixed gliomas (including
Oligoastrocytoma and Anaplastic oligodendroglioma).
The current biomarkers are not decision-making
parameters on a case by case basis, but they have
progressively emerged as cornerstone indicators as
far as stratification in clinical studies is concerned.
As a consequence, during therapeutic trials, it can be
considered that groups of patients whose biomarkers
anticipate a poor response to standard treatment might
be offered the opportunity of innovative therapies.
In this context, the most recent developments of
molecular biology culminating in integrated molecular
analysis of tumors and the possibilities offered by
neurosphere cultures established from glioblastoma
multiforme that recapitulate the tumor allow to
M. Wager ()
Poitiers University Hospital, University of Poitiers Medical
School, EA 3805 Poitiers, France
e-mail: m.wager@chu-poitiers.fr
anticipate a growing importance of a series of new

biomarker in the years to come.
Keywords Biomarkers Gliomagenesis Gliomas
Genomic analysis Kinase Tumor grade
Introduction
This chapter will focus on adult gliomas only. The
authors think that gliomas in children and adults
are quiet different entities that should be considered
separately. This conviction relies on clinical reasons
on the one hand, because, as compared to adult
gliomas, tumors in children are of lower prevalence,
more often low grade at diagnosis, and tumor progression to higher grades arise later in the course
of the disease. Furthermore, their anatomical distribution is different from that in the adult. Available
biologic data also tend to distinguish these two periods of life: three of the main ways of gliomagenesis in adults PI3-kinase/Akt/PTEN, p53/MDM2/p14
ARF, Rb/CyclinD1/CD4/p16 CyclinD1 Rb/CyclinD1/
CD4/p16INK4A CD4 seem only rarely involved
in children gliomas, quite as EGFrs amplification
(Tamber et al., 2006).
Grade I glioma, i.e. pilocytic astrocytomas, is a
truly benign lesion that can be cured by total surgical removal. This tumor will be only quickly evoked
in this chapter. Indeed, every aspect of this tumor
clinical presentation and imaging features, principles
of treatment and outcome distinguish it from adult
diffuse gliomas, and its place in the World Health
Organization (WHO) classification can be considered
a historic inheritance.

11
12
The WHO classification (Louis et al., 2007)

describes, according to malignity criteria, different grades among the diffuse infiltrating gliomas,
from low grade gliomas (LGG, grade II) to high
grade gliomas (grade III or anaplastic gliomas, and
grade IV or glioblastoma multiforme GBM). The
second criterion for classification is morphology,
according to the cellular type considered as predominant, leading to astrocytic, oligodendroglial or mixed
(i.e.oligoastrocytic) tumors. The course of the disease,
therapeutic strategy and outcome are so profoundly
different according to these criteria that it is necessary
to treat the question of biomarkers separately for each
of them. Indeed, even among the biomarkers that have
gained clinical attention, no one covers neither the full
spectrum of the different grades, nor the spectrum of
response to all treatments, regarding an overall survival
prognosis. And the perspective of a classification relying on exclusive biologic criteria appears very far from
us. So, it is necessary to consider each biomarker in its
limited, relevant context.
For each grade (II, II or IV) of each morphotype,
the usefulness of biomarkers can be considered on two
different points of view: diagnostic and outcome prognosis on the one hand, prediction of the response to
treatments on the other hand. It will be seen through
this chapter that each biomarker finds an original place
according to both criterions.
Who and Sainte Anne Hospital

Classifications
It is not conceivable today to study gliomas biomarkers
without referring at least at some point to pathology. Considering that some instances of clear relationship between a tumor type and a biomarker
oligodendroglial phenotype and loss of heterozygosity (LOH) 1p/19q is far the better known example
have been recognized, one understands that the
internal logic of a pathology classification is pivotal.
There are today two classifications for adult gliomas:
the WHO classification, and the Sainte-Anne Hospital
(SAH) classification. It is noteworthy that until today,
the pathology of these tumors remains so debated
that it has been possible to describe them from at
least two profoundly different points of view, reflected
in these two classifications. This remark underlines
M. Wager et al.
the need for biomarkers. Furthermore, classifications

sometimes lack reproductibility (Mittler et al., 1996)
and some samples are misinterpreted. This situation
also stresses the need for relevant biomarkers.
The WHO classification is the most universally
admitted and is the standard in literature. It classifies
gliomas according to their predominant cellular type.
So, it carries the drawback not to distinguish tumor
cells from elements of the infiltrated parenchyma,
and particularly tumor astrocytes, from reactive astrocytes. Furthermore, in the WHO classification, the
word glioblastoma is applied both to a histological type and to a concept i.e. tumors gathered due
to the fact that they are highly malignant, but the
origin of which can be astrocytic or oligendendrocytic. This point can be confusing and very pejorative
when these groups are approached under the angle of
biomarkers.
The so-called Sainte Anne hospital classification
is a completely original approach of these tumors
(Daumas-Duport et al., 2000; Varlet et al., 2005).
Contrasting with the WHO classification based on
the predominant cell type, the Sainte Anne hospital classification is based on the distinction of two
patterns of tumor growth, solid tumour tissue versus
isolated tumor cells. It defines glioblastoma multiforme as tumors of certain and exclusive astrocytic
differentiation, defined by variable intensity of glial
fibrillary acidic protein (GFAP ) staining, the observation of necrosis being mandatory. In the Sainte Anne
hospital classification, the other tumors are oligodendrogliomas, of low (type A oligodendrogliomas)
or high (type B oligodendrogliomas) grade. In those
tumors, encountered astrocytes are considered reactive. The Sainte Anne hospital classification also
relies on clinical and imaging features that are part
of a diagnosis considered as a synthesis combining
these different parameters. Besides, Sainte Anne hospital classification also defines among the high grade
gliomas a subgroup named glioneuronal tumors. These
tumors exhibit non neuronal cells expressing neurofilamentary protein immunostaining (Varlet et al., 2004).
Interest of Biological Fluids Biomarkers

A literature review shows that until now biomarkers
have been looked for mostly on gliomas themselves,
13
and very rarely in the body fluids (serum or

cerebrospinal fluid) of these patients. While the examples of solid tumors are numerous, for which one
observes how valuable body fluids biomarkers can
be, this interest might be less obvious at first sight
as far as gliomas are concerned. Indeed, the sensitivity of magnetic resonance imaging (MRI) is so good
that is allows positive diagnosis of brain tumor in
almost every patient harbouring a glioma. And when
brain MRI discloses a brain tumor, as a rule diagnosis
requires either a biopsy, or a surgical removal. Hence
in both situations samples are available for biomarkers study. That is probably one of the reasons why
body fluid biomarkers received so little attention up
until now.
All the biomarkers presented below were studied
on tumors. There are nevertheless clinical situations
in which the availability of body fluids biomarkers
would be of great value, and particularly: first for
the patients in whom surgery is impossible. Second
if biomarkers could help detecting progression of low
grade gliomas to anaplasia, and the degree of anaplasia. Indeed, one of the greatest difficulties in gliomas
diagnosis is tumor heterogeneity, raising the question
of samples representativeness, with the risk of underestimating the tumor grade, in cases where the sample(s)
do not come from a region reflecting the highest degree
of anaplasia of the tumor. This question remains if
biomarkers are looked for on tumor samples. Body
fluids biomarker solving this difficulty would constitute a progress. Thirdly if they allowed the tumour
growth rate assessment, as it seems that this can be
case (Hormigo et al., 2006). Last but not least, some
patients benefit long survival, and for them access to
body fluids biomarkers would mean possible and even
comfortable repeated workups.
Figure 2.1 presents the potential interest
of different biomarkers including body fluids
markers.
It is possible today to distinguish two types of
biomarker: those which are now considered relevant in
the clinical field, and those which are until now interesting perspectives raised by the recent new findings in
cancer biology, but that have not not yet? gained
clinical interest. Both groups are presented below. As
already mentioned, genotypic and phenotypic features
of low grade gliomas (LGG) and high grade gliomas,
as well as their natural history, response to treatments
and prognosis appear to be so profoundly different
that it is necessary to deal with their biomarkers

distinctly.
Biomarkers in the Clinical Field

Importance of Biomarkers in Low Grade
Gliomas Management
Low grade gliomas have long been considered benign
tumors. It is now a well recognized fact that they are
truly pre malignant tumors carrying a poor mid term
prognosis. These tumors progress to anaplasia over
a few years, until their eventual transformation into
glioblastoma multiforme carrying gloomy prognosis.
In this context, after long controversies it is now widely
admitted that low grade gliomas surgical total or subtotal removal is the cornerstone of their treatment (Sanai
and Berger, 2008). Surgery can nevertheless be insufficient to prevent tumor progression and change of
grade in two scenarios. In the first place when sufficient tumor resection as first treatment is impossible
due to the alliance of volume and location of the tumor.
Secondly if in spite of initial surgery a residual tumor
volume superior to 810 cc remains inaccessible to surgical removal in spite of possibilities offered by awake
surgery for eloquent brain areas (Berger, 1995). In this
context two parameters are useful to determine, that
biomarkers could contribute to assess. The first one is
the tumor volume and even more importantly the tumor
growth rate, parameters influencing directly the therapeutic strategy and particularly on a surgical point
of view. Presently this volume is calculated on serial
repeated follow-up MRIs. One should insist on this
pivotal well documented fact that not only the volume
itself, but also the growth rate carries prognostic value
(Pallud et al., 2006).
Spontaneous Course of the Disease

Until today relatively few biomarkers have received
attention in this context. Regarding low grade
gliomas prognosis, it has been demonstrated that the
Delta Ex2-3 isoform of the p73 gene predicts survival (Wager et al., 2006). But the major achievement
of molecular biology in the characterization of low
14
Fig. 2.1 Anticipated
usefulness of biomarkers for
diagnosis and follow-up of
adult gliomas, in relation with
tumor grade. ( ) +/:
moderate; ( ) + important;
( ) ++ very important.
Grades II, III and IV
according to the WHO
classification
M. Wager et al.
Grade II
Grade III
Grade IV
++ (***)
+ (**)
Prediction of
In case of total or
++
++
response to
subtotal surgical
treatments
resection : +/ (*)
Natural course of
the disease
In case of non
sufficient resection :
++
Body fluid
++
biomarkers
1- because follow up
+ / because as a
+ / because as a
rule, tumor samples
rule, tumor samples
are available
are available
often on long periods

of time.
2 because no tumor
sample is available as
long as there is no
surgery (neither
biopsy nor resection)
3 If they resolve
the difficulty of
surgical samples
representativeness
grade gliomas up until now has been the demonstration that codeletion of the chromosome arms 1p and
19q is generally associated to the oligodendroglial phenotype, while tumours not presenting this codeletion
are generally astrocytomas or mixed oligoastrocytic
tumours (Reifenberger et al., 1994). Molecular studies have shown that loss of chromosomal material
arises through an unbalanced translocation of 19p to 1q
t(1;19)(q10;p10) (Jenkins et al., 2006). The molecular
significance of these deletions relies on the characterization of genes located in the deleted areas. Several
candidate genes have been located in the 1p36 region
but currently, no functional data are available that
could provide interest for clinicians. This finding is

all the more so important that this codeletion is found
in approximately two thirds of oligo tumors (pure
oligodendroglial and anaplastic oligodendroglial, and
mixted oligoastrocytic and anaplastic oligoastrocytic) considered altogether (Reifenberger et al.,
1994). It appears that low grade tumors presenting this
codeletion both carry a better prognosis, and are better
responders to alkylating chemotherapy (Ricard et al.,
2007). Other data in literature support the combination
of these two aspects, and the 1p/19q codeletion is now
recognized as a clinically relevant biomarker of these
tumors.
Prediction of Response to Treatments

For patients harbouring a tumor not amenable to surgical treatment, and demonstrating clinical and/or volumetric tumor progression, the two available therapeutic
modalities are radiation therapy and chemotherapy.
Historically, radiation therapy has been, and remains
today a pivotal treatment for low grade gliomas inaccessible to sufficient surgical resection. There is today
no predictive biomarker of low grade gliomas response
to radiation therapy. A fairly recent strategy proposes
neo adjuvant chemotherapy for low grade gliomas
which are deemed inaccessible to sufficient surgical
removal as initial treatment. This chemotherapy relies
on alkylating agents, and some responding patients
can benefit surgery as a second step, depending on
the tumor response to chemotherapy in terms of volume shrinkage. Regarding the predictive value of
response to chemotherapy, low grade gliomas now
benefit the recent progress of molecular biology initially reported for higher grade oligendendroglioma
which treatment have included chemotherapy for long
(Ricard et al., 2007). The methylation status of the
MGMT (O6-methyl-guanyl-methyl transferase) gene
promoter might also be predictive of low grade gliomas
chemosensitivity to alkylating agents, but this remains
to be demonstrated. Concerning astrocytic tumors,
other biomarkers are under investigation, which are
presented below, but although promising they have not
reached the clinician yet. However, whatever the loss
of portions of chromosomes 1p36 and 19q13.3 and
the therapeutic strategy, all the patients that are not in
remission after initial and even repeated when functionally possible sufficient surgical removal of their
low grade tumor eventually develop a progression to a
higher grade tumor.
Importance of Biomarkers in High Grade

Gliomas Management
15
are homogenous enough as far as overall survival is

concerned. However, observed survival is, for some
patients, different and sometime quiet different from
the anticipated outcome. As previously mentioned,
part of the explanation might be the lack of reproductibility of grading. Another part of explanation
might be the particular behaviour of some subsets of
gliomas (i.e. glioneuronal tumors as described in the
Sainte Anne hospital classification), that might benefit
very long survival after complete surgical removal. In
this context, while numerous biomarkers are identified
as prognostic factors, many of them have no impact on
decision-making due to fact that their value is less relevant than that of already known indicators (e.g., grade,
age, impact of treatment on survival) (Wager et al.,
2008).
Focusing on anaplastic (Grade III) oligos, either
pure oligodendroglial, or mixed tumors, it has
been shown that 1p/19q codeletion is an independent favourable prognostic factor (for a review, see
Idbaih et al., 2007). Methylation status of the MGMT
promoter gene is also prognostic, and of potential interest, its relevance can exceed the predictiveness of the
response to alkylating agents (nitrosoureas and temozolomide) for these tumors. Indeed, a high correlation
was reported between MGMT methylation and 1p/19q
codeletions which are associated to a better prognosis
in anaplastic astrocytomas.
Glioblastoma (grade IV) patients present the shorter
overall survival of all gliomas. While one of the first
successes of molecular biology have been to distinguish primary and secondary glioblastoma, their
clinical outcome and response to treatment are so identical that this distinction had no clinical consequence.
During the last years, the biological breakthrough concerned biomarkers predicting response of these tumors
to chemotherapy (see following paragraph). But several purely prognostic biomarkers are now emerging,
which are on the verge of penetrating the clinical field.
They will be presented further in this chapter (see
Biomarkers outside of the clinical field: perspectives
emerging of recent new findings in cancer biology).
Spontaneous Course of the Disease

Prediction of Response to Treatments
High grade gliomas are today among the most difficult tumors to treat and carry a gloomy prognosis. The WHO classification defines according to
tumor grade (III and IV) groups of patients that
Focusing on grade III tumors, a wide corpus of

retrospective studies suggests that 1p/19q codeletion
predicts better response to alkylating agents
16
(Cairncross et al., 1998, for a review, see Idbaih

et al., 2007). As far as methylation of the MGMT
promoter gene status is concerned, noticeably its
status in anaplastic astrocytomas is only prognostic
but not predictive of response to alkylating agent
chemotherapy as it is in glioblastoma.
Regarding glioblastoma, in 2005 a benefit on overall
survival was demonstrated, real although modest, for
a subpopulation of glioblastoma patients receiving the
concomitant administration of radiation therapy and
chemotherapy by alkylating agents (Stupp et al., 2005).
It appeared that this benefit depends on the methylation
status of the MGMT gene promoter (Hegi et al., 2005).
MGMT is a DNA repair enzyme that catalyses the
elimination of alkyl groups on DNA. This observation
has been at the origin of an important renewal of interest in the medical community in the biologic aspect of
the problem, and methylation status of this promoter
gene has become a biomarker predicting response to
a protocol which rapidly became the first line standard of treatment for newly diagnosed glioblastoma
multiforme.
Because response to this alkylating agent (temozolomide) is strongly correlated with the methylation
status of the MGMT promoter gene, systematic analysis of the MGMT promoter is today considered to
be the strongest predictor of response to alkylating
chemotherapy in newly diagnosed glioblastoma multiforme. Several methods for this analysis are available,
and this constitutes another aspect of this problem
(Karayan-Tapon et al., 2010) see below.
As a conclusion, a review of literature shows that
several relevant biomarkers have recently emerged in
the clinical field. But it is remarkable that there is often
a persisting ambiguity in data, as impact on survival
of genuine response to treatment is not always clearly
distinguished from natural course of these diseases. In
the end, the question is whether clinicians should consider these biomarkers for decision making? The core
of the problem is that those biomarkers are reliable as
far as groups are concerned, but are very disputable
as far as patients are considered on a case by case
basis. This is why the answer to this question is different today for groups of patients, and for patients
considered individually. Stratifications of patients in
clinical studies now widely rely on biomarkers, and
one cannot ignore them anymore when drawing conclusions from clinical trials data. It is also a very
common idea today that these biomarkers can help
M. Wager et al.
to decide which groups of patients should be considered for innovative treatments in clinical trials i.e.
groups whose biomarker(s) anticipate a poor response
to standard treatments. But one has to keep in mind
the persisting uncertainty ambiguity between natural course and genuine response to treatments in
numerous situations. Otherwise the price to pay would
be important bias in these trials. As far as individuals are concerned, it is not considered relevant today
to use those biomarkers for therapeutic decision making regarding administration of alkylating agents
on a individual, case by case basis. It is also noteworthy that many technical and standardization difficulties
remain, which should be resolved before a personalized medicine can be envisioned (Karayan-Tapon et al.,
2010; Weller et al., 2010).
Biomarkers Outside of the Current

Clinical Field: Perspectives Emerging of
Recent New Findings in Cancer Biology
In terms of cancer physiopathology of brain tumors,
the beginning of the twentyfirst century has seen
remarkable progress in the understanding of biology
and molecular mechanisms. One of the most significant concepts is the notion that, similarly to other
cancers, glioblastomas possess a population of cells
with stem cell properties. In addition to their tumorigenic potential, these tumors stem cells (TSC) or tumor
initiating cells (TIC) share two main properties of the
stem cells, namely self-renewal and toti/pluripotency,
that is to say a capacity to differentiate into the different cell lineages that constitute the nervous tissue.
Because of these properties and their resistance to a
variety of stresses (chemotherapy, radiations), these
TSC are capable of escaping to the conventional therapies and maintaining the recurring capability of the
tumors. Furthermore, a few cells inside this pool of
TSC is prone to additional genetic events that favor
their dissemination and increase their resistance.
A second exciting approach is the development of
the so-called omics fields that are currently applied
to achieve a very precise knowledge of the genome
and the abnormalities present in tumors (genomics),
as well as to the acquisition of gene expression patterns (transcriptome) resulting in the possibility of
17
classification of different glioblastoma subtypes, and

the detection in the tumors and in different fluids of
the patients of a number of proteins which could be
of real interest to improve the tumor classification,
help to the choice of treatments and improve followup of the tumor itself. Although these findings open
the way to new approaches in the comprehensive biology of the tumors, they have not still given us keys
to major therapeutic breakthroughs. With regards to
this problem, this presentation is subdivided into three
parts: description of the current alterations that are
found in glioblastomas (primary as well as secondary)
and might constitute in the near future biomarkers
participating in decision making. The tumor stem
cell concept and its utilization for a better understanding of the biology of glioblastomas. The use of
new molecular features recently discovered to develop
future molecularly targeted diagnosis approaches and
therapies.
promoter have already been detailed in the previous

paragraph, as they have received the most important
clinical interest. Several other important biomarkers
have recently emerged. They are presented now.
Present Status of Genetic Alterations That

Drive Gliomagenesis and Progression
Analysis of various cohorts of tumors, the most
informative being that of the Cancer Genome Atlas
Research Network, have got insight into the principal ways that are deregulated in glioblastomas, the
most informative being that of the Cancer Genome
Atlas Research Network: (a) the TP53 pathway (>60%
of tumors) with abnormalities of TP53 gene itself
or one member of the pathway (MDM2, MDM4,
p14ARF); (b) the pRB1 pathway (68% of tumors) that
includes RB1, CD4, CDKN2A (p16INK4A); 3) the
PI3 K-PTEN pathway that involves PIK3CA, PIK3R1,
PTEN, and IRS1. As a rule one abnormality of one
of the genes belonging to a given pathway excludes
aberrations of other genes belonging to this pathway.
Even if this huge number of data has largely contributed to a better understanding of the mechanisms
underlying the biology of gliomas, none of them
presents a real practical interest on a therapeutic point
of view. In contrast, a few abnormalities not implicated in the above-mentioned pathways and more or
less specific to gliomas are currently used in the clinical field. Loss of portions of chromosomes 1p36 and
19q13.3, and methylation status of the MGMT gene
Amplification of the BRAF gene on chromosome

band 7q34 in pilocytic astrocytpomas. A tandem
duplication of BRAF activates the gene by generating a fusion gene between the kinase domain of
BRAF and the adjacent locus KIAA1549 (Jones
et al., 2009). This rearrangement is present in
6080% of pylocytic astrocytomas and infrequent
in diffuse low grade astrocytomas thus providing clues for differential diagnosis. Constitutive
activation of BRAF activates the MEK/ERK pathway leading to increased proliferation of tumor
cells. In view of these data, inhibition of the
BRAF/MEK/ERK cascade might constitute a therapeutic target for the treatment of low grade pediatric astrocytomas. However, a serious caveat has
occurred with a new report showing that the use
of some BRAF inhibitors in the control of tumor
growth actually resulted in stimulatory effects on
proliferation rather than stopping it. According to
this work, BRAF-specific inhibitors (for instance
885-A) facilitate the association to its partner CRAF
that is linked to Ras and induce an over activation
of the MEK/ERK pathway. In contrast, pan-RAF
inhibitors (sorafenib) while they do not block the
association of BRAF to CRAF, inhibit the two partners and abolish the stimulation of the MAP kinase
pathway (Heidorn et al., 2010).
Mutations of IDH1 and IDH2 gene (chromosome band 2q33) as a differential diagnosis tool.
The Isocitrate dehydrogenase (IDH) gene encodes
a metabolic enzyme that converts isocitrate to
-ketoglutarate with the concurrent reduction of
NAD(P)+ into NAD(P)H. The two IDH1 and IDH2
isoforms are cytosolic and mitochondrial respectively. Whereas IDH3, as a part of the tricarboxylic
acide (TCA) cycle, generates NADH for energy
production, IDH1 and 2 participate to shuttling
of electrons between mitochondria and cytosol.
In 2008, it was found that mutations in IDH1
all located on the R132 residue and less frequent mutations on R172 in IDH2 were present
in 12% of gliomas analysed in the series. The
IDH1 mutated tumors were most common to young
18
M. Wager et al.
patients and secondary glioblastomas and associated to longer survival (Parsons et al., 2008). A
subsequent study (Balss et al., 2008) explored the
presence of IDH1 mutations on a series of 685
tumors comprised of most if not all major glioma
subtypes. The mutations were nearly absent in primary glioblastoma and present in 88% of secondary
glioblastomas. The high frequency of IDH1 R132
mutations in oligodendrogiomas (69%) and mixed
oligoastrocytic tumors (78%) were consistent with
an arising of secondary glioblasoma from lower
grade tumors. No mutation was found in pilocytic astrocytomas making possible a molecular
discrimination between this tumor type and infiltrating low-grade gliomas. Analysis of another series
(Watanabe et al., 2009) provided arguments suggesting that IDH mutations are early events in
gliomagenesis.
The biochemical consequences of the IDH1
mutation most likely rely on the substitution of
R132/R172 in IDH1/IDH2 because the two arginine
residues contract hydrophilic interactions that allow
the binding of isocitrate to the enzymes. Because of
the large range of substituting residues, it is probable that R replacement supports tumorigenesis by
impairing isocitrate binding. This leads to a loss of
function and qualifies IDH1 and 2 as tumor suppressor genes (Zhao et al., 2009). In fact new recent
data have been reported indicating that R132 mutant
IDH1 not only does not catalyze the conversion
of isocitrate to -ketoglutarate but instead reduces
-ketoglutarate to the D-2 enantiomer form of
2-hydroxyglutarate (Dang et al., 2009). Consistent
with this gain of function by the mutant enzyme,
human gliomas with the R132 mutation in IDH1
contain 100-fold more 2-hydroxyglutarate than in
tumors with WT IDH1. One possible explanation for an oncogenic function of mutant IDH1
could result from its inhibitory capacity of the
oxygen-sensing enzymes hypoxia-inducible prolylhydroxylases (PHDs). Indeed, PHDs are inhibited
in cells carrying mutant IDH1. PHD inhibition
leads to the activation of the HIF transcription factor but currently, this does not give a clue as to
know the real function of the mutant IDH1 in an
oncogenic process. Pathological increase in the D2-hydroxyglutarate levels are found is associated
with encephalopathy and cardiomyopathy but not
with tumor risk, whereas the L-2-hydroxyglutarate
enantiomer also associated with progressive neural

defects is linked to increased risk of glioma (Aghili
et al., 2009).
A new mechanism relevant to the occurrence of
resistance to alkylating agents has emerged from
data showing an increased frequency of inactivating mutations in the MSH6 mismatch repair gene
in patients who had recurrence following temozolomide treatment. This suggested that MSH6
deficiency may contribute to the emergence of resistance. Thus, MSH6 mutations are likely to be
selected during temozolomide therapy.
Most recently, studies on significant cohorts of
annotated tumors have resulted in the detection
of new genetic abnormalities: the NF1 gene was
found mutated in 14% of glioblastomas; mutations
of PIK3 R1 and ERBB2 genes were also detected
respectively in 10 and 8% of analysed glioblastomas.
However, the practical interest has not yet been proven
and further studies will be necessary before concluding
on their utility for clinicians.
Brain Tumours and Stem Cells

A profusion of data in the literature feeds the notion
that, as for other tumor types, brain tumors and
particularly glioblastoma multiforme contain a fraction of tumor cells that share a number of features
characteristic of neural stem cells and immature precursors (progenitors) more or less engaged in a differentiated lineage. Briefly, these cells are defined by
two main properties: ability to perpetuate themselves
(self-renewal capacity for true stem cells) and capacity to give rise to populations of postmitotic cells
of the cerebral tissues through differentiation. While
the existence of SC has been strongly established in
hematopoietic tissue since the nineties, the presence of
NSC in brain has been delayed until the demonstration
of neural stem cells in subventricular zones of the adult
human brain was reported (Sanai et al., 2004). This
immediately reinforces the hypothesis that recurrence
of GBM is likely to occur through cancer-initiating
precursors with stem-cell properties, involving principally self-renewal in vitro and in vivo. For different
reasons, this model postulates that only this small subset of cells is able to proliferate extensively and has
19
a true tumorigenic potential whereas most if not all of

the cells composing the bulk of the tumor have lost this
potential (for a review see Vescovi et al., 2006). The
model immediately raises one obvious question about
the status of NSC as targets for transforming events.
There are at least two reasons that this may be the case:
first, fewer events are required to impair the phenotype
of cells that have self-activating mechanisms already
operating. In addition, the persistence for long periods
of time of self-renewing SC implies that the opportunity for mutations to accumulate in individual stem
cells is higher than in non proliferating dying cells.
This is of particular interest to explain the resistance
to radiotherapy and chemotherapy acquired by TSC
(Eramo et al., 2006; Liu et al., 2006; Bao et al., 2006).
Two important consequences emerge from the tumor
stem cell paradigm: first, distinct signaling pathways
are activated in self-renewing stem cells and progenitors more advanced in differentiation. This opens the
way to the definition of new targets. Second, the impact
of therapeutics on a tiny subset of tumorigenic stem
cells could reduce the useful targets for efficient treatment of tumors. In addition, the new data provided by
integrated genomic analysis (see next section) could be
useful for defining with accuracy molecular signatures
for each category of TSC or progenitors.
and pathways involved in gliomas physiopathology.

Second they should provide the clinician with new
tools for improving diagnosis of tumors that are
not easily classified by conventional histopathological
analysis. In this respect, the definition of gene signatures specific of tumors types and grades is likely to
introduce new classification schemes that will go further than those resulting from histopathological and
imaging analysis alone. Such large studies require the
establishment of publicly available banks of annotated
tumors. By definition, these banks should contain sufficient numbers of each tumor subtype as to allow
robust statistical analysis. Furthermore, use of efficient
algorithms for achieving successfully specific signatures predictive of a given cellular phenotype. These
conditions have been fulfilled by the creation of large
networks, the Cancer Genome Atlas Research Network
(TCGA, URL:http://cancer genome.nih.gov) being the
most representative. These approaches confirmed some
of the molecular pathways that had been previously
involved in gliomagenesis and, most interestingly,
identified new subtypes of GBM. Among the series of
recent publications on the subject, a few deserve particular attention as they open the way to real progress
in terms of diagnosis, prognosis and treatment.
The work published by the Cancer Genome Atlas
Network in 2008 is the first integrated genomic
analysis implicating several parameters: DNA copy
number, gene expression (trancriptome) and DNA
methylation profiling in 206 human glioblastomas (an
almost simultaneously published study led to the same
general conclusions). This study univocally showed
that most of the GBM carry abnormalities in TP53,
RB1 and tyrosine kinase receptor pathways. One
major conclusion was that these abnormalities implicating significant signalization pathways were a core
requirement for GBM pathogenesis. The most recent
data published by the TCGA start from the idea that
different tumor subtypes may originate from different
causes or different cells of origin, or both (Verhaak
et al., 2010). By combining gene expression classification data, discarding several possible bias and by
demonstrating their reproducibility on an independent
validation set, the authors defined a molecular classification into four primary GBM subtypes, namely
Proneural, Neural, Classical and Mesenchymal subtypes. For each of them, the 210 genes (per subtype)
signature that was established was integrated with
genomic aberrations. Principal characteristics for each
Perspectives Resulting from Integrated

Genomic Analysis to Define a Framework
for a Molecular Classification of Tumours
and for Investigation of Targeted
Therapies
The New Tools of Genomic Analysis
Molecular biology of cancer is more and more dominated by the development of large scale-array-based
profiling methods at the DNA, RNA and protein levels. New technologies that are now commonly used
involve array-based comparative genomic hybridization (array-CGH), SNP (single nucleotide polymorphism) arrays, transcriptomic arrays for gene expression, new sequencing techniques. The interest of such
approaches is double: first from a fundamental point
of view, the findings resulting from these technologies
should culminate in an overview of unregulated genes
20
M. Wager et al.
CLASSICAL SUBTYPE:
Chromosome 7 amplification and chromosome 10 loss
Four-fold increase EGFR amplification (97%)
EGFR point mutation or vII EGFR mutation (12/22 tumors)
Relative absence of TP53 mutations
Frequent deletion of chromosome band 9p21.3 (p16INK4A and p14ARF) accompanying EGFR
alterations (94%). Mutual exclusion with alterations of the RB pathway indicating the implication
of this way by the only CDKN2A deletions.
Overexpression of components of stem cell marker NES, NOTCH (NOTCH3, JAG1, LFNG),
Sonic Hedgehog (SMO, GAS1, GLI2) signaling pathways.
MESENCHYMAL SUBTYPE:
Focal hemizygous deletions of a 17q11.2 chromosome region containing NF1 (53%
lower
NF1 expression levels. No methylation in the NF1 locus

Frequency of NF1 mutations (14/20 tumors classified as Mesenchymal)
Expression of Mesenchymal markers: CHI3L1/YKL40, MET
High expression of members of the tumor necrosis factor superfamily pathway and NF-kb
pathway: TRADD, RELB, TNFRS1A
PRONEURAL:
Focal amplification at 4q12

GBM than other subtypes)
alterations of PDGFRA most frequently found in this subtype of

PDGFRA overexpression almost excusively found in this subtype
Point mutations (R132) in IDH1 (11/12 tumor)

Frequent TP53 mutations and loss of heterozygosity (LOH)
High expression of oligodendrocyitic development genes: PDGFRA, NKX2-2, OLIG2
Mutations of PI3KCA/PI3KR1 (10/16 tumors)
Expression of several neural development genes: DCX, DLL3, ASCL1, TCF4
NEURAL:
Expression of neuron markers: NEFL, GABRA1, SYT1, SLC12A5. Two normal brain tissues
were classified as Neural
Fig. 2.2 Properties of the four glioblastoma subtypes as reported by Verhaak et al. (2010)
subtype are presented in Fig. 2.2. Characteristically,

the four subtypes exhibited molecular expression
features reminiscent of distinct cell type gene sets
associated with neurons, oligodendrocytes, astrocytes and cultured astroglial cells. For instance, the
Proneural subtype was associated to an oligodendrocytic signature but not the astrocytic signature.
A clinical correlation could also be established with

each subtype. The most consistent one was age, with
younger patients overrepresented in the Proneural
subtype. This association survived the multicentric
origin of the patients. Furthermore, a comparison of
intensive treatments (concurrent chemotherapy and
radiotherapy or more than three subsequent cycles
21
of chemotherapy) on survival was done. It showed

a significant reduction in mortality in Classical and
Mesenchymal subtypes but not in Proneural subtype.
Finally, the methylation status of MGMT promoter
gene, previously associated to response to therapy,
was not found associated to subtype.
is that the identification of the small number of MR

generating and maintaining a specific tumor signature
is synonymous of efficient tools for diagnostic and
pharmacological interventions.
Integrative genome analysis has also been performed to the rapidly progressing microRNA (miRs)
field. The importance of the miRs in the oncogenic process and more precisely the metastatic progression has
been particularly explored in various tumor pathologies. In gliomasa number of papers have confirmed the
stimulatory role of certain miRs such as miR-21 in the
progression of metastasis. Down-regulation of miR-21
inhibits the EGFR pathway thus confirming the stimulating activity of this miR on oncongenic genes as
shown previously in other tumors (Zhou et al., 2010).
More recently, Johnsons group in Harvard reported the
characterization of miR-26a as a cooperating component of an amplicon the presence of which in human
glioblastomas is associated with a markedly decreased
survival (Kim et al., 2010). This amplicon contains
CDK4 and CENTG1, two oncogenes, regulating the
RB1 and PI3K/AKT pathways respectively. miR-26a
alone induces in vivo decrease of PTEN, RB1 and
MAP3K2/MEKK2 pathways, thereby promoting AKT
activation, proliferation and decreasing c-Jun terminal
kinase-dependent apoptosis. As a whole, these data
define a functional oncomiR/oncogene DNA cluster
that promotes aggressiveness of glioblastomas by targeting the RB1, PI3K/AKT and JNK pathways. While
these data should be completed by functional studies
that will definitely characterize the pathways targeted
by the different miRs, these results open the way
to a molecular follow-up of the tumors by use of
miR microarrays. The next approach would consist of
designing inhibitors of miRs in order to block some of
the steps in which their tumor stimulatory function is
involved.
Perspectives for Transfer to Patients Care

As a whole, this amazing set of data is going to
provide the clinician with a frame of relevant informations that might be exploited for diagnosis and
treatment. However, it is clear that at the present level,
only clinical departments with an access to specialized laboratories will acquire the capacity to treat their
patients in function of the recent data. Simplification
of analysis is therefore needed in terms of tests and
cost. A recent publication suggests that this could be
achieved (Carro et al., 2010). The authors postulated
that specific transcriptional signatures of tumors are
placed under the regulatory effects of master gene
regulators (MR). In other words, the model of gene
expression profile-based cancer analysis which is currently studied in most cancer pathologies could be
advantageously replaced by a cellular network analysis
of the MR controlling the signatures and the corresponding phenotypes (subtypes). Such an approach
was applied to the search of a transcriptional module of MRs regulating the molecular signature of
the mesenchymal subtype. This analysis identified
two transcriptional genes, C/EBP and STAT3, whose
simultaneous over-expression oriented cell neural stem
cells toward mesenchymal transformation. Most significantly, co-expression of the two factors was necessary and sufficient to reprogram the neural lineage
along the malignant mesenchymal lineage. In contrast, their simultaneous elimination led to reduction of
the mesenchymal signature and diminution of tumor
aggressiveness. These results were corroborated by
data showing a direct relationship between the expression levels of STAT3 and C/EBP in human GBM
and poor clinical outcome. The biological significance of these results is presently unknown as STAT3
induces astrocytic differentiation and counteracts neuronal differentiation while C/EBP induces neurogenesis and inhibits gliomagenesis. The principal conclusion regarding the clinical utility of these results
Neurosphere Cultures: Biomarkers

Themselves, or a Tool for Analysing
Molecular Biomarkers?
On the border of the biomarkers field, we have to
mention the ability for some gliomas to generate neurospheres. It has been shown not only that only a subset
of gliomas can generate neurospheres, but also that
this ability is predictive of overall survival of patients
22
(Laks et al., 2009). On a clinical point of view, several factors preclude clinical use of neurosphere. A
number of papers have shown that glioblastoma contain cells that can generate neurospheres when placed
under stem cell culture conditions. Because these cells
derived from tumors exhibit representative markers of
stemness (usually CD133) and because of their tumorigenic potential in vivo, it is generally admitted that
neurospheres might reflect the tumor cell origin. In this
respect, molecular analysis of individual neurospheres
could be of obvious interest to establish the tumor phenotype of each tumor and predict tumor response to
treatment and outcome. However, several points must
be clarified before achieving this goal. One of the
most recent papers on the subject (Chen et al., 2010)
reports a hierarchy of self-renewing tumor-initiating
cell types in glioblastoma. It has shown that individual glioblastoma multiforme carry a CD133+ /CD133
heterogeneous population of self-renewing cells with
tumor initiation capacity. These points are important
as they raise the crucial question of the representativity of cultures of neurospheres. Furthermore, while
the growth of neurospheres has already been shown to
be limited to a subset of glioblastoma, the new data
establish that PTEN deficiency of the tumor cells is
a necessary requirement for successful neurospheres
propagation. In other words, on the basis of these
two parameters cell heterogeneity and PTEN insufficiency the possibility exist that cell types initially
present in the tumor are not conserved in the cultures.
For these reasons and in spite of the obvious interest
for a better understanding of the biology of glioblastoma, other studies will have to be performed before
neurospheres acquire a biomarker status.
In conclusion, recent advances in molecular biology
have permitted a better understanding of the complex
mechanisms underlying gliomagenesis, and numerous
biomarkers appeared. Some of them LOH 1p/19q,
methylation status of the MGMT promoter gene have
proven such an association to overall survival as far as
groups of patients are concerned that stratifying analysis of clinical trials on these parameters has become a
standard. But uncertainties remain regarding what pertains to natural history of gliomas, and what pertains
to response to treatments, as assessed by biomarkers.
It is also the reason why their importance in the choice
between standard treatment and innovative therapies
during clinical trials on this basis should remain cautious. Due to these uncertainties, and other technical
M. Wager et al.
and standardization reasons, time has not come yet?for biomarkers based personalized medicine.
Many other biomarkers are currently under study,
and some of them IDH1 and IDH2 seem to be
on the verge of reaching the clinical field. Recent data
of integrated genomic analysis show that we are in
the edge of new approaches which will improve tools
given to clinicians for the diagnosis, decision making, and follow-up of adult gliomas. The molecular
characterization of glioblastoma main subcategories
will probably allow innovations in their classification.
It should benefit from data to come on micro-RNAs
(MiRs), as several preliminary results showed that their
status of expression could contribute to tumor signatures. One can conceive a personalized approach to
each tumor generating neurospheres, beginning with
culture. Neurospheres might be valuable for their
capacity at maintaining the original tumor, assessment
of response and study of resistance to treatments.
References
Aghili M, Zahedi F, Rafiee E (2009) Hydroxyglutaric aciduria
and malignant brain tumor: a case report and literature
review. J Neurooncol 91:233236
Balss J, Meyer J, Mueller W, Korshunov A, Hartmann C,
Von Deimling A (2008) Analysis of the IDH1 codon 132
mutation in brain tumors. Acta Neuropathol 116:597602
Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB,
Dewhirst MW, Bigner DD, Rich JN (2006) Glioma stem
cells promote radioresistance by preferential activation of the
DNA damage response. Nature 444:756760
Berger MS (1995) Role of surgery n diagnosis and management
of low grade gliomas. In: Appuzzo MLJ (eds) Benign cerebral glioma, vol II, Chapter 15. Neurosurgical topics. AANS
Publications, Park Ridge, IL, pp 293307
Cairncross JG, Ueki K, Zlatescu MC, Lisle DK, Finkelstein
DM, Hammond RR, Silver JS, Stark PC, Macdonald DR,
Ino Y, Ramsay DA, Louis DN (1998) Specific predictors of chemotherapeutic response and survival in patients
with anaplastic oligodendroglioma. J Natl Cancer Inst 90:
14731479
Carro MS, Lim WK, Alvarez MJ, Bollo RJ, Zhao X, Snyder EY,
Sulman EP, Anne SL, Doetsch F, Colman H, Lasorella A,
Aldape K, Califano A, Iavarone A (2010) The transcriptional
network for mesenchymal transformation of brain tumours.
Nature 463:318325
Chen R, Nishimura MC, Bumbaca SM, Kharbanda S, Forrest
WF, Kasman IM, Greve JM, Soriano RH, Gilmour LL,
Rivers CS, Modrusan Z, Nacu S, Guerrero S, Edgar KA,
Wallin JJ, Lamszus K, Westphal M, Heim S, James CD,
VandenBerg SR, Costello JF, Moorefield S, Cowdrey CJ,
Prados M, Phillips HS (2010) A hierarchy of self-renewing
23
tumor-initiating cell types in glioblastoma. Cancer Cell

17:362375
Dang L, White DW, Gross S, Bennett BD, Bittinger MA,
Driggers EM, Fantin VR, Jang HG, Jin S, Keenan MC,
Marks KM, Prins RM, Ward PS, En KE, Liau LM,
Rabinowitz JD, Cantley LC, Thompson CB, Vander Heiden
MG, Su SM (2009) Cancer- associated IDH1 mutations
produce 2-hydroxyglutarate. Nature 462:739744
Daumas-Duport C, Beuvon F, Varlet P, Fallet-Bianco C (2000)
Gliomas: WHO and Sainte-Anne hospital classifications.
Ann Pathol 5:413428
Hegi ME, Diserens AC, Gorlia T, Hamou MF, DeTribolet N,
Weller M, Kros JMJM, Hainfellner JA, Mason W, Mariani
L, Bromberg JEC, Hau P, Mirimanoff RO, Cairncross JG,
Janzer RC, Stupp R (2005) MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med
352:9971003
Heidorn SJ, Milagre C, Whittaker S, Nourry A, NiculescuDuvas I, Dhomen N, Hussain J, Reis-Filho JS, Springer CJ,
Pritchard C, Marais R (2010) Kinase-dead BRAF and oncogenic RAS cooperate to drive tumor progression through
CRAF. Cell 140:209221
Hormigo A, Gu B, Karimi S, Riedel E, Panageas KS, Edgar MA,
Tanwar MK, Rao JS, Fleisher M, DeAngelis LM, Holland
EC (2006) YKL-40 and matrix metalloproteinase-9 as potential serum biomarkers for patients with high grade gliomas.
Clin Cancer Res 12:56985704
Idbaih A, Omurao A, Ducray F, Hoang-Xuan K (2007)
Molecular genetic markers as predictors of response to
chemotherapy in gliomas. Curr Opin Oncol 19:606611
Jenkins RB, Blair H, Ballman KV, Giannini C, Arusell RM,
Law M, Flynn H, Pass S, Felten S, Brown PD, Shaw
EG, Buckner JC (2006) A t(1;19)(p10;q10) mediates the
combined deletions of 1 p and 19q and predicts a better
prognosis of patients with oligodendroglioma. Cancer Res
66:98529861
Jones DT, Kocialkowski S, Liu L, Pearson DM, Ichimura K,
Collins VP (2009) Oncogenic RAF1 rearrangement and a
novel BRAF mutation as alternatives to KIAA1549:BRAF
fusionin activating the MAPK pathway in pilocytic astrocytoma. Oncogene 28:21192123
Karayan-Tapon L, Quillien V, Guilhot J, Wager M, Fromont G,
Saikali S, Etchevery A, Hamlat A, Loussouarn D, Campion
L, Campone M, Vallette FM, Gratas-Rabbia-R C (2010)
Prognostic value of O6-methylguanine-DNA methyltransferase status in glioblastoma patients, assessed by five different methods. J Neurooncol 97:311322
Kim H, Huang W, Jiang X, Pennicooke B, Par PJ,
Johnson MD (2010) Integrative genome analysis reveals an
oncomir/oncogene cluster regulating glioblastoma survivorship. Proc Natl Acad Sci USA 107:21832188
Laks DR, Masterman-Smiyh M, Visnyel K, Angenieux B,
Orozco N, Foran I, Young WH, Vinters HV, Liau LM,
Lazareff JA, Mischel PS, Cloughesy TF, Horvath S,
Kornblum HL (2009) Neurosphere formation is an independent predictor of clinical outcome in malignant glioma. Stem
Cells 27:980987

Irvin D, Black KL, Yu JS (2006) Analysis of gene expression and chemoresistance of CD133+ cancer stem cells in
Louis DN, Ohgaki H, Wiestler OD, Cavenee WK (eds) (2007)
WHO Classification of tumours of the central nervous system, 3rd ed. IARC Press, Lyon
Mittler MA, Walters BC, Stopa EG (1996) Observer reliability
in histological grading of astrocytoma stereotactic biopsies.
J Neurosurg 85:10911094
Pallud J, Mandonnet E, Duffau H, Kujas M, Guillevin R,
Galanaud D, Taillandier L, Capelle L (2006) Prognostic
value of initial magnetic resonance imaging growth rates
for world health organization grade II gliomas. Ann Neurol
60:380383
Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt
P, Mankoo P, Carter H, Siu IM, Gallia GL, Olivi A,
McLendon R, Rasheed BA, Keir S, Nikolskaya T, Nikolsky
Y, Busam DA, Tekleab H, Diaz LA Jr, Hartigan J, Smith
DR, Strausberg RL, Marie SK, Shinjo SM, Yan H, Riggins
GJ, Bigner DD, Karchin R, Papadopoulos N, Parmigiani G,
Vogelstein B, Velculescu VE, Kinzler KW (2008) An integrated genomic analysis of human glioblastoma multiforme.
Science 321:18071812
Reifenberger J, Reifenberger G, Liu L, James CD, Wechsler W,
Collins VP (1994) Molecular genetic analysis of oligodendroglial tumors shows preferential allelic deletions on 19q
and 1p. Am J Pathol 145:11751190
Ricard D, Kaloshi G, Amiel-Benouaich A, Lejeune J, Marie Y,
Mandonnet E, Kujas M, Mokhtari K, Taillibert S, LaigleDonadey F, Carpentier AF, Omuro A, Capelle L, Duffau H,
Cornu P, Guillevin R, Canson M, Hoang-Xuan K, Delattre
JY (2007) Dynamic history of low-grade gliomas before and
after temozolomide treatment. Ann Neurol 61:484490
Sanai N, Berger MS (2008) Glioma extend of resection and its
impact on patient outcome. Neurosurg 62(4):7534766
Sanai N, Tramontin AD, Quinones-Hinojosa A, Barbaro NM,
Gupta N, Kunwar S, Lawton MT, McDermott MW, Parsa
AT, Manuel-Garcia Verdugo J, Berger MS, Alvarez-Buylla
A (2004) Unique astrocyte ribbon in adult human brain contains neural stem cells but lacks chain migration. Nature
427:740744
Stupp R, Mason WP, Van den Bent MJ, Weller M, Fisher
B, Taphoorn MJ, Belanger K, Brandes AA, Marosi C,
T, Allgeier A, Lacombe D, Cairncross JG, Eisenhauer E,
Mirimanoff ROand the European Organisation for Research
and Treatment of Cancer Brain Tumor and Radiotherapy
Groups; National Cancer Institute of Canada Clinial Trials
Group (2005) Radiotherapy plus concomitant and adjuvant
temozolomide for glioblastoma. N Engl J Med 352:987996
Tamber MS, Bansal K, Liang M-L, Mainprize TG, Salhia B,
Northcott P, Taylor M, Rutka JT (2006) Current concepts in
the molecular genetics of pediatric brain tumors: implication
for emerging therapies. Child Nerv Syst 22:13791394
Varlet P, Jouvet A, Miquel C, Saint-Pierre G, Beuvon F, DaumasDuport C (2005) Criteria of diagnosis and grading of oligodendrogliomas or oligoastrocytomas according to the WHO
and Sainte-Anne classifications. Neurochirurgie 51:239246
Varlet P, Soni D, Miquel C, Roux FX, Meder JF, Chneiweiss
H, Daumas- Duport C (2004) New variants of malignant
24
glioneuronal tumors: a clinicopathological study of 40 cases.
Neurosurg 55:13771391
Verhaak RGW, Hoadley KA, Purdom E, Wang V, Qi Y,
Wilkerson MD, Miller CR, Ding L, Golub T, Mesirov JP,
Alexe G, Lawrence M, OKelly M, Tamayo P, Weir BA,
Gabriel S, Winckler W, Gupta S, Jakkula L, Feiler HS,
Hogson JG, James CD, Sarkaria JN, Brennan C, Kahn A,
Spellman PT, Wilson RK, Speed TP, Gray JW, Meyerson
M, Getz G, Perou CM, Hayes DN, and Cancer Genome
Atlas Research Network (2010) Integrated genomic analysis
identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and
NF1. Cancer Cell 17:98110
Vescovi AL, Galli R, Reynolds BA (2006) Brain tumour stem
cells. Nat Rev Cancer 6:425436
Wager M, Guilhot J, Blanc JL, Ferrand S, Milin S, Bataille
B, Lapierre F, Denis S, Chantereau T, Larsen CJ, KarayanTapon L (2006) Prognostic value of increase in transcript
levels of Tp73 DeltaEx2-3isoforms in low-grade glioma
patients. Br J Cancer 95:10621069
Wager M, Menei P, Guilhot J, Levillain P, Michalak S, Bataille
B, Blanc JL, Lapierre F, Rigoard P, Milin S, Duthe F,
M. Wager et al.
Bonneau D, Larsen CJ, Karayan-Tapon L (2008) Prognostic
molecular markers with no impact on decision-making: the
paradox of glioma based on a prospective study. Br J Cancer
98:18301838
Watanabe T, Nobusawa S, Kleihues P, Ohgaki H (2009)
IDH1 mutations are early events in the development of
astrocytomas and oligodendrogliomas. Am J Pathol 174:
11491153
Weller M, Stupp R, Reifenberger G, Brandes AA, Van Den Bent
MJ, Wick W, Hegi ME (2010) MGMT promoter methylation
in malignant gliomas: ready for personalized medicine? Nat
Rev Neurol 6:3951
Zhao S, Lin Y, Xu W, Jiang W, Zha Z, Wang P, Yu W,
Li Z, Gong L, Peng Y, Ding J, Lei Q, Guan KL, Xiong
Y (2009) Glioma-derived mutations in IDH1 dominantly
inhibit IDH1 activity and induce HIF-1. Science 32:
261265
Zhou X, Ren Y, Moore L, Mei M, You Y, Xu P, Wang G, Jia Z,
Pu P, Zhang W, Kang C (2010) Downregulation of miR-21
inhibits EGFR pathway and suppresses the growth of human
glioblastoma cells independent of PTEN status. Lab Invest
90:144155
Chapter 3
Molecular Subtypes of Gliomas

Lonneke A.M. Gravendeel and Pim J. French
Abstract Gliomas are the most common primary

brain tumors with heterogeneous morphology and variable prognosis. However, differences between histological subclasses and grades are subtle, and classifying gliomas is subject to a large inter-observer
variability. This variability can result in misdiagnosis of gliomas. As treatment decisions in patients
rely mainly on histological classification and clinical
parameters, there is an urgent need of developing a
more accurate and objective classification model. The
identification of specific molecular markers (LOH of
1p19q, IDH1 mutation, MGMT methylation status), as
well as molecular clusters based on gene expression
have been studied extensively. These specific molecular features within gliomas can help diagnosis, can
give a more accurate prognosis, and may also be used
to develop personalized targeted therapy in the future.
Keywords Gliomas Molecular markers Molecular
cluster LOH 1p19q IDH1 MGMT
Introduction
Gliomas are the most common type of primary brain
tumors in adults with an incidence rate of 5.27 per
100,000 patients every year (Ohgaki et al., 2004;
Louis et al., 2007). In 1926, Bailey and Cushing
suggested a classification model based on distinct
P.J. French ()

Department of Neurology, Josephine Nefkens Institute,
Erasmus Medical Centre, 3000 DR Rotterdam,
The Netherlands
e-mail: p.french@erasmusmc.nl
histological morphologies (Bailey and Cushing, 1926),

which forms the basis of the currently used WHO classification (Louis et al., 2007). Two major subtypes are
recognized: Astrocytic (A) and oligodendrocytic (OD)
tumors, the latter including pure OD tumors and mixed
oligoastrocytic (MOA) tumors. Astrocytic tumors are
further separated into grades I (pilocytic astrocytomas
[PA]), II (low grade), III (anaplastic), and IV (glioblastoma [GBM]). Oligodendrocytic tumors are further
separated into grades II (low grade) and III (anaplastic). Patient survival, time to tumor progression, and
response to therapy are all associated with subtype and
grade of the tumor (Louis et al., 2007). This classification model, combined with the patients prognostic
features (e.g. age and Karnofsky Performance Score
[KPS]), guides treatment decisions.
Differences between histological subtypes are very
subtle, and classifying gliomas is subject to a large
interobserver variability (Scott et al., 1995; Murphy
et al., 2002; Kros et al., 2007; Hildebrand et al., 2008).
Clearly, this variability can result in misdiagnosis of
gliomas, for example by assigning a prognostically
favorable lower-grade glioma into a poor prognostic
glioma (e.g. false-positive GBM), and in assigning a
prognostically less favorable higher-grade glioma into
a good prognostic glioma (e.g. false-negative GBM).
Since treatment protocols often depend on the diagnosed histological subtype, accuracy in diagnosis is
very important for patients in order to get optimal treatment (Murphy et al., 2002). Therefore, more accurate
methods to diagnose gliomas are urgently required.
The molecular characteristics of gliomas have been
studied extensively over the last years, in order to
provide more objective and accurate methods of identifying distinct molecular tumor subgroups, and to
identify specific molecular tumor markers that can help

25
26
diagnosis. In the future, these molecular features may

also be used to develop personalized targeted therapy.
Single Molecular Markers in Gliomas

LOH 1p19q
Loss of heterozygosity (LOH) of 1p19q is a chromosomal aberration that is strongly associated with
classical ODs (Griffin et al., 2006; Jenkins et al.,
2006; Cairncross and Jenkins, 2008). Determination
of the 1p19 status in gliomas is clinically relevant
because of two important features. First, gliomas with
LOH of 1p19q generally grow more slowly than most
other gliomas and therefore have a better prognosis. Second, the presence of this mutation is predictive for response to treatment with alkylating agents
such as PCV (procarbazine, lomustine, and vincristine)
(Cairncross et al., 1998; Kouwenhoven et al., 2006).
However, many glial tumors benefit from alkylating
agents at some level. Nevertheless, 1p19q status is a
strong prognostic factor, and the presence or absence
of 1p and 19q is currently tested for in patients with a
tumor containing oligodendroglial features.
IDH1
Recently, somatic mutations in the gene encoding isocitrate dehydrogenase 1 (IDH1) have been identified
in gliomas (Parsons et al., 2008). IDH1 mutations
occur mainly in lower grade gliomas and in secondary
GBMs and are therefore thought to be early events in
glioma genesis (Hartmann et al., 2009). Interestingly,
glioma-specific mutations in IDH1 always affect the
amino acid arginine in position 132 of the amino acid
sequence, which belongs to an evolutionary highly
conserved region located at the binding site for isocitrate (Parsons et al., 2008; Bleeker et al., 2009).
The most frequent mutation occurring in this region
is the R132H mutation (>90%), but other variants
also have been found (R132S, R132C, R132G, and
R132L) (Bleeker et al., 2009; Gravendeel et al., 2010).
Interestingly, non-R132H mutations segregate in distinct histological and molecular subtypes of glioma
(Hartmann et al., 2009; Gravendeel et al., 2010).
L.A.M. Gravendeel and P.J. French
Histologically, non-R132H mutations occur sporadically in classic oligodendrogliomas and at significantly

higher frequency in other grade II and III gliomas.
Genetically, non-p.R132H mutations occur in tumors
with TP53 mutation, are virtually absent in tumors
with LOH on 1p and 19q and accumulate in distinct
(gene-expression profiling based) intrinsic molecular
subtypes (Gravendeel et al., 2009; 2010).
Importantly, IDH1 mutations are associated with
improved prognosis (Parsons et al., 2008; Kloosterhof
et al., 2010). Therefore IDH1 mutation status is likely
to be used as a prognostic molecular marker in the
near future. Additionally, two studies have examined
whether IDH1 mutation status can predict response to
treatment in gliomas. In a group of patients with dedifferentiated low-grade astrocytomas progressive after
radiotherapy response to TMZ did not differ between
IDH1 mutant and wild-type tumors (Dubbink et al.,
2009). In patients with anaplastic oligodendroglial
tumors treated with radiotherapy alone or radiotherapy with adjuvant PCV, IDH1 mutations reported had
no predictive value for response (van den Bent et al.,
2010).
MGMT
The
O6 -methylguanine-DNA
methyltransferase
(MGMT) gene encodes for the nuclear repair enzyme
alkyltransferase, which removes alkylating adducts
from the O6 position of thymine. By doing this,
the enzyme is involved in maintaining the integrity
of the DNA. More specifically, the product of the
MGMT gene protects the cells from being damaged
by alkylating and methylating agents (e.g. BCNU
[N,N[prime]-bis(2-chloroethyl)-N-nitrosourea],
procarbazine and TMZ).
In gliomas, the CpG islands located in the promotor
of the MGMT gene are frequently methylated, causing epigenetic silencing of this gene. Theoretically,
this methylation would result in a greater susceptibility for alkylating and methylating agents. In daily
practice, the meaning and implications of the MGMT
status are more difficult to interpret. Several studies
showed that the epigenetic silencing of the MGMT
gene is of clinical importance, because its association with increased survival and better response to
combined chemoirradiation in GBMs (Gerson, 2004;
Hegi et al., 2005). It was shown that the effect of

the MGMT silencing was especially present when
TMZ was used as chemotherapy (Hegi et al., 2005).
Other studies showed that the time of administering the
TMZ seemed to be crucial. One study showed that a
positive effect of the methylated MGMT gene on survival was only seen if TMZ was administered at the
time of radiation, while other studies showed positive
effects when administering TMZ before or after irradiation (Chakravarti et al., 2006; Everhard et al., 2006;
Criniere et al., 2007). Interestingly, a recent study
showed that MGMT promotor methylation in GBMs
was a predictor for better response to radiotherapy,
and a prognostic marker even in patients not receiving adjuvant alkylating chemotherapy (Rivera et al.,
2010). This study suggests that MGMT might be a general favorable prognostic factor in GBMs, instead of
being a predictive marker for response to alkylating
chemotherapy.
Another issue that makes it difficult to interpret
the role of MGMT, is that there are several different approaches in the assays for measuring the
MGMT promoter methylation status. These different
approaches cause variable results, which may be difficult to compare. MGMT activity can be measured
using immunohistochemistry, but also using Q-pcr, or
promotor methylation assays. Also, differences in outcome arise using assays on frozen tissue as well as on
paraffin samples, as well as by possible interobservervariability. Nevertheless, MGMT status is thought to
be an important biomarker in gliomas.
Genome Wide Molecular Markers

in Gliomas
Several techniques have been developed to perform
genome wide analysis of the tumors epigenome,
genome or transcriptome. Whilst markers for glioma
have been identified on all levels (TCGARN, 2008),
this review will discuss the molecular markers identified by the tumors transcriptome, that are currently
used (or show promise to aid) in clinical decision
making. Gene expression profiling involves the measurement of the activity (the expression) of thousands
of genes at once. In general, gene expression profiling
is performed using microarrays; chips that contain the
complementary sequence to thousands of target mRNA
27
sequences. mRNA isolated from tumor samples is processed and labeled and subsequently hybridized to the
microarray. The signal extracted from the microarray is a measure of gene expression levels, which is
visualized using fluorescence.
Expression profiling can be used to identify molecular subtypes of tumors roughly by two methods:
Supervised and unsupervised. Supervised clustering
uses external information to separate tumors into predefined subgroups (e.g. responders vs. non-responders;
long vs. short-survivors), and then specifically screens
for genes that are differentially expressed between
these groups. In contrast, unsupervised clustering does
not use external information and thus classifies tumors
based on homologies in gene expression profiles.
One of the first large studies that used supervised
clustering (on survival) has identified three large subtypes of glioma with distinct prognosis. Subtypes were
named according to the signature of the genes they predominantly expressed: Proneural, Mesenchymal and
Proliferative subtypes (Phillips et al., 2006). These
subtypes have also been identified in GBM using
unsupervised methods (Verhaak et al., 2010). A different supervised study identified genes associated
with response to treatment (French et al., 2005). Most
often however, supervised clustering has been used to
define gene expression signatures based on histological
subtypes (Nutt et al., 2003).
Thus far, only three groups have performed unsupervised analysis to define intrinsic molecular subgroups of gliomas (Gravendeel et al., 2009; Li et al.,
2009; Verhaak et al., 2010). In all cases, the unsupervised clusters identified more subtypes of gliomas
than histology. The molecular clusters correlate better with survival than histology (Nutt et al., 2003;
Gravendeel et al., 2009; Li et al., 2009). Therefore,
molecular clustering provides an objective and more
accurate method to classify gliomas, and may even
be used to predict patients prognosis. The molecular clusters all contained a wide variety of histological
subtypes. The fact that different histological subtypes
were assigned to the same molecular cluster means that
these phenotypically different tumors have a similar
genetic composition. Indeed, two independent studies have demonstrated that genetic changes segregate
into distinct molecular subtypes indicating that causal
genetic change drives a distinct pattern of gene expression (Gravendeel et al., 2009; 2010; Verhaak et al.,
2010).
28
For example, gliomas of different histological subtypes with LOH of 1p19q are accumulating within one
distinct molecular profile, regardless of their histological appearance, showing significant longer survival
times than other molecular subgroups (Gravendeel
et al., 2009). These findings imply that 1p19q status should be determined in all histological subtypes of gliomas, instead of testing this mutation in
oligodendroglial-like tumors only.
In the future, the specific genetic features of molecular glioma subgroups can be used to improve diagnosis, to give a more accurate prognosis, as well as to
develop personalized therapies. It is likely that each
molecular glioma subgroup will benefit from its own
specific treatment based on the specific underlying
molecular pathways and markers. Novel randomized
controlled trials should take these molecular clusters into account when comparing different therapy
regimens in gliomas.
References
Bailey P, Cushing HW (1926) A Classification of the tumors of
the glioma group on a histogenetic basis with a correlated
study of prognosis. J. B. Lippincott, Philadelphia, PA
Bleeker FE, Lamba S, Leenstra S, Troost D, Hulsebos T,
Vandertop WP, Frattini M, Molinari F, Knowles M, Cerrato
A, Rodolfo M, Scarpa A, Felicioni L, Buttitta F, Malatesta
S, Marchetti A, Bardelli A (2009) IDH1 mutations at
residue p.R132 (IDH1(R132)) occur frequently in high-grade
gliomas but not in other solid tumors. Hum Mutat 301:711
Cairncross JG, Ueki K, Zlatescu MC, Lisle DK, Finkelstein
DM, Hammond RR, Silver JS, Stark PC, Macdonald DR,
Ino Y, Ramsay DA, Louis DN (1998) Specific genetic
predictors of chemotherapeutic response and survival in
patients with anaplastic oligodendrogliomas. J Natl Cancer
Inst 90(19):14731479
Cairncross G, Jenkins R (2008) Gliomas with 1p/19q codeletion:
a.k.a. oligodendroglioma. Cancer J 14(6):352357
Chakravarti A, Erkkinen MG, Nestler U, Stupp R, Mehta
M, Aldape K, Gilbert MR, Black PM, Loeffler JS (2006)
Temozolomide-mediated radiation enhancement in glioblastoma: a report on underlying mechanisms. Clin Cancer Res
12(15):47384746
Criniere E, Kaloshi G, Laigle-Donadey F, Lejeune J, Auger
N, Benouaich-Amiel A, Everhard S, Mokhtari K, Polivka
M, Delattre JY, Hoang-Xuan K, Thillet J, Sanson M
(2007) MGMT prognostic impact on glioblastoma is dependent on therapeutic modalities. J Neuro-Oncol 83(2):
173179
Dubbink HJ, Taal W, van Marion R, Kros JM, van Heuvel
I, Bromberg JE, Zonnenberg BA, Zonnenberg CB, Postma
TJ, Gijtenbeek JM, Boogerd W, Groenendijk FH, Smitt PA,
Dinjens WN, van den Bent MJ (2009) IDH1 mutations in
L.A.M. Gravendeel and P.J. French

low-grade astrocytomas predict survival but not response to
temozolomide. Neurology 73(21):17921795
Everhard S, Kaloshi G, Criniere E, Benouaich-Amiel A, Lejeune
J, Marie Y, Sanson M, Kujas M, Mokhtari K, Hoang-Xuan K,
Delattre JY, Thillet J (2006) MGMT methylation: a marker of
response to temozolomide in low-grade gliomas. Ann Neurol
60(6):740743
French PJ, Swagemakers SM, Nagel JH, Kouwenhoven MC,
Brouwer E, van der Spek P, Luider TM, Kros JM, van den
Bent MJ, Sillevis Smitt PA (2005) Gene expression profiles
associated with treatment response in oligodendrogliomas.
Cancer Res 65(24):1133511344
Gerson SL (2004) MGMT: its role in cancer aetiology and
cancer therapeutics. Nat Rev Cancer 4(4):296307
Gravendeel LA, Kouwenhoven MC, Gevaert O, de Rooi JJ,
Stubbs AP, Duijm JE, Daemen A, Bleeker FE, Bralten LB,
Kloosterhof NK, De Moor B, Eilers PH, van der Spek
PJ, Kros JM, Sillevis Smitt PA, van den Bent MJ, French
PJ (2009) Intrinsic gene expression profiles of gliomas are
a better predictor of survival than histology. Cancer Res
69(23):90659072
Gravendeel LA, Kloosterhof NK, Bralten LB, van Marion R,
Dubbink HJ, Dinjens W, Bleeker FE, Hoogenraad CC,
Michiels E, Kros JM, van den Bent M, Smitt PA, French
PJ (2010) Segregation of non-p.R132H mutations in IDH1
in distinct molecular subtypes of glioma. Hum Mutat
31(3):E1186E1199
Griffin CA, Burger P, Morsberger L, Yonescu R, Swierczynski
S, Weingart JD, Murphy KM (2006) Identification of
der(1;19)(q10;p10) in five oligodendrogliomas suggests
mechanism of concurrent 1p and 19q loss. J Neuropathol Exp
Neurol 65(10):988994
Hartmann C, Meyer J, Balss J, Capper D, Mueller W, Christians
A, Felsberg J, Wolter M, Mawrin C, Wick W, Weller M,
Herold-Mende C, Unterberg A, Jeuken JW, Wesseling P,
Reifenberger G, von Deimling A (2009) Type and frequency
of IDH1 and IDH2 mutations are related to astrocytic and
oligodendroglial differentiation and age: a study of 1,010
diffuse gliomas. Acta Neuropathol 118(4):469474
Hegi ME, Diserens AC, Gorlia T, Hamou MF, de Tribolet N,
Weller M, Kros JM, Hainfellner JA, Mason W, Mariani L,
Bromberg JE, Hau P, Mirimanoff RO, Cairncross JG, Janzer
RC, Stupp R (2005) MGMT gene silencing and benefit
from temozolomide in glioblastoma. N Engl J Med 352(10):
9971003
Hildebrand J, Gorlia T, Kros JM, Afra D, Frenay M, Omuro
A, Stupp R, Lacombe D, Allgeier A, van den Bent MJ,
Investigators, E. B. T. G (2008) Adjuvant dibromodulcitol
and BCNU chemotherapy in anaplastic astrocytoma: results
of a randomised European organisation for research and
treatment of cancer phase III study (EORTC study 26882).
Eur J Cancer 44(9):12101216
Jenkins RB, Blair H, Ballman KV, Giannini C, Arusell RM,
Law M, Flynn H, Passe S, Felten S, Brown PD, Shaw
EG, Buckner JC (2006) A t(1;19)(q10;p10) mediates the
combined deletions of 1p and 19q and predicts a better
prognosis of patients with oligodendroglioma. Cancer Res
66(20):98529861
Kloosterhof NK, Bralten LB, Dubbink HJ, French PJ, van den
Bent MJ (2010) Isocitrate dehydrogenase-1 mutations: a fundamentally new understanding of diffuse glioma? Lancet
Oncol. Epub ahead of print
Kouwenhoven MC, Kros JM, French PJ, Biemond-ter Stege EM,

Graveland WJ, Taphoorn MJ, Brandes AA, van den Bent MJ
(2006) 1p/19q loss within oligodendroglioma is predictive
for response to first line temozolomide but not to salvage
treatment. Eur J Cancer 42(15):24992503
Kros JM, Gorlia T, Kouwenhoven MC, Zheng PP, Collins VP,
Figarella-Branger D, Giangaspero F, Giannini C, Mokhtari
K, Mork SJ, Paetau A, Reifenberger G, van den Bent
MJ (2007) Panel review of anaplastic oligodendroglioma
from European Organization For Research and Treatment of
Cancer Trial 26951: assessment of consensus in diagnosis,
influence of 1p/19q loss, and correlations with outcome. J
Neuropathol Exp Neurol 66(6):545551
Li A, Walling J, Ahn S, Kotliarov Y, Su Q, Quezado M,
Oberholtzer JC, Park J, Zenklusen JC, Fine HA 2009
Unsupervised analysis of transcriptomic profiles reveals six
glioma subtypes. Cancer Res 69(5):20912099
WHO Classification of tumours of the central nervous system, Lyon
Murphy M, Loosemore A, Ferrer I, Wesseling P, Wilkins PR,
Bell BA (2002) Neuropathological diagnostic accuracy. Br J
Neurosurg 16(5):461464
Nutt CL, Mani DR, Betensky RA, Tamayo P, Cairncross JG,
Ladd C, Pohl U, Hartmann C, McLaughlin ME, Batchelor
TT, Black PM, von Deimling A, Pomeroy SL, Golub TR,
Louis DN (2003) Gene expression-based classification of
malignant gliomas correlates better with survival than histological classification. Cancer Res 63(7):16021607
Ohgaki H, Dessen P, Jourde B, Horstmann S, Nishikawa T,
Di Patre PL, Burkhard C, Schuler D, Probst-Hensch NM,
Maiorka PC, Baeza N, Pisani P, Yonekawa Y, Yasargil
MG, Lutolf UM, Kleihues P (2004) Genetic pathways to
glioblastoma: a population-based study. Cancer Res 64(19):
68926899
McLendon R, Rasheed BA, Keir S, Nikolskaya T, Nikolsky
Y, Busam DA, Tekleab H, Diaz LA Jr., Hartigan J, Smith
29
Science 321(5897):18071812
Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH,
Wu TD, Misra A, Nigro JM, Colman H, Soroceanu L,
Williams PM, Modrusan Z, Feuerstein BG, Aldape K (2006)
Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble
stages in neurogenesis. Cancer Cell 9(3):157173
Rivera AL, Pelloski CE, Gilbert MR, Colman H, De La Cruz C,
Sulman EP, Bekele BN, Aldape KD (2010) MGMT promoter
methylation is predictive of response to radiotherapy and
prognostic in the absence of adjuvant alkylating chemotherapy for glioblastoma. Neurooncology 12(2):116121
Scott CB, Nelson JS, Farnan NC, Curran WJ Jr., Murray KJ,
Fischbach AJ, Gaspar LE, Nelson DF (1995) Central pathology review in clinical trials for patients with malignant
glioma. A Report of Radiation Therapy Oncology Group
83-02. Cancer 76(2):307313
TCGARN (2008) Comprehensive genomic characterization
defines human glioblastoma genes and core pathways.
Nature 455(7216):10611068
van den Bent MJ, Dubbink HJ, Marie Y, Brandes AA, Taphoorn
MJ, Wesseling P, Frenay M, Tijssen CC, Lacombe D, Idbaih
A, van Marion R, Kros JM, Dinjens WN, Gorlia T, Sanson
M (2010) IDH1 and IDH2 mutations are prognostic but not
predictive for outcome in anaplastic oligodendroglial tumors:
a report of the European Organization for Research and
Treatment of Cancer Brain Tumor Group. Clin Cancer Res
16(5):15971604
Verhaak RG, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson
MD, Miller CR, Ding L, Golub T, Mesirov JP, Alexe G,
Lawrence M, OKelly M, Tamayo P, Weir BA, Gabriel S,
Winckler W, Gupta S, Jakkula L, Feiler HS, Hodgson JG,
James CD, Sarkaria JN, Brennan C, Kahn A, Spellman
PT, Wilson RK, Speed TP, Gray JW, Meyerson M, Getz
G, Perou CM, Hayes DN Cancer Genome Atlas Research,
N (2010) Integrated genomic analysis identifies clinically
relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell
17(1):98110
Chapter 4
Glioblastoma: Germline Mutation of TP53

Haruhiko Sugimura, Hidetaka Yamada, Shinji Kageyama, Yasuhiro Yamamura, Naoki
Yokota, Hiroki Mori, Moriya Iwaizumi, Kazuya Shinmura, Kiyotaka Kurachi, Toshio
Nakamura, Masaru Tsuboi, Masato Maekawa, and Tomoaki Kahyo
Abstract Various brain tumors are components of

familial cancer syndromes, and Li-Fraumeni syndrome
and Li-Fraumeni-like syndromes are the most famous
entities. A rare, sporadic occurrence of brain tumor
in peculiarly young subjects, however, sometimes provides a clue to understanding of carcinogenesis due to
germline mutations. In this section, two cases of glial
tumors in subjects having germline mutation of TP53
are presented, and genetic etiology of brain tumor is
discussed. The first case is a 41 year-old father of a
fatal adrenocortical carcinoma case of a 4-years-old
daughter, 16 years prior to the occurrence of astrocytoma of himself. The search for family history and
previous clinical records in the four different community hospitals for the last three decades and the followups of his family members disclosed cancer of urinary
bladder, cancer of pancreas, hepatoblastoma, and thymoma in the relatives. Germline mutation E286A of
TP53 was identified in the affected members. Another
case is a 21-year-old male, without any family history
of cancer, who suffered from brain tumor and colorectal cancer. An attending physician was sticky to find
an etiology in this unusual occurrence of cancer. After
several attempts in vain, an I195T germline mutation
was identified and functional analysis was performed.
These anecdotes highlight importance of genetic analysis in case of glial tumors in relatively young adult or
in adolescence whether or not they have family history of cancer. The problems and strategies to find
H. Sugimura ()
Department of Investigative Pathology I, Hamamatsu
University School of Medicine, Higashi-ward, Hamamatsu
431-3192, Japan
e-mail: hsugimur@hama-med.ac.jp
TP53 mutation carriers and to prevent or to delay the

occurrence of the tumors in them are discussed.
Keywords Glioblastoma Germline mutation TP53
Mismatch deficiencies Inhibitors Tumors
Introduction
Traditional epidemiological studies have identified few
environmental risk factors for malignant brain tumors
(Osborne et al., 2001), but genetic components of the
etiology of brain tumors, although rare, have been relatively well-defined for glial tumors (Schwartzbaum
et al., 2006). Brain tumors are often accompanied
by a genetic cancer syndrome such as Li-Fraumeni
syndrome (LFS1, OMIM accession 151623), CHEK2related syndrome (LFS2, OMIM accession 609265),
Li-Fraumeni-like syndrome (LFLS), Maffuci syndrome (OMIM accession 166000), Olier syndrome
(OMIM accession 166000), tuberous sclerosis (OMIM
accession 191100), or von Hippel-Lindau syndrome
(OMIM accession 193300). The last two of these
syndromes are usually related to tumors having specific histopathology, subependymal giant astrocytoma
and hemangioblastoma, respectively, but the other
syndromes accompany a variety of glial tumors and
choroid plexus tumors. TP53 is undoubtedly the most
influential genetic factor related to the occurrence
of human brain tumors. In this review, we report
two cases in which a glial tumors was caused by a
germline mutation of TP53. The first case one was
diagnosed in a family in which several members had
suffered from various malignancies for decades and
consulted different community hospitals in the county
where they lived, which had a population of 700,000.

31
32
A definitive diagnosis of Li-Fraumeni syndrome was

finally made by identifying the germline mutation of
TP53 (Sameshima et al., 1992), 20 years after the
first cancer occurred in the family. The second case
was a sporadic case of synchronous colon cancer and
glioblastoma multiforme. TP53 sequencing is not routine in such cases, and subsequent identification of the
mutation was a very unexpected positive finding. In
this chapter we review the TP53 germline mutations
identified in glial tumor cases that have been reported
in the literature, and we discuss the importance of functional assessment of the variants. We also address the
issue of genetic testing for TP53 mutations in patients
with glial tumors in situations that would arouse suspicion in most experienced clinicians, such as their
occurrence in particularly young persons or clustered
in a family, especially in routine clinical practice, and
we offer perspectives in regard to the management
of TP53 mutant carriers, that may seem bold at the
moment.
Case Reports
Case 1
A 41-year-old male was diagnosed with a gemistocytic astrocytoma (Fig. 4.1a), and a retrospective and
follow-up study of his family over a 40-year period
revealed cases of hepatoblastoma, adrenocortical carcinoma, thymoma, pancreatic cancer, and stomach
cancer (Fig. 4.1b). A germline TP53 E286A mutation was identified (Sameshima et al., 1992), and there
was marked TP53 overexpression in the astrocytoma
(Fig. 4.1c). The functional significance of the E286A
mutation is not definitively recapitulated in vitro, but
the DNA-damage-associated dysregulation of the cell
cycle has been investigated in cells derived from carriers (Goi et al., 1997). The spectrum of cancers in this
family is typical of Li-Fraumeni syndrome.
Case 2
A 21-year-old male with no family history of cancer
simultaneously developed symptoms of colorectal cancer and a brain tumor at the same time. The brain
tumor was diagnosed as a glioblastoma multiforme
H. Sugimura et al.
(Fig. 4.1d, e), and marked TP53 overexpression was

detected in the glioblastoma cells. The histological
diagnosis of the colon cancer was well-differentiated
adenocarcinoma (Fig. 4.1f), and TP53 was detected
in the tumor cells (Fig. 4.1g). The curiosity of the
attending physician was aroused by the clinical phenotype of this patient, and the physician suspected
one of the genetic syndromes. A tentative diagnosis of
Turcot syndrome was made, but the colon cancer did
not exhibit microsatellite instability. The patient and
his family were very cooperative in regard to further
attempts to determine the etiology of his disease. A
genetic counselor explained several possible etiologies
of the patients clinical phenotype and the importance
of conducting a genetic analysis. DNA from the blood
cells of the patient and his healthy parents was used to
sequence the entire germline TP53, including all of the
exons, and an I195T mutation, which has been known
as a hot spot for somatic mutations, was found as de
novo germline mutation in the patients gene but not
in his parents. This missense variant TP53 is a wellknown somatic mutation but has never been recorded
in Li-Fraumeni syndrome. The locus at the codon
195 has been found to alter transcriptional activity in
a yeast system (Petitjean et al., 2007), and deficient
transactivation capacity has been demonstrated in a
mammalian cell system in vitro by co-transfection with
p53 target molecules (Fig. 4.1h) (Yamada et al., 2009).
TP53 and Glial Tumors

Glial tumors of varying biological grades are often
encountered in both Li-Fraumeni syndrome and
Li-Fraumeni-like syndrome (Pearson et al., 1982;
Santibanez-Koref et al., 1991). Lis early collection
included 14 brain tumors, 9 of which were gliomas.
All of the patients were under age 45 years old, and
the mode was 1529 years old (Li et al., 1988).
Germline mutations of TP53 have been one of
the best-known genetic etiologies of human cancers, including brain tumors. Cases of human cancer
consistent with and not consistent with Li-Fraumeni
syndrome have been investigated for germline mutations of TP53. The latest database of TP53 germline
mutations (IARC TP53 database, an R14 release,
November, 2009) indicates that 138 of the 1053
recorded germline mutations of TP53 have occurred
33
Relative luciferase activity
d
Fig. 4.1 (a) Astrocytoma, grade 2, gemistocytic, of Case 1.
Hematoxylin-eosin stain; (b) A family pedigree of the Case 1.
An arrow indicates a case of astrocytoma. 1, pancreas cancer
at age 36. 2, bladder cancer at age 20. 3, astrocytoma at 41.
4, hepatoblastoma at 2. 5, hepatoblastoma at 2. 6, thymoma
at 17. Members 3, 5 and 6 were revealed to be carriers of
the TP53 mutation (Sameshima et al., 1992). Closed circles
and rectangles are carriers identified. The other members were
not tested for TP53 mutation; (c) TP53 immunostaining of
astrocytoma of the Case 1. Nuclear staining is prominent; (d)
Glioblasoma multiforme of the Case 2. Hematoxylin eosin stain;
(e) Glioblastoma multiforme, anti-glial fibrillary acidic protein
1.0
WT
Thr195
Empty
0.5
0.0
p21
BAX
MDM2
(GFAP) (f) Adenocarcinoma of the colon in the Case 2; (g) TP53

immunostaining of colon cancer in the Case 2; (h) Transcription
repression by mutant TP53 in the Case 2. Evaluation of the
transcriptional activation function of p53 by luciferase assay.
p21-, BAX-, and MDM2-luciferase activities were measured in
p53-null H1299 cells transiently transfected with a p53 expression vector, the firefly luciferase reporter vector pGL4.10, and
the transfection control vector pGL4.74. Values for luciferase
activity are means standard deviation of three independent
experiments. For each p53-responsive gene, the luciferase activity of cells transfected with wild-type (wt) p53 was set at 1.0
34
in patients with a brain tumor. Even in the absence of

characteristic tumors, such as breast cancer, soft tissue sarcoma, and adrenocortical cancers, which are
the characteristic tumors of Li-Fraumeni syndrome,
there are the cases of familial aggregation, particularly of brain tumors alone (Kyritsis et al., 1994)
surveyed immortalized lymphocytes derived from the
gliomas of 51 patients and found TP53 mutations in
6 of them. They proposed that multiplicity (in the
brain and other organs) and a familial history implies
the presence of a TP53 germline mutation, and they
pointed out that 6 subjects in 9 pedigrees were in their
30s when their glial tumors were diagnosed. These
reports are instructive when we encounter brain tumor
cases in adolescents and young adults in our practice.
More than 20 different germline missense mutations
have been reported in the germline of patients with
glial tumors of various biological grades, ranging from
oligodendroglioma to astrocytoma and to glioblastoma
multiforme.
Is Any Specific Mutation Spectrum

Associated with Brain Tumors?
Examination of the updated IARC database, suggests two questions in regard to the relation between
germline TP53 mutations and brain tumors. The first
question concerns whether there are germline TP53
mutations that specify the occurrence of a tumor in a
particular organ. Lubbe et al. reported a familial brain
tumor syndrome associated with codon 236 (Lubbe
et al., 1995), and Vital et al. (1998) reported two
families with a cluster of astrocytomas and choroid
plexus tumors caused by codon 248 mutations. The
specific significance of this mutation in these families
seems to have drawn a great deal of attention since
then, but these mutations have never been particularly
questioned as brain tumor-causing germline mutations
(Kleihues et al., 1997) reviewed 91 families with
germline mutations and brain tumors and found an earlier age of onset of their brain tumors than in families
with brain tumors but no germline mutations, but they
were not able to identify any particular spectrum of
germline mutations that tended to be more common in
patients with brain tumors than in patients with tumors
at other sites. The spectrum of somatic TP53 mutations
has been extensively investigated and widely accepted
H. Sugimura et al.
as carcinogen fingerprints of the tumors (Hussain et al.,

2000), but there has been little assessment of the spectrum of germline TP53 mutations in terms of a possible
relationship to the specific phenotype of the proband
and the probands descendants. Varley summarized
the germline mutations in Li-Fraumeni syndrome and
other syndromes, and claimed to have detected a mutation specificity preferentially associated with adrenocortical carcinoma (Varley, 2003). That claim requires
validation by thorough testing of family members, but
the concept of specific germline mutations deserves
further corroboration. In any event, no spectrum of
germline mutations has ever been identified in brain
tumors patients.
The second question concerns whether the brain
tumors that develop in persons with a germline TP53
mutation have any specific characteristics in common.
Bogler (Bogler et al., 1995) reviewed the spectrum of
TP53 mutations in brain tumors and found that loss of
a remaining wild-type allele was the most important
second hit in tumorigenesis and loss of the remaining
wild-type allele was detected in our case 2. His review
focused on the nature of somatic genetic changes in
brain tumors.
Glioblastoma multiforme is one of the targets of The
Atlas of Cancer Genome, prepared in the United States
(Ledford, 2010). Watanabe (Watanabe et al., 2009)
recently reported identifying an R132C mutation of
the IDH1 gene in an astrocytoma in a patient with the
Li-Fraumeni syndrome. The recent next generation
sequencing has generated tremendous information on
genetic changes in glioblastoma (Verhaak et al., 2010).
IDH gene mutations are one of the most remarkable
discoveries of high-throughput sequencing projects in
regard to glioblastoma in general (Yan et al., 2009),
and since 12% of glioblastomas have been found
to harbor a IDH1 mutation (Parsons et al., 2008),
the IDH1 mutations discovered by Watanabe are not
specific for glioblastoma multiforme associated with
Li-Fraumeni syndrome. The TP53 mutations often
coincide found in germline cells are often the same
as those found in somatic cells. With few exceptions,
the genotype-phenotype relationship in genetic cancer
diseases has not been extensively studied. The landscape of somatic genetic changes, including mutations
and copy number amplifications, in tumors that occur
in persons with a particular genetic background will
soon become available, and it will be necessary to wait
until enough cases accumulate to draw conclusions as
to whether any correlation exists between the spectrum

of germline mutations and spectrum of phenotypic
(somatic) mutations in patients with the tumors.
When and How Should Requests Be Made

to Perform Genetic Testing and How and
When Should Testing Be Performed?
It is not easy to obtain genomic information, especially for practicing physicians, when there is no or
little family history of cancer or when the patients
relatives are not very familiar with recent genetic
knowledge, including the significance of DNA. Even
after overcoming the hurdles of the difficulty of obtaining DNA and conducting a genomic analysis, the odds
of obtaining a result that can completely explain the
patients condition are seldom high. Several studies,
including the study by Paunu et al. (Paunu et al.,
2001), did not reveal any germline TP53 mutations in
familial glial tumors. Portwine also reported negative
findings in a search for germline TP53 mutations in
children with sporadic brain tumors (Portwine et al.,
2001). The prevalence of germline TP53 mutations that
have biological significance in asymptomatic populations appears to be extremely low, perhaps only 0.3%
(Palmero et al., 2008), and it is difficult to persuade the
healthy relatives to undergo genetic testing. We consider the likelihood of detection of a germline TP53
mutation in the adolescent cancer patient in Case 2
to have been very small. Most brain tumors are not
thought of as genetic or hereditary diseases, and brain
tumors are still basically classified according to their
morphology and somatic genetic information (Burger
and Scheithauer, 2007). Despite anecdotal reports of
brain tumors as genetic manifestations, the overall
attempt to identify constitutional genetic changes in
brain tumor patients has been tedious until recently.
Li et al. (1995) investigated TP53 germline mutations
in 80 unselected glioma patients and found a TP53
germline mutation in only 1 of the 65 cases, that did not
have a peculiar family history, as opposed in 3 of the 15
cases with a familial or unusual personal history. As
stated in the introduction, taking a complete family history is often difficult, and cases in family members who
would have developed cancer but died at an early age
of some other cause would be missed. When the parents have few siblings, it is very difficult to conclude
35
that the occurrence of a disease in a particular case is

sporadic. Then, how about unusual cases? In many
of the cases in which we persuaded young patients and
family members to undergo DNA testing and obtained
their consent the results were negative, and the reason
the patient suffered this particular disease, which was
unusual for his or her age remained a mystery to the
parents or relatives. There are still no valid scientific
criteria for performing or recommending genetic tests
in clinical settings. The criteria are often subjective and
depend on the feeling of the attending physician or
his or her enthusiasm regarding how far to pursue the
investigation of the cause of the disease in cases that
arouse suspicion in attending clinicians. Unusual clusterings of malignant tumors in a family attract general
practitioners attention, but the absence of a family history of cancer or difficulty in obtaining it often impede
further investigation by genetic analysis. The recent
worldwide trend toward small families further reduces
the possibility of identifying a peculiar pedigree to
analyze for genetic disease. Then, can luck alone be
expected to result in the identification of significant
mutations?
In compensation the era of small family sizes and
the rigorous procedures required to obtain patient
consent, a wide variety of genetic test, technologies
are now available and are easier and friendlier than
ever before. A genome-wide, personal whole genome
sequence strategy is now available, and it will become
less costly in the near future. The introduction of cytogenetic arrays in recent years has facilitated the discovery of unexpected changes in copy numbers, small
deletion/insertion polymorphisms, amplifications, and
rather rare (0.15%) single nucleotide polymorphisms
(SNPs) when we attempt to determine the etiology of a
common disease in an unusual situation (Rieber et al.,
2009; Schwarzbraun et al., 2009). The era of personal
sequencing is coming, and we, as one of the part of
medical systems these days, i.e., as a patient, a family
member of the patient, counselor, physician, insurer,
or scientist, may have to prepare for the time when
germline tests will be routinely performed in every
patient with an unusual brain tumor.
Syndromes Expanding?
Mismatch deficiencies are one of the well known
genetic causes of brain tumors. Paraf proposed the
36
existence of two types of brain tumor-polyposis

(BTP) syndrome (Paraf et al., 1997). The BTP
syndrome type 1 is characterized by non-polypotic
(thus, the name is paradoxical) colorectal tumors and
microsatellite instability, the same as found in hereditary non-polypotic colorectal cancer (HNPCC). BTP1
syndrome corresponds to Turcot syndrome (MIM
accession 270630) and/or Muir-Torre syndrome (MIM
accession 158320), which can also have glioblastoma
multiforme as a component of the syndrome (Park
et al., 2009). BTP syndrome type 2 is characterized
by brain tumors associated with adenomatous polyposis coli. We are not sure whether our Case 2 is a
case of BTP syndrome type 1, because of the absence
of microsatellite instability in the colorectal cancer in
Case 2. It would be inappropriate to call cases like
Case 2 brain tumor-polyposis syndrome. In conclusion, we are compelled to expand the concept of brain
tumor-colorectal tumor syndrome to include what we
refer to here as TP53 disorders.
Problems in Screenings and Preventions

The unusual occurrence of common cancers in subjects harboring a germline mutation of TP53 prompted
us to consider the feasibility of detecting or mass
screening for such carriers in the general population.
An R337H mutation of TP53 was found in two of
750 healthy subjects in a study conducted in southern
Brazil (Palmero et al., 2008). A thorough characterization of this germline mutation (variant) in terms
of its penetrance and the cancer spectrum of the subjects families is under way, and justification of neonatal mass screening for this variant has been debated
(Achatz et al., 2009). To justify this kind of attempts
of neonatal mass screening, we must be ready and feasible for detection of the early occurrence of tumors
in carriers, and we also have to have the tools to treat
the tumors or to reduce the hazard caused by tumors
in the carriers. Endoscopic surveillance of the gastrointestinal (GI) tract for cancer and surveillance for
breast cancer by mammography would be an acceptable choice for many subjects instead of more invasive
measures. Some GI tumors can be treated endoscopically, and some breast tumors can be cured by minimally invasive surgery. The situation in regard to brain
tumors is different. CT, MRI, and positron emission
H. Sugimura et al.
tomography (PET) are powerful diagnostic imaging

methods for detecting small tumors, but is it acceptable for the carriers of germline TP53 mutations to
undergo craniotomy or stereotactic radioneurosurgery?
Since carriers of germline TP53 mutations are thought
to be more sensitive to radiation it is not certain that
the gamma knife is the best choice of treatment.
Many investigators are attempting to find agents
that will reduce the risk associated with a germline
heterogeneous TP53 mutation. One bold, promising
strategy would be to use p53-activating drugs to
treat patients with germline heterozygosity of p53.
Deacetylase inhibitors are promising candidates for
such drugs that would reduce the risk associated
with germline heterogeneous TP53 mutation, because
acetylation of the C-terminal region of TP53 protein up-regulates its transcriptional and transactivational activity. Treatment with a deacetylase inhibitor
is expected to compensate for the depleted function of
mutant p53 in the p53-heterozygous cells by increasing the acetylation level of wild-type p53. One of the
deacetylase inhibitors, the sirtuin (histone deacetylase
type [HDAC] III) -targeting inhibitor, is particularly
noteworthy, because the sirtuin (HDAC III) is catalytically different from the other histone deacetylases.
The members of the sirtuin deacetylase family require
NAD to exert their deacetylase activity, and these characteristics may lead to novel efficacy against cancer
cells and cancer-predisposing cells. There is a mouse
model with haploinsufficiency of p53 that is prone to
develop various tumors (Jacks et al., 1994), and tailless, a nuclear receptor gene has recently been shown
to contribute to brain tumor initiation in the p53 haploinsufficient mouse (Liu et al., 2010). Such models
will provide monitoring systems to test the preventive
efficacy of drugs against tumorigenesis in persons with
greater genetic susceptibility to cancers. Various sirtuin inhibitors, both natural and synthetic, have been
evaluated in vitro (Kahyo et al., 2008), and application
of these drugs to an animal model and then clinically
must be assessed in the near future.
References
Achatz MI, Hainaut P, Ashton-Prolla P (2009) Highly prevalent
TP53 mutation predisposing to many cancers in the Brazilian
population: a case for newborn screening? Lancet Oncol
10:920925
Bogler O, Huang HJ, Kleihues P, Cavenee WK (1995) The p53

gene and its role in human brain tumors. Gila 15:308327
Burger PC, Scheithauer BW (2007) Tumor of the central nervous
system, vol 7. American Registry of Pathology and Armed
Forces Institute of Pathology, Washington, DC
Goi K, Takagi M, Iwata S, Delia D, Asada M, Donghi R,
Tsunematsu Y, Nakazawa S, Yamamoto H, Yokota J, Tamura
K, Saeki Y, Utsunomiya J, Takahashi T, Ueda R, Ishioka
C, Eguchi M, Kamata N, Mizutani S (1997) DNA damageassociated dysregulation of the cell cycle and apoptosis
control in cells with germ-line p53 mutation. Cancer Res
57:18951902
Hussain SP, Hollstein MH, Harris CC (2000) p53 tumor suppressor gene: at the crossroads of molecular carcinogenesis,
molecular epidemiology, and human risk assessment. Ann
NY Acad Sci 919:7985
Jacks T, Remington L, Williams BO, Schmitt EM, Halachmi S,
Bronson RT, Weinberg RA (1994) Tumor spectrum analysis
in p53-mutant mice. Curr Biol 4:17
Jeannot E, Mellottee L, Bioulac-Sage P, Balabaud C, Scoazec
JY, Van Nhieu JT, Bacq Y, Michalak S, Buob D, Genthep
(Inserm network), Laurent-Puig P, Rusyn I, Zucman-Rossi
J (2010) Spectrum of HNF1A somatic mutations in hepatocellular adenoma differs from that in MODY3 patients and
suggests genotoxic damage. Diabetes. 59:18361844
Kahyo T, Ichikawa S, Hatanaka T, Yamada MK, Setou M (2008)
A novel chalcone polyphenol inhibits the deacetylase activity
of SIRT1 and cell growth in HEK293T cells. J Pharmacol Sci
108:364371
Kleihues P, Schauble B, zur Hausen A, Esteve J, Ohgaki H
(1997) Tumors associated with p53 germline mutations: a
synopsis of 91 families. Am J Pathol 150:113
Kyritsis AP, Bondy ML, Xiao M, Berman EL, Cunningham
JE, Lee PS, Levin VA, Saya H (1994) Germline p53 gene
mutations in subsets of glioma patients. J Natl Cancer Inst
86:344349
Ledford H (2010) Big science: the cancer genome challenge.
Nature 464:972974
Li FP, Fraumeni JF Jr, Mulvihill JJ, Blattner WA, Dreyfus MG,
Tucker MA, Miller RA (1988) A cancer family syndrome in
twenty-four kindreds. Cancer Res 48:53585362
Li YJ, Sanson M, Hoang-Xuan K, Delattre JY, Poisson M,
Thomas G, Hamellin R (1995) Incidence of germ-line p53
mutations in patients with gliomas. Int J Cancer 64:383387
Liu HK, Wang Y, Belz T, Bock D, Takacs A, Radlwimmer B,
Barbus S, Reifenberger G, Lichter P, Schtz G (2010) The
nuclear receptor tailless induces long-term neural stem cell
expansion and brain tumor initiation. Genes Dev 24:683695
Lubbe J, von Ammon K, Watanabe K, Hegi ME, Kleihues P
(1995) Familial brain tumour syndrome associated with a
p53 germline deletion of codon 236. Brain Pathol 5:1523
Osborne RH, Houben MP, Tijssen CC, Coebergh JW, van Duijn
CM (2001) The genetic epidemiology of glioma. Neurol
57:17511755
Palmero EI, Schuler-Faccini L, Caleffi M, Achatz MI, Olivier M,
Martel-Planche G, Marcel V, Aguiar E, Giacomazzi J, Ewald
IP, Giugliani R, Hainaut P, Ashton-Prolla P (2008) Detection
of R337H, a germline TP53 mutation predisposing to multiple cancers, in asymptomatic women participating in a breast
cancer screening program in Southern Brazil. Cancer Lett
261:2125
37
Paraf F, Jothy S, Van Meir EG (1997) Brain tumor-polyposis
syndrome: two genetic diseases? J Clin Oncol 15:27442758
Park DM, Yeaney GA, Hamilton RL, Mabold J, Urban N,
Appleman L, Flickinger J, Lieberman F, Mintz A (2009)
Identifying Muir-Torre syndrome in a patient with glioblastoma multiforme. Neuro Oncol 11:452455
P, Mankoo P, Carter H, Kamiyama H, Jimeno A, Hong
SM, Fu B, Lin MT, Calhoun ES, Kamiyama M, Walter K,
Nikolskaya T, Nikolsky Y, Hartigan J, Smith DR, Hidalgo
M, Leach SD, Klein AP, Jaffee EM, Goggins M, Maitra A,
Iacobuzio-Donahue C, Eshleman JR, Kern SE, Hruban RH,
Karchin R, Papadopoulos N, Parmigiani G, Vogelstein B,
Velculescu VE, Kinzler KW (2008) An integrated genomic
analysis of human glioblastoma multiforme. Science 321:
18071812
Paunu N, Syrjakoski K, Sankila R, Simola KO, Helen P, Niemela
M, Matikainen M, Isola J, Haapasalo H (2001) Analysis of
p53 tumor suppressor gene in families with multiple glioma
patients. J Neurooncol 55:159165
Pearson AD, Craft AW, Ratcliffe JM, Birch JM, Morris-Jones P,
Roberts DF (1982) Two families with the Li-Fraumeni cancer
family syndrome. J Med Genet 19:362365
Petitjean A, Mathe E, Kato S, Ishioka C, Tavtigian SV, Hainaut
P, Olivier M (2007) Impact of mutant p53 functional properties on TP53 mutation patterns and tumor phenotype: lessons
from recent developments in the IARC TP53 database. Hum
Mutat 28:622629
Portwine C, Chilton-MacNeill S, Brown C, Sexsmith E,
McLaughlin J, Malkin D (2001) Absence of germline and
somatic p53 alterations in children with sporadic brain
tumors. J Neurooncol 52:227235
Rieber J, Remke M, Hartmann C, Korshunov A, Burkhardt
B, Sturm D, Mechtersheimer G, Wittmann A, Greil J,
Blattmann C, Witt O, Behnisch W, Halatsch ME, Orakcioglu
B, von Deimling A, Lichter P, Kulozik A, Pfister S (2009)
Novel oncogene amplifications in tumors from a family with
Li-Fraumeni syndrome. Genes Chromosomes Cancer 48:
558568
Sameshima Y, Tsunematsu Y, Watanabe S, Tsukamoto T, Kawaha K, Hirata Y, Mizoguchi H, Sugimura T, Terada M, Yokota
J (1992) Detection of novel germ-line p53 mutations in
diverse-cancer-prone families identified by selecting patients
with childhood adrenocortical carcinoma. J Natl Cancer Inst
84:703707
Santibanez-Koref MF, Birch JM, Hartley AL, Jones PH, Craft
AW, Eden T, Crowther D, Kelsey AM, Harris M (1991)
p53 germline mutations in Li-Fraumeni syndrome. Lancet
338:14901491
Schwartzbaum JA, Fisher JL, Aldape KD, Wrensch M (2006)
Epidemiology and molecular pathology of glioma. Nat Clin
Pract Neurol 2:494503, quiz 491 p following 516
Schwarzbraun T, Obenauf AC, Langmann A, Gruber-Sedlmayr
U, Wagner K, Speicher MR, Kroisel PM (2009) Predictive
diagnosis of the cancer prone Li-Fraumeni syndrome by accident: new challenges through whole genome array testing. J
Med Genet 46:341344
Varley JM (2003) Germline TP53 mutations and Li-Fraumeni
syndrome. Hum Mutat 21:313320
MD, Miller CR, Ding L, Golub T, Mesirov JP, Alexe G,
38
Lawrence M, OKelly M, Tamayo P, Weir BA, Gabriel S,
Winckler W, Gupta S, Jakkula L, Feiler HS, Hodgson JG,
James CD, Sarkaria JN, Brennan C, Kahn A, Spellman PT,
Wilson RK, Speed TP, Gray JW, Meyerson M, Getz G,
Perou CM, Hayes DN, and Cancer Genome Atlas Research
Network (2010) Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by
abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer
Cell 17:98110
Vital A, Bringuier PP, Huang H, San Galli F, Rivel J, Ansoborlo
S, Cazauran JM, Taillandier L, Kleihues P, Ohgaki H (1998)
Astrocytomas and choroid plexus tumors in two families with
identical p53 germline mutations. J Neuropathol Exp Neurol
57:10611069
H. Sugimura et al.
Watanabe T, Vital A, Nobusawa S, Kleihues P, Ohgaki H
(2009) Selective acquisition of IDH1 R132C mutations in
astrocytomas associated with Li-Fraumeni syndrome. Acta
Neuropathol 117:653656
Yamada H, Shinmura K, Yamamura Y, Kurachi K, Nakamura T,
Tsuneyoshi T, Yokota N, Maekawa M, Sugimura H (2009)
Identification and characterization of a novel germline p53
mutation in a patient with glioblastoma and colon cancer. Int
J Cancer 125:973976
Yan H, Parsons DW, Jin G, McLendon R, Rasheed BA, Yuan W,
Kos I, Batinic-Haberle I, Jones S, Riggins GJ, Friedman H,
Friedman A, Reardon D, Herndon J, Kinzler KW, Velculescu
VE, Vogelstein B, Bigner DD (2009) IDH1 and IDH2 mutations in gliomas. N Engl J Med 360:765773
Chapter 5
Familial Gliomas: Role of TP53 Gene

Soufiane El Hallani and Ilham Ratbi
Abstract The presence of a personnel or familial

history of cancer in a patient with primary brain
tumor should prompt the search for genetic predisposition. Familial aggregation of brain tumors and
especially gliomas has been reported in 57% of the
cases. Rarely, familial gliomas can be attributed to
hereditary multisystem syndromes. In majority of the
cases, a hereditary syndrome cannot be identified and
genetic alterations predisposing to familial gliomas
are not clearly identified. Studies investigating candidate loci in glioma families have frequently examined
TP53 gene which encodes p53, a checkpoint protein
that plays a crucial role in DNA damage repair and
apoptosis. Although occasional germline TP53 mutations were reported in glioma families, the frequency
remained low (2.56.7%). However, germline TP53
mutations are more incriminated in glioma patients
with multifocality or secondary malignancy, particularly if these factors are combined. Moreover, a functional single nucleotide polymorphism at codon 72 of
TP53 gene was found to be associated with earlier
onset of sporadic glioblastomas, opening new insights
into the role of low-risk variants as genetic susceptibility to develop brain tumors. This chapter reviews
the genetic predisposition in brain tumors and highlights the role of TP53 gene in familial gliomas, and in
other forms of genetic predisposition or susceptibility
to gliomas.
S. El Hallani ()
Cancer Imaging Unit, Integrative Oncology Department, British
Columbia Cancer Research Centre, Vancouver, BC, Canada
V5Z 1L3
e-mail: Soufiane.elhallani@yahoo.ca
Keywords Familial gliomas TP53 Candidate loci

p53 Apoptosis Codon
Introduction
Primary brain tumors are a mixture of different
histopathologies which include both benign and malignant forms. Gliomas are the major subtype accounting for 90% of primary brain tumors. Astrocytomas,
oligodendrogliomas, ependymomas, and tumors of the
choroid plexus all arise from the glial origin of the
central nervous system. In most cases, glioma etiology is not well understood. Ionizing radiation is one
of the few well-recognized exogen factors associated
with gliomas and meningiomas. Other environmental
agents have been suggested (electromagnetic field, cellular phone, professional exposures) but epidemiologic
studies have not been conclusive. Familial aggregation
of gliomas is reported in 57% of the cases, suggesting a genetic etiology or common familial exposure
to environmental agents. A small number of familial gliomas can be attributed to hereditary multisystem syndromes such as Neurofibromatose 1 and 2,
Tuberous Sclerosis, Turcot syndrome and Li-Fraumeni
syndrome. However, in many other cases, a hereditary
syndrome cannot be identified and genetic alterations
predisposing to familial gliomas are not yet clarified.
The study of inherited predisposition to cancer has
been one of the most attractive research areas in the
past two decades. Indeed, the identification of susceptibility genes provides a better understanding of molecular oncogenesis, offering potential targets for therapeutic interventions. Furthermore, the ability to identify
those at increased risk is of immediate clinical relevance in terms of primary and secondary interventions.

39
40
Genetic linkage analysis has led to locate highly penetrant genes for several common cancers, including
breast and ovarian cancers (BRCA1 and BRCA2), colon
cancer with adenomatous polyposis coli (APC), hereditary non-polyposis colon cancer (MSH2 and MLH1),
melanoma (CDNK2A) and testicular cancer (TCG1).
However, reports on familial gliomas remained few.
Segregation analysis ruled out a single gene explanation and strongly incriminated a multifactorial inheritance model with a gene-gene or gene-environmental
interactions. The difficulty to obtain a sufficient number of constitutional DNA from affected members of
informative glioma families limited the approach of
genetic linkage analysis. Studies investigating candidate loci in glioma families have frequently examined
TP53 gene which encodes p53, a checkpoint protein that plays a crucial role in DNA damage repair
and apoptosis. Germline TP53 mutations are found in
7177% of classic Li-Fraumeni syndrome that predisposes individuals to a wide spectrum of cancers
including brain tumors. Interestingly, previous studies
identified germline TP53 mutations in several families with multiple glioma patients but without ascertained Li-Fraumeni syndrome. This chapter reviews
the genetic predisposition in brain tumors and highlights the role of TP53 gene in familial gliomas, and in
other forms of genetic predisposition or susceptibility
to gliomas.
Genetic Predisposition to Familial Cancer

Cancers arise by accumulating structural or functional
alteration in DNA. Majority of the DNA alterations
involve mutations of proto-oncogenes, tumor suppressor genes and DNA repair genes. Cancers may be
sporadic or as a part of hereditary predisposition syndrome. Sporadic cancers are generally characterized
by late onset, unilateral and isolated tumor. Mendelian
cancer syndromes account for 5% of the common
cancer cases. They are characterized by early onset,
multifocality and secondary malignancy. They occur
in multiple generations and their inheritance fits an
autosomal dominant or recessive model.
Different strategies have been used to identify predisposing genes in familial cancers including gliomas.
Linkage analysis consists to identify an eventual locus
of predisposition in one or more informative families.
S. El Hallani and I. Ratbi
Based on polymorphic DNA probes, linkage studies

are to locate the predisposing gene by analyzing the cosegregation of polymorphic markers with the disease.
Two linkage analysis studies have been conducted in
familial gliomas and were able to identify an autosomal dominant locus at 15q23-q26 (Paunu et al., 2002),
and an autosomal recessive locus at 1q21-q25 (Malmer
et al., 2005). The limitation of this approach lies in the
difficulty of obtaining a sufficient number of informative families for such rare cancer syndromes. A second
strategy is to identify a possible locus of predisposition
using fine genetic mapping of the tumors. By comparing tumor DNA with constitutional DNA of familial
glioma patients, molecular techniques known as allelic
loss (or loss of heterozygosity) have led to define frequent deletions of several chromosomal regions: 1p,
6q, 9p, 10q, 11p, 13q, 17p, 19q and 22q. Recurrent
deletions of these chromosomal segments may reflect
the existence within them of one (or more) tumor suppressor gene, whose one copy could be inactivated
by this mechanism, and the other one by a germline
mutation. More recently, analysis of familial glioma by
CGH-array allowed to bring new fields of exploration
of locus predisposition (on 6q, 12p, 16p, 22q) (Paunu
et al., 2000). Testing known candidate genes is alternative when only the blood sample from the proband is
available. Several candidate genes have been explored
in familial gliomas. Germline TP53 tumor suppressor
gene mutations were reported in some cases, whereas
no mutations were identified in the PTEN, P16, CDK4
and CHEK2 genes (Tachibana et al., 2000; El Hallani
et al., 2009a).
Brain Tumors in Predisposition

Syndromes
Hereditary tumor syndromes of the nervous system are
varied and their recognition is critical to provide optimal clinical care and genetic counselling. In the last
decade, many of the genes responsible for these typically autosomal dominant familial tumor syndromes
have been identified.
Neurofibromatosis Type 1
The phenotype of the disease varies widely. Whereas
some patients remain asymptomatic, others develop
multiple tumors of the nervous system (neurofibroma,

optic glioma, plexiform neurofibroma, neurofibrosarcoma, astrocytoma, and meningioma), tumors of other
organs (pheochromocytoma, hypothalamic tumor,
rhabdomyosarcoma, duodenal carcinoid), cafe-au-lait
spots, axillary or inguinal freckling, hamartomas of
the iris, vascular diseases, cognitive impairment, and
behavioral problems. It is caused by mutations in the
NF1 tumor suppressor gene located on chromosome
17 at q11.2.
Bilateral vestibular schwannomas are pathognomonic
of the disorder. Patients often develop schwannomas
of other cranial, spinal, and peripheral nerves, as well
as intracranial and intraspinal meningiomas. Less frequently, they may develop low grade gliomas and
ependymomas. Other features associated with neurofibromatosis type 2 are ocular abnormalities, specifically juvenile posterior subcapsular lenticular opacities
(60% to 80% of patients), retinal hamartomas, hearing
loss, and occassional cafe-au-lait spots. It is caused by
mutations in the NF2 tumor suppressor gene located
on chromosome 22q12.
Tuberous Sclerosis Complex (TSC)

This condition predisposes to hamartomas and benign
tumors in the brain, heart, and kidney. Nervous system
lesions of TSC include tubers in the cerebral cortex,
subependymal nodules, and subependymal giant cell
astrocytomas. Two tumor suppressor genes have been
identified: TSC1 at 9q34 (coding for hamartin) and
TSC2 at 16p13 (coding for tubulin).
Von Hippel-Lindau Syndrome (VHL)

Individuals are predisposed to develop a variety of
malignant and benign neoplasms, most frequently retinal, cerebellar, and spinal hemangioblastomas, renal
cell carcinomas, pheochromocytomas, and pancreatic
tumors. The finding of multiple hemangioblastomas at
young age is highly suggestive of VHL. It is caused
41
by mutations in the VHL gatekeeper tumor suppressor

gene on chromosome 3p25.
Gorlin Syndrome
Also called nevoid basal cell carcinoma syndrome, it
is characterized by the association of developmental
abnormalities and an increased incidence of malignancy. Neoplastic manifestations are mainly basal
cell carcinomas, medulloblastoma, and, occasionally,
meningiomas. Gorlin syndrome is caused by abnormalities in the patched homolog 1 gene (PTCH1) on
chromosome 9q31 in about 85% of cases.
Turcot Syndrome (TS)

Occurrence of primary brain tumor and multiple colorectal adenomas and/or colorectal adenocarcinoma
are the hallmarks of this syndrome. The most frequent
brain tumors are astrocytomas, glioblastomas, medulloblastomas, ependymoma, spongioblastoma, gliosarcoma, and oligodendroglioma. TS type 1 includes
patients with glioblastoma, often associated with
hereditary nonpolyposis colorectal carcinoma and corresponding germline mutations in the DNA mismatch repair genes (MMR): postmeiotic segregation
increased 2 (PMS2), MutL homolog 1 (MLH1), and
MutS homolog 2 (MSH2). TS type 2 includes patients
with medulloblastoma and familial adenomatous polyposis that result from mutations in the adenomatous
polyposis of the colon gene (APC).
Cowden Syndrome and Lhermitte-Duclos

Disease
Cowden syndrome is a rare autosomal dominant disorder, associated with thyroid, breast, and uterine cancer
as well as tricholemmomas, hamartomas, skin tumors,
and meningiomas. Germline mutations in the phosphatase and tensin homolog (PTEN) gene located
on chromosome 10q22-23 were described. LhermitteDuclos disease, also known as dysplastic cerebellar
gangliocytoma, is strongly associated with Cowden
syndrome.
42
Melanoma-Astrocytoma Syndrome
People with this rare syndrome have an increased
risk of developing malignant cutaneous melanoma
and nervous system tumors such as astrocytoma, neurofibroma, schwannoma, and meningioma. To date
all published families documented with melanomaastrocytoma syndrome are linked to the CDKN2 locus
on 9p21.
Li-Fraumeni Syndrome
Li-Fraumeni syndrome (LFS) is a rare but highly
penetrant autosomal cancer predisposition syndrome
that is characterized by a familial clustering of early
onset tumors including sarcomas, breast cancers, brain
tumors and adrenocortical carcinomas (Li et al., 1988;
Malkin et al., 1990). Initially considered as a rare
syndrome, LFS and its variants are increasingly recognized as one of the most frequent and diverse forms
of predisposition to cancer. To fulfill the classical LFS
definition, the family history must include a proband
with sarcoma before 45 years of age; a first degree relative of the proband with any cancer below 45 years
and another first or second degree relative in the same
lineage with any cancer before 45 years or sarcoma at
any age. The most common childhood and adolescent
cancers are soft-tissue sarcomas and osteosarcomas.
Leukemia and brain tumors, including choroid plexus
tumors, are almost exclusively confined to infants and
early childhood, whereas adrenocortical carcinomas
occur from infancy through late. In young adults,
breast cancer is by far the most common malignancy.
Other cancers including early onset melanoma, lung,
gastric, pancreatic, prostate and colorectal cancer were
also described in excess in some families.
Germline TP53 gene mutations are the underlying cause of LFS in 7177% of the cases (Evans
et al., 2002). The TP53 tumor suppressor gene located
on chromosome 17p13 encodes a protein involved in
many overlapping cellular pathways that control cell
proliferation and homeostasis, such as cell cycle, apoptosis, and DNA repair. The p53 protein is a transcription factor constitutively expressed in most cell types
and activated in response to various stress signals. Loss
of p53 function is thought to suppress a mechanism of
protection against accumulation of genetic alterations.
Somatic TP53 genetic alterations are frequent in a

variety of human sporadic cancers, with frequencies
varying from 10 to 60%, depending on the tumor
type or population group. Germline TP53 mutations
may contribute up to 17% of all familial cancer cases.
Both somatic and germline mutations are compiled in
a worldwide database at the International Agency for
Research on Cancer (www-p53.iarc.fr). Most mutations result in missense substitutions that are scattered
throughout the gene but are particularly dense in exons
58, encoding the DNA binding domain. Analysis of
tumor patterns in carriers of a TP53 germline mutation
has demonstrated some genotype/phenotype correlations. Brain tumors were more likely associated with
missense mutations within the DNA-binding surface
of the p53 protein that makes contact with the minor
groove of DNA (Petitjean et al., 2007). The majority
of them were found at codons 175, 245 and 248. These
mutant proteins exhibit loss of transactivation function
and dominant-negative effects. To date, no other gene
has been significantly associated with LFS. Reports
that germline mutations in the CHEK2 gene may predispose to LFS have not been substantiated (Vahteristo
et al., 2001).

Brain tumors are the third most common tumors arising from germline TP53 mutations. It is also estimated that not more than a half of families with
germline TP53 mutations fulfill the definition of classic LFS. Therefore, the possibility that germline TP53
mutations could play an important role in familial
aggregation of gliomas was suggested. Various case
reports were first supportive of this hypothesis. Lbbe
et al. (1995) described clinical, neuropathological and
molecular genetic findings in a Swiss family with four
brain tumors in only two generations. The neoplasms
observed covered a wide range from a slowly growing
lesion already apparent at birth, to anaplastic astrocytoma in a young adult and glioblastomas at the age
of less than 10 years. A germline deletion of codon
236 of the TP53 gene was identified as an underlying cause and detected in all affected family members.
Vital et al. (1998) reported 2 French families with an
identical germline TP53 mutation in codon 248 and
a clustering of brain tumors. The youngest patient
in each family developed a malignant choroid plexus

tumor while several young adults of both kindred
succumbed to low-grade astrocytoma, anaplastic astrocytoma or glioblastoma. To determine whether TP53
gene plays an important role in familial glioma, the
frequency of germline mutations was then investigated.
For this purpose, van Meyel et al. (1994) went
on to screen 26 members of 16 glioma families for
germline TP53 mutations of exons 5 through 9 using a
polymerase chain reaction-single-strand conformation
polymorphism method. However, no germline mutations were found. Tachibana et al. (2000) investigated
15 familial gliomas by screening exons 5 through 8
and found one GBM patient exhibiting germline TP53
mutation. Careful examination of the corresponding
family history showed multiple first degree relatives
with GBM and a variety of other cancers (lung, colon,
uterine and ovarian), many of them occurring before
the age of 45, but not all the criteria for classical LFS
were fulfilled; especially no family members developed a sarcoma. Paunu et al. (2001) identified 18
families with two or more gliomas through questionnaires sent to 369 consecutive glioma patients operated
at Tampere University Hospital during 19831994.
Again, sequencing analysis of exons 4 through 10
of the TP53 gene revealed no germline mutations in
any of the 18 families. To date, the largest cohort of
glioma families was gathered by the Salpetriere group.
El Hallani et al. (2009a) examined 79 blood DNA of
glioma patients who had at least one family member
affected by glioma. Germline TP53 mutation screening
was performed on exons 5 through 8. Five patients with
variant sequences were identified. Two of them had a
known polymorphism at codon 213 and the three other
patients had missense TP53 mutations at codon 141,
251, and 273. Only one out of the mutated patient met
the criteria for classic LFS. Taken together, only 2.5%
(2/78) of familial glioma patients without ascertained
LFS carried a mutated TP53 gene.
In most studies, it was considered adequate to analyze the highly conserved regions of the TP53 gene,
which are exons 58. Earlier analyses covering all
exons, splice junctions and the promoter region of
the TP53 gene suggested that 10% of germline mutations lie outside exons 410. Thus, previous reports
have only provided a fairly thorough approximation
of mutations within the TP53 gene. The fact that the
occurrence of germline TP53 mutations among glioma
families was a rare event speaks against the claim that
43
such mutations would play a significant role in the genesis of familial gliomas. The relatively rare germline
TP53 mutation in glioma families should promptly
alert one to an incomplete form of LFS rather than a
distinct glioma predisposition.
An immunohistochemical case-control analysis of
the tumor samples was performed to compare the
rate of TP53 mutations and consequent aberrant p53
expression in familial and sporadic gliomas (Paunu
et al., 2001). This analysis indicated that overexpressing p53 protein is as common in familial as
in sporadic gliomas. In the absence of germline mutations, p53 accumulation probably arises from somatic
mutations. Thus, somatic mutations of the p53 gene
seem to be similarly involved in the pathogenesis
of familial and sporadic gliomas. Together with the
germline mutation analysis, this suggests that the role
of the TP53 tumor suppressor gene is similar in the
tumorigenesis of most familial and sporadic gliomas.
TP53 in Other Forms of Genetic

Predisposition to Gliomas
In addition to familial aggregation of gliomas, clinicians should be aware of other overlapping situations in which genetic etiology should be suspected:
patients with multifocal glioma, glioma as another
primary malignancy, and glioma associated with a family history of cancer without ascertained Li-Fraumeni
syndrome. Gliomas usually present as a single lesion
and rarely metastasize. However, multifocal gliomas,
which exhibit more than one independent focus of
involvement, were found to occur with a frequency of
911%. Moreover, multifocal gliomas are frequently
associated with other primary malignancies.
Kyritsis et al. (1994) investigated whether germline
TP53 gene mutations may account for this phenomenon. Overall, nine out of 39 patients (23%)
with multifocal glioma, secondary malignancies, or
a family history of cancer exhibited heterozygous
germline TP53 mutations. All these mutations would
have resulted in amino acid changes in the p53 protein. They were detected in six out of 19 patients with
multifocal glioma (including two with family history
of cancer, one with another primary malignancy, and
two with all three risk factors); one out of 4 patients
with unifocal glioma, another primary malignancy and
44
a family history of cancer; and two out of 15 patients

with unifocal glioma and a family history of cancer but
no second malignancies. No mutations were detected
in the patient with unifocal glioma and another malignancy or in the 12 control patients with unifocal glioma
and no second malignancies or family history of cancer. Patients having mutations were younger than other
patients in the same group. Combinations of these
risk factors increased the percentage of germline TP53
gene mutations to approximately 43% if two factors
were present, or 67% in the presence of all three factors. Therefore, screening germline TP53 mutations
is recommended if theses factors are combined in
order to identify high risk relatives for genetic counseling, early cancer detection, and possible enrollment
in chemoprevention trials.
TP53 Polymorphism in Sporadic Gliomas

Inherited cancer syndromes are associated with rare
and highly penetrant monogenic mutations, but genetic
factors also play a role in sporadic cancers. Single
Nucleotide Polymorphisms (SNP) can affect protein
function, promoter activity, messenger RNA stability,
and splice variants and therefore can result in a change
in the cellular ability to cope with DNA damage, which
contributes to cancer susceptibility. Most of the association studies that aimed at identifying low-penetrance
variants for susceptibility to gliomas have been based
on a candidate gene approach.
One of the critical molecular steps of gliomagenesis involves the p53 pathway. A functional SNP at
codon 72 of TP53 gene results in the presence of
either proline (Pro) or arginine (Arg) in the amino acid
sequence of TP53. This change involves DNA-binding
motifs of p53 and the Pro72 variant results in weaker
induction of apoptosis than the Arg72 variant (Pim
and Banks, 2004). Then this polymorphism has been
linked to glioma susceptibility but results from different studies are controversial. The Pro72 variant was
also suspected of altering the tumor sensitivity to irradiation and chemotherapy. However, conducted glioma
studies suggest that the Pro72 variant does not have
any prognostic impact on the survival of GBM patients,
neither on less aggressive histologic subgroups of glial
tumors. Interestingly, El Hallani et al. (2009b) reported
that the Pro/Pro genotype is associated with earlier onset in GBM population (49.1 years) compared
to Arg/Arg (56.6 years) and Arg/Pro (55.9 years).
The Pro/Pro genotype was significantly more represented in young patients (less than 45 years old)
than in elderly. The accelerating effect of functional
p53 insufficiency on tumorigenesis was previously
demonstrated. In a model of mice lacking the xeroderma pigmentosum group A gene (XPA/ mice) and
highly predisposed to tongue tumors when exposed
to 4-nitroquinoline 1-oxide, the p53 haploinsufficiency
(TP53/+) and complete inactivation (TP53/) did
not increase the incidence of cancer but accelerated
dramatically tumor development (Ide et al., 2003).
Genomic alterations and molecular pathways involved
in GBM differ between young and older patients. TP53
inactivation is a prominent mechanism in the GBM
of child and young adult compared to the GBM of
older patient. As a consequence, young patients might
be more sensitive to the functional variation of TP53
codon 72, explaining that Pro/Pro genotype constitutes
a potent risk factor in this population.
In conclusion, clinicians should inquire about any
family history of brain tumors and be familiar with
the most important hereditary syndromes with neoplastic manifestations in the nervous system, so that
these diagnoses are not missed. Genetic alterations predisposing to familial aggregation of gliomas without
ascertained hereditary syndrome are yet to be clarified. Although occasional glioma families carrying
germline TP53 mutations have been identified, current data do not support routine screening of TP53
gene, except for atypical combination of multifocality
and secondary malignancy. As suggested by segregation analysis, a model of familial glioma predisposition
based solely on high-risk mutations seems unlikely,
and much of the inherited risk is likely to be a consequence of the coinheritance of multiple low-risk
variants. However, our knowledge of predisposition
to glioma is developing. The advent of genome-wide
association study will enable researchers to identify
variants that influence an individuals susceptibility to
develop glioma. Such studies are at the vanguard of
the new technologies and started to identify risk loci
in gliomas. Identifying the sequence changes responsible for causal associations should thus provide insight
into the biological mechanisms of glioma, and this
may lead to the development of etiological hypotheses
regarding non genetic risk factors.
References
El Hallani S, Boisselier B, Marie Y, Paris S, Idbaih A, Carpentier
C, Hoang-Xuan K, Delattre JY, Sanson M (2009a) TP53
mutations but no CHEK2 1100DelC variant in familial
gliomas. Cancer Genet Cytogenet 188:126128
El Hallani S, Ducray F, Idbaih A, Marie Y, Boisselier B,
Colin C, Laigle-Donadey F, Rodro M, Chinot O, Thillet
J, Hoang-Xuan K, Delattre JY, Sanson M (2009b) TP53
codon 72 polymorphism is associated with age at onset of
glioblastoma. Neurology 72:332336
Evans DG, Birch JM, Thorneycroft M, McGown G, Lalloo F,
Varley JM (2002) Low rate of TP53 germline mutations in
breast cancer/sarcoma families not fulfilling classical criteria
for Li-Fraumeni syndrome. J Med Genet 39:941944
Ide F, Kitada M, Sakashita H, Kusama K, Tanaka K, Ishikawa
T (2003) p53haploinsufficiency profoundly accelerates the
onset of tongue tumors in mice lacking the xeroderma pigmentosum group A gene. Am J Pathol 163:17291733
Kyritsis AP, Bondy ML, Xiao M, Berman EL, Cunningham
JE, Lee PS, Levin VA, Saya H (1994) Germline p53 gene
mutations in subsets of glioma patients. J Natl Cancer Inst
86:344349
Li FP, Fraumeni JF, Mulvihill JJ, Blattner WA, Dreyfus MG,
Tucker MA, Miller RW (1988) A cancer family syndrome
in twenty-four kindreds. Cancer Res 48:53585362
Lbbe J, von Ammon K, Watanabe K, Hegi ME, Kleihues P
(1995) Familial brain tumour syndrome associated with a
p53 germline deletion of codon 236. Brain Pathol 5:1523
Malkin D, Li FP, Strong LC, Fraumeni JF, Nelson CE, Kim DH,
Kassel J, Gryka MA, Bischoff FZ, Tainsky MA (1990) Germ
line p53 mutations in a familial syndrome of breast cancer,
sarcomas, and other neoplasms. Science 250:12331238
Malmer B, Haraldsson S, Einarsdottir E, Lindgren P, Holmberg
D (2005) Homozygosity mapping of familial glioma in
Northern Sweden. Acta Oncol 44:114119
Paunu N, Lahermo P, Onkamo P, Ollikainen V, Rantala I, Heln
P, Simola KO, Kere J, Haapasalo H (2002) A novel low-
45
penetrance locus for familial glioma at 15q23-q26.3. Cancer
Res 62:37983802
Paunu N, Sallinen SL, Karhu R, Miettinen H, Sallinen P,
Kononen J, Laippala P, Simola KO, Heln P, Haapasalo H
(2000) Chromosome imbalances in familial gliomas detected
by comparative genomic hybridization. Genes Chromosomes
Cancer 29:339346
Paunu N, Syrjkoski K, Sankila R, Simola KO, Heln P, Niemel
M, Matikainen M, Isola J, Haapasalo H (2001) Analysis of
p53 tumor suppressor gene in families with multiple glioma
patients. J Neurooncol 55:159165
Petitjean A, Mathe E, Kato S, Ishioka C, Tavtigian SV, Hainaut
P, Olivier M (2007) Impact of mutant p53 functional properties on TP53 mutation patterns and tumor phenotype: lessons
from recent developments in the IARC TP53 database. Hum
Mutat 28:622629
Pim D, Banks L (2004) p53 polymorphic variants at codon 72
exert different effects on cell cycle progression. Int J Cancer
108:196199
Tachibana I, Smith JS, Sato K, Hosek SM, Kimmel DW,
Jenkins RB (2000) Investigation of germline PTEN, p53,
p16(INK4A)/p14(ARF), and CDK4 alterations in familial
glioma. Am J Med Genet 92:136141
Vahteristo P, Tamminen A, Karvinen P, Eerola H, Eklund C,
Aaltonen LA, Blomqvist C, Aittomki K, Nevanlinna H
(2001) p53, CHK2, and CHK1 genes in Finnish families
with Li-Fraumeni syndrome: further evidence of CHK2 in
inherited cancer predisposition. Cancer Res 61:57185722
van Meyel DJ, Ramsay DA, Chambers AF, Macdonald DR,
Cairncross JG (1994) Absence of hereditary mutations in
exons 5 through 9 of the p53 gene and exon 24 of the
neurofibromin gene in families with glioma. Ann Neuro.
35:120122
Vital A, Bringuier PP, Huang H, San Galli F, Rivel J, Ansoborlo
S, Cazauran JM, Taillandier L, Kleihues P, Ohgaki H (1998)
Astrocytomas and choroid plexus tumors in two families with
identical p53 germline mutations. J Neuropathol Exp Neurol
57:10611069
Chapter 6
The Role of IDH1 and IDH2 Mutations in Malignant Gliomas

Yukihiko Sonoda, Ichiyo Shibahara, Ryuta Saito, Toshihiro Kumabe, and Teiji
Tominaga
Abstract A recent study identified mutations in

the active sites of isocitrate dehydrogenase 1 and 2
(IDH1and IDH2) genes in several types of glioma. All
mutations affected a single amino acid located in the
binding site of isocitrate (R132 of IDH1 and R172
of IDH2). We analyzed the genomic region spanning
wild-type R132 of IDH1 and R172 of IDH2 by direct
sequencing in 125 glial tumors. A total of 39 IDH1
mutations and one IDH2 mutation were observed.
IDH1 and IDH2 mutations were frequently present
in astrocytic and oligodendroglial tumors. However,
primary glioblastomas were characterized by a low
frequency of mutations (5%) at amino acid position
132 of IDH1. Mutations of IDH1 and IDH2 genes
were significantly associated with improved outcome
in patients with anaplastic astrocytomas. IDH1 and
IDH2 mutations seem to play an important role in
early tumor progression of specific types of glioma
and might arise from a common glial precursor. The
infrequency of IDH1 mutation in primary glioblastomas revealed that these subtypes are entities that are
genetically distinct from other glial tumors. Analyses
of IDH1 and IDH2 status have significant utility for
diagnosis and treatment of patients with gliomas.
Keywords IDH1 and IDH2 Mutations Gliomas
Glioblastomas Grades of glioma Arginine
Introduction
Gliomas are the most common type of primary brain
tumor and are grouped into four grades according
Y. Sonoda ()
Department of Neurosurgery, Tohoku University Graduate
School of Medicine, Aoba-ku, Sendai-shi, Miyagi 980-8577,
Japan
e-mail: sono@nsg.med.tohoku.ac.jp
to the World Health Organization (WHO) criteria

(Louis et al., 2007). Gliomas include several specific histological subtypes; the most common are
astrocytomas, oligodendrogliomas, and ependymomas. Glioblastomas (GBMs) (WHO grade IV), the
most malignant type of glioma, may develop very
rapidly de novo (primary glioblastoma) in elderly
patients, or develop more slowly from low-grade diffuse (DA) (WHO grade II) or anaplastic astrocytoma
(AA) (WHO grade III) (secondary glioblastoma) in
younger patients (Ohgaki and Kleihues, 2009).
Malignant gliomas are believed to develop as
the result of stepwise accumulations of genetic
lesions. Several genes, including TP53, PTEN,
CDKN2A/CDKN2B, and EGFR, are altered in gliomas
(Ohgaki and Kleihues, 2009). These alterations tend
to occur in a defined order during the progression
to a high-grade tumor. TP53 mutation appears to
be a relatively early event during the development
of an astrocytoma, whereas the loss or mutation of
PTEN and amplification of EGFR are characteristic of higher-grade tumors (Furnari et al., 2007).
Oligodendrogliomas (O) (WHO grade II) and anaplastic oligodendrogliomas (AO) (WHO grade III) typically show 1p/19q codeletion (Reifenberger et al.,
1994).
Recent genome-wide mutational analysis has
revealed somatic mutations of the cytosolic nicotinamide adenine dinucleotide phosphate (NADP+ )dependent isocitrate dehydrogenase (IDH1) gene at
2q33 in approximately 12% of GBMs (Parsons
et al., 2008). Isocitrate dehydrogenase catalyzes
the oxidative decarboxylation of isocitrate to alphaketoglutarate thereby enabling NADPH production.
Mutations were found that affect the amino acid arginine in position 132 of the amino acid sequence, which
is evolutionarily highly conserved, and is located in

47
48
the binding site of isocitrate (Xu et al., 2004). In the

vast majority of these cases, wild-type arginine in position 132 was replaced by histidine (R132H) (Balss
et al., 2008; Parsons et al., 2008; Ichimura et al., 2009;
Sonoda et al., 2009; Watanabe et al., 2009; Yan et al.,
2009b).
The IDH2 gene is homologous to IDH1, which
uses NADP+ as an electron receptor. Gliomas without IDH1 mutations were often found to have mutations at the analogous amino acid (R172) of the
IDH2 gene (Yan et al., 2009b). Both IDH1and IDH2
mutations reduced enzymatic activity of the encoded
protein.
Methodology
We obtained 125 samples of tumor tissue from surgical
patients diagnosed and treated at Tohoku University
Hospital. Resected specimens were quick-frozen in
liquid nitrogen and kept at 80 C until nucleic acid
extraction (Sonoda et al., 2009). The series included
2 DA, 21 AA, 58 primary glioblastomas (prGBM),
3 secondary glioblastomas (secGBM), 8 O, 14 AO,
4 anaplastic oligoastrocytomas of WHO grade III
(AOA), 8 gangliogliomas of WHO grade II (GG),
and 5 anaplastic gangliogliomas (AG) of WHO grade
III. Exon 4 of the IDH1 gene including codon 132
was amplified using 100 ng each of sense primer
5 -CGGTCTTCAGAGAAGCCATT and antisense
primer 5 -GCAAAATCACATTATTGCCAAC. A
fragment of 219 bp in length spanning the catalytic domain of IDH2 including codon 172 was
amplified using 100 ng each of sense primer 5 CAAGCTGAAGAAGATGTGGAA-3 and antisense
primer 5 -CAGAGACAA GAGGATGGCTA-3 . PCR
was performed using standard buffer conditions,
namely, 100 ng of DNA and Ex-Taq HS DNA
Polymerase (Takara Bio Inc., Shiga, Japan) employed
for 30 cycles with denaturing at 95 C for 30 s,
annealing at 56 C for 30 s, and extension at 72 C for
40 s in a total volume of 50 ul. The PCR products
were purified using a highly pure PCR product purification kit (Roche, Basel, Switzerland). All sequence
reactions were performed using the GenomeLabTM
DTCS quick-start kit (Beckman Coulter, Inc.,
Fullerton, CA). The reactions were carried out in
an automated DNA analyzer (CEQ 8000; Beckman
Coulter).
Y. Sonoda et al.
Results
IDH Mutations in Gliomas
We detected 39 mutations in the IDH1 gene in 125
tumors. All mutations were heterozygous with one
wild-type allele being present. Only codon 132 of
IDH1 was affected by mutations and all mutations
were of the R132H type. In addition, the mutation of
IDH2 at codon 172 was detected in one AA without IDH1 mutation. The type of mutation was R172S
(AGG to AGT).
Mutations of IDH1 or IDH2 were frequently
observed in AA (62%), secGBM (67%), O (67%), AO
(50%), AOA (75%), GG (38%), and AG (60%). Only
a few mutations occurred in prGBM (5%). Despite no
mutations in DA, we could not draw any reliable conclusions because of the fact that there were only two
cases.
Patient Survival and IDH Mutation

Patients with GBM carrying an IDH1 mutation (n = 5)
had a median survival of 66 months, which was longer
than the 17-month survival in patients with wild-type
IDH1 (n = 57). However, there was no significant
difference because there were too few tumors of GBM
with IDH1 mutation (P = 0.1; log-rank test). Mutations
of the IDH1 and IDH2 genes were associated with
prolonged overall survival in AA patients; the median
overall survival was 50 months for patients with mutation (n = 13) and 22 months for those without mutation
(n = 8) (P < 0.001; log-rank test).
Discussion
Detection of IDH1 and IDH2 Mutations
The current routine procedure for assessing IDH gene
status is DNA sequencing. All mutations of the IDH1
gene were somatic and missense mutations at codon
132 (arginine). Of these, almost all mutations were of
the R132H type; however, 5 other mutations leading
to R132C, R132S, R132G, R132L, and R132V were
found. Similarly, all IDH2 mutations were found
in codon 172; these mutations resulted in amino
acid exchanges from arginine to guanine, methionine,
serine, and lysine (Balss et al., 2008; Hartmann et al.,

2009; Sonoda et al., 2009; Yan et al., 2009b).
DNA extraction and sequencing are not available
at every institute; however, recently mutation-specific
IDH1 antibodies for the most frequent mutation of
the R132H type have shown useful properties of high
sensitivity and selectivity (Kato et al., 2009; CameloPiragua et al., 2010; Capper et al., 2010).
IDH1 and IDH2 Mutations are Frequently

Found in Various Grades of Glioma
According to previous reports, no pilocytic astrocytoma with an IDH1 mutation has been found,
with the exception of a single case (Balss et al.,
2008; Korshunov et al., 2009). High frequencies of
IDH1 mutations were found in DA, AO, O, and AO
(Watanabe et al., 2009; Yan et al., 2009b). Similarly,
IDH1 mutations were more frequent in secGBM than
prGBM (Yan et al., 2009b). Among other CNS tumors,
IDH1 mutation has been reported in pleomorphic
xanthoastrocytoma, primitive neuroectodermal tumor,
and GG (Balss et al., 2008; Sonoda et al., 2009; Yan
et al., 2009b). On the other hand, no IDH1 mutation
was found in ependymal tumors, medulloblastomas,
meningeal tumors, and schwannoma (Balss et al.,
2008; Ichimura et al., 2009; Yan et al., 2009b).
Although the frequency of IDH1 mutation in younger
Fig. 6.1 A model of glioblastoma development. Abbreviations:

IDH1/2, isocitrate dehydrogenase 1/2; mut, mutation; 1p19q
codel., 1p 19q codeletion; CDKN2A, B, cyclin-dependent kinase
inhibitor 2A, B; HD, homozygous deletion; PTEN, phosphatase
tensin homolog; EGFR, epidermal growth factor receptor; amp,
49
patients is higher than that in older patients, this mutation has rarely been found in pediatric tumors (Balss
et al., 2008). Among non-CNS tumors, IDH1 mutation was found in two prostate cancers and one B-ALL
(Bleeker et al., 2009; Kang et al., 2009). In addition,
IDH1 mutation has been identified in a subset of acute
myeloid leukemia cases (Mardis et al., 2009). IDH2
mutations were only found in WHO grade II or III
gliomas, usually without IDH1 mutation. According
to previous reports, only one grade III glioma showed
both IDH1 and IDH2 mutations (Wick et al., 2009).
IDH1 and IDH2 Mutations Are Early

Events in Gliomagenesis
IDH1 and IDH2 mutations are commonly found in
both astrocytomas and oligodendrogliomas, but the
pattern of other genetic alterations is different. Usually,
TP53 mutations and 1p/19q codeletion were exclusive,
with the exception of only a few cases (Ichimura et al.,
2009; Watanabe et al., 2009). The majority of lowgrade diffuse astrocytomas have both TP53 and IDH1
mutations, whereas most oligodendrogliomas show
both IDH1 mutations and 1p/19q codeletion (Ichimura
et al., 2009; Watanabe et al., 2009). Therefore, IDH1
and IDH2 mutations are very early events in gliomagenesis and may affect common glial precursors
(Fig. 6.1).
amplification; DA, diffuse astrocytoma; AA, anaplastic astrocytoma; O, oligodendroglioma; AO, anaplastic astrocytoma;
prGBM, primary glioblastoma; secGBM, secondary glioblastoma
50
The pattern of other genetic alterations in gliomas

with IDH mutations is entirely different from that in
gliomas without IDH mutations. Alterations of PTEN,
EGFR, or CDKN2A/CDKN2B are frequently found in
GBMs with wild-type IDH1 and IDH2, but in only
5% of cases of AA and GBM with IDH mutation
(Yan et al., 2009a; b). Similarly, 1p/19q codeletion was
rarely found in oligodendroglial tumors without IDH
mutation (Watanabe et al., 2009). TP53 mutations were
frequently found in AA with IDH1 mutation; however, these were often found in prGBMs without IDH1
mutations (Ohgaki and Kleihues, 2009; Yan et al.,
2009a).
IDH Mutation as a Favorable Prognostic

Factor in Glioma Patients
Clinically, IDH1 mutation is strongly correlated with
good prognosis in patients with several grades of
glioma (Sanson et al., 2009; Sonoda et al., 2009;
Wick et al., 2009; Yan et al., 2009b). Multivariate
analysis has confirmed that IDH1 mutation is an
independent favorable prognostic marker in anaplastic
gliomas and glioblastomas after adjustment for other
factors (age, histology, MGMT promoter methylation
status, and genomic profile) and treatment modality
(extent of resection) (Sanson et al., 2009; Wick et al.,
2009).
Fig. 6.2 The roles of IDH1 and IDH2 in cellular metabolism.

Abbreviations: IDH1/2, isocitrate dehydrogenase 1/2; -KG,
-ketoglutarate; 2-HG, 2-hydroxyglutarate; 2HGD, PHD,
Y. Sonoda et al.
2-Hydroxyglutarate (2-HG) Levels Are

Elevated in Gliomas with IDH1 Mutations
Although the biological function of IDH mutation
is not fully understood, IDH mutations are always
monoallelic and do not result in simple loss of function. IDH1 mutation impairs the enzymes affinity for
its substrate and dominantly inhibits wild-type IDH1
activity through the formation of catalytically inactive heterodimers (Zhao et al., 2009). Consequently,
R132H mutation disrupts the function of IDH1 to convert isocitrate to -ketoglutarate. More recently, IDH1
mutations were found to result in a new ability of
the enzyme to catalyze the NADPH-dependent reduction of -ketoglutarate to R(-)-2-HG (2-HG) (Fig. 6.2)
(Dang et al., 2009). 2-HG levels were remarkably
elevated in human malignant gliomas with IDH1 mutations (Dang et al., 2009). Structural studies demonstrated that the replacement of arginine 132 by histidine results in a shift of the residues in the active site
to produce structural changes consistent with reduced
oxidative decarboxylation of isocitrate and acquisition
of the ability to convert -ketoglutarate to 2-HG.
The accumulation of 2-HG is found in patients
with inherited metabolic disorder 2-hydroxyglutaric
aciduria. This disease is caused by a deficiency
in the enzyme 2-HG dehydrogenase, which converts 2-HG to -ketoglutarate (Struys et al., 2005).
Patients with 2-HG dehydrogenase deficiency have
2-hydroxyglutarate dehydrogenase prolyl hydroxylase; HIF-1,

hypoxia-inducible factor 1; HIF-1-OH, hydroxylated HIF-1
an increased risk of developing brain tumors (Aghili

et al., 2009). The effect of IDH1 mutation on cellular
metabolism requires more investigation, but reduction
of -ketoglutarate by 2-HG or mutant IDH1 results in
a lower level of prolyl hydroxylases and promotes the
accumulation of hypoxia-inducible factor 1 (HIF-1)
(Fig. 6.2). Alterations in HIF-1 may result from mutant
IDH1 protein expression. In addition to IDH mutations
in gliomas, mutations of other metabolic enzymes such
as fumarate hydratase and succinate dehydrogenase
occur in paraganglioma and leiomyoma, respectively
(King et al., 2006). Energy is produced predominantly
by aerobic glycolysis in the cytoplasm of most cancer
cells, rather than by mitochondrial oxidative phosphorylation as in normal differentiated cells, a phenomenon termed the Warburg effect. Aerobic glycolysis may provide cancer cells with a growth advantage
by supplying required metabolites for incorporation
into the biomass to produce new cells. Aerobic glycolysis may also be important for glioma progression.
Both PI3K and tyrosine kinase signaling are involved
in growth control and glycolysis. PI3K activation by
PTEN mutation and/or tyrosine kinase activation by
EGFR alteration are frequently found in prGBMs.
However, PTEN mutation and EGFR alterations are
rarely found in AA and GBM patients with IDH
mutation, suggesting that the mechanism of cellular
metabolism in glioma might depend on IDH status.
Conclusions
IDH mutations seem to play an important role in the
formation of astrocytic and oligodendroglial tumors.
Detection of IDH mutations is easier to perform and
interrupt than the determination of 1p/19q codeletion
and MGMT promoter methylation. Such information
could be useful to improve the diagnostic and therapeutic strategies for gliomas. Furthermore, measurement of 2-HG production will enable identification of
patients with IDH1 mutant brain tumors. Of course,
further analysis of IDH1 and IDH2 in glioma model
systems will be necessary to clarify the genetic mechanisms involved in the initiation and malignant progression of this disease. In addition, extensive genetic
profiling of gliomas may allow the molecular classification of gliomas to replace the current histological classification in the near future. Patients with
51
IDH1 mutation may benefit from treatment modalities designed to inhibit the mutant IDH1 expression.
Inhibition of 2-HG production might also have therapeutic potential in the treatment of gliomas.
References
Aghili M, Zahedi F, Rafiee E (2009) Hydroxyglutaric aciduria
and malignant brain tumor: a case report and literature
Von Deimling A (2008) Analysis of the IDH1 codon 132
Camelo-Piragua S, Jansen M, Ganguly A, Kim JC, Louis DN,
Nutt CL (2010) Mutant IDH1-specific immunohistochemistry distinguishes diffuse astrocytoma from astrocytosis.
Acta Neuropathol 119(4):509511
Capper D, Weissert S, Balss J, Habel A, Meyer J, Jager D,
Ackermann U, Tessmer C, Korshunov A, Zentgraf H,
Hartmann C, Von Deimling A (2010) Characterization of
R132H mutation-specific IDH1 antibody binding in brain
tumors. Brain Pathol 20:245254
Marks KM, Prins RM, Ward PS, Yen KE, Liau LM,
MG, Su SM (2009) Cancer-associated IDH1 mutations produce 2-hydroxyglutarate. Nature 462:739744
Furnari FB, Fenton T, Bachoo RM, Mukasa A, Stommel JM,
Stegh A, Hahn WC, Ligon KL, Louis DN, Brennan C, Chin
L, Depinho RA, Cavenee WK (2007) Malignant astrocytic
glioma: genetics, biology, and paths to treatment. Genes Dev
21:26832710
Reifenberger G, Von Deimling A (2009) Type and frequency
diffuse gliomas. Acta Neuropathol 118:469474
Ichimura K, Pearson DM, Kocialkowski S, Backlund LM, Chan
R, Jones DT, Collins VP (2009) IDH1 mutations are present
in the majority of common adult gliomas but rare in primary
glioblastomas. Neuro-oncology 11:341347
Kang MR, Kim MS, Oh JE, Kim YR, Song SY, Seo SI, Lee JY,
Yoo NJ, Lee SH (2009) Mutational analysis of IDH1 codon
132 in glioblastomas and other common cancers. Int J Cancer
125:353355
Kato Y, Jin G, Kuan CT, Mclendon RE, Yan H, Bigner
DD (2009) A monoclonal antibody IMab-1 specifically
recognizes IDH1R132H, the most common glioma-derived
mutation. Biochem Biophys Res Commun 390:547551
52
King A, Selak MA, Gottlieb E (2006) Succinate dehydrogenase
and fumarate hydratase: linking mitochondrial dysfunction
and cancer. Oncogene 25:46754682
Korshunov A, Meyer J, Capper D, Christians A, Remke M,
Witt H, Pfister S, Von Deimling A, Hartmann C (2009)
Combined molecular analysis of BRAF and IDH1 distinguishes pilocytic astrocytoma from diffuse astrocytoma.
Acta Neuropathol 118:401405
Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC,
Jouvet A, Scheithauer BW, Kleihues P (2007) The 2007
WHO classification of tumours of the central nervous system.
Mardis ER, Ding L, Dooling DJ, Larson DE, Mclellan MD,
Chen K, Koboldt DC, Fulton RS, Delehaunty KD, Mcgrath
SD, Fulton LA, Locke DP, Magrini VJ, Abbott RM, Vickery
TL, Reed JS, Robinson JS, Wylie T, Smith SM, Carmichael
L, Eldred JM, Harris CC, Walker J, Peck JB, Du F, Dukes
AF, Sanderson GE, Brummett AM, Clark E, Mcmichael
JF, Meyer RJ, Schindler JK, Pohl CS, Wallis JW, Shi X,
Lin L, Schmidt H, Tang Y, Haipek C, Wiechert ME, Ivy
JV, Kalicki J, Elliott G, Ries RE, Payton JE, Westervelt P,
Tomasson MH, Watson MA, Baty J, Heath S, Shannon WD,
Nagarajan R, Link DC, Walter MJ, Graubert TA, Dipersio
JF, Wilson RK, Ley TJ (2009) Recurring mutations found
by sequencing an acute myeloid leukemia genome. N Engl J
Med 361:10581066
Ohgaki H, Kleihues P (2009) Genetic alterations and signaling pathways in the evolution of gliomas. Cancer Sci 15:
60026007
Mclendon R, Rasheed BA, Keir S, Nikolskaya T, Nikolsky
Y, Busam DA, Tekleab H, Diaz LA Jr., Hartigan J, Smith
Science 321:18071812
Reifenberger J, Reifenberger G, Liu L, James CD, Wechsler
W, Collins VP (1994) Molecular genetic analysis of
oligodendroglial tumors shows preferential allelic deletions
on 19q and 1p. Am J Pathol 145:11751190
Y. Sonoda et al.
Sanson M, Marie Y, Paris S, Idbaih A, Laffaire J, Ducray F,
El Hallani S, Boisselier B, Mokhtari K, Hoang-Xuan K,
Delattre JY (2009) Isocitrate dehydrogenase 1 codon 132
mutation is an important prognostic biomarker in gliomas.
Sonoda Y, Kumabe T, Nakamura T, Saito R, Kanamori M,
Yamashita Y, Suzuki H, Tominaga T (2009) Analysis of
IDH1 and IDH2 mutations in Japanese glioma patients.
Cancer Sci 100:19961998
Struys EA, Salomons GS, Achouri Y, Van Schaftingen E, Grosso
S, Craigen WJ, Verhoeven NM, Jakobs C (2005) Mutations
in the D-2-hydroxyglutarate dehydrogenase gene cause D-2hydroxyglutaric aciduria. Am J Hum Genet 76:358360
Watanabe T, Nobusawa S, Kleihues P, Ohgaki H (2009)
IDH1 mutations are early events in the development of
astrocytomas and oligodendrogliomas. Am J Pathol 174:
11491153
Wick W, Hartmann C, Engel C, Stoffels M, Felsberg
J, Stockhammer F, Sabel MC, Koeppen S, Ketter R,
Meyermann R, Rapp M, Meisner C, Kortmann RD, Pietsch
T, Wiestler OD, Ernemann U, Bamberg M, Reifenberger
G, Von Deimling A, Weller M (2009) NOA-04 randomized
phase III trial of sequential radiochemotherapy of anaplastic glioma with procarbazine, lomustine, and vincristine or
temozolomide. J Clin Oncol 27:58745880
Xu X, Zhao J, Xu Z, Peng B, Huang Q, Arnold E, Ding J (2004)
Structures of human cytosolic NADP-dependent isocitrate
dehydrogenase reveal a novel self-regulatory mechanism of
activity. J Biol Chem 279:3394633957
Yan H, Bigner DD, Velculescu V, Parsons DW (2009a) Mutant
metabolic enzymes are at the origin of gliomas. Cancer Res
69:91579159
Yan H, Parsons DW, Jin G, Mclendon R, Rasheed BA, Yuan W,
VE, Vogelstein B, Bigner DD (2009b) IDH1 and IDH2
mutations in gliomas. N Engl J Med 360:765773
Zhao S, Lin Y, Xu W, Jiang W, Zha Z, Wang P, Yu W, Li Z,
Gong L, Peng Y, Ding J, Lei Q, Guan KL, Xiong Y (2009)
Glioma-derived mutations in IDH1 dominantly inhibit IDH1
catalytic activity and induce HIF-1alpha. Science 324:
261265
Chapter 7
Malignant Glioma: Isocitrate Dehydrogenases

1 and 2 Mutations
Zachary J. Reitman and Hai Yan
Abstract Mutations in the cytoplasmic NADP+ dependent isocitrate dehydrogenase, IDH1, frequently
occur in gliomas. The mutations are somatic, almost
always heterozygous, and occur at R132, an active site
residue of IDH1 that is important for its catalytic function. IDH1 mutations occur in >70% of WHO grades
II and III astrocytomas and oligodendrogliomas, as
well as WHO grade IV secondary glioblastomas, and
more rarely in other glioma subtypes. IDH2 is the
mitochondrial homolog of IDH1, and mutations in
IDH2 R172, the analogous residue to IDH1 R132,
also occur in these subtypes of gliomas, albeit at a
much lower frequency. R132H and R172K are the
most common IDH1 and IDH2 mutations observed
in gliomas, respectively, though alterations to other
amino acids at these hotspots have also been observed.
IDH1 mutations frequently co-occur with loss of chromosomes 1p and 19q or point mutations of TP53, and
they appear to occur before other genetic alterations
during glioma pathogenesis. Furthermore, IDH1 mutations are associated with a younger age at diagnosis
and a better prognosis for many glioma subtypes. The
frequency of IDH1 and IDH2 mutations in specific
glioma subtypes suggests that testing for these mutations may help to guide clinical decision-making for
glioma patients, and PCR- and antibody-based tests
have been developed to determine tumor mutation
H. Yan ()
The Pediatric Brain Tumor Foundation Institute, The Preston
Robert Tisch Brain Tumor Center, Duke University Medical
Center, Durham, NC 27710, USA; The Department of
Pathology, Duke University Medical Center, Durham, NC
27710, USA
e-mail: Yan0002@mc.duke.edu
status. The IDH1 and IDH2 mutations abolish the

normal function of the encoded enzymes to oxidize
isocitrate to -ketoglutarate and confer a neomorphic
gain of enzymatic activity to reduce -ketoglutarate
to 2-hydroxyglutarate. This gain of function suggests
that IDH1 and IDH2 mutations are oncogenic, but the
mechanism by which this promotes glioma pathogenesis remains unknown.
Keywords Malignant glioma IDH1 Mutation
R132 Enzyme Dehydrogenase
Introduction
The study of genetic alterations that arise in tumors
of the central nervous system represents a decades-old
field. Recent whole-genome analyses have advanced
this field by integrating years of careful observations
and by revealing genetic changes that were not obvious in previous studies. Foremost among discoveries
of novel genetic changes are frequent mutations of
IDH1 and IDH2 in gliomas. These mutations stand
apart from other genetic alterations in cancer due
in part to the fact that they occur with striking frequency in certain glioma subtypes. Also, while the vast
majority of genetic alterations in cancer affect genes
involved in cellular signaling and growth control, the
IDH1 and IDH2 mutations affect cellular metabolic
enzymes without known signaling or growth regulating
properties. Furthermore, rather than simply activate or
inactivate the normal function of the genes as is typically the result for alterations in cancer, the IDH1 and
IDH2 mutations bestow a novel neomorphic activity on
the encoded enzymes. Study of these unique mutations
may further our understanding of the pathology of

53
54
tumors of the central nervous system and possibly

provide avenues to better manage these diseases.
Mutations in IDH1 were discovered in a wholegenome sequencing analysis of coding exons in
glioblastoma samples (Parsons et al., 2008). IDH1
is the cytoplasmic nicotinamide adenine dinucleotide
phosphate (NADP+ )-dependent isocitrate dehydrogenase. IDH1 mutations observed in glioma specifically
alter codon 132, which encodes an arginine (R132) in
the enzymes substrate binding site. The mutations are
somatic and almost always heterozygous. Additionally,
mutations in the analogous R172 codon of IDH2,
which encodes the mitochondrial NADP+ -dependent
isocitrate dehydrogenase which is the only homolog
of IDH1, also occur in many of the same subtypes of
glioma at a much lower frequency (Yan et al., 2009).
The most frequent IDH1 alteration observed in gliomas
is R132H, with other alterations at this codon more
rarely observed (Balss et al., 2008). This chapter considers the epidemiology and clinical characteristics of
central nervous system tumors with IDH mutations,
and briefly discusses the biochemical function of the
mutated enzymes.
Tumor Type Distribution

IDH mutations occur frequently in specific glioma
subtypes. Gliomas are primary tumors of the central nervous system that histologically resemble glia,
the supporting cells of the brain. Like other CNS
tumors, gliomas are classified by the World Health
Organization from grades I to IV, with grades IIIIV
tumors considered malignant gliomas. The IDH mutations occur in gliomas classified as astrocytomas and
oligodendrogliomas, which histologically resemble
astrocytes and oligodendrocytes, respectively. Grade
I pilocytic astrocytomas predominantly occur in children and seldom progress to higher-grade tumors.
Grade II gliomas comprise diffuse astrocytomas, welldifferentiated oligodendrogliomas, and mixed oligoastrocytomas. These can progress to grade III gliomas
of the same histology, with grade II diffuse astrocytomas progressing to grade III anaplastic astrocytomas,
grade II well-differentiated oligodendrogliomas progressing to grade III anaplastic oligodendrogliomas,
and grade II oligoastrocytomas progressing to grade
III anaplastic oligoastrocytomas. Glioblastomas are
Z.J. Reitman and H. Yan
grade IV astrocytomas that have a particularly varied

and dysplastic histology and carry a grim prognosis.
Secondary glioblastomas, by definition, arise from
grade II and III astrocytomas. In contrast to this, the
majority of glioblastomas are primary glioblastomas
which arise de novo, rather than from a previous lowergrade tumor (Louis et al., 2007). Figure 7.1 depicts the
glioma subtypes and pathways of tumor progression
discussed above.
IDH1 mutations are common (>70%) in grades II
and III astrocytomas and oligodendrogliomas, and in
secondary glioblastomas (Balss et al., 2008). They
occur more rarely in primary glioblastomas (316%)
(Balss et al., 2008; Parsons et al., 2008; Bleeker et al.,
2009; Yan et al., 2009), and are seldom found in grade
I pilocytic astrocytomas (02%) (Balss et al., 2008;
Ichimura et al., 2009; Korshunov et al., 2009; Yan
et al., 2009). IDH2 mutations share a similar tumor
type distribution with IDH1 mutations (Yan et al.,
2009). They are generally found in tumors that do
not contain IDH1 mutations, and occur much less frequently than IDH1 mutations (Hartmann et al., 2009;
Yan et al., 2009). As an example, in a group of
1,010 grades II and III gliomas, 71% contained IDH1
mutations while only 3% contained IDH2 mutations
(Hartmann et al., 2009). The percentage of astrocytic and oligodendroglial tumors containing IDH1 or
IDH2 mutations based on two large studies is shown
in Fig. 7.1. Gangliogliomas seldom (17% or less)
contain IDH1 mutations (Sonoda et al., 2009), and
extensive analysis of other CNS tumors such as medulloblastomas, meningiomas, and ependymomas has not
revealed IDH mutations.
IDH1 and IDH2 mutations are found in tumors
but not in normal tissues from the same patient,
and are almost always heterozygous with a wild-type
allele (Yan et al., 2009). IDH1 and IDH2 mutations
in gliomas may be specific for humans, as analogous mutations were not found in grades IIIV
canine gliomas which are histopathologically indistinguishable from human gliomas (Reitman et al.,
2009). However, IDH1 and IDH2 mutations also
occur in up to 23% of human acute myelogenous
leukemias (Mardis et al., 2009; Ward et al., 2010),
and rare cases of IDH1 R132 mutations in colorectal cancer (Sjoblom et al., 2006), prostate cancer (Kang et al., 2009), and acute lymphoblastic
leukemia, B-type (Kang et al., 2009) have been
reported.
Malignant Glioma: Isocitrate Dehydrogenases 1 and 2 Mutations
55
Fig. 7.1 Schematic of pathogenesis and progression of astrocytic and oligodendroglial tumors, frequency of tumors carrying
IDH1 R132 and/or IDH2 R172 mutations for each tumor type,
and other common genetic alterations that occur in the pathogenesis or progression of each tumor type. Tumor grade is
indicated to the left. Arrows indicate de novo pathogenesis
from a normal cell, which has been speculated to be a normal stem cell, normal glial progenitor, or other normal cell, or
progression from the indicated lower-grade tumor type. Grade
I tumors: PA, pilocytic astrocytoma. Grade II tumors: O, welldifferentiated oligodendroglioma; OA, oligoastrocytomas; A,
diffuse astrocytoma. Grade III tumors: AO, anaplastic oligodendroglioma; AOA, anaplastic oligoastrocytoma; AA, anaplastic
astrocytomas. Grade IV tumors: pGBM, primary glioblastoma;
sGBM, secondary glioblastoma; sGBMO, secondary glioblastomas with oligodendroglial component. HD, homozygous deletion. N.S., not studied. Percentages were reported by Yan et al.
(2009), except for OA and AOA which were reported by
Hartmann et al. (2009) since relatively few OA and AOA tumors
were analyzed in Yan et al. Figure reproduced from Reitman and
Yan (2010) with permission and adapted to include percentages
Association with Other Genetic

Alterations
group of a variety of glioma subtypes, 92% of tumors

with IDH1 R132 mutations also contained either
TP53 mutations or 1p/19q loss, while IDH1 wildtype tumors rarely contained these changes (Ichimura
et al., 2009). Groups of grade II diffuse astrocytomas,
grade III anaplastic astrocytomas, primary glioblastomas, and secondary glioblastomas analyzed for both
IDH1 and TP53 in the same tumors often have
a higher prevalence of TP53 mutations among the
IDH1-mutated group than the IDH1wild type group
(Parsons et al., 2008; Ichimura et al., 2009; Watanabe
et al., 2009a; Yan et al., 2009). IDH1 R132 mutation
is significantly associated with the presence of any
alteration in the TP53 pathway, including alterations
in MDM2, MDM4, and p14ARF, in the combined
group of grades II and III astrocytomas (Ichimura
et al., 2009). Interestingly, 5 astrocytoma patients with
Li-Fraumeni syndrome, which is defined by germline
TP53 mutation and predisposes to astrocytomas and
other cancers, had somatic IDH1 mutations (Watanabe
et al., 2009b). Together, these findings illustrate that
IDH mutations generally co-occur with TP53 mutations in the astrocytic tumors, and suggest that both
In gliomas, IDH mutations strongly associate with

TP53 mutation or codeletion of chromosomes 1p and
19q, but are inversely associated with other genetic
changes common in glioma, such as EGFR amplification. TP53 mutations are commonly present in grades
II and III astrocytomas and oligoastrocytomas as well
as secondary glioblastomas, but only about 30% of primary glioblastomas contain TP53 mutations; they are
rare in pure oligodendroglial tumors. Pure oligodendroglial tumors frequently possess total 1p/19q loss
and rarely contain TP53 mutations. 1p/19q loss is rare,
however, in the astrocytic tumors. In grades II and
III oligoastrocytomas, either TP53 mutation or 1p/19q
loss are often present, and these changes are inversely
associated. Thus, TP53 mutation and 1p/19q loss are
generally mutually exclusive and distributed among
the grade II and III astrocytic and oligodendroglial
tumors, respectively.
IDH1 and IDH2 mutations frequently co-occur with
either TP53 mutations or 1p/19q loss. In one large
56
are crucial steps in transformation of normal cells into

astrocytoma cells.
Oligodendroglial tumors frequently contain 1p/19q
codeletions. IDH mutations often co-occur with 1p/
19q loss in oligodendrogliomas, with one study finding
1p/19q loss in 85% of IDH1 or IDH2-mutated oligodendrocytic tumors, but not in oligodendrocytic tumors
with wild-type IDH genes (Yan et al., 2009). In another
study, all oligodendroglial tumors with 1p/19q loss
harbored mutations in either IDH1 or IDH2, while not
all tumors with IDH mutations harbored 1p/19q loss in
this group (Ichimura et al., 2009). Overall, IDH mutations seem to be a common feature of astrocytomas and
oligodendrogliomas, with oligodendrogliomas defined
by 1p/19q loss and astrocytomas defined by TP53
mutations.
IDH mutations are inversely associated with many
of the hallmark changes of primary glioblastomas,
including EGFR amplification, CDKN2A or CDKN2B
deletion, and PTEN mutations. These changes are
uncommon in secondary glioblastomas and lowergrade tumors. IDH mutations and EGFR amplification
are inversely correlated among grades IIIV gliomas
(Sanson et al., 2009), grades IIIV astrocytic tumors
(Ichimura et al., 2009), grade IIIIV astrocytic tumors
(Yan et al., 2009), or glioblastomas (Parsons et al.,
2008; Nobusawa et al., 2009). Similarly, changes in
PTEN and CDKN2A/CDKN2B are common among
IDH wild-type, but not IDH-mutated, grade III astrocytomas and glioblastomas (Yan et al., 2009). IDH1
mutations have been shown to be negatively associated
with PTEN mutations as well (Ichimura et al., 2009).
Secondary glioblastomas with wild-type IDH1 have
genetic features, such as frequent EGFR amplification,
that are characteristic for primary glioblastomas. This
suggests that the IDH1 wild-type secondary glioblastomas are actually primary glioblastomas for which
the glioblastoma was initially mistaken for a lowergrade lesion (Nobusawa et al., 2009). On the other
hand, primary glioblastomas with mutated IDH1 more
frequently have alterations, such as TP53 mutation,
that are more common in secondary glioblastoma, suggesting that these tumors may have progressed from a
lower-grade astrocytoma that was not clinically evident
(Nobusawa et al., 2009).
IDH1 mutations also associate with characteristic
transcriptional and epigenetic profiles. Gliomas with
IDH1 R132 mutations frequently display a pattern of
hypermethylation throughout the genome compared to
gliomas without IDH1 mutations (Noushmehr et al.,

2010). Additionally, gene expression analysis has
revealed that gliomas typically cluster into proneural, neural, classical, or mesenchymal gene expression
signatures (Phillips et al., 2006), and that IDH1mutated gliomas associate with the proneural signature (Verhaak et al., 2010). The proneural signature
resembles the gene expression signature of normal
oligodendrocytes, and has previously been found to
associate with grade II and III gliomas, and with
younger age at diagnosis and better prognosis compared to gliomas with the other signatures (Phillips
et al., 2006). Because IDH1-mutated gliomas have
genetic, epigenetic, and gene expression alterations
that are distinct from gliomas without IDH1 mutations,
it stands to reason that pathogenesis and biology of
IDH1-mutated gliomas may be fundamentally different from other gliomas. Since IDH2 mutations are rarer
in gliomas than IDH1 mutations, it is not clear whether
IDH2 mutated tumors have similar genetic, epigenetic,
and gene expression profiles to the IDH1-mutated
tumors.
Timing of Mutations
IDH1 and IDH2 mutations appear to be early events
in glioma formation. The proportion of IDH1 and
IDH2 mutated tumors does not increase with grade
(Balss et al., 2008), suggesting that the mutations arise
during tumor formation, rather than during progression to a higher grade. The mutations are frequent
in grade II gliomas, indicating that they are important for early steps in tumor formation. In a group of
51 grade II glioma patients with two or more biopsies, 42 (82%) had an IDH1 mutation at the first
biopsy, and had an identical mutation at later biopsies. Two of the remaining nine patients initially did
not have an IDH1 mutation, but developed an IDH1
mutation by a later biopsy (Watanabe et al., 2009a).
This indicates that IDH1 mutation is an early event
in glioma pathogenesis that persists throughout tumor
progression.
IDH mutations also occur before other common
genetic alterations in glioma. Of 51 patients with multiple biopsies, four grade II glioma patients with IDH1
mutations at their first biopsy had TP53 mutations as
well at their last biopsy. Another three of these patients
had IDH1 mutation alone at their first biopsy and

1p/19q loss by the time of their last biopsy (Watanabe
et al., 2009a). However, no cases of IDH1 or IDH2
mutation occurring after acquisition of a somatic TP53
mutation or 1p/19q loss have been reported. The fact
that IDH mutations occur frequently in astrocytomas
containing TP53 mutations and in oligodendrogliomas
with 1p/19q loss, and that IDH mutations occur before
the other genetic alterations, has led to speculation that
the mutations occur in a precursor cell that can give
rise to both the astrocytic and oligodendroglial tumor
types. Other genetic pathways to tumorigenesis that
do not involve frequent IDH mutations likely lead to
gliomas such as grade I pilocytic astrocytomas and
grade IV primary glioblastomas by separate pathways,
as evidenced by a different constellation of genetic
changes observed for those tumor types. The common genetic changes found in grades I-IV gliomas, and
the suggestion that IDH1 and IDH2 mutations arise in
cell that can give rise to both astrocytic and oligodendroglial grade II gliomas, have been incorporated in
Fig. 7.1.
Patient Age
IDH mutations occur rarely in pediatric glioma
patients, and occur in the younger of adult glioma
patients. Pediatric patients that do have IDH1 or IDH2
mutations are generally older teenagers with grades
IIIII tumors rather than young children (De Carli
et al., 2009). In studies of adult patients with grades
II or III astrocytomas, primary glioblastomas, grade III
anaplastic oligoastrocytomas, any grade II glioma, any
grade III glioma, or any glioblastoma, patients with
IDH1 mutations had a younger median age at diagnosis than those with the same tumor type who did
not have IDH1 mutations (Balss et al., 2008; Parsons
et al., 2008; Ichimura et al., 2009; Yan et al., 2009).
For example, Watanabe et al. found that grade II diffuse astrocytoma patients with IDH1 mutations had
a median age of 34 compared to 52 for IDH1 wildtype patients (Watanabe et al., 2009a), and Yan et al.
found that grade III anaplastic astrocytoma patients
with IDH1 or IDH2 mutations had a median age of
34 compared to 53 for the IDH wild-type patients
(Balss et al., 2008). However, a significantly younger
age for IDH1-mutated patients has not been observed
57
for patients with pure oligodendroglial tumors. The

older median age for grade II and III astrocytoma
patients without IDH mutations has been speculated
to be caused by misclassification of some primary
glioblastomas as lower-grade tumors (Balss et al.,
2008). Misclassification of primary glioblastomas as
grade III anaplastic astrocytomas could also contribute
to the lower frequency of IDH1 mutations in grade III
anaplastic astrocytomas compared to grade II diffuse
astrocytomas that has been observed in multiple studies (Balss et al., 2008; Ichimura et al., 2009; Yan et al.,
2009).
Patient Survival
Glioma patients with IDH mutations survive longer
than patients with wild-type IDH1 and IDH2. In one
study, glioblastoma patients with IDH1 mutations survived 3.7 years compared to 1.1 years for IDH1 wildtype glioblastoma patients (Parsons et al., 2008). Other
studies have found that patients with IDH-mutated
grade III anaplastic astrocytomas and groups of grade
II, III, or IV gliomas survive longer than patients with
IDH wild-type tumors of the same type (Dubbink et al.,
2009; Ichimura et al., 2009; Sanson et al., 2009; Yan
et al., 2009). Since IDH-mutated patients are younger
than non-mutated patients, and since younger age is
associated with longer survival, these univariate analyses do not show whether IDH mutation predicts better
survival independently from age. Thus, some of the
large differences in survival reported for these tumors
may be reflect the fact that IDH mutations are markers
for young age, and therefore better prognosis. Despite
the association with survival, no association was found
between IDH1 status and response to temozolomide in
grade II diffuse astocytomas (Dubbink et al., 2009).
Multivariate analyses that take into account age
and other factors shed the most light on IDH mutations as independent predictors of survival. Results
of such analyses have been mixed. In one study, a
Cox regression multivariate analysis including age,
tumor type, grade, TP53 mutation status, and 1p/19q
codeletion status failed to identify IDH1 status as an
independent prognostic factor (Ichimura et al., 2009).
Other studies, however, have found IDH1 status to
be a predictor of good outcome. A Cox multivariate analysis that did not use age as a covariate was
58
performed on a randomized clinical trial of grade III

glioma patients found IDH1 status to be more strongly
associated with longer progression-free survival than
other prognostic factors, such as 1p/19q status and
MGMT promoter methylation (Wick et al., 2009). In
another large group of patients, multivariate analysis
adjusting for grade, age above or below 48, MGMT
status, genomic profile, and treatment, confirmed that
IDH1 mutation as an independent favorable prognostic
marker (Sanson et al., 2009). A multivariate analysis
of glioblastoma patients that adjusted for age found a
significant association of IDH1 mutation with longer
survival (Nobusawa et al., 2009).
Clinical Testing
The specificity of IDH mutations for certain tumor
types and their association with outcome makes testing for IDH mutation status potentially useful for the
diagnosis and prognosis of gliomas. The high frequency of IDH mutations in grades IIIII gliomas
and secondary glioblastomas makes them highly sensitive and specific for these tumors compared to other
CNS tumors. This has been exploited to show that
IDH1 status can help distinguish between grade II diffuse astrocytomas and grade I pilocytic astrocytomas,
which may be helpful in cases for which scant material
is available for histopathological analysis (Korshunov
et al., 2009). In another study, the sensitivity and
specificity of IDH1 mutations were 73.3% and 96.3%,
respectively, for secondary glioblastomas compared to
primary glioblastomas (Nobusawa et al., 2009).
Given the potential diagnostic utility for determining IDH mutation status, several methods have been
developed to determine whether IDH1 or IDH2 are
mutated in tumor samples. Polymerase chain reaction
(PCR)-based assays have been developed to accurately detect IDH1 and IDH2 mutation in the DNA of
tumor cells (Meyer et al., 2009), and these methods
can determine IDH status in formalin-fixed, paraffinembedded tissues even in histologically normal tissue
that contains little tumor material (Horbinski et al.,
2009). Additionally, monoclonal antibodies have been
developed to specifically stain tumors containing the
IDH1 R132H protein (Capper et al., 2009; Kato et al.,
2009), which could be put to use in typical diagnostic

pathology laboratories.
Properties of Mutated Enzymes

The frequency, specificity, and early timing of IDH
mutations provide strong evidence for their importance
in glioma tumorigenesis, and functional studies have
begun to reveal the properties of the mutated enzymes.
The IDH1 and IDH2 mutations observed in cancer
diminish the normal isocitrate dehydrogenase activity
of the enzymes, shown on the left side of Fig. 7.2, as
demonstrated by in vitro assays of either recombinant
mutant enzymes (Zhao et al., 2009) or lysates of cells
overexpressing the mutant enzymes (Yan et al., 2009).
Additionally, recombinant IDH1 R132H can bind to
wild-type IDH1, and the resulting heterodimers have
diminished isocitrate dehydrogenase activity (Zhao
et al., 2009). Since most IDH mutations are heterozygous with a wild-type allele, this has led to speculation
that the IDH1 mutants function in cancer to dominant negatively inactivate wild-type IDH1. However,
the formation of wild-type:mutant heterodimers has
not been demonstrated in vivo, and overexpression of
IDH1 R132 and IDH2 R172 mutants has not been
shown to lower cellular isocitrate dehydrogenase activity in glioma cell lines, as would be expected for
a dominant negative enzyme. Additionally, Park and
colleagues have demonstrated that IDH1 and IDH2
protect of the cell from noxious insults (Kim et al.,
2007; Park et al., 2008). It seems reasonable that one
wild-type allele of either gene could continue to provide protective benefits to the cell when the other allele
is mutated in cancer, but it seems unlikely that the
major function for IDH mutations cancer would be to
inactivate these protective enzymes.
The specificity of the IDH mutations for hotspot
codons is reminiscent of oncogenes such as KRAS
and PIK3CA, which are also activated by specific point mutations. This resemblance suggests
that IDH mutations also confer a gain of function. Consistent with this, IDH1 R132 and IDH2
R172 mutants gain the neomorphic activity to reduce
-ketoglutarate to (R)-2-hydroxyglutarate as shown
on the right side of Fig. 7.2 (Dang et al., 2009;
59
Fig. 7.2 Wild-type IDH1 and IDH2 catalyze the oxidative

decarboxylation of isocitrate to form -ketoglutarate, but IDH1
R132 and IDH2 R172 mutants catalyze the reduction of
-ketoglutarate to (R)-2-hydroxyglutarate. Wild-type IDH1 and
IDH2 normally catalyze a two step reaction, starting with isocitrate on the upper left and resulting in -ketoglutarate on the
bottom center. First, they catalyze the NADP+ -dependent oxidation of the -hydroxyl of isocitrate (blue) to produce an
intermediate, oxalosuccinate. Then, they catalyze the decarboxylation of the -carboxyl (green) of this intermediate to produce
-ketoglutarate and CO2 . This reaction is reversible: under
appropriate conditions, the enzymes can catalyze the addition
of CO2 to -ketoglutarate to form oxalosuccinate. Then, they
catalyze the NADPH-dependent reduction of the -ketone to a
hydroxyl to form isocitrate. R132 of IDH1 is analogous to R172
of IDH2. In each enzyme, this arginine residue forms hydrogen

bonds with the -carboxyl of isocitrate. This presumably coordinates removal of the -carboxyl to release CO2 in the forward
reaction. Conversely, the arginine coordinates the addition of
CO2 to form the -carboxyl in the reverse reaction. As expected
for enzymes with mutations in a residue that aids in substrate
binding and catalysis, IDH1 R132 and IDH2 R172 mutants are
impaired in their normal ability to bind isocitrate or convert
it to -ketoglutarate. However, the mutated enzymes still bind
-ketoglutarate. Nevertheless, the enzymes do not coordinate
the addition of CO2 to -ketoglutarate followed by reduction of
the -ketone as normally occurs in the reverse reaction. Instead,
only reduction of the -ketone occurs. This results in conversion
of -ketoglutarate and NADPH to (R)-2-hydroxyglutarate and
NADP+ as shown on the right
Gross et al., 2010; Ward et al., 2010). IDH1mutated gliomas contain a 100-fold elevated concentration of (R)-2-hydroxyglutarate (also known
as R(-)-2-hydroxyglutarate or D-2-hydroxyglutarate)
compared to IDH1-wild-type gliomas (Dang et al.,
2009). This neomorphic activity apparently leads to
the accumulation of (R)-2-hydroxyglutarate observed
in the tumors, and also may alter the cellular
NADPH/NADP+ ratio and cause flux away from ketoglutarate. One or more of these metabolic changes
could confer a selective advantage to glioma cells
and promote glioma formation through a yet-unknown
mechanism. If the presence of mutated IDH1 or IDH2
is essential for the malignant properties of glioma cells,
it may be possible to design molecular therapies that

target these tumor-specific mutants. Further research
will be needed to determine the downstream effects of
the gain of function of the IDH mutations.
Mutation Types
A spectrum of different mutated codons has been
observed at IDH1 R132 and IDH2 R172 in glioma.
IDH1 R132H is by far the most common IDH
mutation in gliomas, accounting for 88.292.8%
of IDH1-mutated tumors, followed by R132C
(3.44.6%), R132S (0.82.5%), R132G (0.63.9%),
60
R132L (04.5%), and R132V (one case) (Balss

et al., 2008; Hartmann et al., 2009; Ichimura et al.,
2009; Watanabe et al., 2009a; Yan et al., 2009).
Only a handful of the rarer IDH2 mutations have
been reported, with a total of 24 R172K, 9 R172M,
5 R172W, and 2 R172G cases found in two large
studies (Hartmann et al., 2009; Yan et al., 2009).
All of the observed mutations result from a single
base pair change in the R132 or R172 codon, except
for the single observed case of IDH1 R132V, which
results from alteration of two base pairs (Balss et al.,
2008). Different mutations have a range of impact
on enzyme function. For instance, IDH1 R132H has
a greater loss of isocitrate dehydrogenase activity
and more rapidly reduces -ketoglutarate than IDH1
R132C, R132L, or R132S (Dang et al., 2009). If these
changes in activity provide a selective advantage to
a glioma or pre-glioma cell, then perhaps the greater
potency for the R132H mutation explains its higher
observed frequency in gliomas compared to the other
mutations. Interestingly, IDH1 R132C is observed
more frequently than IDH1 R132H in non-glioma
cancers, including acute myelogenous leukemia and
case reports of prostate cancer, colorectal cancer,
and acute lymphoblastic leukemia, B-type (Sjoblom
et al., 2006; Kang et al., 2009; Mardis et al., 2009).
Additionally, all five cases of IDH1 mutations reported
in glioma patients with Li-Fraumeni syndrome were
R132C mutations (Watanabe et al., 2009b). Also,
patients with IDH1 R132C and R132S mutations
were significantly younger than those with R132H
mutations in one large study (Hartmann et al., 2009).
Differences have also been noted between patients
with IDH1- and IDH2-mutated tumors. One large
study noted a trend for IDH2 mutations to occur
more frequently in grade III tumors compared to
grade II tumors, and that IDH2 mutations were significantly more frequent in oligodendroglial tumors
compared to astrocytic tumors (Hartmann et al., 2009).
Though IDH1 and IDH2 mutations are inversely associated, several cases of grade III gliomas contained
both alterations (Hartmann et al., 2009). Interestingly,
acute myeloid leukemias exhibit a strikingly different distribution of IDH1 and IDH2 mutations than
gliomas. Acute myeloid leukemias contain IDH2
mutations slightly more frequently than IDH1 mutations (Ward et al., 2010). Furthermore, almost half
of IDH1 and IDH2 the mutations in acute myeloid
leukemias occur at IDH2 R140 which also apparently
confers (R)-2-hydroxyglutarate production to the

enzyme (Ward et al., 2010). Of note, this site that has
never been found to be mutated in gliomas (Yan et al.,
2009). The biological explanation for the distribution
of the different IDH mutations in different cancers
remains to be elucidated.
Conclusion
IDH1 and IDH2 mutations occur frequently in grades
II and III gliomas and secondary glioblastomas. These
changes associate with other genetic alterations such as
TP53 mutation and 1p/19q loss, as well as a younger
patient age and longer patient survival among most
tumor types. PCR and immunohistochemical techniques can determine mutation status in surgical tumor
samples and hold promise as powerful diagnostic and
prognostic tests to aid in the clinical management of
this difficult set of diseases. The mutations likely occur
early in tumorigenesis and their genetic profile suggests that they act as oncogenes. The mutated enzymes
have greatly reduced normal enzymatic activity but
gains the novel, possibly oncogenic, ability to produce
(R)-2-hydroxyglutarate. The frequency and specificity
of the IDH1 and IDH2 mutations in glioma suggests
that these alterations are central to glioma formation
and/or maintenance, but further study is needed to
understand the mechanism by which they interact with
glioma biology.
References
von Deimling A (2008) Analysis of the IDH1 codon 132
Capper D, Zentgraf H, Balss J, Hartmann C, von Deimling
A (2009) Monoclonal antibody specific for IDH1 R132H
mutation. Acta Neuropathol 118:599601
Marks KM, Prins RM, Ward PS, Yen KE, Liau LM,
MG, Su SM (2009) Cancer-associated IDH1 mutations produce 2-hydroxyglutarate. Nature 462:739744

De Carli E, Wang X, Puget S (2009) IDH1 and IDH2 mutations
in gliomas. N Engl J Med 360:2248 author reply 2249
Dubbink HJ, Taal W, van Marion R, Kros JM, van Heuvel
I, Bromberg JE, Zonnenberg BA, Zonnenberg CB, Postma
TJ, Gijtenbeek JM, Boogerd W, Groenendijk FH, Smitt PA,
Dinjens WN, van den Bent MJ (2009) IDH1 mutations in
low-grade astrocytomas predict survival but not response to
temozolomide. Neurology 73:17921795
Gross S, Cairns RA, Minden MD, Driggers EM, Bittinger MA,
Jang HG, Sasaki M, Jin S, Schenkein DP, Su SM, Dang
L, Fantin VR, Mak TW (2010) Cancer-associated metabolite 2-hydroxyglutarate accumulates in acute myelogenous
leukemia with isocitrate dehydrogenase 1 and 2 mutations.
J Exp Med 207:339344
Reifenberger G, von Deimling A (2009) Type and frequency
diffuse gliomas. Acta Neuropathol 118:469474
Horbinski C, Kofler J, Kelly LM, Murdoch GH, Nikiforova MN
(2009) Diagnostic use of IDH1/2 mutation analysis in routine
clinical testing of formalin-fixed, paraffin-embedded glioma
tissues. J Neuropathol Exp Neurol 68:13191325
Ichimura K, Pearson DM, Kocialkowski S, Backlund LM, Chan
R, Jones DT, Collins VP (2009) IDH1 mutations are present
in the majority of common adult gliomas but are rare in
primary glioblastomas. Neuro Oncol 11:341347
Kang MR, Kim MS, Oh JE, Kim YR, Song SY, Seo SI, Lee JY,
Yoo NJ, Lee SH (2009) Mutational analysis of IDH1 codon
132 in glioblastomas and other common cancers. Int J Cancer
125:353355
Kato Y, Jin G, Kuan CT, McLendon RE, Yan H, Bigner DD
(2009) A monoclonal antibody IMab-1 specifically recognizes IDH1R132H, the most common glioma-derived mutation. Biochem Biophys Res Commun 390:547551
Kim SY, Lee SM, Tak JK, Choi KS, Kwon TK, Park JW (2007)
Regulation of singlet oxygen-induced apoptosis by cytosolic NADP+-dependent isocitrate dehydrogenase. Mol Cell
Biochem 302:2734
Korshunov A, Meyer J, Capper D, Christians A, Remke M,
Witt H, Pfister S, von Deimling A, Hartmann C (2009)
Combined molecular analysis of BRAF and IDH1 distinguishes pilocytic astrocytoma from diffuse astrocytoma.
Mardis ER, Ding L, Dooling DJ, Larson DE, McLellan MD,
Chen K, Koboldt DC, Fulton RS, Delehaunty KD, McGrath
SD et al. (2009) Recurring mutations found by sequencing an acute myeloid leukemia genome. N Engl J Med
361:10581066
Meyer J, Pusch S, Balss J, Capper D, Mueller W, Christians A,
Hartmann C, von Deimling A (2010) PCR- and Restriction
Endonuclease-Based Detection of IDH1 Mutations. Brain
Pathol 20:298300
61
Nobusawa S, Watanabe T, Kleihues P, Ohgaki H (2009) IDH1

mutations as molecular signature and predictive factor of
secondary glioblastomas. Clin Cancer Res 15:60026007
Noushmehr H, Weisenberger DJ, Diefes K, Phillips HS, Pujara
K, Berman BP, Pan F, Pelloski CE, Sulman EP, Bhat KP,
Verhaak RG, Hoadley KA, Hayes DN, Perou CM, Schmidt
HK, Ding L, Wilson RK, Van Den Berg D, Shen H,
Bengtsson H, Neuvial P, Cope LM, Buckley J, Herman JG,
Baylin SB, Laird PW, Aldape K (2010) Identification of
a CpG island methylator phenotype that defines a distinct
subgroup of glioma. Cancer Cell 17:510522
Park SY, Lee SM, Shin SW, Park JW (2008) Inactivation of
mitochondrial NADP+-dependent isocitrate dehydrogenase
by hypochlorous acid. Free Radic Res 42:467473
Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt P,
Mankoo P, Carter H, Siu IM, Gallia GL et al. (2008) An integrated genomic analysis of human glioblastoma multiforme.
Science 321:18071812
Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano
RH, Wu TD, Misra A, Nigro JM, Colman H, Soroceanu
L, Williams PM, Modrusan Z, Feuerstein BG, Aldape K
(2006) Molecular subclasses of high-grade glioma predict
prognosis, delineate a pattern of disease progression, and
resemble stages in neurogenesis. Cancer Cell 9:157173
Reitman ZJ, Olby NJ, Mariani CL, Thomas R, Breen M, Bigner
DD, McLendon RE, Yan H (2010) IDH1 and IDH2 hotspot
mutations are not found in canine glioma. Int J Cancer
127(1):245246
Reitman ZJ, Yan H (2010) Isocitrate dehydrogenase 1 and 2
mutations in cancer: alterations at a crossroads of cellular
metabolism. J Natl Cancer Inst 102:932941
Sanson M, Marie Y, Paris S, Idbaih A, Laffaire J, Ducray F,
El Hallani S, Boisselier B, Mokhtari K, Hoang-Xuan K,
Delattre JY (2009) Isocitrate dehydrogenase 1 codon 132
mutation is an important prognostic biomarker in gliomas.
Sjoblom T, Jones S, Wood LD, Parsons DW, Lin J, Barber
TD, Mandelker D, Leary RJ, Ptak J, Silliman N, Szabo S,
Buckhaults P, Farrell C, Meeh P, Markowitz SD, Willis J,
Dawson D, Willson JK, Gazdar AF, Hartigan J, Wu L, Liu
C, Parmigiani G, Park BH, Bachman KE, Papadopoulos N,
Vogelstein B, Kinzler KW, Velculescu VE (2006) The consensus coding sequences of human breast and colorectal
cancers. Science 314:268274
Sonoda Y, Kumabe T, Nakamura T, Saito R, Kanamori M,
Yamashita Y, Suzuki H, Tominaga T (2009) Analysis of
IDH1 and IDH2 mutations in Japanese glioma patients.
Cancer Sci 100:19961998
MD, Miller CR, Ding L, Golub T, Mesirov JP et al (2010)
Integrated genomic analysis identifies clinically relevant
subtypes of glioblastoma characterized by abnormalities
in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell 17:
98110
Ward PS, Patel J, Wise DR, Abdel-Wahab O, Bennett BD, Coller
HA, Cross JR, Fantin VR, Hedvat CV, Perl AE, Rabinowitz
JD, Carroll M, Su SM, Sharp KA, Levine RL, Thompson CB
(2010) The common feature of leukemia-associated IDH1
and IDH2 mutations is a neomorphic enzyme activity converting alpha-ketoglutarate to 2-hydroxyglutarate. Cancer
Cell 17:225234
62
Watanabe T, Nobusawa S, Kleihues P, Ohgaki H (2009a) IDH1
mutations are early events in the development of astrocytomas and oligodendrogliomas. Am J Pathol 174:11491153
Watanabe T, Vital A, Nobusawa S, Kleihues P, Ohgaki H
(2009b) Selective acquisition of IDH1 R132C mutations in
astrocytomas associated with Li-Fraumeni syndrome. Acta
G, von Deimling A, Weller M (2009) NOA-04 randomized

Phase III trial of sequential radiochemotherapy of anaplastic glioma with procarbazine, lomustine, and vincristine or
Zhao S, Lin Y, Xu W, Jiang W, Zha Z, Wang P, Yu W, Li Z, Gong
L, Peng Y, Ding J, Lei Q, Guan KL, Xiong Y (2009) Gliomaderived mutations in IDH1 dominantly inhibit IDH1 catalytic
activity and induce HIF-1alpha. Science 324:261265
Chapter 8
Metabolic Differences in Different Regions of Glioma

Samples
Francisca M. Santandreu, Jordi Oliver, and Pilar Roca
Abstract Gliomas are a heterogeneous disease both

under a clinical and a pathological point of view.
The differential expression of genotypic and metabolic
alterations presents a regional distribution within the
tumor mass. Even the coexistence of different subpopulations of cancer cells, differing in their sensitivity
to apoptosis, autophagy and chemotherapy, has been
proved in gliomas. Metabolic alterations are important
since they confer adaptive, proliferative, and survival
advantages on glioma cells. Thereby, metabolic reprogramming provides substrates for biosynthetic pathway, induces apoptosis resistance. The heterogeneous
distribution of oxygen within the tumor determines
metabolic differences between regions. Glioma cells
located at the periphery of the tumor have higher proliferative capacity, which is accompanied by a greater
respiration and mitochondrial oxidative capacity compared to cells located at the center of the tumor; such
differences could be related to differential degree of
hypoxia or oxidative stress between regions of the
tumor. Regional differences within glioma have diagnostic and therapeutic implications, and hinder the
prediction of the tumors biologic aggressiveness and
the patients response to standard treatment.
Keywords Metabolism Cancer Glioma
Antioxidant system Oncogenes Hypoxia
P. Roca ()
Department de Biologia Fonamental i Cincies de la Salut.
Ed. Guillem Colom, Universitat de les Illes Balears,
Carretera Valldemossa Km 7.5, Palma de Mallorca, 07122,
Balearic Islands, Spain
e-mail: pilar.roca@uib.es
Introduction
Gliomas are primary tumors of the central nervous
system. There are several classifications for gliomas
according to cellular type, grade, and location. The
World Health Organization subdivides astrocytomas in
four grades (IIV), according to the increasing malignancy determined by pathological evaluation of the
tumor. Low-grade gliomas are localized tumors, welldifferentiated and portend a better prognosis for the
patient. High-grade gliomas are anaplastic, and have
a high growth and invasive capacity; grade IV astrocytomas or glioblastomas constitute the most common
and agressive forms. Glioblastoma patients have a
median survival expectancy of only 14 months on
the current standard treatment of surgical resection to
the extent feasible, followed by adjuvant radiotherapy plus temozolomide, given concomitantly with and
after radiotherapy. Histological grade, tumor type, age,
Karnofsky performance status, tumor location and the
magnitude of surgical resection are prognostic factors
for gliomas (Hentschel and Sawaya, 2003; Lamborn
et al., 2004). Treatment for brain gliomas depends
on the location, the cellular type and the grade of
malignancy, and often, is a combined approach, using
surgery, radiation therapy, and chemotherapy. Highgrade gliomas almost always grow back even after
complete surgical excision due to their high tendency
to infiltrate (Laerum et al., 1984; Kaba and Kyritsis,
1997). The recurrence of tumor growth, which is the
major cause of mortality for patients with gliomas,
has been recently associated to the existence of a
heterogeneous population of cancer cells within the
tumor.
From a clinical point of view, gliomas are heterogeneous tumors. The heterogeneity also exists inside

63
64
the tumor mass since gliomas express metabolic and

histological differences between individual regions.
Metabolic alterations in gliomas have a great
importance because they promote tumors biologic
aggressiveness, stimulating cellular proliferation and
invasion. In this sense, a better knowledge of the heterogeneous nature of gliomas and their metabolism
will facilitate the development of new diagnostic and
therapeutic strategies to fight against the disease, overcoming the important limitations of standard therapy.
This chapter will describe the heterogeneous nature
of glioma, pointing out the presence of regional differences. For their implications on tumor growth and
maintenance, we will emphasize the main features of
metabolic reprogramming in tumor cells and, in particular, we will describe some metabolic differences
between individual regions within glioma tissue.
Human Gliomas: Tumors

of Heterogeneous Nature
Gliomas are a heterogeneous disease both under a clinical and a pathological point of view. Intratumoral
heterogeneity could be defined as a genotypic and/or
phenotypic variability within glioma, and it has been
recognized in karyotypic and histological analyses, by
a variable expression of proliferative markers, growth
factors and their receptors, and intrinsic resistance
to chemotherapy and radiation therapy (Coons and
Johnson, 1993). The differential expression of these
cellular features can be diffuse or present a regional
distribution. Regional heterogeneity implies that individual cells tend to be similar to their immediate
neighbors, whereas more distant cells exhibit different
features.
Histological analyses indicate that malignant
gliomas contain tumor cells with different degrees
of dedifferentiation, endothelial cells with variable
morphology and blood vessels that contain leukocytes
and erythrocytes. Regional differences in proliferative activity, histological features and genotypic
alterations have been identified in human gliomas.
However, intratumoral heterogeneity in ploidy is not
arbitrary, Coons and Johnson described the clustering
of regions with similar percentages of aneuploid
cells, hypothesizing that glioma progression might be
the result of local mutations and subsequent clonal
F.M. Santandreu et al.
expansion. Additionally, the degree of heterogeneity in proliferative markers as well as genotypic

heterogeneity increases with the tumor histological
grade; most importantly, the higher the intratumoral
heterogeneity in these markers of malignancy, the
higher the adverse impact in the usefulness of their
analysis to predict survival of glioma patients (Coons
and Johnson, 1993). Other studies using cell colonies
and human paraffin-embedded sections suggest a
heterogeneous spatial distribution of cancer cells
inside the tumor mass, where the most proliferative
cells are located at the periphery of the tumor (Bru
et al., 2003). Similarly, magnetic resonance imaging
studies evidence a structural heterogeneity within the
tumor, where different areas can be distingished from
the periphery, which is the most proliferative area of
the tumor, to the center. This structural heterogeneity
is also correlated with heterogeneity in vascularization
(Le Duc et al., 1999).
At an additional level of complexity, we must
consider that different subpopulations of cancer cells
coexist within the tumor mass. The presence of more
than one subclone in the primary tumor has been
related with tumor recurrence and intratumoral genotypic heterogeneity in gliomas. In relation to this, Ito
and colleagues analyzed one case of glioblastoma with
oligodendroglial components and determined using a
primary cell culture the possible relationship between
genetic modifications and chemosensitivity (Ito et al.,
2007). While the primary tumor exhibited 1p/19q/10q
losses, the recurrent glioblastoma only showed a loss
of heterozygosity (LOH) on chromosome 10q. In that
study, the chemosensitivity was related to cells with
1p/19q LOH, whereas cells exhibiting 10q LOH were
more resistant to chemotherapeutic agents. Shapiro
and Shapiro suggested that the karyotypic alterations
are associated to cancer cell phenotype. Specifically
that hyperdiploid cells, which are unstable in culture,
tend to have short doubling times and are usually sensitive to anticancer drugs such as 1,3-bis(2-chloroethyl)1-nitrosourea (BCNU); whereas, cancer cells that are
nearly diploid have a normal appearance, are stable in
culture, grow more slowly and are more likely to be
resistant to nitrosoureas; therefore, it was suggested
that the latter population could include stem tumor
cells which could be involved in the repopulation of
tumor mass (Shapiro and Shapiro, 1985).
More recently, the existence of a small fraction of
cells, exhibiting features of primitive neural progenitor
8 Metabolic Differences in Different Regions of Glioma Samples
cells, with capacity to induce tumor formation, has

been proved in gliomas. These quiescent cells with
self-renewal capacity are known as cancer stem cells,
which constitute a reservoir of self-sustaining cells
with the ability to self-renew and to cause the heterogeneous lineages of cancer cells that comprise the tumor.
In 2004, Singh and colleagues succesfully isolated cancer stem cells from different types of brain tumors.
Stem cell population was found exclusively in the fraction of tumor cells expressing CD133. The injection of
only one hundred CD133+ cells into the NOD-SCID
mouse brain led to the growth of a tumor with identical histological features as those displayed by the
parenteral tumor. Conversely, the fraction of CD133
tumor cells failed to form tumors, even when 1000-fold
more CD133 cells were injected into the brains of the
mice. It is important to consider that the existence of
tumor cell subpopulations could provide an explanation to the high rate of refractory malignant glioma. In
addition, conventional therapeutic approaches targeting the overall population of glioma cells may select
the more resistant tumor cells owing to their idiosyncratic properties. In this sense, several studies have
found that CD133+ population has an increased resistance to apoptosis (type I programmed cell death), and
autophagy (type II programmed cell death) induced
by the treatment with chemotherapeutic agents such
as temozolomide (Fu et al., 2009). Additionally, gene
expression studies in gliomas reveal that CD133+
tumor cells have an elevated expression of multi-drug
resistance gene BCRP1, DNA repair genes such as
O6-methylguanine-DNA methyltransferase (MGMT)
and genes that inhibit apoptosis (Liu et al., 2006; Kang
and Kang, 2007).
Interrelationship Between Energy

Metabolism and Hallmarks of Cancer
Tumors follow the same basic metabolic pathways as normal tissue, but changes in the tumors
microenvironment and the tumor cells themselves can
have important effects on their metabolism. There
are two interrelated factors that crucially influence
tumor metabolism: oxygen concentration and altered
metabolic pathways/enzymes. Hypoxia, a common
condition in tumors, determines the way in which
cellular energy is produced since the metabolism of
65
most energy substrates such as fatty acids, ketone bodies, amino acids (especially glutamine) and lactate is
oxygen dependent (Bouzier et al., 1998). However,
glucose can be used for energy production under
both normoxic and hypoxic conditions in tumor cells.
The presence of oxygen in gliomas, unlike in normal brain tissue, does not inhibit glycolysis (Warburg
effect). Additionally, in the presence of oxygen, glioma
cells preferentially metabolize the excess of glucose
by glycolysis, which provokes a decrease in tumor
respiration (Crabtree effect). In tumor cells, the
metabolic alterations such as enhanced glycolysis,
inhibited tricarboxylic acid (TCA) cycle and enhanced
lactate production would result in a net loss of carbon
that could have been used for anabolic reactions; however, cancer cells can avoid this by means of a much
higher net consumption of glucose than normal cells.
Which Advantages Confer Metabolic

Alterations to Tumor Cells?
Metabolic alterations promote tumor growth for several reasons (Hsu and Sabatini, 2008; Kroemer and
Pouyssegur, 2008). Firstly, conditions of fluctuating
oxygen tension, due to inconstant hemodynamics of
distant blood vessels, would be lethal for cells that
rely on oxidative phosphorylation (OXPHOS) to generate ATP. Secondly, lactate produced as end product
of aerobic glycolysis causes acidification of the extracellular milieu, supporting tumor invasion and the
suppression of anticancer immune effectors. Thirdly,
a major advantage to the cancer cell is that intermediates from the glycolytic pathway can be redirected
toward anabolic reactions linked to cellular growth and
proliferation. Figure 8.1 summarizes how glycolytic
intermadiates are used for the synthesis of macromolecules in the cancer cell. For example, cancer cells
enhance their biosynthetic capabilities by expressing a
tumor-specific form of pyruvate kinase, M2-PK, which
is less active in the conversion of phosphoenolpyruvate (PEP) to pyruvate and thus less efficient at ATP
production but facilitates the incorporation of glucose
carbons into lipids, amino acids and nucleic acids.
In this sense, glucose 6-phosphate becomes available
for the synthesis of glycogen and ribose 5-phosphate
and dihydroxyacetone phosphate for the synthesis of
triacylglycerides and phospholipids. Fourthly, tumor
66
Fig. 8.1 Metabolic reprogramming constitutes a major advantage to cancer cells since intermediates from the glycolytic
pathway can be redirected toward anabolic reactions linked
to cellular growth and proliferation. NADPH, nicotinamide
adenine dinucleotide phosphate; DHAP, dihydroxyacetone
phosphate; PEP, phosphoenolpyruvate; OXPHOS, oxidative
phosphorylation
cells can metabolize glucose through pentose phosphate pathway to generate nicotinamide adenine
dinucleotide phosphate (NADPH), which ensures the
antioxidant systems of cancer cells and facilitates their
resistance to chemotherapeutic agents. In summary,
actively dividing cells not only need great amounts
of ATP but also macromolecules such as nucleotides,
lipids and proteins, synthesis of which is facilitated by
metabolic reprogramming. Additionally, recent studies have shown that several steps in lipid synthesis are required for and may even actively promote
tumorigenesis.
The molecular mechanisms that underlie metabolic
reprogramming of cancer cells are complex. Tumor
microenvironment favors a specific metabolic profile.
Oxygen levels within a tumor fluctuate both temporally and spatially, and almost always are insufficient
to satisfy tumor cell growth, leading to hypoxia and the
stabilization of the hypoxia inducible factor 1 (HIF-1),
which initiates a transcriptional program that provides
multiple solutions to low oxygen availability by
decreasing the dependence on aerobic respiration. In
parallel, HIF-1 stimulates angiogenesis by upregulating several factors among which vascular endothelial
growth factor (VEGF) is included. When hypoxic
stress is detected by HIF-1 protein, cell metabolism is

shifted toward glycolysis by the increased expression
of inhibitors of mitochondrial metabolism, glucose
transporters and glycolytic enzymes such as lactate
dehydrogenase A (LDH-A). Lactate dehydrogenase is
the enzyme that catalyzes the last step of glycolysis.
This tetrameric protein is formed by subunits A
and B, encoded by two different genes, which are
differentially regulated. The inhibition of LDH-A
prevents the Warburg effect and forces cancer cells to
use oxidative phosphorylation to oxidize NADH and
produce ATP (Fantin et al., 2006). In this way, LDH-A
overexpression facilitates anaerobic metabolism of
glucose and conversion of pyruvate to lactate. During
tumorigenesis, a shift from A/B isoenzyme pattern
to the LDH-A pattern has been suggested. Besides it
has been described that tumor growth is attenuated
while cells depend on respiration to obtain energy,
suggesting that aerobic glycolysis might be essential
for cancer progression.
Metabolic alterations in the tumor are also promoted
by the activation of oncogenes and the loss of tumor
suppressors. The major oncogenic events (such as constitutive activation of growth factor pathways, constitutive activation of HIF-1, and inactivation of p53) could
constitute the common cause of metabolic reprogramming and hallmarks of cancer such as autonomous
growth, resistance against apoptosis, limitless replication, and angiogenesis (Kroemer and Pouyssegur,
2008). A diagram of the intimate relationship between
tumor metabolism and the main events of cellular
transformation is shown in Fig. 8.2. The activation
of oncogenes such as Myc and Akt is involved in the
Warburg effect, which inhibits oxidative phosphorylation, stimulates glycolysis and increases lactate
production. On the other hand, the loss of the tumor
suppressor protein p53 prevents SCO2 (synthesis of
cytochrome c oxidase 2) gene expression, which is
required for assembling cytochrome c oxidase, and
thus, interferes in the mitochondrial respiratory chains
function; it also prevents the activation of one of its
cellular effectors, TIGAR (TP53-induced glycolysis
and apoptosis regulator), which inhibits glycolysis by
decreasing the levels of fructose-2,6-bis-phosphate.
Regardless of whether the tumor microenvironment or oncogene activation plays a more important
role in driving the development of a distinct cancer
metabolism, it is likely that the resulting alterations
confer adaptive, proliferative, and survival advantages
67
Fig. 8.2 Tumor altered metabolism is caused by hypoxic stress

but also by oncogene activation and tumor suppressor loss.
In turn, metabolic reprogramming in cancer cells stimulates
tumor growth, apoptosis resistance, angiogenesis, avoidance of

immunosurveillance, tissue invasion and metastasis. OXPHOS,
oxidative phosphorylation; HIF-1, hypoxia inducible factor 1
on the cancer cell. Metabolic alterations provide

substrates for biosynthetic pathways, induce apoptosis
resistance, and stimulate signaling pathways involved
in the pro-oncogenic process.
Cancer cells are characterized for their ability to
support metabolic alterations, chronic hypoxia as well
as oxidative stress (Dang and Semenza, 1999). The
mitochondrion of tumor cell constitutes an important
point of control of cellular processes such as proliferation/apoptosis, oxidative metabolism and production
of reactive oxygen species (ROS). The mitochondria of glioma cells exhibit important functional and
ultrastructural alterations (Oudard et al., 1997), which
could be related with the high dependence of glucose
that glioma cells show. Mitochondrial DNA mutations
can appear as a result of tumor progression, in turn,
some of them could actively contribute to tumor progression due to concomitant stimulation of ROS production, tumor growth and aerobic glycolysis (Zhou
et al., 2007). Primary mitochondrial dysfunctions have
direct effects on tumor proliferation and metabolism;
mutations of the mitochondrial matrix protein fumarate
hydratase and the inner mitochondrial membrane
protein succinate dehydrogenase could have causal
implications in tumor development (Rustin, 2002).
Besides, these mutations induce HIF-1 due to the
accumulation of TCA cycle intermediates (fumarate
or succinate), which competitively inhibit the alphaketoglutarate-dependent HIF-1 prolyl hydroxylase,
the enzyme that usally targets HIF-1 for oxygendependent destruction. Thus, total o partial defects in
OXPHOS can lead to increased glycolysis and inherent
resistance to apoptosis.
Mitochondria are the main cellular source of ROS in
tumor cells. Reactive oxygen species are released as a
consequence of incomplete reduction of oxygen during
respiration, and it is a price that a cell working in the
presence of oxygen has to pay in favor of more efficient
bioenergetics. Endogenous production of ROS contributes to intracellular signaling transduction (Benhar
et al., 2002). It appears that signals linked to proliferation and survival need ROS for efficient transmission
to the nucleus. In this way, ROS act as second messengers stimulating cellular proliferation through the
activation of redox-sensitive signal transduction pathways; changes in redox state cause structural modifications and, in consequence, modulate the function of
cytosolic enzymes and transcription factors that control, among others, the gene expression of proteins that
participate in the protection against oxidative stress.
Among the proteins and signaling pathways sensitive
to redox changes, some signaling cascades play a crucial role in the regulation of gene expression and prevention of apoptosis such as Ras/Raf/MEK/ERK and
PTEN/PI3K/Akt/mTOR/NF-kappaB, components of
which are mutated or aberrantly expressed in glioblastomas. Oncogenes such as PI3K (phosphatidylinositol
3-kinase) and Akt have direct effects on cellular
68
metabolic deregulation (Kroemer and Pouyssegur,

2008). Protein kinase B or Akt is a downstream effector of insulin signaling that stimulates the uptake
and consumption of glucose through the glycolytic
pathway. The activation of PI3K causes the transcriptional downregulation of carnitine palmitoyltransferase
1A (CPT1A), an enzyme located in the outer mitochondrial membrane that esterifies long-chain fatty
acids to carnitine, thereby initiating the mitochondrial
import of fatty acids and channeling them to betaoxidation. Therefore, PI3K activation inhibits betaoxidation and contributes to the increased dependency
of cancer cells on glucose.
Metabolic Differences Between Regions

of Human Glioma
Malignant cells in a solid tumor are known to be
exposed to a heterogeneous environment, with tumor
regions characterized by a variable gradient of essential metabolites, nutrients, oxygen and growth factors.
Hypoxia in tumors results from an imbalance between
oxygen delivery and oxygen demand, and it is heterogeneous within the tumor mass. Oxygen consumption
within tumors varies and depends largely on the number of proliferative cells, since these cells consume the
most oxygen. A heterogeneous distribution of oxygen
within the tumor can be determined by multiple factors
including: variations in vascular density/orientation;
altered erythrocyte flux and function; longitudinal
oxygen gradients along vessels and radial perivascular oxygen gradients. Depending on the oxygenation
status, different areas are established within a solid
tumor. This oxygen gradient can influence the different metabolic behavior between regions. Essentially,
lactate concentrations are higher in hypoxic areas in
relation to the more oxygenated ones, thus supporting a switch from aerobic to anaerobic metabolism
concomitant with hypoxia. Lactate tumor concentrations are important since this metabolite is a prognostic
factor for several human cancers: high lactate levels
increase the likelihood of metastases and have a negative impact on disease free survival and overall survival
(Walenta et al., 2000). Hypoxia is specially relevant
in human gliomas since it influences the tumors clinical behavior, increasing its biologic aggressiveness and
decreasing its response to therapy (Evans et al., 2004).
Hypoxic cells tend to be quiescent and chemotherapy

agents are more effective in actively dividing cells.
Hypoxia also decreases the response of tumors to
radiotherapy.
Concerning the intratumoral metabolic differences,
it has been postulated that depending on the oxygenation status inside the tumor, a symbiotic relationship
between two cancer cell populations could exist. In
2008, Sonveaux and colleagues described a metabolic
symbiosis between hypoxic and aerobic cancer cells:
cancer cells under hypoxic conditions convert glucose to lactate and extrude it, whereas aerobic cells
take up lactate and utilize it for oxidative phosphorylation. In the case of limited glucose availability for
the tumor, this symbiosis would facilitate a more efficient use of glucose. At the molecular level, it has
been described that a key component of this symbiotic
relationship could be the monocarboxylate transporter
1 (MCT1), this isoform is repressed by hypoxia and
transports lactate into, rather than out of, cancer cells.
Oxygen-dependent expression of MCT1 allow aerobic cancer cells to efficiently take up and, in concert
with the oxygen-dependent expression of lactate dehydrogenase B (LDH-B), utilize lactate as an energy
substrate, thereby freeing these cells from the need
to take up large quantities of glucose. Interestingly,
a similar hypothesis was proposed three years before
by Griguer and colleagues (2005), on the basis of
their observations of glycolytic metabolism in cultured glioma cells, suggesting that lactate produced
by anaerobic glycolisis within the hypoxic region of
glioma could also be used as an oxidative substrate
in another region where oxygen supply was higher.
These authors proved that glioma cells from different established cell lines showed a heterogeneous
glucose metabolism. In this sense, there are highly glycolytic glioma cells whereas others preferentially use
OXPHOS to obtain energy. The use of OXPHOS to
obtain energy enhances cellular survival in the presence of low glucose concentrations, and this metabolic
phenotype is related to the expression of lactate dehydrogenase A and B isoenzymes. In glioma cells, the
ratio of LDH reductive/oxidative activity is about 7
for the A form and 1 for the B form, suggesting
that lactate oxidation is more favorable for cells that
express LDH-B. Whereas the expression of LDH-B
isoform is independent of extracellular glucose concentration in glioma cells, the expression of LDH-A
isoform is highly glucose-dependent. Thereby, LDH-A
expression is lost in the presence of low glucose

concentrations.
Taking the aforementioned considerations into
account, we characterized the possible differences
in energy metabolism and ROS metabolism between
the central and peripheral regions of human glioma
(Santandreu et al., 2008). In our study, central region
refers to the tumor core, where biopsy sampling
is usually performed, whereas peripheral region
refers to a region closer to surgical resection margins.
Histological analyses showed that each region contained similar cellularity and number of tumor cells
and, additionally, that the presence of residual nontumoral tissue was minimal and equal in both regions.
Mitochondrial abundance between regions was similar,
and the biochemical composition of isolated mitochondria was equal in terms of protein content and
mitochondrial DNA copy number.
Immunohistochemical analysis using Ki-67 proliferation marker revealed that the highest number
of proliferative cells was located at the periphery
of the glioma mass. In contrast with the apparent
mitochondrial structural homogeneity between glioma
regions, mitochondria from the periphery of the tumor
showed a significantly higher resting respiration (nonphosphorylating conditions) than mitochondria from
the center of the tumor. Both the drop in oxygen consumption and cytochrome c oxidase activity reflected
a lower mitochondrial oxidative capacity in the center as opposed to the periphery of the tumor. It is
important to mention that the decline in mitochondrial
oxidative capacity in the center of the tumor was not
caused by variations in mitochondrial abundance, in
mitochondrial DNA copy number or in mitochondrial
DNA lesions between regions. Therefore, metabolic
differences between the center and the periphery of the
tumor could be better explained by intratumoral variations in the degree of oxidative stress or hypoxia. In our
study an important trend toward a greater lipid oxidation was identified in the center of the tumor, which
was accompanied by a significant adaptive increase
in mitochondrial enzymatic antioxidant systems (manganese superoxide dismutase or MnSOD) and tissular
non-enzymatic antioxidant systems (an increase in the
levels of reduced glutathione and increased relationship between reduced vs oxidized glutathione).
During resting respiration the maximum hydrogen peroxide (H2 O2 ) production capacity tends to be
higher in isolated mitochondria from the periphery of
69
the tumor in relation to those isolated from the center

of the tumor, and it is consistent with the notion of an
increased mitochondrial functionality in the peripheral
region of glioma. The capacity to produce ROS does
not reflect the intracellular levels of ROS to which cells
are exposed. In contrast, the observations of a trend
toward a higher oxidative damage, greater respiratory
dysfunction and the adaptive increase in antioxidant
systems in the center of the tumor, appear to indicate a
greater degree of chronic oxidative stress in this region.
Considering that cancer cells located at the periphery
of the tumor have higher proliferative activity, there is
a higher energy demand in this region. In this situation,
it is assumed that oxidative stress is low because cells
display a respiratory control in which respiration is
activated by ADP, and this stimulation stops after the
conversion of ADP to ATP.
In summary, recent studies support the concept of
metabolic heterogeneity within glioma. This metabolic
heterogeneity appears not to be random but to be
ascribed to defined regions in the spatial structure
of the tumor. In this sense, differences in mitochondrial oxidative capacity exist between the central and
peripheral region of glioma, and are associated with a
differential proliferative capacity; such metabolic differences might be caused by a heterogeneous distribution of oxygen within the tumor. Figure 8.3 represents
a hypothetic diagram of metabolic differences between
individual regions of glioma and the possible factors
involved.
Possible Therapeutic and Diagnostic

Implications
In this chapter, we have tried to illustrate, in a simplified form, the relevance of cancer cell metabolism,
not only for anabolic pathways associated with tumor
growth but also for the existence of intimate links
between the main oncogenes and metabolic reprogramming in tumors. It should be noted that the
knowledge of the high glycolytic rate in tumor cells
has led to the development of highly sensitive and
specific diagnostic tests for cancer using the glucose
analog 2-deoxy-2-[18 F]fluoro-D-glucose by positron
emission tomography (FDG-PET). The diagnosis of
gliomas using this imaging technique is based on the
70
Fig. 8.3 Possible metabolic differences between the central and

peripheral region of human glioma. Mitochondria from the center of the tumor show a decline in respiration compared to those
from the periphery of the tumor which it is probably caused by
a higher degree of oxidative and hypoxic stress. Cancer cells
located at the periphery of the tumor have higher proliferative
activity, which is coupled to a better mitochondrial oxidative

capacity and better maintenance of redox homeostasis, compared with cancer cells located at the center of the tumor. In
the graphic, the width of the arrow indicates the intensity of
the process. OXPHOS, oxidative phosphorylation; ROS, reactive
oxygen species; HIF-1, hypoxia inducible factor 1
preferential incorporation and storage of fludeoxyglucose F18 in tumor tissue compared to healthy tissue.
Several studies performed with gliomas point out
that differences in energy metabolism between normal cells and cancer cells, could be used as a biochemical basis to develop new therapeutic strategies
selectively targeted against cancer cells (Seyfried and
Mukherjee, 2005; Nebeling et al., 1995). The inhibition of enzymes and pathways that participate in
metabolic reprogramming could have a stong effect
on tumor growth, not only limiting bioenergetic flux
and anabolic reactions in cancer cells but also reverting the neoplastic phenotype by inducing apoptosis
or by blocking invasion and angiogenesis. In other
words, interventions such as the inhibition of the
PI3K/Akt/mTOR pathway (that would inhibit tumor
growth), reestablishment of p53 function (that would
restore apoptosis and senescence) or inhibition of
transcription factor HIF-1 (that would inhibit angiogenesis), would also normalize metabolic functions
in gliomas. Preclinical and clinical evaluation of
metabolic inhibitors is still in its early stages, with
the exception of mTOR antagonists. The mammalian
target of rapamycin (mTOR) is a serine/threonine
kinase that enhances cellular growth while inhibiting catabolic reactions mediated by autophagy (Faivre
et al., 2006). Although the lack of enzymatic inhibitors
with an acceptable degree of specificity is one of the
main obstacles at this time, the discovery of selective metabolic inhibitors for anabolic and bioenergetic
pathways in gliomas will provide, alone or in combination therapy, a completely new arsenal with which to
combat central nervous system tumors.
Intratumoral heterogeneity in gliomas is also important because of its relevance in several clinical aspects,
notably, histological grading, patients therapeutic
response and refractory disease. Since variability in
specific histological features (cellular density, necrosis, cytologic and nuclear pleomorfism, mitotic activity and microvascular proliferation) produces different
histological grades between individual regions within
glioma (Coons and Johnson, 1993), it is likely that
errors in histological classification induced by regional
heterogeneity could be minimized avoiding limited
biopsies (localized sample site(s) and small sample
size).
Recent studies support the idea that malignant
gliomas might be considered as microecosystems
where tumor cells, microenvironment, vasculature and
cancer stem cells are all interrelated. Thus, the most
malignant cells could be selected under adverse conditions such as the pressure from the immune system, radiotherapy or chemotherapy. The increase in
antioxidant defense systems in glioma cells has been
associated with a higher resistance to radiotherapy
and chemotherapy (Lee et al., 2004). If one accepts
that metabolic differences and differences in antioxidant capacity do exist between individual regions
of glioma, then it may follow that standard therapy
could retard tumor growth in the short term but might
facilitate recurrence in the long term by means of a
selective pressure in vivo, which would promote the
survival of the more resistant cellular population. In
most clinical trials, in which monoclonal antibodies
or low molecular-weight kinase inhibitors have been
used to control the dysregulation in glioma growth,
monotherapies have failed to show a survival benefit for patients. If we accept the coexistence of more
than one subtype of cancer cells within glioma, it is
likely that combination therapy might be the more efficient strategy to eliminate the distinct subpopulations
contained in the tumor mass. Additionally, functional
imaging techniques for diagnosis such as FDG-PET
could help to assess in vivo metabolic heterogeneity in human gliomas (Goldman et al., 1997). In all
likelyhood, a better understanding of some aspects of
tumor cellular biology such as metabolic heterogeneity could clarify the problem of why glioma patients
diagnosed with the same histological grade have different evolutions and respond differentially to standard
treatment.
Acknowledgements We gratefully acknowledge the support
from the Conselleria dEconomia, Hisenda i Innovaci del
Govern de les Illes Balears and the Fondo de Investigaciones
Sanitarias del Ministerio de Sanidad y Consumo del Gobierno
Espaol (PI060266).
References
Benhar M, Engelberg D, Levitzki A (2002) ROS, stress-activated
kinases and stress signaling in cancer. EMBO Rep 3:
420425
Bouzier AK, Voisin P, Goodwin R, Canioni P, Merle M (1998)
Glucose and lactate metabolism in C6 glioma cells: evidence
for the preferential utilization of lactate for cell oxidative
metabolism. Dev Neurosci 20:331338
Bru A, Albertos S, Luis Subiza J, Garcia-Asenjo JL, Bru I
(2003) The universal dynamics of tumor growth. Biophys J
85:29482961
Coons SW, Johnson PC (1993) Regional heterogeneity in the
DNA content of human gliomas. Cancer 72:30523060
Dang CV, Semenza GL (1999) Oncogenic alterations of
metabolism. Trends Biochem Sci 24:6872
Evans SM, Judy KD, Dunphy I, Jenkins WT, Hwang WT, Nelson
PT, Lustig RA, Jenkins K, Magarelli DP, Hahn SM, Collins
RA, Grady MS, Koch CJ (2004) Hypoxia is important in the
biology and aggression of human glial brain tumors. Clin
Cancer Res 10:81778184
Faivre S, Kroemer G, Raymond E (2006) Current development
of mTOR inhibitors as anticancer agents. Nat Rev Drug
Discov 5:671688
Fantin VR, St-Pierre J, Leder P (2006) Attenuation of LDH-A
expression uncovers a link between glycolysis, mitochondrial physiology, and tumor maintenance. Cancer Cell
9:425434
Fu J, Liu ZG, Liu XM, Chen FR, Shi HL, Pangjesse CS,
Ng HK, Chen ZP (2009) Glioblastoma stem cells resistant to temozolomide-induced autophagy. Chin Med J 122:
12551259
71
Goldman S, Levivier M, Pirotte B, Brucher JM, Wikler D,

Damhaut P, Dethy S, Brotchi J, Hildebrand J (1997) Regional
methionine and glucose uptake in high-grade gliomas: a
comparative study on PET-guided stereotactic biopsy. J Nucl
Med 38:14591462
Griguer CE, Oliva CR, Gillespie GY (2005) Glucose metabolism
heterogeneity in human and mouse malignant glioma cell
lines. J Neurooncol 74:123133
Hentschel SJ, Sawaya R (2003) Optimizing outcomes with maximal surgical resection of malignant gliomas. Cancer Control
10(2):109114
Hsu PP, Sabatini DM (2008) Cancer cell metabolism: warburg
and beyond. Cell 134:703707
Ito M, Wakabayashi T, Natsume A, Hatano H, Fujii M,
Yoshida J (2007) Genetically heterogeneous glioblastoma
recurring with disappearance of 1p/19q losses: case report.
Neurosurgery 61:E168E169
Kaba SE, Kyritsis AP (1997) Recognition and management of
gliomas. Drugs 53:235244
Kang MK, Kang SK (2007) Tumorigenesis of chemotherapeutic
drug-resistant cancer stem-like cells in brain glioma. Stem
Cells Dev 16:837847
Kroemer G, Pouyssegur J (2008) Tumor cell metabolism: cancers Achilles heel. Cancer Cell 13:472482
Laerum OD, Bjerkvig R, Steinsvag SK, de Ridder L (1984)
Invasiveness of primary brain tumors. Cancer Metastasis Rev
3:223236
Lamborn KR, Chang SM, Prados MD (2004) Prognostic factors for survival of patients with glioblastoma: recursive
partitioning analysis. Neuro Oncol 6:227235
Le Duc G, Peoch M, Remy C, Charpy O, Muller RN, Le Bas
JF, Decorps M (1999) Use of T(2)-weighted susceptibility
contrast MRI for mapping the blood volume in the gliomabearing rat brain. Magn Reson Med 42:754761
Lee HC, Kim DW, Jung KY, Park IC, Park MJ, Kim MS, Woo
SH, Rhee CH, Yoo H, Lee SH, Hong SI (2004) Increased
expression of antioxidant enzymes in radioresistant variant
from U251 human glioblastoma cell line. Int J Mol Med
13:883887
Nebeling LC, Miraldi F, Shurin SB, Lerner E (1995) Effects of
a ketogenic diet on tumor metabolism and nutritional status
in pediatric oncology patients: two case reports. J Am Coll
Nutr 14:202208
Oudard S, Boitier E, Miccoli L, Rousset S, Dutrillaux B, Poupon
MF (1997) Gliomas are driven by glycolysis: putative roles
of hexokinase, oxidative phosphorylation and mitochondrial
ultrastructure. Anticancer Res 17:19031911
Rustin P (2002) Mitochondria, from cell death to proliferation.
Nat Genet 30:352353
Santandreu FM, Brell M, Gene AH, Guevara R, Oliver J, Couce
ME, Roca P (2008) Differences in mitochondrial function
and antioxidant systems between regions of human glioma.
Cell Physiol Biochem 22:757768
Seyfried TN, Mukherjee P (2005) Targeting energy metabolism
in brain cancer: review and hypothesis. Nutr Metab 2:30
Shapiro JR, Shapiro WR (1985) The subpopulations and isolated cell types of freshly resected high grade human
72
gliomas: their influence on the tumors evolution in vivo
and behavior and therapy in vitro. Cancer Metastasis Rev
4:107124
Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J,
Hide T, Henkelman RM, Cusimano MD, Dirks PB (2004)
Identification of human brain tumour initiating cells. Nature
432:396401
Sonveaux P, Vegran F, Schroeder T, Wergin MC, Verrax J,
Rabbani ZN, De Saedeleer CJ, Kennedy KM, Diepart C,
Jordan BF, Kelley MJ, Gallez B, Wahl ML, Feron O,
Dewhirst MW (2008) Targeting lactate-fueled respiration
selectively kills hypoxic tumor cells in mice. J Clin Invest
118:39303942

Walenta S, Wetterling M, Lehrke M, Schwickert G, Sundfor
K, Rofstad EK, Mueller-Klieser W (2000) High lactate
levels predict likelihood of metastases, tumor recurrence, and
restricted patient survival in human cervical cancers. Cancer
Res 60:916921
Zhou S, Kachhap S, Sun W, Wu G, Chuang A, Poeta L,
Grumbine L, Mithani SK, Chatterjee A, Koch W, Westra
WH, Maitra A, Glazer C, Carducci M, Sidransky D, McFate
T, Verma A, Califano JA (2007) Frequency and phenotypic
implications of mitochondrial DNA mutations in human
squamous cell cancers of the head and neck. Proc Natl Acad
Sci USA 104:75407545
Chapter 9
Glioblastoma Patients: Role of Methylated MGMT

Giulio Metro and Alessandra Fabi
Abstract The O6 -methylguanine-DNA methyltransferase (MGMT) protein is a DNA repair enzyme that
antagonizes the anti-tumor effects of alkylating agents,
particularly temozolomide. Consistent with this mechanism of action, MGMT silencing by gene promoter
methylation has been shown to be both predictive
and prognostic in clinical trials of newly diagnosed
glioblastoma patients treated with temozolomide in
combination with radiotherapy and as adjuvant treatment. However, assessment of methylation of the
MGMT gene promoter still requires standardization
and prospective validation in clinical trials in order to
best define the role of this biomarker for individualization of treatment. Nevertheless, even patients with
temozolomide-sensitive glioblastoma cannot avoid
eventual recurrence. Moreover, the optimal treatment
strategy for patients with tumors lacking methylation
of the MGMT gene promoter has yet to be determined.
Here, we discuss the predictive and prognostic value of
MGMT silencing, focusing on the importance of standardizing MGMT assessment as a crucial prerequisite
for achieving personalization of treatment according to
MGMT status in the next future.
Keywords Glioblastoma MGMT Methylation
Alkylating agents Silencing Pseudoprogression
Introduction
The
O6 -methylguanine-DNA
methyltransferase
(MGMT) protein is a DNA repair enzyme which is
A. Fabi ()
Division of Medical Oncology, Regina Elena National Cancer
Center Institute, 00144 Rome, Italy
e-mail: alessandra.fabi@virgilio.com
ubiquitously expressed in normal tissues (Gerson,

2004). Its high conservation throughout evolution suggests that MGMT plays a crucial role in maintaining
cell physiology and genome integrity (Gerson, 2004).
Although the levels of MGMT expression may vary
considerably among different organs, tumors usually
exhibit higher levels of expression compared with
their tissue of origin (Gerson, 2004). Importantly,
MGMT has been implicated in resistance to anticancer therapy with alkylating agents, particularly
temozolomide. In fact, MGMT acts by antagonizing
temozolomide-induced DNA methylation at the
O6 position of guanine by transferring the methyl
group to the active site of the enzyme itself. If left
unrepaired, O6 -methylguanine undergoes incorrect
pairing with thymine instead of cytosine during DNA
replication, eventually leading to cell death (Gerson,
2004). Consistent with this mechanism of action,
low levels/loss of function of MGMT have been
shown to be associated with increased sensitivity to
temozolomide treatment in malignant gliomas (Chinot
et al., 2007; Friedman et al., 1998; Hegi et al., 2004,
2005). Importantly, loss of function of MGMT in
glioblastoma multiforme (GBM) can be attributed
almost exclusively to hypermethylation of the MGMT
gene promoter, a phenomenon known to determine
epigenetic silencing of the MGMT corresponding
protein (Weller et al., 2010). Although the enzyme
activity of MGMT appears to be the most important
mechanism underlying resistance to temozolomide
and alkylating agents in general, other mechanisms
may exist such as disturbances of the mismatch repair
system and increased expression of the chromatinassociated gene poly(ADP-ribose) polymerase-1
(PARP-1), which is involved in nucleotide excision
repair system (Curtin et al., 2004; Yip et al., 2009).

73
74
G. Metro and A. Fabi
Assessing the Status of MGMT in Tumor

Tissue
MGMT Methylation in Newly Diagnosed

Glioblastoma: Predictive Role
The potential clinical utility of MGMT as a biomarker

in malignant gliomas has led to an ongoing debate
on how MGMT status should be assessed and which
specific procedure is best suited for routine clinical
applications. A test aimed at assessing MGMT status
needs to be standardized and reproducible in different laboratories, and must have a clinically relevant
cut-off point. This, in order to allow patient selection and individualize therapy. Immunohistochemistry
(IHC) detects the levels of the MGMT protein in tumor
tissue, while the enzymatic activity of MGMT can be
measured by high-performance liquid chromatography
(HPLC) (Weller et al., 2010). By contrast, epigenetic
silencing of the MGMT gene by promoter methylation can be assessed using methylation-specific PCR
(MSP) (Weller et al., 2010). The MSP assay can be
performed on DNA extracted from paraffin-embedded
tissue samples despite the fact that fixation may cause
deterioration of the DNA so that the best results are
usually obtained on frozen tissue. Currently, the gelbased MSP assay represents the most widely used
technology and is the only test that has been shown
to be both predictive and prognostic in clinical trials of GBM patients treated with temozolomide (Hegi
et al., 2004, 2005; Stupp et al., 2009). Recently, a
real-time PCR MSP assay has been developed for
quantitative assessment of the MGMT methylation status (qMSP), proving more amenable to definition of
technical cut-off and quality control compared with the
qualitative gel-based MSP assay (Vlassenbroeck et al.,
2008). Real-time PCR MSP assay is being prospectively validated in the phase III randomized CENTRIC
trial (Clinicaltrials.gov NTC00689221) which investigates cilengitide, a new integrin inhibitor, in addition to temozolomide plus radiotherapy for patients
with newly diagnosed GBM with methylation of the
MGMT gene promoter (Stupp et al., 2010). Given
the large variability of MGMT methylation (3060%)
reported in the literature for GBM (Weller et al.,
2010), prospective validation of cut-offs for optimal
MGMT assessment is crucial for future individualization of treatment according to MGMT methylation
status.
The first observation on a potential predictive role of

MGMT in malignant glioma patients was made more
than 10 years ago (Belanich et al., 1996). Patients
with low protein levels of MGMT as assessed by
immunofluorescence microscopy showed greater benefit from carmustine (BCNU) compared with patients
with high MGMT levels (Belanich et al., 1996).
Similarly, low protein levels of MGMT as detected
by IHC were found to be predictive of prolonged
progression-free survival in GBM patients treated with
first-line temozolomide (Friedman et al., 1998) as
well as of increased survival in inoperable newly
diagnosed GBM treated with neoadjuvant temozolomide (Chinot et al., 2007). Importantly, a correlation
with survival was also demonstrated when malignant glioma patients harbouring methylation of the
MGMT gene promoter were treated with nitrosoureas
(Esteller et al., 2000) or temozolomide plus radiotherapy (Hegi et al., 2004), strongly suggesting that
MGMT gene silencing through methylation of the
MGMT gene promoter as assessed by gel-based MSP
assay could provide both a predictive and prognostic
biomarker for increased sensitivity to alkylating agent
chemotherapy with or without radiotherapy. This was
subsequently confirmed in the randomized European
Organization for Research and Treatment of Cancer
(EORTC) 26981-22981 National Cancer Institute of
Canada (NCIC) trial, where surgically treated patients
for newly diagnosed GBM were randomized to radiotherapy or radiotherapy plus concomitant and adjuvant
temozolomide (Stupp et al., 2005). In this trial, consistent with the idea that methylation could predict
the benefits of alkylating agent chemotherapy, methylation of the MGMT gene promoter was found to
be associated with a strikingly longer progressionfree survival in patients treated with temozolomide
plus radiotherapy compared with methylated patients
treated with radiotherapy alone (10.3 months versus
5.9 months, respectively, P = 0.001) (Hegi et al.,
2005). By contrast, in unmethylated patients only a
slight trend was observed towards longer progressionfree survival with temozolomide plus radiotherapy versus radiotherapy alone (5.3 months versus 4.4 months,
respectively, P = 0.02), suggesting little or no benefit

from the addition of temozolomide to radiotherapy in
the unmethylated group (Hegi et al., 2005). The modest
effect of temozolomide in patients lacking methylation
of the MGMT gene promoter has raised an ongoing
discussion as to whether assessment of the MGMT
status should be made mandatory, and whether temozolomide should be avoided in patients with tumors
lacking methylation of the MGMT gene promoter. To
this regard, it is important to note that the qualitative gel-based MSP assay used in the EORTC-NCIC
trial divided patients into methylated and unmethylated
(45% versus 55%, respectively) (Stupp et al., 2005).
On the other hand, quantitative assays (e.g. qMSP)
suggest the presence of a subgroup of patients with
intermediate methylation who may still benefit from
temozolomide, thus representing the gray zone in
the test results. For this reason, at the present time
MGMT testing should be considered more informative rather than decision-making in routine clinical
practice, and all patients with newly diagnosed GBM
should be treated with standard temozolomide plus
radiotherapy followed by adjuvant temozolomide irrespective of the methylation status of the MGMT gene
promoter.
MGMT Methylation in Newly Diagnosed

Glioblastoma: Prognostic Role
Besides being predictive of sensitivity to alkylating agent chemotherapy in newly diagnosed GBM,
the methylation status of the MGMT gene promoter
appears to be also a strong prognostic factor for outcome (Hegi et al., 2004, 2005; Stupp et al., 2009).
In the EORTC-NCIC trial methylation of the MGMT
gene promoter was found to be significantly associated
with longer survival irrespective of treatment compared with absence of methylation (18.2 months versus
12.2 months, respectively, P = 0.001), which corresponds to a risk reduction of 55% (hazard ratio for
death = 0.45; 95% confidence interval, 0.320.61).
In an updated analysis, the 2 and 5 year survival
rates in patients with methylated MGMT were 49 and
14%, respectively, while the corresponding figures for
75
methylated patients treated initially with radiotherapy

alone were 24 and 5%. On the other hand, patients
with an unmethylated MGMT promoter treated with
temozolomide plus radiotherapy had a 2 and 5 year
survival of 15 and 8%, respectively, compared with
2 and 0% of the unmethylated patients who received
radiotherapy alone (Stupp et al., 2009). The small
improvement in survival observed in unmethylated
with the addition of temozolomide to radiotherapy
could be attributed to the somewhat arbritary separation of patients into methylated and unmethylated (the
so-called gray zone) as well as to other situations
such as differences in post-progression therapy and
individual variability due to other unknown prognostic
factors.
Recently, the prognostic relevance of methylation
of the MGMT gene promoter has been confirmed in
elderly GBM patients treated with concomitant and
adjuvant temozolomide (Brandes et al., 2009). In addition, the observation that 74% of patients with GBM
who survive longer than 5 years have MGMT promoter methylation, as opposed to less than 50% in
an unselected population of GBM patients, underlines again the prognostic value of MGMT methylation
(Krex et al., 2007). Nevertheless, long-term survival
can be observed even in the absence of methylation of
the MGMT gene promoter, indicating that MGMT is
only one aspect of a more complex biological system
(Krex et al., 2007).
The Informative Role of MGMT

Methylation in Newly Diagnosed
Glioblastoma
Recent evidence suggests that the importance of knowing the methylation status of the MGMT gene promoter may go beyond its predictive and prognostic
role. In fact, MGMT status could help predict the
occurrence of a phenomenon known as pseudoprogression (Brandes et al., 2008). Pseudoprogression
occurs in up to 30% of GBM patients treated with
radiotherapy or concomitant temozolomide and radiotherapy, and consists of a radiologically-detected
increase in lesion (with or without worsening of
76
neurological symptoms) which subsequently subsides

or even decreases in size without further treatment
(Brandes et al., 2008; Taal et al., 2008). That is because
pseudoprogression likely reflects a treatment-related
local tissue reaction with inflammation, oedema and
increased abnormal vessel permeability rather than
true tumor progression. In a series of 103 glioblastoma patients treated with concomitant temozolomide
and radiotherapy, lesion enlargement at MRI occurred
in 48.5% of patients, of whom 64% were later on
considered as having pseudoprogression (31% of the
total) (Brandes et al., 2008). Importantly, in this study
MGMT methylated tumors were found to be significantly associated with pseudoprogression, as within
the group showing tumor enlargement 91% of the
methylated patients had pseudoprogression compared
with only 41% of the unmethylated GBMs (P =
0.0002) (Brandes et al., 2008). Considering the diagnosis of pseudoprogression and knowing that it occurs
more often in MGMT methylated patients is crucial
for planning optimal treatment strategies and designing
clinical trials with novel agents.
MGMT Methylation for Predicting

Clinical Outcome of Recurrent
Glioblastoma
In contrast to newly diagnosed GBM, the predictive
value of MGMT methylation has been questioned in
recurrent glioblastomas. Changes in MGMT methylations at recurrence occur in approximately 40% of
patients and appear to be more frequent in patients
with methylated tumors at first surgery (Brandes et al.,
2010). Importantly, the absence of a strong predictive effect in recurrent GBM with the use of various
schedules of administration of temozolomide (Brandes
et al., 2006; Wick et al., 2007) might suggest the
predominance of MGMT-independent mechanisms of
resistance in the setting of recurrent disease. Escape
from MGMT methylation-mediated sensitivity to the
alkylating drug by selection for mismatch repair deficiency might be involved in this loss of prediction
(Yip et al., 2009).
Few studies investigated the relationship between
MGMT methylation and sensitivity to drugs other than
temozolomide in recurrent glioblastoma. In 19 patients
with recurrent malignant gliomas a study showed
that single-agent treatment with a nitrosurea drug,
namely fotemustine, was associated with a considerable rate of disease control in patients with methylated
MGMT (66.5%) compared with those with unmethylated MGMT (0%) (Fabi et al., 2009). Despite the low
number of patients analyzed and the heterogeneity of
patients histotypes included in this study, these data
suggest that the presence of MGMT promoter methylation could predict response to fotemustine. However,
assessment of MGMT status was performed at the time
of initial surgery and no re-assessment of MGMT was
conducted at the time of tumor recurrence. This might
be important, since changes in the status of MGMT
promoter methylation may occur after primary treatment for newly diagnosed glioblastoma (Brandes et al.,
2010).
MGMT Methylation in Grade III

Anaplastic Gliomas: Only Prognostic?
The possible predictive value of methylation of the
MGMT gene promoter in GBM patients has recently
been challenged in grade III anaplastic gliomas. In fact,
in the NOA-04 trial a significantly longer progressionfree survival and overall survival were shown for
anaplastic glioma with MGMT methylation irrespective of initial treatment with radiotherapy or alkylatingbased chemotherapy with temozolomide or PCV (procarbazine, lomustine and vincristine) (Wick et al.,
2009). Similarly, in the EORTC trial 26951 testing
radiotherapy alone versus radiotherapy followed by
adjuvant PCV, progression-free and overall survival
were found to be significantly longer in patients whose
tumors showed methylation of the MGMT gene promoter irrespective of the administration of chemotherapy (van den Bent et al., 2009). These findings suggest that methylation of the MGMT gene promoter is
only prognostic but not predictive for chemotherapy
in anaplastic gliomas, thus hypothesizing that other
markers might be involved in sensitivity to chemotherapy in anaplastic gliomas. Also, these observation
underline the importance of studying grade III and IV
gliomas as two distinct entities.
MGMT-Depleting Strategies
Alternative, more protracted dosing regimens of temozolomide might have a role in increasing sensitivity to
treatment through optimal depletion of MGMT activity

(Weller et al., 2010). Indirect comparisons of treatment
efficacy in recurrent disease suggest that the use doseintense schedules of temozolomide might be superior
to conventional temozolomide 150200 mg/m2 administered 5 days every 4 weeks (Brandes et al., 2006;
Perry et al., 2010; Wick et al., 2007; Yung et al., 1999).
However, whether this presumed superiority is to be
attributed to enhanced MGMT depletion is not known.
Unfortunately, in a recent trial randomizing chemonave patients with recurrent anaplastic astrocytoma or
GBM to conventional temozolomide or continuous
temozolomide 100 mg/m2 for 21 days every 4 weeks,
an inferior outcome was observed for the continuous dose-intense schedule (Brada et al. 2010). At the
present time, studies are underway to test dose-intense
temozolomide as upfront adjuvant treatment for newly
diagnosed GBM following concomitant temozolomide
plus radiotherapy.
Several MGMT inhibitors are under active investigation with the goal of modulating resistance to alkylating agent chemotherapy. O6-benzylguanine (O6BG) is a potent MGMT-depleting agent that when used
in combination with temozolomide has the potential of
restoring temozolomide sensitivity in recurrent malignant gliomas (Quinn et al., 2009). However, O6-BG
depletes MGMT nonselectively, thus lowering MGMT
levels in normal cells as well, therefore resulting
in substantial toxicity, particularly myelosuppression.
Similarly to O6-BG, Lomeguatrib, O4-benzylfolates
and 2-amino-O4-benzylpteridine are other MGMTdepleting agents that are being investigated in combination with temozolomide or BCNU in recurrent
disease progressive to alkylating agent chemotherapy.
77
a prospective manner, and it is presently not recommended to use the MGMT promoter methylation
assay to determine who should receive temozolomide
and who should not. On the other hand, the knowledge so far gained about MGMT status should be
exploited to design thoughtful clinical studies aimed
at improving the overall outcome of glioblastoma
patients, eventually leading to individualization of
treatment. For instance, an interesting approach would
be that of trying to overcome resistance to alkylating agent chemotherapy in patients with unmethylated MGMT gene promoter. On this basis, a recent
trial has closed accrual after enrolling 1153 patients
with the aim of showing whether temozolomide
dose-intensification (temozolomide for 21 days every
4 weeks) after chemo-radiotherapy would improve
outcome of patients with newly diagnosed glioblastoma (Clinicaltrials.gov NCT00304031). In this study,
assessment of MGMT status by qMSP assay has been
made mandatory for trial inclusion and used as a stratification factor for assigning patients to the control
and experimental arms. Alternative strategies may also
be useful in unmethylated patients for enhancing the
antitumor effect of radiotherapy during the concomitant phase of treatment (Metro et al., 2010). To this
regard, given the synergistic activity shown by enzastaurin, a protein kinase C-beta inhibitor, in conjunction
with radiotherapy (Tabatai et al., 2007), a phase II
trial is evaluating enzastaurin given with radiotherapy
and as adjuvant treatment in newly diagnosed glioblastoma patients without methylation of the MGMT gene
promoter (Clinicaltrials.gov NCT00509821).
References
Conclusions and Future Directions
In patients with malignant gliomas several molecular markers other than MGMT may have clinical
relevance, including overexpression of the epidermal
growth factor receptor, presence of the epidermal
growth factor receptor vIII mutation, loss or mutation of PTEN gene. However, MGMT methylation
status seems to be the most valuable marker for prediction of outcome and prognosis of patients treated
with alkylating agent chemotherapy. Nevertheless, to
date no biomarker has been definitively validated in
Belanich M, Pastor M, Randall T, Guerra D, Kibitel J, Alas L,

Li B, Citron M, Wasserman P, White A, Eyre H, Jaeckle
K, Schulman S, Rector D, Prados M, Coons S, Shapiro
W, Yarosh D (1996) Retrospective study of the correlation
between the DNA repair protein alkyltransferase and survival
of brain tumor patients treated with carmustine. Cancer Res
56:783788
Brada M, Stenning S, Gabe R, Thompson LC, Levy D, Rampling
R, Erridge S, Saran F, Gattamaneni R, Hopkins K, Beall
S, Collins VP, Lee SM (2010) Temozolomide versus procarbazine, lomustine, and vincristine in recurrent high-grade
glioma. J Clin Oncol 28:46014608
Brandes AA, Franceschi E, Tosoni A, Bartolini S, Bacci A, Agati
R, Ghimenton C, Turazzi S, Talacchi A, Skrap M, Marucci
G, Volpin L, Morandi L, Pizzolitto S, Gardiman M, Andreoli
78
A, Calbucci F, Ermani M (2010) O(6)-methylguanine DNAmethyltransferase methylation status can change between
first surgery for newly diagnosed glioblastoma and second
surgery for recurrence: clinical implications. Neuro Oncol
12:283288
Brandes AA, Franceschi E, Tosoni A, Benevento F, Scopece
L, Mazzocchi V, Bacci A, Agati R, Calbucci F, Ermani M
(2009) Temozolomide concomitant and adjuvant to radiotherapy in elderly patients with glioblastoma: correlation
with MGMT promoter methylation status. Cancer 115:3512
3518
Brandes AA, Franceschi E, Tosoni A, Blatt V, Pession A,
Tallini G, Bertorelle R, Bartolini S, Calbucci F, Andreoli
A, Frezza G, Leonardi M, Spagnolli F, Ermani M (2008)
MGMT promoter methylation status can predict the incidence and outcome of pseudoprogression after concomitant radiochemotherapy in newly diagnosed glioblastoma
patients. J Clin Oncol 26:21922197
Brandes AA, Tosoni A, Cavallo G, Bertorelle R, Gioia V,
Franceschi E, Biscuola M, Blatt V, Crin L, Ermani M
(2006) Temozolomide 3 weeks on and 1 week off as firstline therapy for recurrent glioblastoma: phase II study from
gruppo italiano cooperativo di neuro-oncologia (GICNO). Br
J Cancer 95:11551160
Chinot OL, Barri M, Fuentes S, Eudes N, Lancelot S,
Metellus P, Muracciole X, Braguer D, Ouafik L, Martin PM,
Dufour H, Figarella-Branger D (2007) Correlation between
O6 -methylguanine-DNA methyltransferase and survival in
inoperable newly diagnosed glioblastoma patients treated
with neoadjuvant temozolomide. J Clin Oncol 25:1470
1475
Curtin NJ, Wang LZ, Yiakouvaki A, Kyle S, Arris CA, CananKoch S, Webber SE, Durkacz BW, Calvert HA, Hostomsky
Z, Newell DR (2004) Novel poly(ADP-ribose) polymerase-1
inhibitor, AG14361, restores sensitivity to temozolomide in
mismatch repair-deficient cells. Clin Cancer Res 10:881889
Esteller M, Garcia-Foncillas J, Andion E, Goodman SN,
Hidalgo OF, Vanaclocha V, Baylin SB, Herman JG (2000)
Inactivation of the DNA-repair gene MGMT and the clinical response of gliomas to alkylating agents. N Engl J Med
343:13501354
Fabi A, Metro G, Russillo M, Vidiri A, Carapella CM, Maschio
M, Cognetti F, Jandolo B, Mirri MA, Sperduti I, Telera S,
Carosi M, Pace A (2009) Treatment of recurrent malignant
gliomas with fotemustine monotherapy: impact of dose and
correlation with MGMT promoter methylation. BMC Cancer
9:101
Friedman HS, McLendon RE, Kerby T, Dugan M, Bigner SH,
Henry AJ, Ashley DM, Krischer J, Lovell S, Rasheed K,
Marchev F, Seman AJ, Cokgor I, Rich J, Stewart E, Colvin
OM, Provenzale JM, Bigner DD, Haglund MM, Friedman
AH, Modrich PL (1998) DNA mismatch repair and O6alkylguanine-DNA alkyltransferase analysis and response to
temodal in newly diagnosed malignant glioma. J Clin Oncol
16:38513857
Gerson SL (2004) MGMT: its role in cancer aetiology and
cancer therapeutics. Nat Rev Cancer 4:296307
Hegi ME, Diserens AC, Godard S, Dietrich PY, Regli L,
Ostermann S, Otten P, Van Melle G, de Tribolet N,
Stupp R (2004) Clinical trial substantiates the predictive value of O-6-methylguanine-DNA methyltransferase

promoter methylation in glioblastoma patients treated with
temozolomide. Clin Cancer Res 10:18711874
RC, Stupp R (2005) MGMT gene silencing and benefit from
temozolomide in glioblastoma. N Engl J Med 352:9971003
Krex D, Klink B, Hartmann C, von Deimling A, Pietsch T,
Simon M, Sabel M, Steinbach JP, Heese O, Reifenberger
G, Weller M, Schackert G (2007) German glioma network.
Long-term survival with glioblastoma multiforme. Brain
130:25962606
Lee SM, Brada M, Stenning S, Thompson L, Gabe R (2008).
for BR12 Collaborators; NCRI Brain Tumour Clinical
Studies Group/GB. A randomised trial of procarbazine,
CCNU and vincristine (PCV) vs temozolomide (5-day or
21-day schedule) for recurrent high grade glioma (MRC
BR12, ISRCTN83176944). 33rd meeting of the European
Society for Medical Oncology, Stockholm, Sweden, 1216
September
Metro G, Fabi A, Mirri MA, Vidiri A, Pace A, Carosi M,
Russillo M, Maschio M, Giannarelli D, Pellegrini D, Pompili
A, Cognetti F, Carapella CM (2010) Phase II study of fixed
dose rate gemcitabine as radiosensitizer for newly diagnosed
glioblastoma multiforme. Cancer Chemother Pharmacol
65:391397
Perry JR, Blanger K, Mason WP, Fulton D, Kavan P, Easaw J,
Shields C, Kirby S, Macdonald DR, Eisenstat DD, Thiessen
B, Forsyth P, Pouliot JF (2010) Phase II trial of continuous
dose-intense temozolomide in recurrent malignant glioma:
RESCUE study. J Clin Oncol 28:20512057
Quinn JA, Jiang SX, Reardon DA, Desjardins A, Vredenburgh
JJ, Rich JN, Gururangan S, Friedman AH, Bigner DD,
Sampson JH, McLendon RE, Herndon JE 2nd, Walker A,
Friedman HS (2009) Phase II trial of temozolomide plus
O6-benzylguanine in adults with recurrent, temozolomideresistant malignant glioma. J Clin Oncol 27:12621267
Stupp R, Hegi ME, Mason WP, van den Bent MJ, Taphoorn
MJ, Janzer RC, Ludwin SK, Allgeier A, Fisher B, Belanger
K, Hau P, Brandes AA, Gijtenbeek J, Marosi C, Vecht CJ,
Mokhtari K, Wesseling P, Villa S, Eisenhauer E, Gorlia
T, Weller M, Lacombe D, Cairncross JG, Mirimanoff RO
(2009) European Organisation for Research and Treatment
of Cancer Brain Tumour and Radiation Oncology Groups;
National Cancer Institute of Canada Clinical Trials Group.
Lancet Oncol 10:459466
Stupp R, Hegi ME, Neyns B, Goldbrunner R, Schlegel U,
Clement PM, Grabenbauer GG, Ochsenbein AF, Simon M,
Dietrich PY, Pietsch T, Hicking C, Tonn JC, Diserens AC,
Pica A, Hermisson M, Krueger S, Picard M, Weller M
(2010) Phase I/IIa study of cilengitide and temozolomide
with concomitant radiotherapy followed by cilengitide and
temozolomide maintenance therapy in patients with newly
diagnosed glioblastoma. J Clin Oncol 27:58815886
Mirimanoff RO (2005) European Organisation for Research
and Treatment of Cancer Brain Tumor and Radiotherapy
Groups; National Cancer Institute of Canada Clinical Trials
Group. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352:987996
Taal W, Brandsma D, de Bruin HG, Bromberg JE, SwaakKragten AT, Smitt PA, van Es CA, van den Bent MJ (2008)
Incidence of early pseudo-progression in a cohort of malignant glioma patients treated with chemoirradiation with
temozolomide. Cancer 113:405410
Tabatabai G, Frank B, Wick A, Lemke D, von Krthy G,
Obermller U, Heckl S, Christ G, Weller M, Wick W (2007)
Synergistic antiglioma activity of radiotherapy and enzastaurin. Ann Neurol 61:153161
van den Bent MJ, Dubbink HJ, Sanson M, van der Lee-Haarloo
CR, Hegi M, Jeuken JW, Ibdaih A, Brandes AA, Taphoorn
MJ, Frenay M, Lacombe D, Gorlia T, Dinjens WN, Kros
JM (2009) MGMT promoter methylation is prognostic but
not predictive for outcome to adjuvant PCV chemotherapy in anaplastic oligodendroglial tumors: a report from
EORTC Brain Tumor Group Study 26951. J Clin Oncol
27:58815886
Vlassenbroeck I, Califice S, Diserens AC, Migliavacca E, Straub
J, Di Stefano I, Moreau F, Hamou MF, Renard I, Delorenzi
M, Flamion B, DiGuiseppi J, Bierau K, Hegi ME (2008)
Validation of real-time methylation-specific PCR to determine O6 -methylguanine-DNA methyltransferase gene promoter methylation in glioma. J Mol Diagn 10:332337
Weller M, Stupp R, Reifenberger G, Brandes AA, van den Bent
MJ, Wick W, Hegi ME (2010) MGMT promoter methylation
79
in malignant gliomas: ready for personalized medicine? Nat
Rev Neurol 6:3951
Wick A, Felsberg J, Steinbach JP, Herrlinger U, Platten M,
Blaschke B, Meyermann R, Reifenberger G, Weller M, Wick
W (2007) Efficacy and tolerability of temozolomide in an
alternating weekly regimen in patients with recurrent glioma.
Yip S, Miao J, Cahill DP, Iafrate AJ, Aldape K, Nutt CL, Louis
DN (2009) MSH6 mutations arise in glioblastomas during
temozolomide therapy and mediate temozolomide resistance.
Yung WK, Prados MD, Yaya-Tur R, Rosenfeld SS, Brada M,
Friedman HS, Albright R, Olson J, Chang SM, ONeill AM,
Friedman AH, Bruner J, Yue N, Dugan M, Zaknoen S, Levin
VA (1999) Multicenter phase II trial of temozolomide in
patients with anaplastic astrocytoma or anaplastic oligoastrocytoma at first relapse. Temodal Brain Tumor Group. J Clin
Oncol 17:27622771
Chapter 10
Brain Tumor Angiogenesis and Glioma Grading: Role

of Tumor Blood Volume and Permeability Estimates
Using Perfusion CT
Rajan Jain
Abstract Perfusion imaging of brain tumors has

been done using various tracer and non-tracer modalities and can provide additional physiologic and hemodynamic information, which is not available with
routine morphologic imaging. Tumor vascular perfusion parameters obtained using CT or MR perfusion
have been used for tumor grading, prognosis and
treatment response in addition to differentiating treatment/radiation effects from recurrent neoplasms. This
chapter is an overview of utility of perfusion CT for
assessment of brain tumors, description of the technique, its advantages and limitations.
Keywords Perfusion CT Glioma grading
Angiogenesis Tumor blood volume Permeability
Introduction
Gliomas, the most common primary brain neoplasms
in adults are very heterogeneous tumors. High-grade
gliomas can be highly invasive and extremely vascular tumors. Two of the most important factors in
determining malignancy of gliomas are their ability
to infiltrate the brain parenchyma and to recruit or
synthesize vascular networks for further growth i.e.,
neoangiogenesis (Folkman, 1992; Jain et al., 2002).
R. Jain ()
Division of Neuroradiology, Department of Radiology and
Department of Neurosurgery, Henry Ford Health System,
Detroit, MI 48202, USA
e-mail: rajanj@rad.hfh.edu
Malignant brain tumors are characterized by neovascularity and increased angiogenic activity with a higher
proportion of immature and highly permeable vessels.
Glioma grading is currently based on the histological assessment of the tumor, which is achieved by
either brain biopsy or cytoreductive surgery; however,
there are inherent limitations with these techniques
and their interpretation (Law et al., 2008). In vivo
perfusion imaging techniques provide additional information regarding tumor physiology and hemodynamics, which may help in better characterizing glioma
and may also overcome some of the limitations of
histopathologic grading and conventional morphologic
imaging. Perfusion imaging has been used to assess
tumor grade, prognosis, and recently to assess treatment response, which has caught more attention due to
advent of newer therapeutic options including antiangiogenic agents. Traditionally, perfusion imaging of
brain tumors has been done with magnetic resonance
imaging (MRI), using various perfusion imaging techniques and estimating tumor blood volume, blood flow,
and permeability (Roberts et al., 2000; Law et al.,
2004, 2008). However, perfusion CT (PCT), which
has also been used recently for glioma grading (Ellika
et al., 2007; Jain et al., 2008), provides a linear relationship between tissue signal and tissue concentration
of a contrast agent unlike perfusion MR and, hence,
probably provides a more robust and less biased estimation of physiologic and hemodynamic parameters.
In view of the wider availability, faster scan times, and
low cost combined with its ease of quantification of
various perfusion parameters as compared to MR perfusion, PCT is potentially well suited to study brain
tumors and monitoring tumor response to antiangiogenic agents (Ellika et al., 2007; Jain et al., 2008).

81
82
In Vivo Perfusion Imaging Versus

Histopathology
The current standard for tumor grading is histopathologic assessment of tissue, which has inherent limitations such as sampling error, interobserver variation and wide variety of classification systems that
are available, the most commonly of which used is
the WHO grading system (Law et al., 2008). Most
of the gliomas, especially high grade gliomas, show
a high degree of regional heterogeneity within the
tumor based on tumor cellularity, edema, and necrosis. Many of these factors are inherently dependent
on the blood supply. Histopathological evaluation of
tumor angiogenesis using various markers such as
microvascular density (MVD), microvascular cellular
proliferation (MVCP), and total vascular area (TVA)
is also limited by this regional heterogeneity, and its
confounding effect is worsened by its small size and
a limited number of samples obtained with surgical
biopsy. This can frequently result in inaccurate classification and grading of gliomas due to sampling errors.
Hence, there is a need for noninvasive in vivo clinical imaging tools which can study perfusion in the
entire tumor, can be used to assess much larger volumes than small biopsy samples, and probably guide
biopsy and excision sites for better results. Brain tumor
angiogenesis is a continuously evolving process which
can also be affected by various treatment modalities.
Hence, in vivo perfusion imaging that can be repeated,
unlike invasive procedures such as surgical excision or
biopsy, can help assess continued evolution of these
tumors as well as treatment response. Another important limitation of histopathological grading system is
that gliomas having similar grades respond differently
to similar treatment regimens as has been noted with
different molecular and genetic markers (Cairncross
et al., 1998), suggesting that there is a role for other
biomarkers such as perfusion parameters in predicting
progression or survival apart from histopathological
grading (Law et al., 2008).
Perfusion Parameters and Their

Importance
Dynamic contrast-enhanced imaging techniques using
MRI or CT have been used to obtain measures of
R. Jain
tumor vascular physiology and hemodynamics. After

the rapid administration of a contrast agent and the
acquisition of serial images at short intervals (seconds), an analysis that uses a pharmacokinetic model
of the time dependence of contrast can produce imaging biomarkers such as tumor blood volume, blood
flow, vascular permeability, and size of extravascular
extracellular space. Many of these parameters have
been correlated with tumor grade, aggressiveness, and
prognosis (Law et al., 2004, 2006, 2008; Jain et al.,
2008).
Tumor Blood Volume

Regional tumor blood volume measurements reflect
an assessment of tumor vasculature and perfusion,
and have been correlated with glioma grading as well
as prognosis. Measurement of tumor blood volume
is a good surrogate marker for MVD, a measure of
angiogenesis and an important prognostic indicator
(Leon et al., 1996; Li et al., 1994; Weidner, 1995) in
many human cancers. The association between MVD
and tumor aggressiveness can be explained by the
following: (1) solid tumors are composed of two interdependent components which are the malignant cells
and the stroma that they induce, and MVD could be
a measure of the success that a tumor has in forming this stromal component; (2) endothelial cells in
this stromal component stimulate the growth of tumor
cells; thus, the more intratumoral vessels there are, the
more endothelial cells and paracrine growth stimulation; (3) intratumoral MVD is a direct measure of the
vascular window through which tumor cells pass to
spread to distant sites (Weidner, 1995). Tumoral MVD,
however, does not distinguish new blood vessels from
the native ones, does not mark actively proliferating
endothelial cells, and does not correlate with the degree
of endothelial cell proliferation. However, these limitations do not seem to diminish the clinical value of this
measure. Cha et al. (2003) showed a strong correlation of CBV (cerebral blood volume) measurements in
mouse gliomas with MVD and suggested that rCBV
(regional CBV) may be elevated due to increase in
vessel size or total number of vessels or both. Aronen
et al. (2000) also showed a strong correlation of CBV
and tumor energy metabolism with MVD using MR
perfusion and FDG-PET imaging, respectively.
10 Brain Tumor Angiogenesis and Glioma Grading
Tumor Vascular Permeability

It is a fact that tumor blood vessels have defective
and leaky endothelium. Hypoxia or hypoglycemia that
occurs in rapidly growing tumors increases the expression of vascular endothelial growth factor (VEGF)
which is not only a potent angiogenic factor, but
also a potent permeability factor (Plate et al., 1992;
Shweiki et al., 1992). VEGF leads to the development of neoangiogenic vessels which are immature,
tortuous (Jain et al., 2002) and also have increased
permeability to macromolecules due to large endothelial cell gaps, incomplete basement membrane, and
absence of smooth muscles. These abnormal tumor
vessels can be used as potential markers to assess the
tumor grade. Thus, in vivo measurement of tumor vessel permeability is important for various reasons: (1) it
can be used for grading of tumors because increased
permeability is associated with immature blood vessels, which is seen with neoangiogenesis; (2) it can
be used to study the response of tumors to various
therapies especially antiangiogenic therapy (Bhujwalla
et al., 2003; Raatschen et al., 2008). (3) understanding
the concept of permeability can help in understanding
the mechanism of entry of therapeutic agents into the
central nervous system and; (4) development of methods to selectively alter the blood brain barrier (BBB)
to enhance drug delivery (Provenzale et al., 2005).
In Vivo Blood Volume and Permeability

Quantification: Imaging Techniques,
Limitations and Controversies
MR (Roberts et al., 2000; Cha et al., 2003; Law et al.,
2004, 2008) and CT perfusion (Cenic et al., 2000;
Purdie et al., 2001; Ellika et al., 2007; Jain et al., 2008)
have been used for in vivo characterization of tumor
angiogenesis in various animal and human studies.
One of the major limitations of clinically used imaging techniques is low spatial resolution which limits
assessment of the complex and heterogenous vascular
microenvironment in detail (McDonald and Choyke,
2003; Vajkoczy and Menger, 2000). Another limitation is the current use of low molecular weight contrast
agents for human subjects. Because permeability measurements depend on the leakage of particles from the
83
blood to the interstitial space, it is hypothesized that

low molecular weight contrast agents leak relatively
easily from the blood into the interstitium. This might
also lead to overestimation of the permeability which
will be influenced by and will approximate tumor
blood flow (de Lussanet et al., 2005). High molecular
weight contrast agents leak from the vessels and move
through the interstitium with relative difficulty and are
flow independent. Thus, macromolecular permeability
and vascular volumes may be best measured by high
molecular weight contrast agents (de Lussanet et al.,
2005). However, currently used CT or MRI contrast
agents are of low molecular weight (0.50.9 kDa)
which might limit the accurate differentiation of vascular permeability and blood volume (Choyke, 2005).
Even though CT and MR contrast agents do not differ
much in their molecular weight, different charges of
nonionic CT contrast as compared to ionic MR contrast may also be responsible for the differences in
permeability measured with these two modalities.
Another controversial aspect of measuring permeability with various perfusion imaging techniques is
the scanning time. Delayed permeability due to slow
leakage of contrast from a leaky blood vessel may not
be accurately measured with first pass of the contrast
agent using a 45 or 60 s scanning time (Goh et al.,
2005), and can be only measured with longer acquisition times; however, there is no real consensus regarding the optimal acquisition time. Jain et al. (2008) used
PCT for tumor vascular permeability quantification
and reported slightly higher permeability values than
those reported in literature using other techniques, and
this could be explained by the mathematical model.
Longer acquisition time of 170 s could also explain
the higher values that they obtained because the scanning sequence was optimized to include delayed permeability also as described above. Gliomas, particularly high-grade gliomas, can have extremely variable
and heterogeneous blood flow due to the complex
tumor vasculature which can influence permeability
(McDonald and Choyke, 2003; Vajkoczy and Menger,
2000). Other factors which can also influence permeability include luminal surface area, interstitial, hydrostatic, and osmotic pressure across the endothelium.
Slow blood flow or low osmotic gradients which can
occur in the high-grade tumor with considerable vasogenic edema and also in the central parts of large
tumors can lead to a larger component of delayed
permeability, requiring longer acquisition times.
84
Quantification of vessel permeability with various

perfusion techniques requires two or more compartment pharmacokinetic models with an arterial input
function, making these studies more complex than
CBV estimation (Tofts et al., 1999), and is dependent
on the imaging technique employed and the mathematical model that is used. Post gadolinium T1 weighted
MRI gives a rough estimate of the disruption of BBB,
and has been used in the past for quantitative estimation of permeability but dynamic imaging acquisition
provides a better estimate of vascular permeability
(Tofts et al., 1999). Dual echo gradient echo perfusion weighted imaging (Uematsu and Maeda, 2006)
is based on a simple two compartment kinetic model
(Boxerman et al., 2006), and has more recently been
used to measure vascular permeability and to correctly estimate blood volume in brain tumors with
deficient or absent BBB. There are several studies
which have found a correlation between increased
vascular permeability and higher tumor grade (Law
et al., 2004; Johnson et al., 2004; Jain et al., 2008;
Roberts et al., 2001). However, MR perfusion techniques have certain limitations because of the nonlinear relationship of the signal intensity with the
contrast, both for dynamic contrast enhanced imaging with T1-weighting (Yankeelov et al., 2005), and
for dynamic susceptibility contrast imaging with T2or T2 -weighting. In the latter case, when the contrast agent remains intravascular, the method is widely
accepted as a relative estimate of CBF (cerebral blood
flow) and CBV, although there is a possibility for artifacts because of difficulties in assessing the shape and
timing of the arterial input function (Conturo et al.,
2005).
In the event that substantial leakage of a contrast
agent from intra- to extra-vascular space takes place,
a strong and competing T1 contrast effect is often
noticed in the areas of pathology because of the necessity of short (1 s) repetition times needed to estimate
CBF. As a first-order tactic to minimize the competing
T1 contrast, preloading with contrast agent has been
proposed with some success (Boxerman et al., 2006).
However, this approach does not allow an estimate of
Ktrans . An alternative has also been proposed (Johnson
et al., 2004). To decrease the T1 effect, this approach
used a slower repetition time, lengthening the repetition time of the experiment, and undermining the estimation of CBF; thus, yielding estimates of only CBV
and Ktrans . A further refinement, allowing the estimate
R. Jain
of blood volume and producing an index of transfer

constant, has been suggested (Boxerman et al., 2006),
and a dual-echo gradient echo sequence (Uematsu and
Maeda, 2006) also shows some potential for an index
of blood volume and transfer constant. Despite the
partial success of these rapid imaging studies, in contrast to CT perfusion, there does not appear to be an
MRI technique that will reliably quantify CBF, CBV,
and Ktrans in one experiment. PCT also has the advantage of providing absolute measures of these perfusion
parameters whereas MR estimates are mostly relative
to the normal brain parenchyma. Another major disadvantage of MR perfusion is susceptibility artifacts
due to hemorrhage and various mineral depositions,
which can be a major issue in posttreatment tumor
patients. Despite these facts, MR perfusion has been
more often utilized because MRI is the standard of
care for brain tumor patients; whereas, PCT requires
a separate examination with iodinated contrast agent
and moderate exposure to ionizing radiation.
Perfusion CT Tracer Kinetics Theory

The model proposed by Johnson and Wilson (1966)
is a convenient model for tumor application and is
based on the assumption that the distribution volume of CT contrast medium in tissue consists of the
extra-cellular space in capillaries and interstitial space.
Central Volume Principle relates tissue (tumor) blood
flow (TBF), tissue (tumor) blood volume (TBV) and
mean transit time (MTT) in the following equation:
TBF = TBV/MTT (24). Perfusion studies are obtained
by monitoring the passage of an iodinated contrast
bolus through the cerebral vasculature. There is a linear relationship between contrast agent concentration
and attenuation, with the contrast agent causing a transient rise in attenuation proportional to the amount
of contrast in a given region. Contrast agent timeconcentration curves are generated in an arterial region
of interest (ROI), a venous ROI as well as in each
pixel. Deconvolution of arterial and tissue enhancement curves, according to the adiabatic approximation
of the Johnson and Wilson model (St. Lawrence and
Lee, 1998), gives the blood flow scaled impulse residue
function (IRF), from which TBF, TBV, and MTT can
be determined as prescribed in the reference (Lee et al.,
2003). Specifically, blood flow is equal to the height
of the blood flow scaled IRF. CBV is calculated as an
area under the initial plateau (intravascular phase) of

the blood flow scaled IRF.
Permeability surface-area product characterizes the
diffusion of some of the contrast agent from the blood
vessels into the interstitial space due to deficient or
leaky BBB. Permeability is related to the diffusion
coefficient of contrast agent in the assumed water filled
pores of the capillary endothelium. The diffusion flux
of contrast agent across the capillary endothelium is
dependent on both the diffusion coefficient and the
total surface area of the pores. PS is computed from
the IRF. Contrast agent diffusion appears in the IRF
as a residual enhancement that occurs after the initial
impulse response and which decreases exponentially
with time. The IRF is used to estimate the first-pass
fraction of contrast agent that remains in the tissue,
the extraction fraction, E (Purdie et al., 2001). The
extraction fraction is related to the rate at which contrast leaks out of the vasculature via the following
relationship:
PS
E =1eF ,
where PS is the permeability surface-area product and
F is flow. The PS product has the same dimensions
as flow, and thus the ratio PS
F is dimensionless. In
physiological terms, PS is the rate at which contrast
agent flows into the extravascular tissues; it is related
to another commonly stated parameter of vascular
leakage, the transfer constant by the following:
K trans = EF,
Ktrans
is the transfer constant with, again, the

where
same dimensions as flow. It is easily demonstrated
trans PS. In normal cerethat, if PS
F << 1, then K
bral vasculature, PS is negligible for all contrast agents
presently in use.
Perfusion CT Technique
Perfusion studies can be performed using multidetector row CT scanners. Currently available 16-slice
CT scanners can cover 2 cm of the brain which is
increased to 4 cm using a 64-slice CT scanner. A
low radiation dose non-contrast CT head study is
usually performed to localize the ROI before obtaining
a perfusion scan. For the perfusion scan, 50 ml of
nonionic contrast is injected at a rate of 45 ml/s
through an IV line using an automatic power injector.
85
At 5 s into the injection, a cine (continuous) scan is

initiated with the following technique: 80 kVp, 100
120 mA, and 1 s per rotation for duration of 50 s. After
the initial 50 s cine scan, eight more axial images can
be acquired, one image every 15 s for an additional
2 min, thus giving a total acquisition time of 170 s
to assess for delayed permeability (Jain et al., 2008).
Four 5 mm thick axial slices are acquired with the
16-slice CT scanner, whereas for 64-slice CT scanner,
eight 5 mm thick axial slices are acquired resulting in
a total coverage area of 4 cm instead of 2 cm with the
16-slice scanner. Perfusion maps of perfusion parameters can be obtained using many of the commercially
available software. We use an Advantage Windows
workstation using CT perfusion 3.0 software (General
Electric Medical Systems, Milwaukee, WI) and two
compartment model to generate CBV, CBF, MTT, and
PS maps in our patients. Superior sagittal sinus is generally used as the venous output function and the artery
with the greatest peak and slope on time-attenuation
curves as the arterial input function (AIF). An ROI is
drawn within the confines of a large vessel and the
automatic function of the software picks the pixels
with greatest peak and slope on the time-attenuation
curve for analysis.
Relationship Between Glioma Grade

and Perfusion CT Parameters
Most of the literature regarding utility of perfusion
imaging for glioma grading is based on various MR
perfusion techniques. Recently, perfusion CT has also
been used to grade gliomas based on perfusion parameters (Ellika et al., 2007; Jain et al., 2008). Ellika et al.
(2007) were able to differentiate low and high grade
gliomas with a high sensitivity (85.7%) and specificity (100%) using PCT and nCBV (CBV normalized
relative to normal appearing contra-lateral white matter) threshold of 1.92. This relationship between CBV
and histological grade is intuitive, as pathology studies show higher MVD in higher grade tumors. Jain
et al. (2008) apart from differentiating low and high
grade gliomas (Figs. 10.1 and 10.2) could also differentiate high grade tumor group into grade III and
grade IV based on PS measurements which showed
a stronger predictability as compared to CBV and
especially could differentiate enhancing grade III from
grade IV based on the differences in PS (Fig. 10.3).
86
R. Jain
Fig. 10.1 (a) PS and (b) CBV perfusion CT maps in a 32year-old woman with WHO grade II astrocytoma showing low
permeability (PS = 0.7 ml/100 g/min) and low blood volume
(CBV = 1.01 ml/100 g) within the tumor. Inset: Post-contrast
Fig. 10.2 Perfusion CT maps in a 55-year-old man with

glioblastoma multiforme: (a) PS maps showing very high permeability (PS = 5.14 ml/100 g/min) along the enhancing nodular
margins of the tumor and (b) CBV maps showing high blood volume (CBV = 3.49 ml/100 g) within the periphery of the tumor.
T1-weighted image showing a non-enhancing right frontal

tumor with no surrounding perilesional edema. (Printed with
permission from Jain et al., 2008)
Inset: Postcontrast T1-weighted image showing heterogenously

enhancing mass with irregular central necrosis in the right
peritrigonal region. (Printed with permission from Jain et al.,
2008)
87
as the enhancing grade III tumors show higher PS,

CBV and CBF, which may be more aggressive tumors
with higher recurrence rate and shorter survival periods
as compared to the nonenhancing grade III, which
remains to be determined. Previous studies using MR
perfusion have described various rCBV threshold values for glioma grading. Lev et al. (2004) described
a threshold of 1.5 in discriminating between patients
with low and high grade glioma with a sensitivity and
specificity of 100 and 69%, respectively. Law et al.
(2003) showed a sensitivity and specificity of 95 and
57.5%, respectively, by using 1.75 as the threshold
value for rCBV.
Heterogeneity of Glioma Angiogenesis

and Perfusion Imaging
Fig. 10.3 Perfusion CT PS map in a 46-year-old man with
enhancing grade III glioma involving the right frontal lobe
with extension across the genu of the corpus callosum showing
high permeability (PS = 2.04 ml/100 g/min)., however, significantly less as compared to a grade IV (Fig. 10.1b). Inset:
Postcontrast T1-weighted showing an enhancing lesion involving the right frontal lobe with extension across the genu of the
corpus callosum. (Printed with permission from Jain et al., 2008)
This is in keeping with the current WHO guidelines

of including MVCP as a diagnostic criterion of grade
IV, but not for grade III astrocytic tumors, suggesting
that PS measurements could show better correlation
with MVCP, and hence an imaging biomarker of more
immature and leaky blood vessels. Increased angiogenesis in grade IV tumors is characterized not only
by an increased number of vessels as compared to
grade III astrocytic tumors, but by association also
with disproportionate lengthening, increased pliability, endothelial cell proliferation, and irregular shape,
which can explain the difference in perfusion parameters for grade IV as compared to grade III tumors.
Non-enhancing grade III tumors have also been shown
to have lower mean PS, CBV, and CBF as compared
to the enhancing grade III group (Jain et al., 2008).
This difference is probably due to higher MVD seen
in enhancing tumors as compared to the nonenhancing group. This could have prognostic implications
Angiogenesis involves a multitude of controlled signaling cascades and structural changes that occur in
a defined order and continue until a new vasculature has been formed. Tumor cellular growth usually
outgrows its blood supply leading to hypoxia which
leads to the formation of angiogenic mediators such
as VEGF. VEGF initially leads to the formation of
immature and leaky blood vessels, which results in
increased permeability, leading to extravasation of
plasma, plasma proteins, and deposition of proangiogenic matrix proteins. Later, as these pericyte-poor
new vessels called mother vessels enlarge and give
rise to daughter vessels through a complex series of
endothelial rearrangements, MVD and TVA increase
with continued increase in permeability. Finally, with
vessel maturation, the total number and area of blood
vessels continue to increase more than the vessel
leakiness; hence, evolving into a very heterogeneous
tumor with various regions probably showing different mixtures of vessel characteristics and angiogenesis,
which can be seen as regions with high CBV but
not very high PS and vice versa on perfusion imaging (Fig. 10.4). Correlation of various tumor perfusion
estimates with histological angiogenesis markers could
thus be very useful as far as in vivo identification of
these different regions of angiogenesis is concerned,
and using various perfusion parameters as imaging
biomarkers.
88
R. Jain
Fig. 10.4 Perfusion CT maps in a 39-year-old woman with

heterogeneously enhancing grade III glioma involving the left
temporal lobe. (a) CBV and (b) PS maps showing enhancing region laterally ROI #7 demonstrating the highest PS
(PS = 8.52 ml/100 g/min), however, not showing the highest CBV(CBV = 4.73 ml/100 g) within the tumor; whereas
ROI #8 shows relatively lower PS (PS = 1.14 ml/100 g/min);
however, shows the highest CBV within the tumor (CBV =

11.65 ml/100 g), suggesting there is marked heterogeneity with
in high grade gliomas which could be due to the heterogeneity of tumor angiogenesis. Inset: Post-contrast T1-weighted axial
image showing a heterogeneously enhancing left temporal lobe
tumor
Role of Perfusion CT in Differentiating

Recurrent Tumor from Radiation
Necrosis
MR imaging features and MR spectroscopic imaging have been used to differentiate radiation necrosis
from recurrent tumors with mixed success (Kumar
et al., 2000; Cheronov et al., 2005). Various forms
of metabolic imaging techniques have been utilized
in the past with limited results. FDG-PET (Langleben
and Segall, 2000), which is based on tumor glucose
metabolism, has shown variable sensitivity and specificity in differentiating recurrent tumors from radiation necrosis and also has limited spatial resolution.
Posttreatment recurrent enhancing lesions have also
been evaluated with MR perfusion imaging, showing
increased CBV in recurrent tumors as compared to
non-neoplastic lesions (Covarrubias et al., 2004). Jain
et al. (2007) used perfusion CT to differentiate the
two entities and recurrent tumors showed higher CBV,
CBF, and lower MTT (Fig. 10.5a) as compared to radiation necrosis (Fig. 10.5b). MR perfusion techniques
also have been successfully used; however, PCT could
have slight edge as most of these patients after having undergone multiple various combination therapies
have some components of hemorrhage and mineralization, which could produce susceptibility artifacts
complicating the perfusion analysis especially if using
Recent advances in brain tumor treatment have led to

aggressive management strategies with combinations
of surgery, chemotherapy, and radiation therapy based
on the location and histologic type of the tumor. In
particular, various forms of radiation therapy, including stereotactic radiosurgery, high-dose external beam
radiation, and brachytherapy, have become important
therapeutic adjuncts. Patient survival and quality of life
are correlated with response to therapy, tumor recurrence and also adverse effects of radiation therapy
such as radiation necrosis. Differentiating recurrent
tumors from radiation necrosis on imaging studies has
always been an important clinical imperative because
the management of these two entities is different.
Time interval from radiation therapy to the development of both of these entities after treatment does not
help in differentiating one from the other. In addition,
the problem is confounded by the fact that there is
often a mixture of tumor with necrosis. Conventional
89
Fig. 10.5 (a) Perfusion CT CBV map in a 21-year-old male

who was previously treated for a left posterior temporal lobe
astrocytoma, showing a recurrent enhancing lesion in the left
peritrigonal region which showed high CBV, suggesting recurrent tumor (b) as compared to radiation necrosis in a 49-year-old
male who underwent treatment for a left frontal lobe anaplastic

astrocytoma showing low CBV in a recurrent enhancing lesion
in left frontoparietal region. Inset: Corresponding post-contrast
T1-weighted images. (Printed with permission from Jain et al.,
2007)
dynamic susceptibility weighted imaging (Jain et al.,

2007).
In conclusion, clinically available perfusion imaging tools whether using MRI or CT can provide
additional information regarding brain tumor vascular
estimates which could be useful imaging biomarkers for preoperative glioma grading and angiogenesis
assessment and could also be useful for treatment planning and response assessment. PCT has the advantage
of providing two of the most important tumor vascular
estimates i.e., blood volume and permeability in one
single experiment as well as a linear relationship of tissue signal with tissue contrast agent unlike most of the
available MR perfusion techniques and hence, has the
potential to be a useful tool for brain tumor assessment.
macromolecular contrast magnetic resonance imaging after

treatment with antiangiogenic agent TNP-470. Clin Cancer
Res 9:355362
Boxerman JL, Schmainda KM, Weisskoff RM (2006) Relative
cerebral blood volume maps corrected for contrast agent
extravasation significantly correlate with glioma tumor
grade, whereas uncorrected maps do not. AJNR Am J
Neuroradiol 27:859867
Cairncross JG, Ueki K, Zlatescu MC, Lisle DK, Finkelstein DM,
Hammond RR, Silver JS, Stark PC, Macdonald DR, Ino
Y, Ramsay DA, Louis DN (1998) Specific genetic predictors of chemotherapeutic response and survival in patients
with anaplastic oligodendrogliomas. J Natl Cancer Inst 90:
14731479
Cenic A, Nabavi DG, Craen RA, Gelb AW, Lee TY (2000) A
CT method to measure hemodynamics in brain tumors: validation and application of cerebral blood flow maps. AJNR
Am J Neuroradiol 21:462470
Cha S, Johnson G, Wadghiri YZ, Jin O, Babb J, Zagzag D,
Turnbull DH (2003) Dynamic, contrast-enhanced perfusion
MRI in mouse gliomas: correlation with histopathology.
Magn Reson Med 49:848855
Chernov M, Hayashi M, Izawa M, Ochiai T, Usukura M, Abe
K, Ono Y, Muragaki Y, Kubo O, Hori T, Takakura K (2005)
Differentiation of the radiation-induced necrosis and tumor
recurrence after gamma knife radiosurgery for brain metastases: importance of multivoxel proton MRS. Minim Invasive
Neurosurg 48:228234
Choyke PL (2005) Contrast agents for imaging tumor angiogenesis: is bigger better?. Radiology 235:12
Conturo TE, Akbudak E, Kotys MS, Chen ML, Chun SJ,
Hsu RM, Sweeney CC, Markham J (2005) Arterial input
References
Aronen HJ, Pardo FS, Kennedy DN, Belliveau JW, Packard SD,
Hsu DW, Hochberg FH, Fischman AJ, Rosen BR (2000)
High microvascular blood volume is associated with high
glucose uptake and tumor angiogenesis in human gliomas.
Bhujwalla ZM, Artemov D, Natarajan K, Solaiyappan M,
Kollars P, Kristjansen PE (2003) Reduction of vascular and permeable regions in solid tumors detected by
90
functions for dynamic susceptibility contrast MRI: requirements and signal options. J Magn Reson Imaging 22:
697703
Covarrubias DJ, Rosen BR, Lev MH (2004) Dynamic magnetic
resonance perfusion imaging of brain tumors. Oncologist
9:528537
De Lussanet QG, Langere S, Beets-Tan RG, van Genderen MH,
Griffioen AW, van Engelshoven JM, Backes WH (2005)
Dynamic contrast-enhanced MR imaging kinetic parameters
and molecular weight of dendritic contrast agents in tumor
angiogenesis in mice. Radiology 235:6572
Ellika SK, Jain R, Patel SC, Scarpace L, Schultz LR, Rock JP,
Mikkelsen T (2007) Role of perfusion CT in glioma grading and comparison with conventional MR imaging features.
AJNR Am J Neuroradiol 28:19811987
Folkman J (1992) The role of angiogenesis in tumor growth.
Semin Cancer Biol 3:6571
Goh V, Halligan S, Hugill JA, Gartner L, Bartram CI
(2005) Quantitative colorectal cancer perfusion measurement using dynamic contrast-enhanced multidetector-row
computed tomography: effect of acquisition time and implications for protocols. J Comput Assist Tomogr 29:5963
Jain R, Ellika SK, Scarpace L, Schultz LR, Rock JP, Gutierrez J,
Patel SC, Ewing J, Meikkelsen T (2008) Quantitative estimation of permeability surface-area product in astroglial brain
tumors using perfusion CT and correlation with histopathological grade. AJNR Am J Neuroradiol 29:694700
Jain RK, Munn LL, Fukumura D (2002) Dissecting tumor pathophysiology using intravital microscopy. Nat Rev Cancer
2:266276
Jain R, Scarpace L, Ellika S, Schultz LR, Rock JP, Rosenblum
ML, Patel SC, Lee TY, Mikkelsen T (2007) First-pass perfusion computed tomography: initial experience in differentiating recurrent brain tumors from radiation effects and
radiation necrosis. Neurosurgery 61:778786
Johnson G, Wetzel SG, Cha S, Babb J, Tofts PS (2004)
Measuring blood volume and vascular transfer constant from
dynamic, T(2) -weighted contrast-enhanced MRI. Magn
Reson Med 51:961968
Johnson JA, Wilson TA (1966) A model for capillary exchange.
Am J Physiol 210:12991303
Kumar AJ, Leeds NE, Fuller GN, Van Tassel P, Maor MH,
Sawaya RE, Levin VA (2000) Malignant Gliomas: MR
imaging spectrum of radiation therapy-and chemotherapyinduced necrosis of the brain after treatment. Radiology
217:377384
Langleben DD, Segall GM (2000) PET in differentiation of
recurrent brain tumor from radiation injury. J Nucl Med
41:18611867
Law M, Oh S, Johnson G, Babb JS, Zagzag D, Golfinos J, Kelly
PJ (2006) Perfusion magnetic resonance imaging predicts
patient outcome as an adjunct to histopathology: a second
reference standard in the surgical and nonsurgical treatment
of low-grade gliomas. Neurosurgery 58:10991107
Law M, Yang S, Babb JS, Knopp EA, Golfinos JG, Zagzag
D, Janohnson G (2004) Comparison of cerebral blood volume and vascular permeability from dynamic susceptibility contrast- enhanced perfusion MR imaging with glioma
grade. AJNR Am J Neuroradiol 25:746755
Law M, Yang S, Wang H, Babb JS, Johnson G, Cha S, Knopp
EA, Zagzag D (2003) Glioma grading: sensitivity, specificity,
R. Jain
and predictive values of perfusion MR imaging and proton
MR spectroscopic imaging compared with conventional MR
imaging. AJNR Am J Neuroradiol 24:19891998
Law M, Young RJ, Babb JS, Peccerelli N, Chheang S,
Gruber ML, Miller DC, Golfinos JG, Zagzag D, Johnson
G (2008) Gliomas: predicting time to progression or survival with cerebral blood volume measurements at dynamic
susceptibility-weighted contrast-enhanced perfusion MR
imaging. Radiology 247:490498
Lee TY, Purdie TG, Stewart E (2003) CT imaging of angiogenesis. Q J Nucl Med 47:171187
Leon SP, Folkerth RD, Black PM (1996) Microvessel density is a
prognostic indicator for patients with astroglial brain tumors.
Cancer 77:362372
Lev MH, Ozsunar Y, Henson JW, Rasheed AA, Barest GD,
Harsh GR 4th, Fitzek MM, Chiocca EA, Rabinov JD, Csavoy
AN, Rosen BR, Hochberg FH, Schaefer PW, Gonzalez
RG (2004) Glial tumor grading and outcome prediction
using dynamic spin-echo MR susceptibility mapping compared with conventional contrast-enhanced MR: confounding effect of elevated rCBV of oligodendrogliomas. AJNR
Li VW, Folkerth RD, Watanabe H, Yu C, Rupnick M, Barnes
P, Scott RM, Black PM, Sallan SE, Folkman J (1994)
Microvessel count and cerebrospinal fluid basic fibroblast growth factor in children with brain tumours. Lancet
344:8286
McDonald DM, Choyke PL (2003) Imaging of angiogenesis:
from microscope to clinic. Nat Med 9:713725
Plate KH, Breier G, Weich HA, Risau W (1992) Vascular
endothelial growth factor is a potential tumour angiogenesis factor in human gliomas in vivo. Nature
359:845848
Provenzale JM, Mukundan S, Dewhirst M (2005) The role of
blood-brain barrier permeability in brain tumor imaging and
therapeutics. AJR Am J Roentgenol 185:763767
Purdie TG, Henderson E, Lee TY (2001) Functional CT imaging
of angiogenesis in rabbit VX2 soft-tissue tumour. Phys Med
Biol 46:31613175
Raatschen HJ, Simon GH, Fu Y, Sennino B, Shames DM,
Wendland MF, McDonald DM, Brasch RC (2008) Vascular
permeability during antiangiogenesis treatment: MR imaging assay results as biomarker for subsequent tumor growth
in rats. Radiology 247:391399
Roberts HC, Roberts TP, Bollen AW, Ley S, Brasch RC,
Dillon WP (2001) Correlation of microvascular permeability
derived from dynamic contrast enhanced MR imaging with
histologic grade and tumor labeling index: a study in human
brain tumors. Acad Radiol 8:384391
Roberts HC, Roberts TP, Brasch RC, Dillon WP (2000)
Quantitative measurement of microvascular permeability
in human brain tumors achieved using dynamic contrastenhanced MR imaging: correlation with histologic grade.
Shweiki D, Itin A, Soffer D, Keshet E (1992) Vascular endothelial growth factor induced by hypoxia may mediate hypoxiainitiated angiogenesis. Nature 359:843845
St Lawrence KS, Lee TY (1998) An adiabatic approximation
to the tissue homogeneity model for water exchange in the
brain: I. Theoretical derivation. J Cereb Blood Flow Metab
18:13651377

Tofts PS, Brix G, Buckley DL, Evelhoch JL, Henderson E,
Knopp MV, Larsson HB, Lee TY, Mayr NA, Parker GJ,
Port RE, Taylor J, Weisskoff RM (1999) Estimating kinetic
parameters from dynamic contrast-enhanced T(1)-weighted
MRI of a diffusible tracer: standardized quantities and symbols. J Magn Reson Imaging 10:223232
Uematsu H, Maeda M (2006) Double-echo perfusion-weighted
MR imaging: basic concepts and application in brain tumors
for the assessment of tumor blood volume and vascular
permeability. Eur Radiol 16:180186
91
Vajkoczy P, Menger MD (2000) Vascular microenvironment in
gliomas. J Neurooncol 50:99108
Weidner N (1995) Intratumor microvessel density as a prognostic factor in cancer. Am J Pathol 147:919
Yankeelov TE, Rooney WD, Huang W, Dyke JP, Li X, Tudorica
A, Lee JH, Koutcher JA, Springer CS Jr (2005) Evidence
for shutter-speed variation in CR bolus-tracking studies of
human pathology. NMR Biomed 18:173185
Chapter 11
Vasculogenic Mimicry in Glioma

Zhong-Ping Chen and Yin-Sheng Chen
Abstract For years, sprouting angiogenesis has been

considered an exclusive mechanism of tumor vascularization. However, over the last years, several other
mechanisms have been identified, including vesselcooption, intussusception, recruitment of endothelial
precursor cells (EPCs) and even mechanisms that do
not involve endothelial cells, a process called vasculogenic mimicry (VM). The latter describes a mechanism by which highly aggressive tumor cells can form
vessel-like structures themselves, by virtue of their
high plasticity. Vasculogenic mimicry is known as nonendothelial tumor cell-lined microvascular channels in
aggressive tumors, which may function as blood supply networks. There is also evidence of a physiological
connection between the endothelial-lined vasculature
and VM. Micro-vascular density (MVD) was comparably less in VM positive tumors than in VM negative
tumors in high-grade gliomas. Thus, the existence of
VM may provide a complementation to ensure tumor
blood supply, especially in regions of the gliomas with
less MVD. The presence of VM has been associated
with more aggressive tumor biology and increased
tumor-related mortality. Patients with VM positive
gliomas survived a shorter period of time than those
without VM. Though the detailed molecular mechanisms for VM are not fully understood, glioma stem
cells (GSC) might play a key role since GSC have been
shown to be involved in tumor tissue remodeling. In the
future, successful management of gliomas and other
aggressive cancers could involve the targeting of both
vasculogenic mimicry and angiogenesis.
Z.-P. Chen ()
Department of Neurosurgery, Cancer Center, Sun Yat-sen
University, Guangzhou 510060, China
e-mail: chenzp57@mail.sysu.edu.cn
Keywords Vasculogenic mimicry Angiogenesis

Tumor vascularization Endothelial cells MVD
GSC
Introduction
Solid tumors require an adequate blood supply for survival, growth, and metastasis. Conventionally, blood
vessels are assembled by two processes: (1) vasculogenesis, the reorganization of randomly distributed cells
into a blood vessel network, and (2) angiogenesis, the
sprouting of new vessels from preexisting vasculature
in response to external chemical stimulation.
Gliomas are the most frequent and malignant
primary brain tumors in adults and have a poor
prognosis despite surgery and conventional radiochemotherapy. Histologically, gliomas are highly
angiogenic and characterized by microvascular proliferations (Fig. 11.1b) (Louis et al., 2007). However,
although anti-vascular endothelial growth factor therapy has had significant efficacy in gliomas with nearly
50% of responders, the clinical benefit remains unsatisfactory (Vredenburgh et al., 2007; Kreisl et al., 2009).
Can tumors acquire blood supply through other ways
to escape conventional antiangiogenisis therapy?
Maniotis et al. (1999) reported that blood vessels
of highly aggressive uveal melanomas are formed
by tumor cells instead of endothelial cells. He
termed this novel concept in tumor vascularization as vasculogenic mimicry (VM). Vasculogenic
mimicry describes the ability of aggressive tumor
cells to express endothelium-associated genes and
form extracellular matrix (ECM)-rich vasculogeniclike networks in three-dimensional culture. These
networks recapitulate embryonic vasculogenesis, and

93
94
Z.-P. Chen and Y.-S. Chen
Fig. 11.1 Vasculogenic mimicry in glioma. Endothelial cells

are detected with anti-CD34 immunohistochemistry staining
(dark brown) and vascular basement membrane with PAS staining (purple magenta) in tubular blood vessel (a). (b) microvascular proliferation of gliomas. Seven morphological patterns
of PAS-positive channels can been seen in (c) and (e). Large

cross section (d) and longitudinally section (f) vessels containing red blood cells stained positively with PAS but negatively
with CD34 (VM). (CD34 and PAS dual staining a, b, d and f,
400; c, 100; e, 200)
they have been observed in human aggressive tumors.

Vasculogenic mimicry has been seen in several
malignant tumor types such as breast cancer, liver
cancer, ovarian cancer, melanoma, prostate cancer,
and bidirectional differentiated malignant tumors.
Vasculogenic mimicry of the patterned matrix type was
previously also reported in human glioblastoma tissues (Yue and Chen, 2005; EI Hallani et al., 2010)
and human glioma cell-line xenografts (Niclou et al.,
2008). Subsequent studies have suggested that VM
channels may function as blood supply networks, and
thus could be another target for anti-cancer therapy.
11 Vasculogenic Mimicry in Glioma
However, the relations between VM and vasculogenesis or angiogenesis are not fully understood yet, and
the molecular mechanisms of these processes remain
a puzzle. Thus, a better understanding of tumor vascularization is needed to optimize antivascular therapy.
The Concept and Characteristic of VM

Vasculogenic mimicry is known as non-endothelial
tumor cell-lined microvascular channels in aggressive tumors (Fig. 11.1d, f). Under appropriate matrix
microenvironment conditions, highly aggressive and
metastatic tumor cells can form highly patterned vascular like channels (Fig. 11.1c, e). The generation of
such microvascular channels by genetically deregulated, aggressive tumor cells is termed VM to emphasize its de novo generation that is independent of
angiogenesis. The channel formed by VM is composed of a basement membrane and lined by tumor
cells, and no endothelial cells are found on its inner
wall. Blood plasma and red blood cells (RBCs) can
flow through the channel (Folberg and Maniotis, 2004;
Yue and Chen, 2005). Tumor cells are aligned with
the external superficies of the channel. No inflammatory cells or necrosis are found around the channel.
The endothelial-lined vasculature is closely apposed
to the tumor cell-formed fluid conducting meshwork,
and hypothetically, it is presumed that as the tumor
remodels, the vasculature becomes leaky, resulting in
the extravascular conduction of plasma. Several observations reported that tumour cells are located in the
walls of tumor blood vessels and form a part of the
vessel surface while the remaining part is covered
by endothelium. This is known as mosaic vessels
where tumor cells undergo intravasation into the lumen
and stay temporally in the vessel wall (Chang et al.,
2000). The tumor cell-lined vessels could be the result
of a complete invasion of the vessel wall by tumor
cells, perhaps the endpoint of mosaicism. Since there
has been evidence strongly suggesting that glioma
stem cell (GSC) are capable of forming blood vessels as well as tubular vasculogenic mimicry de
novo. There is also evidence of a physiological connection between the endothelial-lined vasculature and
the VM channel (Fig. 11.2a). In other words, VM is
defined by the unique ability of aggressive tumor cells
to express an endothelial cell (EC) phenotype and to
form vessel-like networks in three-dimensional culture
95
(Fig. 11.3a), mimicking the pattern of embryonic

vascular networks and recapitulating the patterned networks seen in patients aggressive tumors correlating
with a poor prognosis (Maniotis et al., 1999). There
are two distinctive VM types that have been reported in
tumors. (1) The patterned matrix type VM: the channels are composed of a basement membrane, lined
by tumor cells in their external superficies and no
endothelial cells are found on their inner wall despite
blood plasma and red blood cells flowing through
the channels (Maniotis et al., 1999). (2) The tubular type VM: Non-endothelial cell-lined channels and
only tumor cells as the lining cells of their luminal surface (Tmr and Tth, 2000; Shirakawa et al.,
2002a; Folberg and Maniotis, 2004; Van der Schaft
et al., 2005). Morphologically, because VM is PASpositive channel without endothelial cells, CD34 (or
other blood vessel endothelial molecular markers such
as CD31, CD105 etc.) and periodic acid-Schiff (PAS)
dual staining can be used as histological identification
of VM.
Vasculogenic mimicry has been observed in several
human tumors: breast, prostate, ovarian, chorio-, and
lung carcinomas; synovial-, rhabdomyo-, and Ewing
sarcomas; and phaeochromocytoma (Hendrix et al.,
2003). We previously (Yue and Chen, 2005) and subsequently EI Hallani et al. (2010) reported the presence
of VM in high grade gliomas. We used CD34 staining to identify the endothelium in glioblastoma tissue
sections and PAS staining to determine the basement
membrane of tumor blood vessels. Real tumor vessels
showed positive reaction for CD34 on their luminal surface and PAS-positive reaction in their walls.
However, in a subset of gliomas, PAS-positive tubular structures containing red blood cells but lined by
CD34 cells in the luminal surface, i.e. VM, could be
clearly found (Fig. 11.1d, f). Meanwhile, VM was also
found in human glioma cell-line xenografts (Niclou
et al., 2008).
As described by Folberg et al. (1993), seven morphological patterns of PAS-positive channels were
identified in uveal melanoma: straight channels,
arrangements of parallel straight channels, straight
channels that cross-link, arcs (incompletely closed
loops), arcs with branching, closed loops, and networks (networks were defined arbitrarily as at least
three back-to-back closed PAS-positive loops). All
these VM paterns have been observed in gliomas
(Fig. 11.1c, e).
96
Fig. 11.2 Connection between the endothelial-lined vasculature and the VM. (a), Longitudinally sectioned blood vessel
presents distinctive CD34+ and CD34 portions. (Cite from EI
Hallani et al., 2010). (b), Injected via the jugular vein, absorbite
particles could be seen at the hGSC/hGPC-derived intracranial
tumor vessels of various sizes. The inner lumens of these tumor

vessels, composed of transdifferentiated hGSPCs, were irregular
and discontinuous, and absorbite particles aggregated with red
blood cells. (Cite from Dong et al., 2010). (A, CD34 and PAS
dual staining 400; B, H & E staining 400)
Fig. 11.3 Vessel-like networks in three-dimensional culture. (a) Highly aggressive glioma cell line SKMG-4 forms tubular
structures in 3D culture on Matrigel while another glioma cell line SHG44 does not (b). (a and b, 100)
Biological and Clinical Significance of VM (1999) demonstrated that the patterned channels genVasculogenic mimicry is a special biological feature of tumor cells to form a PAS-positive pattern. Morphological analysis showed that PAS-positive
patterned networks, which were found in aggressive tumors, seemed to converge with blood vessels.
It was thus proposed that some types of anastomosis occur between the tumor-cell-lined networks
and the endothelium-lined vasculature, which contributes to the accumulation of erythrocytes in the
network infrastructure (Fig. 11.2). Maniotis et al.
erated by aggressive uveal melanoma cells in vitro

were capable of conducting dye over a short distance. He suggested that VM PAS-positive pattern in
uveal melanoma was indeed a form of tumor microcirculation. Using a mouse xenograft glioma model,
Dong et al. (2010) did functional assays to see if
these VM channels connecting with real blood vessels. They found that vessels from transdifferentiated human glioma stem cells (hGSCs) in intracranial
tumors can conduct absorbite particles which were
injected through the peripheral vein (Fig. 11.2b). These
observations strongly suggested that hGSCs might
generate vascular channels that facilitate tumor perfusion independent from tumor angiogenesis. Yue et al.
(2000) did bloodbrain barrier (BBB) ultrastructure
observation of 18 gliomas and two brain-metastases
originating from lung cancer under transmission electron microscopy. They found the perfect tight conjunction between endothelial cells: complete basement
membrane. Endothelial fenestrae, intercellular gaps,
and vesiculovacuolar organelles could not be found in
18 gliomas; however, quite a few endothelial fenestrae
and vesiculovacuolar organelles could be easily found
in the two brain metastases. The BBB ultra-structure
observation of these astrocytomas does not suggest
the extravasation of erythrocytes through the abnormal endothelial cell and the secondary PAS- positive
pattern formation after intra-tumor micro-hemorrhage
formation. In our previous study (Yue and Chen, 2005),
erythrocyte shadows could be easily found in the
PAS-positive CD34-negative channels of gliomas, supporting that the PAS-positive pattern in gliomas was
microcirculation.
Studies involving a combination of intravenous
tracers, together with confocal and immuno-electron
microscopy, have shown that fluid can be conducted by
the endothelium-lined vasculature, as well as extravascularly along the channel-like spaces created by the
PAS-positive patterned loops and networks that encase
clusters of tumor cells (Potgens et al., 1996; Clarijs
et al., 2002; Maniotis et al., 2002). Laser-captured
micro-dissection was performed in regions that exhibited VM without endothelial cells, central necrosis, or
fibrosis. The results revealed that there is a connection
between VM and angiogenesis.
There are several possible explanations for the functional relevance of VM. The fluid-conducting meshwork might provide a site for nutritional exchange
for aggressive tumors, and might therefore prevent
necrosis of the tumor. Alternatively, it might be analogous to an oedematous inflammatory response, in
which increased blood pressure leads to the escape
of fluid along connective-tissue pathways in intratissue spaces. The complex geometry of the laminincontaining extracellular matrix (ECM) covering that
encases the spheroidal clusters of tumor cells could
also form a suppressive shield against immune surveillance.
Vasculogenic mimicry is associated with poor clinical prognosis in tumor patients. The unique structure of
VM channels facilitates the metastasis of glioma cells.
97
Glioma cells, which line the inner surface of VM channels, are directly exposed to blood flow. Glioma cells
that leak out can migrate through the blood stream and
metastasize to other regions. Furthermore, glioma cells
that line the VM channel are highly malignant, poorly
differentiated, and have high plasticity. These cells can
degrade adjacent connective tissue and penetrate the
basement membrane of blood vessels by secreting proteins that mediate tumor invasion and metastasis. This
phenomenon has been confirmed in liver cancer, breast
cancer, hepatocellular carcinoma and gastrointestinal
stromal tumors (Shirakawa et al., 2002b; Folberg and
Maniotis, 2004; Guzman et al., 2007; Sun et al., 2006,
2008).
We performed a retrospective analysis on 101
glioma patients. The tumor samples were dual stained
for CD34 as well as PAS, and immunohistochemical staining (IHC) for Ki-67, COX-2 and MMP-9.
The association between VM and clinical characteristics was analyzed. The VM were revealed in 13 of
101 samples. The higher grade gliomas had a higher
incidence of VM than that of lower grade gliomas.
VM channels were associated with the expression of
COX-2 and MMP-9. There was no association between
the existence of VM and the sex, age and preoperative epilepsy of the patients or expression of Ki-67.
However, the patients with VM positive tumors survived a shorter period of time than those without it
(Fig. 11.4). Interestingly, Our results showed that the
micro-vascular density (MVD) was comparably less in
VM positive tumors than in VM negative tumors in
high-grade gliomas. Thus, the existence of VM may
provide a complementation to ensure tumor blood supply without involvement of endothelial cells, and may
serve as another mechanism for obtaining nutrients to
survive, especially in regions of the gliomas with less
MVD.
A follow-up blinded study that involved tissue analysis of ovarian carcinoma by two independent pathologists has also shown a strong clinical correlation
between the presence of VM, advanced-stage disease
and poor outcome.
There is a consensus that the microcirculation of
aggressive tumors is complex, and depending on the
time of observation, could consist of mosaic vessels
(which consist of both tumor cells and endothelial
cells) (Chang et al., 2000), co-opted vessels (Dme
et al., 2002) and/or angiogenic vessels (Kerbel, 2000).
There is also strong evidence for the existence of
98
Fig. 11.4 Patients whose tumor had vasculogenic mimicry survived shorter than those without it (P = 0.027) by Kaplan-Meier
survival analysis
an intratumoral, tumor-cell-lined, ECM-rich, patterned

network that can provide an extra-vascular fluid pathway, now known as VM (Maniotis et al., 2002; Clarijs
et al., 2002). The entire microcirculation in aggressive tumors seems to be made up of a combination of
these elements, and it is the result of destructive tumor
growth and remodeling at the same time associated
with high tumor grade and poorer prognosis. Zhang
et al. (2006) proposed a three-stage phenomenon
among vasculogenic mimicry channels, mosaic blood
vessels, and endothelium-dependent blood vessels,
wherein all three patterns participate in tumor blood
supply.
Molecular Mechanisms Underlying VM

There are three factors governing VM channel formation: plasticity of highly malignant tumor cells, remodeling of the ECM, and the connection of VM and host
microcirculation. Many studies have found that VM
can mainly be observed in highly malignant tumors.
A cDNA microarray study of 6000 genes from over
30 human cutaneous melanoma cell lines revealed that

there was a differential expression in 92 genes (2.5fold), including some genes associated with the phenotypes of endothelial and hematopoietic stem cells
(Seftor et al., 2002). A representative example of the
confirmation of selected differentially expressed genes
(in highly aggressive cutaneous melanoma C8161 cells
versus poorly aggressive C81-61 cells isolated from
the same patient) from the microarray analysis was
also determined by semiquantitative RT-PCR measurement. Hendrix et al. (2000) proposed that tumor
cells with an embryonic phenotype have high plasticity. Extracellular matrix remodeling provides the space
needed for VM and is associated with the MMP secretion by tumor cells. In the absence of blood supply
from endothelium-dependent vessels, tumor cells form
VM channels to acquire enough blood supply for their
sustenance.
Vasculogenic mimicry in aggressive tumors
involves several signaling molecules that are also
involved in embryonic vasculogenesis, including
vascular endothelial (VE)-cadherin, erythropoietinproducing hepatocellular carcinoma-A2 (EPHA2),
phosphatidylinositol 3-kinase, focal adhesion kinase,
matrix metaloproteinases and laminin 52-chain.

Studies (Hess et al., 2001, 2006) have indicated that
the main signalling cascade of VM is initiated by
transcriptional up-regulation of VE-cadherin through
HIF-2. VE-cadherin induces a re-localization of
EphA2 to the cell membrane where both proteins
then co-localize resulting in the phosphorylation of
EphA2. Activated EphA2 can phosphorylate FAK,
which results in activation of ERK1/2 and subsequently PI3K. However, PI3K can also be activated by
VE-cadherin/EphA2 signaling independent of FAK.
Active PI3K regulates the transition of pro-MT1-MMP
to active MT-MMP, which subsequently activates proMMP2. Both MT1-MMP and MMP-2 promote the
cleavage of laminin 52 to the pro-migratory fragments 52 and 52x. The release of these fragments
in the extracellular microenvironment may eventually
result in the formation of VM networks. Galectin-3
(Gal-3) positively influences this process by repression of the EGR-1 transcription factor, which causes
transcriptional repression of VE-cadherin and IL-8.
The latter stimulates the expression of MMP-2. Cyclic
AMP can also influence VM network formation in
multiple ways. By stimulation of Epac/Rap1, cAMP
inhibits tube formation. Additionally, it inhibits
ERK1/2, thereby manipulating the main signaling
cascade. However, the stimulatory effect of cAMP
on (Notch-1 and) Notch-4 may cause increased
expression of Nodal, which subsequently positively
regulates VE-cadherin expression. By this latter effect
cAMP may also have a positive effect on VM network
formation (Paulis et al., 2010). However, the detailed
molecular mechanisms for VM are not yet fully
understood.
99
inducing cell death. A great deal of intellectual capital has been devoted to targeting angiogenesis and
lymph-angiogenesis in patients with cancer. The heterogeneity of the tumor vasculature presents an opportunity, as well as a clinical challenge, aside from
issues of drug resistance. Recent in vitro studies have
shown that endostatin inhibits endothelial-cell driven
angiogenesis, but not the formation of melanoma-cell
vascular networks.
Scientists are now focusing on VM. Many investigators involved in basic research on VM are trying to find
an anti-VM therapy. Therapies targeting VM have been
performed in vitro. Suppressing tyrosine kinase activity and knocking out EphA2 gene, downregulating
VE-cadherin, using antibodies against human MMPs
and the laminin 52 chain, and using anti-PI3K therapy
are strategies used to inhibit VM. Chemically modified tetracycline COL-3 can suppress expression of
VM-related genes and reduce VM channel formation.
Pharmacological studies have shown that COL-3 may
depress MMP activity. Hwu (2000) used thalidomide
as an antitumor agent and found that it has antiangiogenesis and other biological effects. In clinical phase
I and phase II trials, thalidomide in combination with
temozolomide has a remarkable antimelanoma effect,
especially in cases of metastatic melanoma in the brain.
However clinical benefit from anti-VM therapy has
not yet been determined. Successful management of
gliomas and other aggressive cancers could involve the
targeting of one or more stages in the vasculogenic
mimicry signaling cascade and/or targets from both
vasculogenic mimicry and angiogenesis.
Advances and Challenges

Clinical Therapeutic Value of VM
Folkman (1971) reported that tumors require a blood
supply for survival, growth, and metastasis, he first
argued that anti-angiogenic therapy would have significant efficacy in tumors. Over the last few years, antiangiogenesis therapy has been emerging as a promising strategy in the treatment of cancer. Treatment
modalities are aimed at inhibiting endothelial proliferation and/or migration of endothelial cells to
hypoxic tumor regions, thereby hampering the supply of oxygen and nutrition to tumor cells and
Vasculogenic mimicry (VM), a newly discovered way

of blood supply to tumors, has attracted the attention
of many researchers, but studies on VM are just at
an early stage, many phenomena unique to VM channel formation remain to be elucidated. As a functional
tumor microcirculation, VM channels need to be studied with regard to their connection with endothelium
dependent vessels, their relationship with lymphatic
tubes, and their dual function as vessels and lymphatic
tubes.
Cells capable of VM display a high degree of
plasticity, causing them to resemble dedifferentiated
100
cell types. The most well known dedifferentiated cell

is the stem cell, as it holds the capacity to generate various novel cell types. Cancer stem cells have been identified in brain tumors, including glioblastomas (Singh
et al., 2003; Galli et al., 2004; Yuan et al., 2004). Like
physiological stem cells, glioblastoma stem cells are
capable of self-renewal and differentiation, and indefinite proliferation, the latter being critical for tumor
growth. Dong et al. (2010) reported that GSCs were
involved in tumor tissue remodeling in a xenograft
model. However, the relationship of glioblastoma stem
cells and vasculogenic mimicry in glioblastomas is still
unclear and the vasculogenic capacity of glioblastoma
tumor stem cells has yet to been proven.
References
Chang YS, di Tomaso E, McDonald DM, Jones R, Jain RK,
Munn LL (2000) Mosaic blood vessels in tumors: frequency
of cancer cells in contact with flowing blood. Proc Natl Acad
Sci USA 97:1460814613
Clarijs R, Otte-Holler I, Ruiter DJ, de Waal RMW (2002)
Presence of a fluid-conducting meshwork in xenografted
cutaneous and primary human uveal melanoma. Invest
Ophthalmol Vis Sci 43:912918
Dme B, Paku S, Somlai B, Tmr J (2002) Vascularization
of cutaneous melanoma involves vessel co-option and has
clinical significance. J Pathol 197:355362
Dong J, Zhang QB, Huang Q, Chen H, Shen YT, Fei XF, Zhang
TY, Diao Y, Wu ZC, Qin ZH, Lan Q, Gu XS (2010) Glioma
stem cells involved in tumor tissue remodeling in a xenograft
model. J Neurosurg 113:249260
EI Hallani S, Boisselier B, Peglion F, Rousseau A, Colin C,
Idbaih A, Marie Y, Mokhtari K, Thomas JL, Eichmann A,
Delattre JY, Maniotis AJ, Sanson M (2010) A new alternative mechanism in glioblastoma vascularization: tubular
vasculogenic mimicry. Brain 133:973982
Folberg R, Maniotis AJ (2004) Vasculogenic mimicry. APMIS
112:508525
Folberg R, Rummelt V, Parys-Van Ginderdeuren R, Hwang T,
Woolson RF, Peer J, Gruman LM (1993) The prognostic
value of tumor blood vessel morphology in primary uveal
melanoma. Ophthalmology 100:13891398
Folkman J (1971) Tumor angiogenesis: therapeutic implications.
N Engl J Med 285:11821186
Galli R, Binda E, Orfanelli U, Cipelletti B, Gritti A, De Vitis S,
Fiocco R, Foroni C, Dimeco F, Vescovi A (2004) Isolation
and characterization of tumorigenic, stem-like neural precursors from human glioblastoma. Cancer Res 64:70117021
Guzman G, Cotler SJ, Lin AY, Maniotis AJ, Folberg R (2007)
A pilot study of vasculogenic mimicry immunohistochemical expression in hepatocellular carcinoma. Arch Pathol Lab
Med 131:17761781
Hendrix MJSE, Kirschmann DA, Seftor RE (2000) Molecular
biology of breast cancer metastasis. Molecular expression of

vascular markers by aggressive breast cancer cells. Breast
Cancer Res 2:417422
Hendrix MJC, Seftor EA, Hess AR, Seftor REB (2003)
Vasculogenic mimicry and tumor-cell plasticity: lessons
from melanoma. Nat Rev Cancer 3:411421
Hess AR, Seftor EA, Gardner LM, Carles-Kinch K, Schneider
GB, Seftor RE, Kinch MS, Hendrix MJ (2001) Molecular
regulation of tumor cell vasculogenic mimicry by tyrosine
phosphorylation: role of epithelial cell kinase (Eck/EphA2).
Hess AR, Seftor EA, Gruman LM, Kinch MS, Seftor RE,
Hendrix MJ (2006) VE-cadherin regulates EphA2 in aggressive melanoma cells through a novel signaling pathway:
implications for vasculogenic mimicry. Cancer Biol Ther
5:228233
Hwu WJ (2000) New approaches in the treatment of
metastatic melanoma: thalidomide and temozolomide.
Oncology (Williston Park) 14(12 Suppl 13):2528
Kerbel RS (2000) Tumor angiogenesis: past, present and the near
future. Carcinogenesis 21:505515
Kreisl TN, Kim L, Moore K, Duic P, Royce C, Stroud I,
Garren N, Mackey M, Butman JA, Camphausen K, Park J,
Albert PS, Fine HA (2009) Phase II trial of single-agent
bevacizumab followed by bevacizumab plus irinotecan at
tumor progression in recurrent glioblastoma. J Clin Oncol
27:740745
Jouvet A, Scheithauer BW, Kleihues P (2007) 2007 WHO
classification of tumours of the central nervous system. Acta
Maniotis AJ, Chen X, Garcia C, DeChristopher PJ, Wu D, Peer
J, Folberg R (2002) Control of melanoma morphogenesis,
endothelial survival, and perfusion by extracellular matrix.
Lab Invest 82:10311043
Maniotis AJ, Folberg R, Hess A, Seftor EA, Gardner LM, Peer
J, Trent JM, Meltzer PS, Hendrix MJ (1999) Vascular channel formation by human melanoma cells in vivo and in vitro:
vasculogenic mimicry. Am J Pathol 155:739752
Niclou SP, Danzeisen C, Eikesdal HP, Wiig H, Brons NH, Poli
AM, Svendsen A, Torsvik A, Enger P, Terzis JA, Bjerkvig
R (2008) A novel eGFP-expressing immunodeficient mouse
model to study tumorhost interactions. FASEB J 22:
31203128
Paulis YW, Soetekouw PM, Verheul HM, Tjan-Heijnen VC,
Griffioen AW (2010) Signalling pathways in vasculogenic
mimicry. Biochim Biophys Acta 1806:1828
Potgens AJG, van Altena MC, Lubsen NH, Ruiter DJ, de
Waal RMW (1996) Analysis of the tumor vasculature and
metastatic behavior of xenografts of human melanoma cell
lines transfected with vascular permeability factor. Am J
Pathol 148:12031217
Seftor EAMP, Schatteman GC, Gruman LM, Hess AR,
Kirschmann DA, Seftor RE, Hendrix MJ (2002) Expression
of multiple molecular phenotypes by aggressive melanoma
tumor cells: role in vasculogenic mimicry. Crit Rev Oncol
Hematol 44:1727
Shirakawa K, Kobayashi H, Heike Y, Kawamoto S, Brechbiel
MW, Kasumi F, Iwanaga T, Konishi F, Terada M, Wakasugi
H (2002a) Hemodynamics in vasculogenic mimicry and
angiogenesis of inflammatory breast cancer xenograft.

Shirakawa K, Wakasugi H, Heike Y, Watanabe I, Yamada
S, Saito K, Konishi F (2002b) Vasculogenic mimicry and
pseudo-comedo formation in breast cancer. Int J Cancer
99:821828
Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire
J, Dirks PB (2003) Identification of a cancer stem cell in
human brain tumors. Cancer Res 63:58215828
Sun B, Qie S, Zhang S, Sun T, Zhao X, Gao S, Ni C, Wang
X, Liu Y, Zhang L (2008) Role and mechanism of vasculogenic mimicry in gastrointestinal stromal tumors. Hum
Pathol 39:444451
Sun B, Zhang S, Zhang D, Du J, Guo H, Zhao X, Zhang
W, Hao X (2006) Vasculogenic mimicry is associated with
high tumor grade, invasion and metastasis, and short survival
in patients with hepatocellular carcinoma. Oncol Rep 16:
693698
Tmr J, Tth J (2000) Tumor sinuses-vascular channels. Pathol
Oncol Res 6:8386
van der Schaft DW, Hillen F, Pauwels P, Kirschmann DA,
Castermans K, Egbrink MG, Tran MG, Sciot R, Hauben
E, Hogendoorn PC, Delattre O, Maxwell PH, Hendrix MJ,
101
Griffioen AW (2005) Tumor cell plasticity in Ewing sarcoma,
an alternative circulatory system stimulated by hypoxia.
Cancer Res 65:1152011528
Vredenburgh JJ, Desjardins A, Herndon JE 2nd, Dowell
JM, Reardon DA, Quinn JA, Rich JN, Sathornsumetee
S, Gururangan S, Wagner M, Bigner DD, Friedman AH,
Friedman HS (2007) Phase II trial of bevacizumab and
irinotecan in recurrent malignant glioma. Clin Cancer Res
13:12531259
Yuan X, Curtin J, Xiong Y, Liu G, Waschsmann-Hogiu S,
Farkas DL, Black KL, Yu JS (2004) Isolation of cancer
stem cells from adult glioblastoma multiforme. Oncogene
23:93929400
Yue WY, Chen ZP (2005) Does vasculogenic mimicry exist in
astrocytoma? J Histochem Cytochem 53:9971002
Yue WY, Zhong ST, Liang HZ (2000) The BBB morphology features of brain tumor. Pract J Cancer (Chinese) 15:
273275
Zhang S, Guo H, Zhang D, Zhang W, Zhao X, Ren Z, Sun
B (2006) Microcirculation patterns in different stages of
melanoma growth. Oncol Rep 15:1520
Chapter 12
Newly Diagnosed Glioma: Diagnosis Using Positron

Emission Tomography with Methionine
and Fluorothymidine
Nobuyuki Kawai, Yoshihiro Nishiyama, and Takashi Tamiya
Abstract The purpose of this prospective study was

to clarify the individual and combined role of METPET and FLT-PET in tumor detection, noninvasive
grading, and assessment of the cellular proliferation
rate in newly diagnosed histologically verified gliomas
of different grades. Sixty-two patients with newly
diagnosed gliomas were investigated with MET-PET
before surgery. Thirty-six patients were also examined
with FLT-PET. MET and FLT uptakes were assessed
by the standardized uptake value of the tumor showing the maximum uptake (SUVmax), and the ratio to
uptake in the normal brain parenchyma (T/N ratio).
All tumors were graded by the WHO grading system
using surgical specimens and the proliferation activity
of the tumors was determined by measuring the Ki-67
index obtained by immunohistochemical staining. On
visual analysis, MET exhibited a slightly higher sensitivity (88.7%) in tumor detection than FLT (86.1%)
and both tracers were 100% sensitive for malignant
gliomas. Low-grade gliomas that were false negative
on MET-PET were also false negative on FLT-PET.
Although the difference in the MET SUVmax and
T/N ratio between grade II and grade III gliomas was
statistically significant (P < 0.05), there was a significant overlap of MET uptake in the tumors. The
difference in the MET SUVmax and T/N ratio between
grade III and grade IV gliomas was not statistically
significant. Low-grade gliomas with oligodendroglial
components had relatively high MET uptake compared
to grade II astrocytoma. The difference in the FLT
N. Kawai ()
Department of Neurological Surgery, Faculty of Medicine,
Kagawa University, Kita-gun, Kagawa 761-0793, Japan
e-mail: nobu@kms.ac.jp
SUVmax and T/N ratio between grade III and grade

IV gliomas was statistically significant (P < 0.001).
However, the difference in the FLT SUVmax and T/N
ratio between grade II and grade III gliomas was not
statistically significant. Grade III gliomas with noncontrast enhancement on MR images had very low
FLT uptake. In 36 patients who underwent PET examination with both tracers, a significant but relatively
weak correlation was observed between the individual SUV max of MET and FLT (r = 0.52, P < 0.05)
and T/N ratio of MET and FLT (r = 0.49, P < 0.05).
FLT SUVmax in the tumor had a higher correlation
(r = 0.81, P < 0.001) with the Ki-67 proliferation index
than MET SUVmax (r = 0.40, P < 0.05). PET studies
using MET and FLT are useful for tumor detection
in newly diagnosed gliomas. However, there is no
complimentary information in tumor detection with
simultaneous measurements of MET- and FLT-PET in
low grade gliomas. FLT-PET seems to be superior to
MET-PET in noninvasive tumor grading and assessment of proliferation activity in gliomas of different
grade.
Keywords Glioma Positron Emission Tomography
Methionine Fluorothymidine MET FLT
Introduction
Morphological imaging using magnetic resonance
imaging (MRI) with gadolinium (Gd) contrast
enhancement is the most widely used method for
diagnosis of gliomas. This technique has the drawback, however, of frequent false negatives in tumor
regions without blood-brain barrier (BBB) breakdown.

103
104
The contrast enhancement only reflects regions

of BBB breakdown, whereas glioma cells usually
infiltrate the brain structure without significant BBB
breakdown. Positron emission tomography (PET) with
an appropriate metabolic tracer plays an important
role in characterizing and delineating gliomas. Among
several metabolic tracers, the glucose analogue FDG
(2-deoxy-2-[18 F]fluoro-D-glucose) is most commonly
used to measure glucose metabolism, and most malignant tumors show high FDG uptake due to increased
FDG uptake and accelerated phosphorylation of FDG
by hexokinase in malignant cells. One problem with
the use of FDG for the diagnosis of gliomas is the
high background uptake of FDG in glucose-dependent
brain tissue. Thus, accurate evaluation of gliomas with
FDG-PET is difficult especially in low grade gliomas.
The rate of protein synthesis is increased in
proliferating brain tumors and increased uptake of
labeled amino acid has been demonstrated in gliomas.
The PET examination with L-methyl-11 C-methionine
(MET) has been reported to delineate gliomas more
accurately than CT or MRI (Herholz et al., 1998),
and is generally considered to provide the most reliable images for grading glioma (Ceyssens et al.,
2006; Hatakeyama et al., 2008; Kaschten et al., 1998).
MET readily crosses the intact BBB via the neutral amino acid transporter system and is incorporated into the active tumor cells. MET-PET is also a
well-established imaging tool for prediction of prognosis (Kaschten et al., 1998; Ribom et al., 2001;
Van Laere et al., 2005), evaluation of response to
treatment (Galldiks et al., 2006; Nariai et al., 2005;
Nuutinen et al., 2000; Ribom et al., 2002), malignant
progression (Ullrich et al., 2009), and differentiation between tumor recurrence and radiation necrosis
in patients with gliomas (Tsuyuguchi et al., 2004).
However, increased MET uptake in non-neoplastic
brain lesions including inflammation, infarction, or
hemorrhage may cause false-positive. Also, the short
half-life of 11 C (20 min) and rapid in vivo degradation
make MET-PET less useful for routine clinical use.
3 -deoxy-3 -[18 F]fluorothymidine (FLT), a fluorinated thymidine analog, has emerged as a promising
PET tracer for evaluating tumor proliferating activity in various malignant brain tumors (Chen et al.,
2005; Choi et al., 2005; Hatakeyama et al., 2008;
Saga et al., 2006). FLT is phosphorylated by thymidine kinase-1 (TK1), a principle enzyme in the salvage
pathway of DNA synthesis, which is trapped inside
the cells. Phosphorylated FLT appears resistant to
N. Kawai et al.
degradation and is suitable for imaging with PET. The

application of FLT phosphorylation as a marker of cell
proliferation is based on the assumption that cellular
FLT trapping is a representation of thymidine incorporation into DNA (Been et al., 2004; Shiels et al.,
1998). As FLT uptake in the normal brain tissue is
very low, FLT-PET provides a low-background brain
image, and thus is considered to be an ideal PET tracer
for the imaging of brain tumors. FLT-PET has been
found useful for noninvasive grading of gliomas (Chen
et al., 2005; Choi et al., 2005; Hatakeyama et al., 2008;
Jacobs et al., 2005; Saga et al., 2006). In addition, FLTPET has been utilized in the prognostic assessment
(Chen et al., 2005, 2007) and evaluation of treatment
response (Chen et al., 2007) in malignant gliomas.
Jacobs et al. (2005) compared MET-PET and FLTPET findings and contrast enhanced MRI to determine DNA metabolism and amino acid uptake as well
as the integrity of the blood-brain barrier (BBB) in
23 patients with gliomas. They conclude that FLTPET provides information on the extent and activity
of a glioma additional to that provided by MET-PET
and MRI (Jacobs et al., 2005). However, the patients
used in the study were quite heterogeneous and 15 of
23 patients were previously treated cases. Treatment
with radiotherapy damages the integrity of the cerebral
endothelial cells (Cao et al., 2005). Endothelial cell
damage may increase the passive influx of the radioisotope through the disrupted BBB and mislead the
interpretation of PET findings. Therefore, the purpose
of this prospective study was to clarify the individual and combined role of MET-PET and FLT-PET in
tumor detection, noninvasive grading and assessment
of the cellular proliferation rate in 62 newly diagnosed
histologically verified gliomas of different grades.
Materials and Methods

Patients
Sixty-two patients with newly diagnosed gliomas
(33 men and 29 women; mean age, 49.8 18.3 year;
range, 289 year) were enrolled in this study. MET has
been available from the beginning of this prospective
study; however, FLT became available later. Therefore,
MET-PET was performed in all 62 patients and
FLT-PET was performed in only 36 of the patients. No
patients had previous treatment except steroid therapy
12 Newly Diagnosed Glioma
before PET. Histopathology was performed on tissue

specimens obtained by biopsy or resection. All tumors
were graded by the World Health Organization (WHO)
grading system (malignancy scale) for CNS tumors.
Tumor types and grades were distributed as follows:
WHO grade II astrocytoma (n = 13), WHO grade
II oligoastrocytoma (n = 3), WHO grade II oligodendroglioma (n = 1), WHO grade II ependymoma
(n = 1), WHO grade III astrocytoma (n = 16), WHO
grade III anaplastic oligodendroglioma (n = 2), WHO
grade III gliomatosis cerebri (n = 1), WHO grade
III anaplastic ependymoma (n = 1), WHO grade IV
glioblastoma (n = 23), WHO grade IV gliosarcoma
(n = 1). The cellular proliferation activity of the tumor
was determined by measuring the Ki-67 labeling index
obtained by immunohistochemical staining with antiKi-67/MIB-1. Immunostained slides were examined at
a high-power magnification (400). The percentage of
tumor cells that stained positively for Ki-67 antigen
was measured in the area containing the largest number
of positive tumor cells and was regarded as representative of the tumor proliferation activity. All patients
underwent contrast-enhanced MRI with gadoliniumdiethylenetriaminepentaacetic acid (Gd-DTPA) within
1 week of PET examination.
PET Examinations and Data Analysis

PET examinations were performed using an ECAT
EXACT HR+ scanner (Siemens/CTI, Knoxville, TN,
USA) in the three-dimensional acquisition mode. The
image system enabled the simultaneous acquisition of
51 transverse per field of view (FOV), with intersection spacing of 3 mm, for a total axial FOV of 15 cm.
The in-plane reconstructed resolution was 4.7 mm
full-width at half maximum in the brain FOV. In
most patients, PET examinations were performed several days before the surgery. When both tracers were
examined in one patient (n = 36), studies were not performed on the same day and the time interval between
the two examinations was usually 2472 h. No special
dietary instructions were given to the patients before
PET examination. Images were acquired with patients
in the supine position, resting, with their eyes closed.
Using 68 Ge rod sources rotating around the head, transmission images of the brain were obtained for 3 min
for MET-PET and for 5 min for FLT-PET. A dose of
113389 MBq (mean, 218 59 MBq) of 11 C-MET
105
or 116336 MBq (mean, 187 54 MBq) of 18 F-FLT

was injected intravenously. Regional emission images
of the brain were obtained for 5 min, beginning 10 min
after the 11 C-MET injection and for 10 min, beginning
40 min after the 18 F-FLT injection.
MET and FLT uptakes in the brain lesion were
semiquantitatively assessed by evaluating the standardized uptake value (SUV, activity concentration/injected
dose/body weight). A region of interest (ROI) was set
manually by an observer around the hottest area of
each lesion or its center located by MRI, if increased
MET and FLT uptake was absent. The maximum value
of SUV (SUVmax) was regarded as the representative
value of each tumor. To calculate the tumor-to-normal
tissue count density (T/N) ratios, ROI was set on the
normal brain parenchyma (usually contralateral normal cerebral tissue excluding ventricles) and the mean
value of SUV (SUVmean) was calculated. The T/N
ratio was determined by dividing the SUVmax of the
tumor with the SUVmean of the normal brain tissue.
A lesion was considered to be non-visualized if the
T/N ratio was < 1.2. This study was divided into three
parts. In the first part, the measured SUVmax and T/N
ratios of MET and FLT were compared with the histological diagnoses obtained by biopsy or resection.
In the second part, the relationship between MET and
FLT uptake in the tumor was evaluated in 18 patients
who underwent PET examination with both tracers.
Finally, the relationship between the tumor proliferation activity determined by Ki-67 labeling index and
tracer uptake was examined.
Statistical Analysis
All parametric data were expressed as mean SD.
Statistical analyses were performed to compare the
tumor grade and SUVmax and T/N ratios of each tracer
by using analysis of variance and post hoc comparisons
with Bonferroni correction. Linear regression analysis,
including Spearman rank correlation coefficient test,
was used to determine whether MET or FLT SUVmax
was related to the proliferative Ki-67 labeling index.
Linear regression analysis was also performed to evaluate the relationship between MET and FLT uptake
in the tumor. Sensitivity and specificity were calculated based on the PET data compared with the
subsequent pathology. Two-tailed probability values of
< 0.05 were considered statistically significant.
106
N. Kawai et al.
Results
Tumor Detection of High- and Low-Grade
Gliomas with MET and FLT
MET-PET detected all 44 high grade (III and IV)
gliomas by visual analysis (T/N ratios > 1.2). Seven of
18 low grade (II) gliomas (38.9%) that did not show
contrast enhancement on MRI with Gd-DTPA were
not detected in MET-PET. Thus, of all the 62 gliomas
studied, 55 (88.7%) were imaged by MET-PET. On
the other hand, FLT-PET detected all 27 high grade
gliomas. Five of 9 low grade gliomas (55.6%) were
not visually detected in FLT-PET. Thus, of all the 36
gliomas studied, 31 (86.1%) were imaged by FLTPET. Tumors that were false-negative on MET-PET
were also false-negative on FLT-PET. No tumors were
detected by FLT-PET only. By using both tracers, the
sensitivity rate in tumor detection was 88.9% (32 of
36 cases) in our series, because one false negative
low grade glioma with FLT had a positive finding in
MET-PET.
MET-PET and FLT-PET can detect all high-grade

gliomas (100% sensitivity).
MET-PET cannot detect about 40% of grade II
gliomas and FLT-PET cannot detect about 56% of
grade II gliomas.
MET-PET exhibits a slightly higher sensitivity in
tumor detection than FLT-PET.
Tracer Uptake and Tumor Grade

Normal brain parenchyma had low MET uptake with
the SUVmean of 1.46 0.33 (range, 0.932.62). The
average MET SUVmax in WHO grade II (n = 18),
III (n = 20) and IV (n = 24) gliomas was 2.80
1.37, 3.94 1.36 and 4.54 1.31, respectively, and
the average MET T/N ratio in grade II, III and IV
gliomas was 1.83 0.84, 2.81 0.98 and 3.31
0.94, respectively (Fig. 12.1a). There was a highly significant correlation between individual MET SUVmax
and T/N ratio (r = 0.89, P < 0.001). The difference
in the MET SUVmax and T/N ratio between grade II
and grade IV gliomas was statistically significant
10
FLT SUVmax
MET SUVmax
8
6
*
4
#
1
0
Grade II
Grade III
Grade II
Grade IV
18
16
14
FLT TN ratio
MET TN ratio
5
4
3
10
6
4
2
Grade III
(n = 20)
Grade IV
(n = 24)
Fig. 12.1 PET tracer uptake in different WHO tumor grade
1
Grade II
(n = 18)
Grade IV
12
Grade III
Grade II
(n = 9)
Grade III
(n = 9)
Grade IV
(n = 18)
107
(P < 0.001). Again, the difference in the MET

SUVmax and T/N ratio between grade II and grade III
gliomas was statistically significant ( : P < 0.05). The
difference in the MET SUVmax and T/N ratio between
grade III and grade IV gliomas was not statistically
significant. A scatterplot figure demonstrated that there
was a significant overlap of MET uptake in the tumors
despite the significant difference between grade II and
grade III gliomas (Fig. 12.1a). This is due to relatively
high MET uptake in grade II oligodendroglioma and
oligoastrocytoma (arrows in Fig. 12.1a) compared to
that in grade II astrocytoma.
Normal brain parenchyma had very weak FLT
uptake with the SUVmean of 0.19 0.04 (range, 0.11
0.28). The average FLT SUVmax in WHO grade II
(n = 9), III (n = 9) and IV (n = 18) gliomas was
0.27 0.08, 0.69 0.42 and 1.98 0.80, respectively, and the average FLT T/N ratio in grade II, III
and IV gliomas was 1.48 0.61, 3.87 2.11 and
10.22 3.79, respectively (Fig. 12.1b). There was a
highly significant correlation between individual FLT
SUVmax and T/N ratio (r = 0.93, P < 0.001). The differences in the FLT SUVmax and T/N ratio between
grade III and grade IV gliomas were statistically significant (#: P < 0.001). However, the differences in
the FLT SUVmax and T/N ratio between grade II
and grade III gliomas were not statistically significant probably due to relatively high variation of FLT
uptake in grade III gliomas. Four patients with grade III
anaplastic astrocytoma (Ki-67 index, 510%) had no

contrast enhancement on MR images with Gd-DTPA.
FLT uptake in these patients was similar to the low levels observed in grade II gliomas (SUVmax, 0.280.37)
(arrows in scatterplot, Fig. 12.1b).
MET and FLT uptakes in gliomas increase with
tumor grade.
Average MET uptake in grade III gliomas is significantly higher than that in grade II gliomas
(P < 0.05).
MET uptake in grade II oligodendroglioma and
oligoastrocytoma is relatively high compared to that
in grade II astrocytoma.
Average FLT uptake in grade IV gliomas is significantly higher than that in grade III gliomas
(P < 0.001).
FLT uptake in non-contrast enhanced grade III
gliomas is relatively low compared to that in contrast enhanced grade III gliomas.
MET and FLT Uptake in Individual Tumors

In 36 patients who underwent PET examination with
both tracers, a significant but relatively weak correlation was observed between the individual SUVmax
of MET and FLT (r = 0.52, P < 0.05) (Fig. 12.2a)
20
n = 36
16
14
FLT T/N ratio
FLT SUVmax
n = 36
18
12
10
8
6
r = 0.52, P < 0.05

0
1
MET SUVmax
Fig. 12.2 Correlation between MET uptake and FLT uptake
r = 0.49, P < 0.05
2
10
0
0
MET T/N ratio
108
and T/N ratio of MET and FLT (r = 0.49, P < 0.05)

(Fig. 12.2b). The most distant three points (arrows
in the figures) from the 95% confidence interval line,
which exhibited high FLT uptake with moderate MET
uptake, were glioblastoma cases with the highest Ki-67
indices (5065%) in this series.
A significant but weak correlation is observed
between the individual uptake of MET and FLT, and
it seems to be difficult to precisely predict one result
from the other.
Tracer Uptake and Tumor Proliferation

Activity
Histopathology was obtained from all 62 patients by
surgery after the PET study. In all tumors examined,
linear regression analysis showed a significant correlation between tracer uptake and tumor proliferation
activity determined by Ki-67 labeling index. The analysis indicated a more significant correlation of the
Ki-67 index with FLT SUVmax (r = 0.81, P < 0.001)
and T/N ratio (r = 0.91, P < 0.001) (Fig. 12.3b) than
with MET SUVmax (r = 0.40, P < 0.05) and T/N ratio
(r = 0.48, P < 0.001) (Fig. 12.3a).
A significant correlation is observed between tracer
uptake and tumor proliferation activity determined
by Ki-67 labeling index.
Ki-67 labeling index is more significantly correlated
with FLT uptake than that with MET uptake.
Discussion
Positron emission tomography (PET) with 2-deoxy2-[18 F]fluoro-D-glucose (FDG) is a well established
method in the diagnosis and management of patients
suffering from brain tumors. FDG uptake is generally
associated with histological tumor grade in gliomas.
FDG uptake in low-grade gliomas (which are mostly
grade II in adults) is usually similar to that of the normal white matter, whereas most grade III anaplastic
gliomas have an FDG uptake exceeding that of the
normal white matter or similar to that of the normal
gray matter. Untreated glioblastomas show high FDG
N. Kawai et al.
uptake, which might be higher than that of the normal

gray matter. One problem with the use of FDG for the
diagnosis of gliomas is the high background uptake of
FDG in glucose-dependent brain tissue. Thus, accurate evaluation of gliomas with FDG-PET is difficult
especially in low grade gliomas.
MET-PET possesses high specificity in tumor detection (Herholz et al., 1998), tumor delineation (Borbly
et al., 2006), and differentiation of benign from malignant lesions (Ogawa et al., 1995). Various studies have
found a correlation between tumor grade and MET
uptake in gliomas (Ceyssens et al., 2006; Hatakeyama
et al., 2008; Kaschten et al., 1998). Amino acids,
including MET readily cross the intact BBB through
neutral amino acid transporters (large amino acid transporter 1: LAT-1) and are incorporated into the area
with active tumor. Our recent study has demonstrated
that the grade of LAT1 immunostaining increases with
glioma grade and the expression of LAT1 is significantly correlated with MET uptake in newly diagnosed
gliomas (Okubo et al., 2010). MET uptake in gliomas
significantly correlates with cell proliferation, in vitro
Ki-67 labeling index (Hatakeyama et al., 2008; Kim
et al., 2005), and proliferating cell nuclear antigen
expression (Sato et al., 1999). Increased MET uptake
seems to be caused by increased carrier-mediated and
passive transport rather than elevated protein synthesis and is highly correlated with microvessel density
(angiogenesis) in gliomas (Kracht et al., 2003; Nojiri
et al., 2009). Many previous studies enrolled patients
with newly diagnosed and previously treated recurrent cases in their materials. Radiation therapy used
as an adjuvant therapy of gliomas can cause loosening
of the endothelial tight junction, vascular leakage, or
endothelial cell death and increase vascular permeability (Cao et al., 2005). Radiation could act to increase
vascular permeability not only in the BBB but also
in the blood-tumor barrier (BTB) (Cao et al., 2005),
and thus, potentially increasing passive, non carriermediated transport of MET through the endothelial
cells to tumor cells. In an earlier study on suspected
gliomas recurrence, there was no significant relationship between the primary tumor histopathology (grade)
and MET uptake ratio (Van Laere et al., 2005), in
contrast to the findings of this study performed in a
pre-therapeutic setting. Also, no significant differences
were found in the SUVmax and T/N ratio of MET
between recurrent malignant glioma and radionecrosis following stereotactic radiosurgery (Tsuyuguchi
b 80
80
n = 62
70
r = 0.40, P < 0.05
60
Ki-67 labeling index
109
50
40
30
20
10
0
n = 36
70
60
50
40
30
20
10
r = 0.81, P < 0.001
0
10
10
1
MET SUVmax
80
80
n = 62
70
r = 0.48, P < 0.001
60
50
40
30
20
10
FLT SUVmax
60
50
40
30
20
10
10
10
0
MET T/N ratio
n = 36
70
r = 0.91, P < 0.001
10
12
14
16
18
FLT T/N ratio
Fig. 12.3 Correlation between PET tracer uptake and Ki-67 index
et al., 2003). As for the mechanism of tracer uptake,

disruption of the BBB and BTB might be responsible for increased availability and transport, and such
rupture might also be present in low grade glioma
and radionecrosis without active tumor cells and mislead the interpretation of PET findings. In this study,
therefore, we only enrolled patients with newly diagnosed glioma to reconfirm the usefulness of MET- and
FLT-PET in tumor detection, noninvasive grading and
assessment of the proliferation rate.
Our study revealed a significant difference in the
MET SUVmax and T/N ratio between grade II and
grade III gliomas. Several studies have demonstrated
the impossibility of differentiation between grade II
and grade III gliomas with MET-PET only (Ceyssens
et al., 2006; Hatakeyama et al., 2008; Kaschten
et al., 1998). We found that 4 patients with grade II
oligodendroglioma and oligoastrocytoma had a higher
MET uptake (SUVmax, 3.456.8) (Fig. 12.1a, arrows
in scatterplot) than did grade II astrocytoma (average
SUVmax, 2.21 0.76, n = 13). Higher MET uptake

in grade II oligodendroglioma compared with grade II
astrocytoma has also been reported in a recent study
(Nojiri et al., 2009). This difference might reflect
specific metabolic properties of oligodendroglial cells
(such as myelin synthesis) and different cellular densities and rates of cell turnover of oligodendroglioma.
A recent study has also shown that an increase in the
microvessel area in oligodendroglioma may contribute
to the higher MET uptake among low-proliferative
grade II gliomas (Nojiri et al., 2009). Increased MET
uptake was observed not only in pure oligodendroglioma but also in mixed oligoastrocytoma (n = 3)
in our study. This finding is contrary to the finding of a
previous MET-PET study on increased MET uptake in
mixed oligoastrocytoma. This could be explained by
the relatively high oligodendroglial tumor components
in our cases. Our results indicate that there was a significant overlap of MET uptake in the tumors despite
a significant difference between grade II and grade III
110
gliomas. Therefore, MET-PET is not an ideal tool to

precisely predict tumor malignancy and a surrogate
for pathologic grading. Also, caution must be paid to
evaluate the tumors with oligodendroglial components
in MET-PET. Some studies have concluded that
MET-PET has more clinical usefulness in assessing
the tumor invasion (Kim et al., 2005; Ogawa et al.,
1995) but not the tumor grade. Although we have not
compared the MET-PET and MRI findings, a previous
study reported that the gadolinium-enhanced area on
MR images covered, on average, only 60% of the
MET accumulation area in malignant gliomas (Miwa
et al., 2004).
In this study we did not include the case with
pilocytic astrocytoma. Recently, Galldiks et al. (2009)
showed that pilocytic astrocytomas demonstrated relatively high MET uptake when comparison was made
with WHO grade II astrocytomas. The histopathological features of pilocytic astrocytoma include microvascular proliferation and infiltration, but are not signs of
malignancy in this tumor entity. MET transport may
be increased by an increased number of microvessels combined with a higher density of transporters
(LAT-1) in the endothelial cells in pilocytic astrocytomas (Okubo et al., 2010). Again, caution must be
paid to evaluate the tumors with pilocytic astrocytoma
in MET-PET.
FLT is a newly developed PET tracer, which allows
for noninvasive assessment of tumor proliferation
(Chen et al., 2005; Choi et al., 2005; Hatakeyama et al.,
2008; Saga et al., 2006). In contrast to MET, which
provides only an indirect measure of the proliferation
status as amino acid uptake in the tumor, FLT allows
the direct measurement of cellular TK1 activity, which
has been reported to be proportional to the proliferation activity of the tumor. In fact, our linear regression
analysis indicated a much more significant correlation of the Ki-67 labeling index with FLT uptake than
with MET uptake in gliomas. As uptake of FLT is
low in intact brain tissue, FLT-PET provides a lowbackground cerebral image, and is thus considered to
be an ideal and attractive PET tracer for the imaging of
gliomas. However, the sensitivity for the detection of
tumors with FLT-PET has been reported to be lower
compared with MET-PET, especially for low-grade
gliomas. This is the case in our small series, absolute uptake of FLT in all tumors examined is lower
than that of MET (SUVmax, 1.23 0.98 vs. 3.84
1.50), and the FLT uptake ratio to normal brain tissue
N. Kawai et al.
is higher than that of MET (T/N ratio, 6.45 4.84

vs. 2.72 1.10) because of extremely low uptake of
FLT in normal brain tissue. Although sensitivity for
tumor detection is lower for FLT-PET (86.1%) than
for MET-PET (88.7%) in our series, the difference
was not statistically significant. By using both tracers, the sensitivity rate in tumor detection was 88.9%
(32 of 36 cases) in our series, because one falsenegative low grade glioma with FLT had a positive
finding in MET-PET. Some studies have investigated
several parameters simultaneously and concluded that
different tracers, such as FDG and MET, can provide
different and complimentary information in individual
cases (Van Laere et al., 2005). However, low grade
gliomas that were false negative on MET-PET were
also false-negative on FLT-PET and no tumors were
detected by FLT-PET only in our series. Therefore,
there is no complimentary information in tumor detection with simultaneous measurements of FLT and MET
uptake in cases where the tumor metabolism is only
slightly increased (low grade gliomas) from that of the
normal brain tissue.
When comparison was made between individual
MET and FLT uptake in the tumor, there was a significant but relatively weak correlation between the
uptake of MET and FLT. The most distant three points
from the 95% confidence interval area, which exhibited high FLT uptake with moderate MET uptake, were
glioblastoma cases with a highly proliferative nature.
This could be explained by the relatively high rate of
metabolic trapping of FLT by TK1 compared to the
influx effect across the BBB in these patients. Kinetic
analysis of FLT uptake could clarify whether the transport effect or metabolic trapping largely contributes
to the increased accumulation of FLT in the tumor
(Jacobs et al., 2005).
The FLT-PET study revealed that the differences in
the SUVmax and T/N ratio between grade III and grade
IV gliomas were statistically significant, but the differences between grade II and grade III gliomas were
not statistically significant in our cases. This may have
been due to the relatively small number of grade III
gliomas and the relatively wide range of FLT uptake in
these tumors (T/N ratio, 1.527.10). Four patients with
grade III anaplastic astrocytoma (Ki-67 index, 510%)
had no contrast enhancement on MR images with
Gd-DTPA. FLT uptake was similar to the low levels
observed in normal brain tissue in these patients (T/N
ratio, 1.522.18, arrows in scatterplot, Fig. 12.1b).
These patients exhibited moderately increased MET

uptake in the tumor (T/N ratio, 2.833.27). This discrepancy might be due to the selectivity of TK1 as a
target of FLT for the salvage pathway of DNA synthesis. FLT accumulation is a reliable measure of the
salvage pathway of DNA synthesis (Schwartz et al.,
2003). In gliomas in which de novo DNA pathways
are predominantly used in pyrimidine biosynthesis,
the proliferation activity might be underestimated by
this tracer (Schwartz et al., 2003). Muzi et al. (2006)
reported a similar case in an FLT-PET study and
concluded that FLT may be less useful in assessing
proliferation in noncontrast-enhancing tumors regardless of histopathology grading. Low tracer access to the
tumor tissue may limit the metabolic trapping of FLT
by TK1 even in proliferative tumors.
In conclusion, we have shown here that PET studies using MET and FLT are useful for tumor detection
in 62 newly diagnosed histologically verified gliomas.
MET has a slightly higher sensitivity in tumor detection than FLT (but not statistically significant) in newly
diagnosed gliomas; and both tracers are 100% sensitive for malignant gliomas. However, there is no
complimentary information with simultaneous measurements of MET and FLT in tumor detection of
low grade gliomas. Our results reveal that MET-PET
and FLT-PET appear to be a valuable tool for noninvasive tumor grading in newly diagnosed gliomas,
however, caution must be paid to evaluate the tumors
with oligodendroglial components for MET-PET and
with noncontrast-enhancement for FLT-PET. FLT-PET
seems to be superior to MET-PET in assessment of
cellular proliferation activity in gliomas of different grade. This study could be adequately powered
to make clear-cut distinctions for glioma grading in
newly diagnosed gliomas when the sample size of FLTPET is appropriately increased. Further studies might
identify the individual role of MET and FLT in gliomas
of different types and grades.
References
Been LB, Suumeijer AJH, Cobben DCP, Jager PL, Hoekstra
HJ, Elsinga PH (2004) [18 F]FLT-PET in oncology: current
status and opportunities. Eur J Nucl Med Mol Imaging 31:
16591672
Borbly K, Nyry I, Tth M, Ericson K, Gulys B (2006)
Optimization of semi-quantification in metabolic PET
studies with 18 F-fluorodeoxyglucose and 11 C-methionine in
111
the determination of malignancy of gliomas. J Neurol Sci
246:8594
Cao Y, Tsien CI, Shen Z, Tatro DS, Ten Haken R, Kessier
ML, Chenevert TL, Lawrence TS (2005) Use of magnetic
resonance imaging to assess blood-brain/blood-gliomas barrier opening during conformal radiotherapy. J Clin Oncol
23:41274136
Ceyssens S, Van Laere K, de Groot T, Goffin J, Bormans G,
Mortelmans L (2006) [11 C]methionine PET, histopathology,
and survival in primary brain tumors and recurrence. AJNR
Chen W, Cloughesy T, Kamdar N, Satyamurthy N, Bergsneider
M, Liau L, Mischel P, Czernin J, Phelps ME, Silverman DHS
(2005) Imaging proliferation in brain tumors with 18 F-FLT
PET: comparison with 18 F-FDG. J Nucl Med 46:945952
Chen W, Delaloye S, Silverman DHS, Geist C, Czernin J, Sayre
J, Satyamurthy N, Pope W, Lai A, Phelps ME, Cloughesy T
(2007) Predicting treatment response of malignant gliomas
to bevacizumab and irinotecan by imaging proliferation with
[18 F] fluorothymidine positron emission tomography: a pilot
study. J Clin Oncol 25:47144721
Choi SJ, Kim JS, Kim JH, Oh SJ, Lee JG, Kim CJ, Ra
YS, Yeo JS, Ryu JS, Moon DH (2005) [18 F]3 -deoxy-3 fluorothymidine PET for the diagnosis and grading of brain
tumors. Eur J Nucl Med Mol Imaging 32:653659
Galldiks N, Kracht LW, Berthold F, Miletic H, Klein JC, Herholz
K, Jacobs AH, Heiss WD (2009) [11 C]-L-methionine
positron emission tomography in the management of children and young adults with brain tumors. J Neurooncol
96:231239
Galldiks N, Kracht LW, Burghaus L, Thomas A, Jacobs AH,
Heiss WD, Herholz K (2006) Use of 11 C-methionine PET to
monitor the effects of temozolomide chemotherapy in malignant gliomas. Eur J Nucl Med Mol Imaging 33:516524
Hatakeyama T, Kawai N, Nishiyama Y, Yamamoto Y, Sasakawa
Y, Ichikawa T, Tamiya T (2008) 11 C-methionine (MET) and
18 F-fluorothymidine (FLT) PET in patients with newly diagnosed glioma. Eur J Nucl Med Mol Imaging 35:20092017
Herholz K, Hlzer T, Bauer B, Schrder R, Voges J, Ernestus
RI, Mendoza G, Weber-Luxenburger G, Lttgen J, Thiel
A, Wienhard K, Heiss WD (1998) 11 C-methionine PET
for differential diagnosis of low-grade gliomas. Neurology
50:13161322
Jacobs AH, Thomas A, Kracht LW, Li H, Dittmar C, Garlip
G, Galldiks N, Klein JC, Sobesky J, Hilker R, Vollmar
S, Herholz K, Wienhard K, Heiss WD (2005) 18 Ffluoro-L-thymidine and 11 C-methylmethionine as markers of
increased transport and proliferation in brain tumors. J Nucl
Med 46:19481958
Kaschten B, Stevenaert A, Sadzot B, Deprez M, Degueldre
C, Fiore GD, Luxen A, Reznik M (1998) Preoperative
evaluation of 54 gliomas by PET with fluorine-18fluorodeoxyglucose and/or carbon-11-methionine. J Nucl
Med 39:778785
Kim S, Chung JK, Im SH, Jeong JM, Lee DS, Kim DG, Jung
HW, Lee MC (2005) 11 C-methionine PET as a prognostic
marker in patients with glioma: comparison with 18 F-FDG
PET. Eur J Nucl Med Mol Imaging 32:5259
Kracht LW, Friese M, Herholz K, Schroeder R, Bauer B, Jacobs
A, Heiss WD (2003) Methyl-[11 C]-L-methionine uptake as
measured by positron emission tomography correlates to
112
microvessel density in patients with glioma. Eur J Nucl Med
Miwa K, Shinoda J, Yano H, Okumura A, Iwama T, Nakahashi
T, Sakai N (2004) Discrepancy between lesion distributions
on methionine PET and MR images in patients with glioblastoma multiforme: insight from a PET and MR fusion image
study. J Neurol Neurosurg Psychiatry 75:14571462
Muzi M, Spence AM, OSullivan F, Mankoff DA, Wells JM,
Grierson JR, Link JM, Krohn KA (2006) Kinetic analysis
of 3 -deoxy-3 -18 F-fluorothymidine in patients with gliomas.
J Nucl Med 47:16121621
Nariai T, Tanaka Y, Wakimoto H, Aoyagi M, Tamaki M,
Ishikawa K, Senda M, Ishii K, Hirakawa K, Ohno K (2005)
Usefulness of L-[methyl-11 C] methionine-positron emission
tomography as a biological monitoring tool in the treatment
of glioma. J Neurosurg 103:498507
Nojiri T, Nariai T, Aoyagi M, Senda M, Ishii K, Ishiwata K,
Ohno K (2009) Contributions of biological tumor parameters to the incorporation rate of L-[methyl-11 C] methionine
into astrocytomas and oligodendrogliomas. J Neurooncol
93:233241
Nuutinen J, Sonninen P, Lehikoinen P, Sutinen E, Valavaara R,
Eronen E, Norrgrd S, Kulmala J, Ters M, Minn H (2000)
Radiotherapy treatment planning and long-term follow-up
with [11 C]methionine PET in patients with low-grade astrocytoma. Int J Radiat Oncol Biol Phys 48:4352
Ogawa T, Hatazawa J, Inugami A, Murakami M, Fujita H,
Shimosegawa E, Noguchi K, Okudera T, Kanno I, Uemura
K, Hadeishi H, Sasajima T (1995) Carbon-11-methionone
PET evaluation of intracerebral hematoma: distinguishing
neoplastic from non-neoplastic hematoma. J Nucl Med 36:
21752179
Okubo S, Zhen HN, Kawai N, Nishiyama Y, Haba R, Tamiya
T (2010) Correlation of L-methyl-11 C-methionine (MET)
uptake with L-type amino acid transporter 1 in human
Ribom D, Engler H, Blomquist E, Smits A (2002) Potential
significance of 11 C- methionine PET as a marker for the
radiosensitivity of low grade gliomas. Eur J Nucl Med
29:632640
Ribom D, Eriksson A, Hartman M, Engler H, Nilsson A,
Lngstrm B, Bolander H, Bergstrm M, Smits A
(2001) Positron emission tomography 11 C-methionine and
N. Kawai et al.
survival in patients with low-grade gliomas. Cancer 92:
25412549
Saga T, Kawashima H, Araki N, Takahashi JA, Nakashima Y,
Higashi T, Oya N, Murai T, Hojo M, Hashimoto N, Manabe
T, Hiraoka M, Togashi K (2006) Evaluation of primary brain
tumors with FLT-PET: usefulness and limitations. Clin Nucl
Med 31:774780
Sato N, Suzuki M, Kuwata N, Kuroda K, Wada T, Beppu T,
Sera K, Sasaki T, Ogawa A (1999) Evaluation of the malignancy of glioma using 11 C-methionine positron emission
tomography and proliferating cell nuclear antigen staining.
Neurosurg Rev 22:210214
Schwartz JL, Tamura Y, Jordan R, Grierson JR, Hrohn KA
(2003) Monitoring tumor cell proliferation by targeting DNA
synthetic processes with thymidine and thymidine analogs.
J Nucl Med 44:20272032
Shiels AF, Grierson JR, Dohmen BM, Machulla HJ, Stayanoff
JC, Lawhorn-Crews JM, Obradovich JE, Muzik O, Mangner
TJ (1998) Imaging proliferation in vivo with [F-18]FLT and
positron emission tomography. Nat Med 4:13341336
Tsuyuguchi N, Sunada I, Iwai Y, Yamanaka K, Tanaka K,
Takami T, Otsuka Y, Sakamoto S, Ohata K, Goto T, Hara M
(2003) Methionine positron emission tomography of recurrent metastatic brain tumor and radiation necrosis after
stereotactic radiosurgery: is a differential diagnosis possible?. J Neurosurg 98:10561064
Tsuyuguchi N, Takami T, Sunada I, Iwai Y, Yamanaka K, Tanaka
K, Nishikawa M, Ohata K, Torii K, Morino M, Nishio A,
Hara M (2004) Methionine positron emission tomography
for differentiation of recurrent brain tumor and radiation
necrosis after stereotactic radiosurgery -in malignant glioma. Ann Nucl Med 18:291296
Ullrich RT, Kracht L, Brunn A, Herholz K, Frommolt P, Miletic
H, Deckert M, Heiss WD, Jacobs AH (2009) Methyl-L11 C-methionine PET as a diagnostic marker for malignant progression in patients with glioma. J Nucl Med 50:
19621968
Van Laere K, Ceyssens S, Van Calenbergh F, de Groot T, Menten
J, Falmen P, Bormans G, Mortelmans L (2005) Direct comparison of 18 F-FDG and 11 C-methionine PET in suspected
recurrence of glioma: sensitivity, inter-observer variability
and prognostic value. Eur J Nucl Med Mol Imaging 32:3951
Chapter 13
Role of Diffusion Tensor Imaging in Differentiation

of Glioblastomas from Solitary Brain Metastases
Sumei Wang, Harish Poptani, Elias R. Melhem, and Sungheon Kim
Abstract Differentiation between glioblastomas and

solitary brain metastases is an important clinical problem as the treatment strategy can significantly differ
depending on the tumor type. However, due to their
similar appearance on conventional magnetic resonance imaging, accurate distinction between glioblastomas and brain metastases remains challenging and
often necessitates an invasive surgical biopsy for a
definitive diagnosis. Recently, diffusion tensor magnetic resonance imaging (DTI) has been applied in
solving this problem and has demonstrated potential
as a noninvasive imaging biomarker to differentiate
glioblastomas from solitary brain metastases. This
chapter provides a brief review on the fundamental
concepts of DTI and how this new imaging method can
be applied in differentiation between glioblastomas
and solitary brain metastases. It also includes discussions on some practical issue to further improve the
reliability of the method and potential metrics sensitive
to tumor that can be derived from DTI data.
Keywords Glioblastomas DTI Metastases
Neoplasms MRI ADC and FA
Introduction
Glioblastomas and brain metastases are the two most
common brain neoplasms in adults. The management
of these two neoplasms is vastly different and can
S. Kim ()
Department of Radiology, Center for Biomedical Imaging,
New York University School of Medicine, New York,
NY 10016, USA
e-mail: Sungheon.Kim@nyumc.org
potentially affect the clinical outcome (Giese and

Westphal, 2001; Soffietti et al., 2002). For example,
patients with glioblastomas almost always undergo
surgical resection (Giese and Westphal, 2001), while
patients with suspected brain metastases without clinical history of systemic cancer undergo a complicated
systemic staging to determine the site of primary carcinoma and evaluation for distant metastases before
any surgical intervention or medical therapy (Soffietti
et al., 2002). In some cases, clinical history and presence of multiple peripheral enhancing lesions in the
brain makes the diagnosis of brain metastases relatively straightforward. However, appearance of the
solitary brain metastases on magnetic resonance imaging (MRI) can be nonspecific. Similarly, glioblastomas
can also occasionally present as multiple peripheral
enhancing lesions. Although glioblastomas typically
present as a solitary mass, a solitary brain metastasis may be the first manifestation of disease in about
30% of patients with systemic cancer. Hence, accurate
distinction between glioblastomas and brain metastases remains challenging, which often necessitates an
invasive surgical biopsy for a definitive diagnosis.
Conventional MRI is limited in making an accurate
distinction between glioblastomas and brain metastases due to their similar appearance. Both neoplasms
may exhibit ring-enhancement and extensive edema.
Over the past few years, diffusion tensor imaging
(DTI) has been increasingly used to study pathologic changes in brain tumors (Murakami et al., 2009;
Yamasaki et al., 2005). DTI has also been investigated in differentiating glioblastomas from metastases
(Lu et al., 2004; Tsuchiya et al., 2005; Wang et al.,
2009). In this chapter, we will briefly explain the
DTI technique and its application in differentiating
glioblastomas from solitary brain metastases.

113
114
Pathophysiology of Brain Tumors

Glioblastoma is the most common and most malignant primary brain tumor. According to the World
Health Organization (WHO) classification, it is also
listed as Grade IV astrocytoma. Pathologically, it
is composed of highly cellular anaplastic cells with
marked nuclear atypia and mitotic activity, accompanied by marked regional heterogeneity and cellular
polymorphism as well as microvascular proliferation
and necrosis. Glioblastoma is associated with a high
proliferation rate (Ki-67 labeling about 1520%). As
glioblastomas are biologically aggressive tumors, they
tend to grow in an infiltrative manner, invading the
surrounding tissues, especially the white matter tracts
(Rees et al., 1996).
Brain metastases are tumors originating from other
organs of the body which spread into the brain via
hematogenous routes. Neoplastic cells from the primary site can break away and enter the bodys circulatory blood stream. These neoplastic cells circulate,
travel and take up residence in the brain. The malignant cells multiply in the endothelial cell layer and
grow into the brain parenchyma in a non-infiltrating
Fig. 13.1 (a) Geometric interpretation of the tensor using the

diffusion ellipsoid. The re-parameterized tensor obtained from
a set of diffusion-weighted measurements has principal, intermediate, and minor eigenvalues (1 , 2 , 3 ) and corresponding
eigenvectors (e1 , e2 , e3 ). These values define the relative magnitude and direction of diffusion along three spatial dimensions
within each voxel respectively. Depending on the underlying
probability of intravoxel diffusion, the tensor may take on one
of three cardinal shapes prolate (CL), spherical (CS), or oblate
S. Wang et al.
advancing pattern. The most common brain metastases

are from lung, breast and melanoma. Histological
appearance and cellularity of the brain metastases are
identical to those of the primary tumor (Zhang and
Olsson, 1997), and they may vary markedly among the
primary tumor types.
Methodology-Diffusion Tensor Imaging

Diffusion imaging probes the self-diffusional movement of water molecules, which can be detected by the
attenuation of MRI signal using a diffusion weighted
sequence. When unimpeded, water molecules move in
a random manner (isotropic diffusion). However, the
presence of obstacles, such as axonal membranes and
myelin sheaths in white matter fiber tracts, restricts
and/or hinders this molecular motion in a particular
direction resulting in anisotropic diffusion. Apparent
diffusivity of water is generally higher in directions
parallel to fiber tracts than in the perpendicular direction (Beaulieu, 2002). Three dimensional probability
distribution of diffusivity can be described by a diffusion tensor ellipsoid with three eigenvectors and the
corresponding eigenvalues (1 and 2 3 , Fig. 13.1a).
(CP). Reprinted with permission from Hess CP et al. (2007).

(b) Bivariate normal distribution of the tissue shape measurement in 3P space. The distribution is highly overlapping in three
groups of tissues: (1) corpus callosum (CC) and internal capsule
(IC), (2) arcuate fasciculus (AF) and subcortical white matter
(SCW), and (3) gray matter (GM), lentiform nucleus (LEN) and
thalamus (TH). Reprinted with permission from Alexander et al.
(2000)
13 Role of Diffusion Tensor Imaging in Differentiation of Glioblastomas
The eigenvector associated with the largest eigenvalue

denotes the predominant orientation of fibers in a given
imaging voxel. If a particular voxel has a high degree
of anisotropy, one of the eigenvalues will be much
higher than the other two.
Most commonly used indices for diffusion tensor
are apparent diffusion coefficient (ADC) and fractional
anisotropy (FA) (Basser and Pierpaoli, 1996), which
can be calculated according to Eqs. (13.1) and (13.2),
respectively
ADC = (1 + 2 + 3 )/3
(13.1)

3 (1 )2 + (2 )2 + (3 )2
FA =
2
21 + 22 + 23
(13.2)
where denotes mean of the three eigenvalues. ADC
is a measure of the directionally averaged magnitude
of diffusion and is related to cell density, size and
parenchyma permeability. FA represents the degree
of diffusion anisotropy, and reflects the degree of
alignment of cellular structure (Basser and Pierpaoli,
1996).
Although FA is a good indicator of diffusion
anisotropy, it does not provide information on the
shape of the diffusion ellipsoid. For example, it cannot
distinguish a flat ellipsoid from an oblong one. Westin
et al. (2002) have modeled diffusion anisotropy using
a set of three basic metrics that depend on the shape of
the diffusion tensor: linear anisotropy coefficient (CL)
where diffusion is mainly along the direction corresponding to the largest eigenvalue; planar anisotropy
coefficient (CP) where diffusion is mainly restricted
to the plane spanned by the two eigenvectors corresponding to the two largest eigenvalues; and spherical
anisotropy coefficient (CS), which indicates isotropic
diffusion (Hess and Mukherjee, 2007) (Fig. 13.1a).
The CL, CP and CS values can be calculated using the
following equations:
CL = (1 2 )/(1 + 2 + 3 )
(13.3)
CP = 2(2 3 )/(1 + 2 + 3 )
(13.4)
CS = 33 /(1 + 2 + 3 ).
(13.5)
The CL, CP and CS values lie in the range from 0 to 1

and the sum of these three metrics is equal to 1 for a
given voxel:
CL + CP + CS = 1.
(13.6)
115
Information on the geometric nature of diffusion

tensor can aid in quantitative characterization of the
local structure in tissue and provide further tensor
shape differentiation in comparison to FA (Alexander
et al., 2000; Westin et al., 2002; Zhang et al., 2004).
The DTI features of white matter are quite different
for each of the anisotropy measures. These differences
arise from the relative contribution of the linear, planar and spherical shape components of the diffusion
tensor. Figure 13.1b shows the distribution of tensor shape measurements from the gray matter, deep
nuclei and white matter regions in three phase (3P)
tensor shape diagram. The commissural and deep projection (internal capsule and corticospinal) tracts are
dominated by linear shape, whereas arcuate fasciculus and peripheral radiation fibers have a significant
planar feature. The gray matter appears isotropic with
high CS and mainly clustered near the top of the
plot. Deep nuclei (thalamus and lentiform) regions
contain some axon connections and are slightly more
anisotropic than the gray matter (Westin et al., 2002;
Zhang et al., 2004). These studies suggest that tensor
shape measurements allow one to explore the tissue
microstructural difference.
Diffusion Tensor Imaging in Tumor

Classification
The diffusion properties of water molecules are
directly related to the microstructure of the medium in
which they reside, which determines the DWI and DTI
contrast for diagnostic purposes. As such, diffusion
MR imaging has been used extensively in clinical practice because of its exquisite sensitivity to cellular status, cytotoxic edema, cellular density, and directional
organization of the tissue (Chenevert et al., 2006).
In brain tumors, DTI provides a promising tool for
detecting microscopic difference in tissue properties.
Diagnostic brain tumor imaging involves not only
the characterization of the tumor, but also detection of
reactive and infiltrative changes surrounding the tumor.
The neoplastic mass can be generally subdivided into
two regions; the contrast-enhancing region representing the solid part of the tumor and the central area with
no or slight enhancement representing the necrotic or
cystic part of the tumor. Similarly, the peritumoral
edematous region can be separated into two regions;
proximal region surrounding the enhancing part of
116
the tumor potentially including infiltrative tumor cells,

and more distal region mainly comprised of vasogenic
edema. These four sub-regions of a tumor lesion can be
substantially different from each other in terms of their
DTI indices. Hence, when comparing the observations
of different studies, it is often important to note where
the measurements were made. In addition, systematic
analysis of DTI parameters from these different areas
may provide a robust way for characterization of brain
neoplasms.
Correlation of ADC with Tumor Cellularity

ADC reflects the rate of diffusional motion of the water
molecules. Of all the histologic features used in tumor
classification, cellularity has been the main target of
assessment with DTI. The higher the tumor cellularity
(and hence the higher volume of intracellular space),
the lower the ADC value due to decreased water diffusivity caused by a relative reduction in extracellular
space for the water molecules to move about (Beaulieu,
2002; Chenevert et al., 2006). This inverse correlation
between ADC and cellularity has been reported in both
glial (Yamasaki et al., 2005) and nonglial tumors (Guo
et al., 2002) (Fig. 13.2).
Fig. 13.2 (a) Scatterplot of FA versus ADC measured in solid

enhancing tumors and contralateral NAWM. Measurements
from lymphomas tend to cluster in the area corresponding
with FA and ADC decreases, whereas measurements from
GBMs tend to cluster in area corresponding with FA and ADC
increases. Reprinted with permission from Toh et al. (2008).
(b) Scatterplot of FA versus MD measured in solid tumors. FA
plotted as a function of MD. FA and MD showed an inverse
S. Wang et al.
ADC values have also been used in differentiating

glioblastomas from metastases, however, with mixed
results (Calli et al., 2006; Lu et al., 2004; Morita
et al., 2005; Oh et al., 2005; Tsuchiya et al., 2005;
Yamasaki et al., 2005). Some reports have suggested
that ADC (Lu et al., 2004; Morita et al., 2005) is helpful for the differentiation, while others indicated the
limited use of ADC in the differentiation of neoplasms
(Calli et al., 2006; Oh et al., 2005; Yamasaki et al.,
2005). In our previous study (Wang S et al., 2009), we
did not observe a significant difference in the ADC
values from the enhancing areas between glioblastomas and metastases, indicating limited sensitivity
and specificity of this parameter in tumor differentiation (Fig. 13.3b). Besides cellularity, other factors such
as extracellular matrix, viscosity and mucins may also
affect the measurement of ADC (Zamecnik, 2005).
Role of FA from the Enhancing Region

of the Tumor
FA reflects the orientation of tissue microstructure, and
as such its use may not be limited to the white matter
tracts alone (Beaulieu, 2002). Regions of relatively
high anisotropy have been reported in brain abscesses
(Kumar et al., 2007), glioblastomas (Beppu et al.,
correlation (R = 0.70). Measurements from lymphomas tend

to cluster in the area corresponding with FA increases and MD
decreases, whereas measurements from high grade tumors tend
to cluster in area corresponding with FA decreases and MD
increases. G2; grade 2 glioma, G3 and G4; grade 3 and grade
4 glioma, ML; malignant lymphoma. Reprinted with permission
from Kinoshita et al. (2008)
Fig. 13.3 (a) A 68 year old female with a glioblastoma in the

left frontal lobe. No hemorrhage was noted based on T1 and
T2-weighted images (not shown). Transverse contrast-enhanced
T1-weighted (1760/3.1) (A) and FLAIR (9420/141) (B) images
show ring-enhancement and extensive edema. ADC map (C)
shows restricted diffusion of the enhancing part. FA (D), CL
(E) and CP (F) from the enhancing part are lower than normal
appearing white matter. (b) Box plot of imaging characteristics in glioblastomas (white box) and brain metastases (gray
box). Boxes represent the median, the 25th and the 75th percentiles, bars indicate the range of data distribution. Circles
117
represent outliers (values more than 1.5 box length from the
75th/25th percentile). indicates significant differences (p <
0.05) between glioblastomas and brain metastases. CR: central region. ER: enhancing region. IPR: immediate peritumoral
region. DPR: distant peritumoral region. (c) Receiver operative
characteristic (ROC) curves for FA, CL, CP, ADC and logistic regression model (LRM) from the enhancing region of the
tumor. ADC+FA+CP is the best predictor for differentiation of
glioblastomas from brain metastases with area under the curve
(AUC) 0.98. Reprinted with permission from Wang S et al.
(2009)
118
2003) and areas of hemorrhage (Kumar et al., 2007),

indicating that FA is also related to structural orientation of the tissue/cells other than the white matter.
In contrast to ADC, the relationship between FA
and tumor cellularity is unclear, as both positive
(Beppu et al., 2003; Kinoshita et al., 2008) and negative (Toh et al., 2008) correlation has been reported
(Fig. 13.2). While Wang W et al. (2009) and Reiche
et al. (2010) reported lower FA from the enhancing
regions of glioblastomas compared with brain metastases, we observed higher FA in the enhancing regions
of glioblastomas than in those of metastases (Wang S
et al., 2009) (Fig. 13.3a). One likely reason for these
contradictory results is the lack of standardized methods, both for acquisition as well as postprocessing
and selection of region of interest (ROI). Our recent
work (Wang et al., 2010) demonstrated high FA in
glioblastomas compared with both brain metastases
and primary cerebral lymphomas. Among these three
tumor types, lymphomas have the highest cellularity, followed by glioblastomas and brain metastases
(Koeller et al., 1997; Rees et al., 1996; Zhang and
Olsson, 1997). These findings indicate that diffusion
anisotropy may not directly correlate with tumor cellularity. It has been reported that FA of tumor can be
affected by several factors including extracellular to
intracellular space ratio, extracellular matrix, tortuosity
and vascularity (Zamecnik, 2005).
Shape Based Diffusion Tensor Metrics

in Differentiation of Brain Tumors
FA is a scalar metric, which indicates the degree of
anisotropy, regardless of the shape of the diffusion
ellipsoid. Diffusion tensor can represent tubular, planar, or spherical shape of diffusion pattern, measured
by CL, CP and CS, respectively (Alexander et al.,
2000; Westin et al., 2002; Zhang et al., 2004). Both
CL and CP values contribute to FA observed in tissue
and their relative values indicate the shape of diffusion
ellipsoid (Alexander et al., 2000).
Tensor shape measurements provide further differentiation of tumors in comparison to that based on
FA and/or ADC (Westin et al., 2002; Zhang et al.,
2004). Zhang et al. (2004) reported lower CL in brain
metastases than contralateral normal brain. Higher CP
values in fibroblastic meningiomas were reported in
S. Wang et al.
comparison with other subtypes, such as endothelial meningiomas (Tropine et al., 2007). Kumar et al.
(2007) reported high CP and low CL in the abscess
cavity compared with normal white matter thus distinguishing true from pseudo white matter tracts. It has
also been reported that epidermoid cysts have high CP
(Santhosh et al., 2009) and tuberculomas showed lower
CL, CP and higher CS (Gupta et al., 2008) compared
with normal white matter. We have earlier demonstrated higher FA, CL and CP from the enhancing
part of glioblastomas in comparison to brain metastases (Wang S et al., 2009) (Fig. 13.3b). These results
suggest that tensor shape measurements provide additional information about tissue characteristics, which
may further aid in tumor classification.
A ring with high CP has been reported in glioblastomas, brain metastases and meningiomas. While the
potential reason for the observation of this ring remains
speculative, its presence may reflect compression of
surrounding tissue by the tumor (Tropine et al., 2007;
Zhang et al., 2004).
DTI in Assessing Tumor Infiltration

Generally speaking, peritumoral region is defined as
the area of abnormality surrounding the enhancing part
of the tumor. In metastatic brain tumors, peritumoral
edema is widely regarded as vasogenic edema. In this
region, increased extracellular water is present due
to leakage of plasma from altered tumor capillaries.
Also this region does not include any tumor cells.
In glioblastomas, on the other hand, the peritumoral
region includes both vasogenic edema and infiltrating
tumor cells.
Discrimination of tumor-infiltrated edema from
pure vasogenic edema may be beneficial for accurate
preoperative diagnosis of glioblastomas and metastases. Most DTI studies on this subject have focused on
the peritumoral regions of the tumor (Lu et al., 2004;
Morita et al., 2005; Tsuchiya et al., 2005; Zhang and
Olsson, 1997). In particular, Lu et al. (2004) reported a
significant difference between tumor-infiltrated edema
and pure vasogenic edema using a parameter called
tumor infiltration index, which measures departure
from a linear relationship between ADC and FA.
These authors also reported higher ADC in metastasis
compared to glioblastomas. However, another study
demonstrated lower ADC in the peritumoral region of

metastases compared to that of glioblastomas (Morita
et al., 2005). In contrast, van Westen et al. (2006)
reported no difference in ADC and FA values in
the peritumoral region of glioblastomas, metastases
and meningiomas. Recently, Kinoshita et al. (2010)
claimed that tumor infiltration index could not differentiate vasogenic edema from tumor infiltrated edema.
The difference in defining the ROIs for the peritumoral region in these studies may in part be responsible for the discrepancy. A number of studies have
focused on the area close to the enhancing region
(peritumoral region) by either manually placing a number of small ROIs around the tumor (Morita et al.,
2005) or by using a band of arbitrarily chosen thickness around the tumor (Law et al., 2002; Oh et al.,
2005). In our previous study (Wang S et al., 2009),
the peritumoral areas were further subdivided into
immediate peritumoral region and distant peritumoral
region with the hypothesis that the immediate peritumoral region may have a higher degree of tumor
infiltration in glioblastomas. We observed a significant
difference in FA, CL, and CP between glioblastomas
and metastases in the immediate peritumoral region. In
the distant peritumoral region, only FA and CP measurements reached significant difference between the
two tumor types (Wang S et al., 2009). While statistical significance was observed, the overall sensitivity,
specificity and accuracy for all the DTI metrics in the
peritumoral areas were lower than in the enhancing
part of the tumor (Fig. 13.3c). Since the edematous
region contains areas of increased extracellular water,
tumor infiltration and varying fractional composition
of normal white/gray matter, it is difficult to determine which factor dominates the DTI metrics. These
confounding factors may further explain the conflicting reports of DTI characteristics in the peritumoral
regions.
119
limited role in tumor classification. We have previously

reported that the single best predictor for differentiation between glioblastomas and brain metastases is
FA, with a sensitivity 89%, specificity 80% and AUC
of 0.90 (Wang S et al., 2009). Accurate characterization of complicated tissue, such as a tumor, may
require two or more imaging parameters. To date, only
a limited number of studies have investigated the role
of DTI parameters in combination for tumor classification. In a recent publication, it was suggested
that a combination of minimum ADCs and difference
of ADC facilitates more accurate grading of astrocytomas than either parameter measured individually
(Murakami et al., 2009). We have previously reported
that a multivariate logistic regression analysis can
determine an optimal combination of DTI parameters
to differentiate glioblastomas from brain metastases
(Wang S et al., 2009). From a study with sixty three
patients, we reported that a combination of ADC, FA
and CP from the enhancing part of the tumor provides
most accurate tumor classification, with a sensitivity of
92%, specificity of 100% and AUC 0.98 (Fig. 13.3c).
Conclusion
In this chapter, we have discussed how glioblastomas
and metastases can be characterized by DTI metrics,
such as ADC, FA, CL, CP and CS. These DTI metrics can be used individually or in combination, to
differentiate glioblastomas from metastases. Further
investigations on a larger patient population and histological validation will be necessary to determine the
robustness of these parameters in differentiating tumor
types. Combined with rapidly growing MRI technology for faster imaging and higher resolution, DTI holds
a great promise to elucidate morphological and functional characteristics of brain tumors noninvasively.
Combined DTI Metrics for Classification

References
As mentioned previously in the introduction, in vivo
DTI data provides a number of parameters that provide information about the shape, magnitude and
degree of diffusion anisotropy, which may in turn be
used to differentiate different tumor types. However,
these parameters, by themselves individually, have a
Alexander AL, Hasan K, Kindlmann G, Parker DL, Tsuruda

JS (2000) A geometric analysis of diffusion tensor measurements of the human brain. Magn Reson Med 44:283291
Basser PJ, Pierpaoli C (1996) Microstructural and physiological
features of tissues elucidated by quantitative-diffusion-tensor
MRI. J Magn Reson B 111:209219
120
Beaulieu C (2002) The basis of anisotropic water diffusion
in the nervous system a technical review. NMR Biomed
15:435455
Beppu T, Inoue T, Shibata Y, Kurose A, Arai H, Ogasawara K,
Ogawa A, Nakamura S, Kabasawa H (2003) Measurement of
fractional anisotropy using diffusion tensor MRI in supratentorial astrocytic tumors. J Neurooncol 63:109116
Calli C, Kitis O, Yunten N, Yurtseven T, Islekel S, Akalin T
(2006) Perfusion and diffusion MR imaging in enhancing
malignant cerebral tumors. Eur J Radiol 58:394403
Chenevert TL, Sundgren PC, Ross BD (2006) Diffusion imaging: insight to cell status and cytoarchitecture. Neuroimaging
Clin N Am 16:619632
Giese A, Westphal M (2001) Treatment of malignant glioma: a
problem beyond the margins of resection. J Cancer Res Clin
Oncol 127:217225
Guo AC, Cummings TJ, Dash RC, Provenzale JM (2002)
Lymphomas and high-grade astrocytomas: comparison of
water diffusibility and histologic characteristics. Radiology
224:177183
Gupta RK, Haris M, Husain N, Saksena S, Husain M, Rathore
RK (2008) DTI derived indices correlate with immunohistochemistry obtained matrix metalloproteinase (MMP-9)
expression in cellular fraction of brain tuberculoma. J Neurol
Sci 275:7885
Hess CP, Mukherjee P (2007) Visualizing white matter pathways in the living human brain: diffusion tensor imaging and
beyond. Neuroimaging Clin N Am 17:407426
Kinoshita M, Goto T, Okita Y, Kagawa N, Kishima H,
Hashimoto N, Yoshimine T (2010) Diffusion tensor-based
tumor infiltration index cannot discriminate vasogenic edema
from tumor-infiltrated edema. J Neurooncol 96:409415
Kinoshita M, Hashimoto N, Goto T, Kagawa N, Kishima
H, Izumoto S, Tanaka H, Fujita N, Yoshimine T (2008)
Fractional anisotropy and tumor cell density of the tumor
core show positive correlation in diffusion tensor magnetic
resonance imaging of malignant brain tumors. NeuroImage
43:2935
Koeller KK, Smirniotopoulos JG, Jones RV (1997) Primary
central nervous system lymphoma: radiologic-pathologic
correlation. Radiographics 17:14971526
Kumar M, Gupta RK, Nath K, Rathore RK, Bayu G, Trivedi R,
Husain M, Prasad KN, Tripathi RP, Narayana PA (2007) Can
we differentiate true white matter fibers from pseudofibers
inside a brain abscess cavity using geometrical diffusion
tensor imaging metrics?. NMR Biomed 21:581588
Law M, Cha S, Knopp EA, Johnson G, Arnett J, Litt AW (2002)
High-grade gliomas and solitary metastases: differentiation
by using perfusion and proton spectroscopic MR imaging.
Radiology 222:715721
Lu S, Ahn D, Johnson G, Law M, Zagzag D, Grossman RI
(2004) Diffusion-tensor MR imaging of intracranial neoplasia and associated peritumoral edema: introduction of the
tumor infiltration index. Radiology 232:221228
Morita K, Matsuzawa H, Fujii Y, Tanaka R, Kwee IL, Nakada
T (2005) Diffusion tensor analysis of peritumoral edema
using lambda chart analysis indicative of the heterogeneity of the microstructure within edema. J Neurosurg 102:
336341
Murakami R, Hirai T, Sugahara T, Fukuoka H, Toya R,
Nishimura S, Kitajima M, Okuda T, Nakamura H, Oya N,
S. Wang et al.
Kuratsu J, Yamashita Y (2009) Grading astrocytic tumors
by using apparent diffusion coefficient parameters: superiority of a one- versus two-parameter pilot method. Radiology
251:838845
Oh J, Cha S, Aiken AH, Han ET, Crane JC, Stainsby JA, Wright
GA, Dillon WP, Nelson SJ (2005) Quantitative apparent diffusion coefficients and T2 relaxation times in characterizing
contrast enhancing brain tumors and regions of peritumoral
edema. J Magn Reson Imaging 21:701708
Rees JH, Smirniotopoulos JG, Jones RV, Wong K (1996)
Glioblastoma multiforme: radiologic-pathologic correlation.
Radiographics 16:14131438
Reiche W, Schuchardt V, Hagen T, Ilyasov KA, Billmann P,
Weber J (2010) Differential diagnosis of intracranial ring
enhancing cystic mass lesions-role of diffusion-weighted
imaging (DWI) and diffusion-tensor imaging (DTI). Clin
Neurol Neurosurg 112:218225
Santhosh K, Thomas B, Radhakrishnan VV, Saini J, Kesavadas
C, Gupta AK, Kapilamoorthy TR, Nair SN (2009) Diffusion
tensor and tensor metrics imaging in intracranial epidermoid
cysts. J Magn Reson Imaging 29:967970
Soffietti R, Ruda R, Mutani R (2002) Management of brain
metastases. J Neurol 249:13571369
Toh CH, Castillo M, Wong AM, Wei KC, Wong HF, Ng SH, Wan
YL (2008) Primary cerebral lymphoma and glioblastoma
multiforme: differences in diffusion characteristics evaluated with diffusion tensor imaging. AJNR Am J Neuroradiol
29:471475
Tropine A, Dellani PD, Glaser M, Bohl J, Ploner T, Vucurevic G,
Perneczky A, Stoeter P (2007) Differentiation of fibroblastic meningiomas from other benign subtypes using diffusion
tensor imaging. J Magn Reson Imaging 25:703708
Tsuchiya K, Fujikawa A, Nakajima M, Honya K (2005)
Differentiation between solitary brain metastasis and highgrade glioma by diffusion tensor imaging. Br J Radiol
78:533537
Wang S, Kim S, Chawla S, Wolf R, Knipp D, Vossough
A, Orourke D, Judy K, Melhem ER, Poptani H (2010)
Differentiation between glioblastomas, brain metastases and
primary cerebral lymphomas using diffusion and perfusion
weighted imaging. Proc ISMRM 4287
Wang S, Kim S, Chawla S, Wolf RL, Zhang WG, ORourke
DM, Judy KD, Melhem ER, Poptani H (2009) Differentiation
between glioblastomas and solitary brain metastases using
diffusion tensor imaging. NeuroImage 44:653660
Wang W, Steward CE, Desmond PM (2009) Diffusion tensor
imaging in glioblastoma multiforme and brain metastases:
the role of p, q, L, and fractional anisotropy. AJNR Am J
Westin CF, Maier SE, Mamata H, Nabavi A, Jolesz FA, Kikinis
R (2002) Processing and visualization for diffusion tensor
MRI. Med Image Anal 6:93108
Yamasaki F, Kurisu K, Satoh K, Arita K, Sugiyama K, Ohtaki
M, Takaba J, Tominaga A, Hanaya R, Yoshioka H, Hama
S, Ito Y, Kajiwara Y, Yahara K, Saito T, Thohar MA (2005)
Apparent diffusion coefficient of human brain tumors at MR
Zamecnik J (2005) The extracellular space and matrix of
gliomas. Acta Neuropathol 110:435442
Zhang S, Bastin ME, Laidlaw DH, Sinha S, Armitage PA,
Deisboeck TS (2004) Visualization and analysis of white

matter structural asymmetry in diffusion tensor MRI data.
Magn Reson Med 51:140147
Zhang M, Olsson Y (1997) Hematogenous metastases of the
human brain-characteristics of peritumoral brain changes: a
121
van Westen D, Latt J, Englund E, Brockstedt S, Larsson EM

(2006) Tumor extension in high-grade gliomas assessed
with diffusion magnetic resonance imaging: values and
lesion-to-brain ratios of apparent diffusion coefficient and
fractional anisotropy. Acta Radiol 47:311319
Chapter 14
131 I-TM-601 SPECT imaging of Human Glioma
Adam N. Mamelak and David Hockaday
Abstract Primary malignant brain tumors (gliomas)

typically infiltrate deeply into surrounding normal
brain tissue. Currently available methods to determine the true extent of glioma infiltration are limited.
For example, T1 weighted MRI tends to underestimate the extent of infiltration, while T2 weighted
imaging tends to overestimate it. Most radio-labeled
ligands are not tumor specific. Accurate determination
of extent of infiltration is important to guide treatment
and to measure response to therapy. TM-601 (chlorotoxin), a 36-amino-acid peptide first identified in the
venom of a scorpion, selectively binds to glioma cells
but not normal brain parenchyma. A phase I/II clinical trial of intracavitary 131 I-TM601 in adult patients
with recurrent high grade glioma was performed to
determine the bio-distribution and toxicity of this
potential therapy. We evaluated the SPECT imaging
and bio-distribution data from this trial to determine
if 131 I-TM601 might also be useful in determining
tumor extent. Adult patients with recurrent high-grade
gliomas underwent tumor resection, implantation of
an intra-cavitary reservoir, and a single-dose injection of 370 MBq (10 mCi) 131 I-TM601 (0.251.0 mg
of 131 I-TM601) 24 weeks after surgery. Total-body
planar scans and whole-brain SPECT scans were
obtained on days 0, 1, 2, 3, and 68 after injection. Post-resection MRI images were co-registered
to the SPECT scans using image analysis software.
Analysis of the rate of radioactive decay and biologic elimination from the body and at the cavity site
A.N. Mamelak ()

Department of Neurosurgery, Cedars-Sinai Medical Center,
Los Angeles, CA, USA
e-mail: Adam.Mamelak@cshs.org
was performed. T1-weighted with gadolinium contrast

(T1-Wc), T2-weighted (T2), and SPECT volumes
were estimated by stereological Cavalieri sections and
compared for overlap. Nonbound 131 I-TM601 was
eliminated by 48 h after injection with the remaining
radio-labeled peptide bound to tumor for at least 6
8 day. Biologic decay rates from 24 to 168 h after
injection were only slightly shorter than the physical decay of 131 I(6.3 vs. 8.0 day), indicating steady
state binding. A comparison of tumor volume estimates using all three imaging parameters indicated that
131 I-TM601determined tumor volumes more closely
paralleled T2 volumes than T1-Wc volumes. Overlap
between co-registered MRI and SPECT scans corroborated the presence of radio-labeled peptide in the
vicinity of infiltrating tumor up to 168 h after injection.
131 I-TM601 provides a reliable estimate for primary
tumor extent. The poor spatial resolution of 131 Iodine
remains a major limitation to this tool. Further modification of this radio-peptide with better imaging isotopes such as 123 I (SPECT) or 124 I (PET), may provide
clinicians with an important tool for determining tumor
extent and differentiating regions of viable tumor from
necrosis.
Keywords Gliomas 131 I-TM601 Ligand
Chlorotoxin Decay rates SPECT imaging
Introduction
Accurate imaging of human malignancies in vivo
requires methods that define the full extent of tumor,
exclude normal tissues, and are spatially resolved
enough to permit accurate delineation of tumor size
and location. This ideal objective is very rarely

123
124
achieved by most currently available imaging methods, and represents a particular challenge for imaging
of gliomas in the human brain.
Gliomas are the most common primary brain tumors
diagnosed annually, comprising approximately 16,500
cases and accounting for nearly 13,000 deaths (Central
Brain Tumor Registry of the United States (CBTRUS),
19921997; American Brain Tumor Association,
Primer of Brain Tumors (ABTA), 2001). The most
lethal gliomas are the high-grade gliomas (HGGs):
grade III anaplastic astrocytoma, anaplastic oligoendroglioma, and grade IV gliomablastoma multiforme
(GBM). The ability to adequately image tumor volume extent is critical for treatment because more than
85% of HGG patients have recurrences near the original tumor site, and extent of tumor resection is one
of the most important predictors of long-term survival
(Brady et al., 1992; Natali et al., 1991; Debinski et al.,
1995). Magnetic Resonance Imaging (MRI) is the primary method for imaging gliomas in vivo. Magnetic
Resonance Imaging carries a very high spatial resolution that can be easily achieved, and is relatively
inexpensive. Unfortunately T1 and T2 weighted MRI
pulse sequences do not reliably differentiate infiltrating
tumor from cytotoxic edema or radiation necrosis. The
addition of gadolinium-based contrast agents can aide
in defining regions of increased capillary permeability
vascularity, but these are indirect measures of tumor
and often underestimate the true extent of glioma invasion. Magnetic Resonance Spectroscopy (MRS) can
detect biochemical abnormalities that are predictive
of tumor, but this approach lacks the spatial resolution needed to reliably determine tumor extent, and is
inherently limited by magnetic field inhomogeneities.
Radio-isotope based imaging methods such as
positron emission tomography (PET) and single photon emission computed tomography (SPECT) rely
on cell surface receptor mediated binding or cellular
uptake of a ligand, with subsequent imaging of the
radio-isotope bound to the ligand. The most prominent
example is the widespread use of 18 Flouro deoxyglucose (18 FDG). See other chapters in this volume for
a more detailed description of the role of 18 FDG for
imaging gliomas. 18 FDG is taken up by metabolically
active cells utilizing glucose for energy production.
Because cancer cells are generally more metabolically
active than normal tissues, they preferentially take up
the 18 FDG, providing an indirect measure of tumor.
A.N. Mamelak and D. Hockaday
While 18 FDG has been quite useful for staging of

metastatic disease in several systemic cancers, it has
been less useful for gliomas because glucose is the
sole source of metabolic energy in the brain, and often
the metabolic activity of normal brain exceeds that
of low grade or intermediate grade glioma. 18 FDG
PET is only 80% accurate in differentiating glioma
from radiation necrosis (Barker et al., 1997; Thompson
et al., 1999). 11 C-Methionine is a similar PET agent,
although it may be slightly more accurate than 18 FDG
in differentiating tumor from necrosis (Bigner et al.,
1998). A comparison of these methods is addressed in
another chapter in this volume. Spectroscopy has much
less intrinsic spatial resolution than PET, and so in general has been even less useful as an imaging platform.
Distinguishing viable tumor from vasogenic edema
or necrosis, however, remains difficult, even with the
use of metabolic imaging agents due to the significant
amount of angiogenesis present in high-grade gliomas
and the breaches of the blood-brain barrier (BBB) at
the edema site.
An imaging agent with the ability to bind to only
tumor cells would provide a novel platform for determination of tumor extent. An ideal compound would
be rapidly diffusible, tumor-specific, and has high resolution. Most of the currently available radioligands
do not meet both of these criteria; either they are
too large to negotiate the BBB and the interstices of
the extracellular matrix (ECM) or they are not specific in their binding to glioma. Monoclonal antibodies
(mAbs) such as 81C6, for instance, are too large to pass
the BBB and diffuse from the surgically created resection cavity into the parenchyma, where residual tumor
may thrive (Mamelak et al., 2006). Moreover, many
mAbs lack glioma specificity and have a high degree
of catabolism of the radiolabel.
Chlorotoxin (CTX), a scorpion-derived 36 amino
acid peptide, has been shown to specifically bind
to a phosphatidyl inositide, a phosphorylated lipid
on lamellipodia of tumor cells. Several immunohistochemical in vitro and in vivo experiments indicate
that CTX does not bind to normal brain parenchyma
or endothelial cells but specifically binds to tumescent glial cells. TM-601, a synthetic version of CTX,
has been shown to selectively label human gliomas
in vivo and in vitro (Soroceanu et al., 1998). Unlike
many other radioligands, TM-601 is relatively small
at 4 kDa: it passes the BBB and diffuses readily
14
through the brain parenchyma and may spread across

the volume of the whole tumor.
These properties make TM-601 potentially useful for imaging, as well as targeting local therapy. Several immunohistochemistry and cell-binding
assays demonstrated a positive correlation between
peptide-glioma binding and increasing tumor malignancy (Soroceanu et al., 1998). Intracranial injections
in animals reliably showed long-term tumor-specific
uptake. These findings and supporting in vivo efficacy
and toxicology data in animals justified a Phase I clinical trial of TM-601 conjugated to 131 I in patients with
recurrent high-grade glioma (Rajapakse et al., 2000).
The primary goals of this study were to determine the
safety profile and biodistribution of 131 I-TM-601.
To preliminary evaluate the imaging potential of
TM-601, the radiographic studies from subset of
patients in this trial were analyzed in detail. Of note,
131 I-TM-601 was initially developed as a potential
therapeutic agent, and that is why 131 I was chosen as
the radioligand. As such the extremely poor spatial resolution of 131 I was a known shortcoming for imaging
purposes, but more spatially resolved imaging agents
could not provide any potential therapy, rendering their
use unfeasible in this initial trial.
Methods
Nine adult patients with recurrent high-grade gliomas
underwent tumor resection implantation of an intracavitary reservoir and a single dose injection of 10
(1) mCi 131 I-TM-601 (0.251.0 mg) 24 weeks after
surgery. Protocol, eligibility criteria, and details of the
clinical trial are presented elsewhere (Mamelak et al.,
2006).
125
Prior to delivery of 131 I-TM-601 to the patient, a

1 cc volume of 0.11 mCi 111 In-DTPA was injected
into the reservoir to assure catheter potency. At
the time of 111 In-DTPA administration, a 256
256 matrix image was taken with a gamma camera
equipped with a high resolution, low-energy (247 keV,
15% window) collimator to monitor delivery of the
111 In-DTPA via catheter and observe any leakage.
Upon completion, dynamic planar images of the head
were continuously acquired for 15 min.
A 57 Co transmission scan was performed with and
without the patient on the scanner bench to connect
subsequent imaging data acquired for total body planar images. Prior to injection, the 57 Co source was
placed 3 cm below the scanner table with the anterior
camera 3 cm above the patient. A high-energy collimator centered at 122 keV with a 20% window was
used and images were acquired with a gantry speed of
10 cm/min and 256 1024 matrix size.
Injection of Radiolabeled Peptide

One to two hours after labeling, patients received 131 ITM-601 intracranially via a 25 gauge butterfly needle
and a shielded 5 cc syringe with a volume of 25 cc.
Initially, 25% of the total radiolabeled peptide dose
was given. After 5 min. the remaining dose was administered, followed by a 12 cc saline flush. Blood and
urine samples were collected for assessment of serum
radioactivity. Urine samples were also collected over
the period of days 12, 23, 34, and 46 or 8, while
blood samples were collected 1, 2, and 4 h after the
completion of the infusion, and days 2, 3, 4, and the
time of imaging on day 6 or 8.
Planar Images
Preparation of 131 I-TM-601
and Preliminary Scans
Presealed, sterilized vials containing lyophilized TM601 were radiolabeled with 10 mCi I131 via the
Iodogen bead method. Extent of labeling was determined via immediate thin layer chromatography
(ITLC) with >92% efficiency required to justify
human use.
After intracranial injection of 131 I-TM-601, anterior

and posterior whole-body planar images of matrix size
1024 256 were acquired on a dual-detector gamma
camera equipped with high-energy collimator set at
364 keV (15% window). Image acquisition included a
20 mL, 200 Ci calibrated 131 I source, which had been
placed 10 cm from the feet of the patient. Subsequent
images were acquired days 2, 3, 4, and either day 6 or
8 post injection.
126
Spectroscopy
A 180 whole brain spectroscopy scan was performed
on a dual headed gamma camera (Toshiba 7200 GCA),
with fiducial markers containing10 Ci 131 I-TM-601
placed on each zygoma and on the nasion to orient
SPECT scans with regard to subsequent MR images.
A 1 mCi 131 I-TM-601 standard was also placed 10 cm
above the patient head. Data was acquired at 3 degrees
per step at 30 s per step. For these images, a 128
128 matrix was employed, the high-energy collimator
was kept centered at 364 keV with a 15% window,
and a ramp filter was applied for 3D reconstruction.
Data were acquired in a proprietary format and converted to Interfile 3.0 for data analysis. Subsequent
SPECT scans were performed with the total body
planar images on days 2, 3, 4, and 6 or 8 after injection.
Magnetic Resonance Imaging

Postoperative MRI scans were also acquired for each
patient prior to 131 I-TM-601 injections. For each
patient, 20 transverse slices were acquired on a 16
bits per pixel scale with matrix size of 256 256.
Sequences included coronal, axial, and sagittal T1weighted images (with and without gadolinium contrast), T2 axial, FLAIR, and 3D-SPGR images. All
MRI scans were stored as DICOM files for future
image processing and analysis.
Spectroscopy Image Processing

All image analysis was performed with Analyze 5.0
(AnalyzeDirect, Lenexa, KS). For each patient, the
SPECT scan, comprised of 128 128 68 voxels of 79.5-mm3 , for each day (18) was loaded on
Analyze 5.0, with the slices containing the standard
excluded. For each SPECT scan, the intensity range
on a 32 bit per pixel scale was calculated and recorded
with the expectation that the magnitude (the range of
intensity values) of the brightest voxel should correspond roughly to the decay and/or the dispersion of
radioisotope. Each value was normalized for the duration of the scans, and the normalized intensities for
each patient were then averaged for all patients on each
day of scanning. To check the accuracy of this method,
the magnitude of the brightest voxel of the standard

was recorded for five patients and measured against
the physical decay of I131 , accomplished by defining
a volume of interest (VOI) which only included the
standard (the slices removed from the cranial VOI) and
measuring the calculated intensities within this VOI.
Once normalized and averaged, the intensities were
fitted exponentially.
Formation of Magnetic Resonance

Imaging Volumes
For each patient, T1 weighted with gadolinium contrast (T1-Wc) and T2 volumes were formed from 2-D
images, on the Analyze 5.0 platform. Resolution of
the coronal and sagittal images was considerably lower
than the axial resolution: voxel height was 7 mm;
length and width, 0.83 mm. Therefore, the Force Cubic
option (under Load As) was used to load the volumes
as linearly interpolated, isotropic voxels of 0.83 mm
0.83 mm 0.83 mm or 0.57 mm3 . The volumes were
primarily analyzed in the axial (transverse) plane.
Image Analysis: Stereological Estimation

of Volume
Tumor volumes for each patient were determined
from the T1-Wc, T2, and SPECT scans. Given the
arbitrary nature of gray-scale threshold methods and
the substantial differences that result in volumes of
extracted objects (Roberts et al., 2000) stereological
Cavalieri sections were used in Analyze 5.0 (Shen
et al., 2005). One of every three slices was sampled
for each scan, and the highlighted or tumor associated
region demarcated.
A reasonable minimum volume was determined
via the exclusion of ambiguous sample points; a reasonable maximum volume was determined via those
sample points inclusion. The mean was then calculated
for each volume with the accompanying error equal
to half the spread between the recorded minimum and
maximum. In addition, for SPECT scans (where limited resolution of 131 I rounds-off tumor extent) a full
width at half-maximum (FWHM) was intermittently
calculated across the cross-sectional diameter (transverse plane) using the Line Profile option, with the
14
127
assumption that intensity falloff would more or less

follow a normal (Gaussian) curve. The diameter of the
stereological volume was then checked by the sampled
FWHM for accuracy. For MR images, two volumes
were determined for each patient: T1-Wc and T2. For
SPECT, the tumor volumes for the second day post
injection and the final day were catalogued for each
patient.
Co-registration
After separate image processing, the isotropic MR volumes were co-registered to the SPECT scan-volumes
using Analyze 5.0. Under the 3-D Registration program, SPECT scans were matched manually and fused
to MR images for each patient, using the alignment of
fiducial markers on each SPECT volume to anatomical structures of the nasion and zygoma on the MR
image. Gray-scale intensities were converted to a 24bit red, green, blue (RGB) Hot-Metal colormap for
visual affect.
Results
Bioelimination
Total body planar-scans had been previously used by
Shen et al. (2005) to estimate effective half-lives in
various organs; in the present nine patient subset, these
half-lives were recalculated as 475 h for bodily elimination, 455 h for elimination from the brain, and
396 h for elimination from the cavity. Qualitatively,
these results are seen in the total-body planar scans on
day 2 and day 68 (Fig. 14.1). For all nine patients,
retention of radiolabeled peptide was observed in the
tumor cavity region for the duration of scans.
All SPECT scans also showed clear evidence of
131 I-TM-601 remaining at the tumor site on Day 8
(168 h post injection) at an activity comparable to that
of the 1 mCi standard. For the standard, the decay rate
of the value of the brightest voxel roughly followed
the theoretical (physical) rate of decay. The calculated
exponential of the standard yielded a half-life of 6.31
days, comparable to the known half-life for 131 I of 8.04
days.
Fig. 14.1 Total body planar scans of a single patient on days

1 and 8. Note the intense uptake and long-term retention of
radioisotope in the right frontal lobe injection site, and minimal uptake anywhere else in the body. The 100 Ci standard
is evident at the patients foot
Tumor Volume
Stereological estimation of tumor volumes in the nine
patients showed 131 I-TM-601 estimated volumes were
intermediate between the T1-Wc estimated volumes on
the low end and T2 determined volumes on the high
end. Over the course of scanning, the estimated volume
24 h post-injection was significantly larger and above
or equal to T2 estimates for most patients; by 168 h
post-injection most peptide determined volumes are
smaller, better centered between T1-Wc and T2 volumes, and considerably less uncertain. Overall, SPECT
determined volumes on Day 2 and Day 8 were always
larger than the T1-Wc determined volumes, and most
were smaller than the T2 determined volumes by Day
8. Moreover, FWHM measures across the span of the
SPECT volume corresponded to the smaller volumes
of Day 8 stereological estimates.
Fusion of Images
The fusion of SPECT volumes to MR images via the
use of fiducial markers yielded high-overlap and wellcentered regions of tumor extent 168 h post-injection
(Fig. 14.2). As the SPECT-estimated volumes were
rounded, they tended to approximate the area around
the edema (highlighted region in T1-Wc) and necrotic
128
Fig. 14.2 Co-registration of MRI and SPECT images demonstrate excellent overlap and permit better definition of region of
tracer uptake. Day 8 imaging has better resolution and closer
approximates the true tumor extent than day 2 estimates
region (dark center region in T1-Wc) and resection

area (also dark) and extend to regions in the T2determined volume. Generally, the SPECT-determined
volume was centered in this region and agreed with the
highlighted T1-Wc and T2 areas.
Spectroscopy images performed on Day 1 exhibited overwhelming scatter and thus were not useful for
image analysis. By Day 2, fused images showed considerable isolation of the highlighted SPECT region to
that of the tumor region on the MR image. Further
differentiation was evident on Day 3, with SPECTdetermined volume better centered and shifted from
the site of injection and the cavity towards the necrotic
core. By Day 8, residual glow from the cavity and
injection site had vanished and remaining radiolabeled isotope had an intensity level comparable to the
standard.
Discussions
Tumor Volumes
For the present study, 131 I was chosen for its potential
therapeutic benefit compared with less energetic but
better imaging radioisotopes such as 123 I or 124 I. With
the limited resolution of the current isotope, considerable difficulty arises in defining exact tumor extent.
Stereological methods offer better statistical estimates
and more reproducible results than traditional threshold methods when the exact extent of the region of
interest is unknown (Roberts et al., 2000; Shen et al.,
2005; Gadeberg et al., 1999). As a practical measure, stereology also allows the sampling of a smaller
number of slices with the same or better accuracy than
ROI drawing. Therefore, as far as 131 I is concerned,
volume of distribution of TM-601 is best estimated
from stereology. Of course, higher resolution radioisotopes may allow more precise methods to be used in
the foreseeable future.
As it stands, spectroscopy-estimated volumes of
distribution must be approached with several factors in mind. Spectroscopy-estimated volumes appear
rounded-off due to the limited resolution of the
radioisotope. Smaller, more peripheral features are lost
among the scatter. Their shape is primarily spherical or
ellipsoidal, despite their more irregular appearance in
T1-Wc or T2 MRI, and the uncertainty of the estimate
is generally large.
Similar qualifications exist for MRI-determined
volumes, where the tumor per se is not measured but
rather the associated enhanced regions of increased
capillary permeability associated with tumor but also
possibly postoperative or radiation-induced effects, or
inflammatory clogs. Generally for T1-Wc, gliomatumor volume is underestimated as invasive fringe
tumor cells lack the increased angiogenesis necessary to register in this modality. In contrast, T2 signal
hypersensitivity may represent tumor or radiationinduced, vasogenic edema, infiltrating tumor, and/or
cytotoxic effects, and as such, T2 estimates tend to
exaggerate the true extent of tumor.
In the present study, SPECT determined volumes
using 131 I-TM-601 appear to provide better estimates
than T1-Wc or T2 determined estimates. The larger
volume and uncertainties in Day 2 stereological estimates corresponded with the presence of more diffusely highlighted voxels than in Day 8, where steep
falloff allowed the highlighted SPECT volume to be
readily discernible. Scattering effects diminished resolution beyond practicality for the initial day, and no
distinguishable trends could be registered within the
interval of a day, i.e., the volumes of consecutive days
were effectively the same. By Day 8, this uncertainty
had diminished substantially and the volumes were
more easily estimated, because a sample marker could
be readily identified as belonging to the set of volume points. Therefore, lower energy imaging isotopes
of higher spatial resolution may provide substantially
better tumor definition.
14
On the whole, estimated SPECT volumes were

smaller on Day 8 suggesting that the non-bound radiolabeled peptide is rapidly eliminated from the brain,
while bound peptide persists. The biodistribution data
also support this contention. The volumes are hedged
between the stereologically estimated T1-Wc and the
T2 volumes. Most SPECT volumes were in fact closer
to the T2 estimate than the T1-Wc by Day 8, suggesting that 131 I-TM-601 diffuses and binds to the full
extent of primary tumor, while the limited resolution
may explain overestimation.
Total body planar images and SPECT scans frequently showed some hot spots of increased tracer
uptake removed from the primary tumor volume.
These hot spots may indicate diffusion of the peptide
to sites distal from the main tumor mass. This finding,
while speculative and preliminary, would be very valuable for determining true tumor extent and predicting
sites of recurrence or treatment failure.
Overlap of Fusion Volumes

The detailed anatomical clarity of MRI coregistered
to SPECT scans allows more accurate differentiation
of actual tumor and imaging artifacts then seen in the
SPECT scans alone. The increased utility of image
fusion is well documented and likely to be important
in determining the utility of this peptide for longterm use. Other imaging technologies such as Digital
Subtraction may also prove useful in increasing the
imaging sensitivity with TM-601.
Indications of Decay Rate

Qualitative evidence in the SPECT scans supports the
elimination calculated by biodistribution in the planar
scans. Although calibration was performed for planarscans, quantitative results were unavailable for the
SPECT scans. No calibration between count-rate and
intensity was available for the particular SPECT scanner used, and the attenuation and scatter corrections
remain unknown. Even if these factors were evident,
defining the VOI (the exact extent of enhanced region)
with the limited resolution of 131 I would be practically
ineffective in the clinical setting.
129
Nevertheless, the maximum and minimum intensities for each scan had been calculated, the difference
of which yielded the intensity value of the brightest
voxel. An ad-hoc approach was taken to compare
the intensity of the injected radioisotope to that of
the standard, though with admittedly limited statistics.
These values show substantial slowing of decay 24
48 h post-injection to rates towards that expected of
purely physical decay, in agreement with the calculated
elimination by biodistribution in the planar images,
396 h for elimination from the cavity. Rates continue mild slowing through 168 h post-injection as
they approach physical decay rates. If the above methods are indeed reliable, these rates may represent a
shift from radioisotope dispersion to physical decay
2448 h post-injection.
Future Directions
TM-601 appears to represent a very attractive ligand for radioisotope-based imaging of gliomas, and
potentially many other malignancies. It meets the ideal
criteria of being tumor-specific, non-toxic and rapidly
diffusible in solid organs. It appears to exhibit longterm binding to tumor at low doses. Several steps
are required to determine the ultimate utility of this
approach. First and foremost, less energetic and more
spatially selective imaging isotopes must be tested in
both animal models and humans. 123 I is an obvious first
choice due to the relative ease of labeling. However,
it is likely that PET sensitive isotopes such as 124 I or
64 Cu may be even more appealing. Methods to maximize peptide-isotope binding conditions still need to
be worked out.
In conclusion, the specificity and stability of 131 ITM-601 allow adequate imaging of high-localization
and limited resolution in the clinical setting, which
is promising for the development of higher resolution radioligands of TM-601 such as 124 I or 64 Cu for
PET imaging. Although SPECT determined volumes
using 131 I-TM-601 appear to provide better estimates
than T1-Wc or T2 determined estimates the overlap
of fusion images suggests 131 I-TM-601 fully estimates
the extent of primary brain tumor. Imaging studies following intravenous injection will also be important in
determining the ultimate utility of this novel peptide in
the clinical setting.
130
References
American Brain Tumor Association, Primer of Brain Tumors
(2001) Chapter 3: facts and statistics, 7th edn. ABTA,
DesPlaines, IL
Barker FGII, Chang SM, Valk PE, Pounds TR, Prados
MD (1997) 18-Flourodeoxyglucose uptake and survival of
patients with suspected recurrent malignant glioma. Cancer
79(1):115126
Bigner DD, Brown MT, Friedman AH, Coleman RE, Akabani
G, Friedman HS, Thorstad WL, McLendon RE, Bigner SH,
Zhao X-G, Pegram CN, Wikstrand CJ, Herndon JEII, Vick
NA, Paleologos N, Cokgor I, Provenzale JM, Zalutsky MR
(1998) Iodine-131-labeled antitenascin monoclonal antibody
81C6 treatment of patients with recurrent malignant gliomas:
phase I trial results. J Clin Oncol 16:22022212
Brady LW, Miyamoto C, Woo DV, Rackover M, Emrich J,
Bender H, Dadparvar S, Steplewski Z, Koprowski H, Black
P, Lazzaro B, Nair S, McCormack T, Nieves J, Morabito
M, Eshleman J (1992) Malignant astrocytomas treated with
iodine-125 labeled monoclonal antibody 425 against epidermal growth factor receptor: a phase II trial. Int J Radiat Oncol
Biol Phys 22(1):225230
Central Brain Tumor Registry of the United States (CBTRUS),
19921997 data. From www.cbtrus.org. 3
Debinski W, Obiri NI, Powers SK, Pastan I, Puri RK (1995)
Human glioma cells overexpress receptors for interleukin
13 and are extremely sensitive to a novel chimeric protein
composed of interleukin 13 and pseudomonas exotoxin. Clin
Cancer Res 1(11):12531258

Gadeberg P, Gundersen HJG, Tgehj F (1999) How Accurate
are Measurements on MRI? A Study on Multiple Sclerosis
Using Reliable 3D Stereological Methods. J MRI 10:7279
Mamelak AN, Rosenfeld S, Bucholz R, Raubitschek A,
Nabors LB, Fiveash JB, Shen S, Khazaeli MB, Colcher
D, Liu A, Osman MH, Guthrie B, Schade-Bijur S, Hablitz
DM, Alvarez VL, Gonda MA (2006) Phase I single-dose
study of intracavitary administered 131 I-TM-601 in adults
with recurrent high-grade glioma. J Clin Oncol 24(22):
36443649
Natali PG, Nicotra MR, Bigotti A, Botti C, Castellani P, Risso
AM, Zardi L (1991) Comparative analysis of the expression of the extracellular matrix protein tenascin in normal
human fetal, adult and tumor tissues. Int J Cancer 47(6):
811816
Rajapakse JC (2000) Random-grid stereologic volumetry of MR
head scans. J MRI 12:833841
Roberts N, Puddephat MJ, McNulty V (2000) The benefit
of stereology for quantitative radiology. Br J Radiol 73:
679697
Shen S, Khaaeli MB, Gillespie G, Alvarez VL (2005) Radiation
dosimetry of 131 I-chlorotoxin for targeted radiotherapy in
glioma-bearing mice. J Neurooncol 71:113119
Soroceanu L, Gillespie Y, Khazaeli MB, Sontheimer H (1998)
Use of chlorotoxin for targeting of primary brain tumors.
Cancer Res 58(12):48714879
Thompson TP, Lunsford LD, Kondziolka D (1999)
Distinguishing recurrent tumor and radiation necrosis
with positron emission tomography versus stereotactic
biopsy. Stereot Funct Neurosurg 73(14):914
Chapter 15
Assessment of Biological Target Volume Using Positron

Emission Tomography in High-Grade Glioma Patients
Habib Zaidi and Srinivasan Senthamizhchelvan
Abstract High-grade gliomas (HGG) are the most

challenging brain tumors to treat. Even though various sophisticated options exist to treat patients with
gliomas, the disease invariably leads to death over
months or years. The major obstacles encountered in
treating gliomas are in determining the exact location,
extent, and metabolic activity of the tumor. Molecular
imaging of energy metabolism, amino acid transport,
cell proliferation and cell death have been found helpful in identifying the biologically active tumor tissues for therapy. It allows a better understanding of
pathology at the molecular level. This ability is especially useful in brain tumors where tissue sampling
in vivo is associated with significant risks. Positron
emission tomography (PET) is one of the most prominent molecular imaging modalities utilized for imaging
pathophysiology of tumors at an early stage. In this
chapter, the applicability of PET in assessing the biologically active tumor volumes in high-grade glioma
patients for radiation therapy treatment planning and
therapy monitoring will be reviewed. We will focus
on the concept of biological target volume (BTV) and
associated methods of image segmentation available
for delineating tumor volumes in connection with their
applicability in high-grade gliomas.
Keywords High-grade gliomas PET Molecular
imaging BTV Necrosis Dose painting
H. Zaidi ()
Division of Nuclear Medicine, Geneva University Hospital,
CH-1211 Geneva 4, Switzerland
e-mail: Habib.zaidi@hcuge.ch
Introduction
The success of cancer treatment depends on multiple factors. One of the most important factors is the
accuracy of the information about tumor location,
extent and magnitude of disease. Traditionally this
information is obtained, through anatomical imaging
methods such as x-ray computed tomography (CT),
magnetic resonance imaging (MRI), and ultrasound
(US). However, it has become clear now, that the
acquisition of molecular and physiological information by noninvasive molecular imaging modalities such
as positron emission tomography (PET) could vastly
enhance our ability to fight cancer at an early stage
(Weissleder, 2006). Molecular imaging has the potential to detect physiological alterations that signal the
existence of cancer when it is still at a curable stage.
Advances in genomics and proteomics technologies
have shown the potential to transform the way in which
cancer is clinically managed today. Molecular imaging
is poised to play a key role in this transformation, since
it will allow the integration of molecular and physiological information specific to each individual case
with anatomical information obtained through conventional imaging methods. As a noninvasive molecular
imaging method PET exploits the unique decay characteristics of positron-emitting isotopes. The isotopes
of fluorine, oxygen, carbon, and others have been routinely used in the development of diagnostically useful
biological tracers that are available for PET imaging
of functional and/ or metabolic assessment of normal
tissues or disease state.
Conventional stand-alone PET has now been
replaced by PET/CT for improved patient throughput and most importantly for the availability of

131
132
complementary information of molecular PET images

and anatomical CT images in one imaging session.
There has been a tremendous expansion of clinical
applications of PET in oncology for the diagnosis,
staging and restaging of cancer patients. More than half
of all patients with cancer receive radiation therapy
(RT) at some stage during the course of their disease
management. Applications of PET in RT have been
reported in lung, head and neck, breast, lymphoma,
prostate and many other cancers (Zaidi et al., 2009).
Studies have also found that PET has advantages over
CT in the standardization of tumor volume delineation
in the reduction of the risk for geometrical misses,
in the minimization of radiation dose to the normal
organs, and in the assessment of tumor burden, blood
flow, tissue inflammation, and hypoxia. Integration
of functional PET data with anatomical CT data has
become a standard in RT particularly in treatment planning of various cancers. Imaging plays a key role in
the state-of-the-art high-precision RT techniques like
three-dimensional conformal radiation therapy (3DCRT), intensity modulated radiation therapy (IMRT),
image guided radiation therapy (IGRT), tomotherapy,
and stereotactic radiation therapy/surgery (SRT/SRS).
These high-precision radiation delivery methods allow
better dose distributions within the targeted tumor
volume while sparing a larger portion of adjacent normal tissues. Success of these RT techniques requires
accurate tumor volume delineation, tumor characterization, and response assessment during and after
treatment. Conventionally these tasks are achieved
through anatomical imaging (CT, MRI, and US). Of
late PET has been increasingly used in high-precision
RT for tumor volume delineation and characterization, because PET brings in the crucial functional and
molecular information which enable the direct evaluation of tumor metabolism, cell proliferation, apoptosis,
hypoxia and angiogenesis. This is a significant advance
in cancer imaging with great potential for optimizing
RT treatment planning and to the management of cancer patients. The availability of PET and CT in a single
imaging system (PET/CT) to obtain a fused anatomical and functional dataset has made the applicability
of PET in radiation oncology clinics much easier (Yap
et al., 2004). The most recent introduction of PET-MR
instrumentation dedicated for concurrent high resolution brain imaging is now revolutionizing the use of
multimodality imaging in tumor brain imaging (Boss
et al., 2010). Numerous reports are available in the
H. Zaidi and S. Senthamizhchelvan
literature in support of the routine use of PET for

RT target volume delineation in non-small cell lung
cancer (NSCLC), head and neck cancers, lymphoma
and in esophageal cancers, with promising preliminary data in many other cancers (Gregoire et al., 2007;
Nestle et al., 2009; Zaidi et al., 2009). The focus of
this chapter is to update the readers on the potential
use of PET imaging in the management of high-grade
gliomas (HGG), with particular emphasis being given
to the role of PET in the assessment of biological target
volumes (BTVs) for RT.
PET-Guided Biological Target Volume

(BTV) Delineation
Over the past two decades radiation oncology community has seen a paradigm shift from 2D treatment planning to 3D conformal treatment planning for RT. One
of the major advantages of 3D treatment planning and
associated treatment delivery techniques is the leverage offered for the dose escalation to tumor volumes
while preserving tolerance doses for normal structures.
The state-of-the-art 3D-CRT techniques (IMRT, IGRT,
tomotherapy, volumetric arc therapy and SRS/SRT
etc.) developed to deliver highly conformal radiation
beams directed towards targets warrant equally precise imaging modalities for accurate delineation of the
tumor extent. Customizing dose delivery to various
parts of the treatment volumes (dose painting and/or
dose sculpting) based on their dose requirements are
possible by 3D-CRT techniques today (Ling et al.,
2000). However it is important to know what needs
to be painted, and how much paint is required
to take complete advantage of these highly conformal radiation delivery techniques (Rickhey et al.,
2010).
The principal objective of all form of radiation
delivery techniques is to achieve highest possible
tumor dose without exceeding the dose level of surrounding normal tissue toxicity. This is achieved in
RT today by selective dose escalation also known as
dose painting with sharp dose gradient along the
tumor boundary. The rationale for dose painting is that,
treating a selective tumor region to high dose should
result in higher tumor control probability and lower
normal tissue complication probability. The expected
outcome of 3D-CRT is minimal complications and side
15 Assessment of Biological Target Volume Using PET in HGG Patients
effects, while achieving maximum possible dose to

tumor tissues. Accurate delineation of the tumor volumes and assessment of the tumor response to the
ongoing treatment regimens are necessary for successful implementation of modern RT techniques. In order
to achieve the above said goals it is highly desirable to precisely locate and visualize metabolic tumor
extensions and delineate boundaries.
Volumetric patient imaging forms the basis for
3D-CRT treatment planning and also helps in the
design of radiation fields and dose distributions. Since
3D imaging is used to delineate target volumes and
normal structures, the quality of the imaging and information obtained from the images have a direct impact
on the patient treatment and potentially on the outcomes and complications. Traditionally, anatomical
imaging like CT and MRI are used for radiotherapy
treatment planning, monitoring and follow-up evaluation. The major advantage of anatomical imaging is
their high resolution, which enables clear visualization of morphological changes. However, the precise
delineation of tumor regions with anatomical imaging has some significant limitations. The CT and
MRI measure the differential density and magnetic
properties of the tissues respectively, both of which
may not necessarily be tumor specific characteristics.
The morphological changes similar to malignancy can
occur due to other confounding factors like infection,
treatment (radiotherapy, chemotherapy and surgery
etc.) induced inflammation, which makes the distinction between tumor biology from other pathological
conditions difficult. Moreover biological changes manifest first and the time frame for the development of
detectable morphological changes is too long, during which the disease might have progressed to an
advanced stage.
The applicability of anatomical imaging for precise delineation of tumor extension is hampered, when
tumor harboring sites have unchanged morphology,
density and magnetic properties similar to normal
tissue. Successful treatment planning requires information about tumor biological characteristics such as
proliferation and hypoxia and the ability to distinguish treatment related scar, edema and necrosis from
malignant cells (Sun et al., 2011). Anatomical imaging modalities lack the aforementioned qualities which
make them insufficient for the delineation of target volume in most of the clinical presentations and as well
treatment response evaluation. Even though PET has
133
lower spatial resolution in comparison to its anatomical

imaging counterparts the sensitivity of PET for detecting tumor biology is high with required activity concentration of nmol to pmol range to detect biological
signals (Nestle et al., 2009). A typical example where
PET plays an important role in defining biologically
active tumor volume over anatomical imaging is shown
in Fig. 15.1.
The term biological imaging was coined by
(Ling et al., 2000) and has become popular in radiation oncology practice. Biological imaging for cancer
detection, staging, therapy monitoring and follow up
has gained vast interest as the evidence of treatment
success with this modality is keep accumulating. The
diagnostic utility of PET as the most prominent biological imaging system is evident from the increased
interest and utility of PET imaging for cancer imaging and therapy planning for the last few years. The
term biological target volume (BTV) was also first
proposed by (Ling et al., 2000). It is a new method of
defining tumor volumes based on metabolism, physiology and molecular biology of tissues. The definition
of BTV, apart from the standard definitions of tumor
volumes as gross tumor volume (GTV), clinical target volume (CTV) and planning target volume (PTV),
is aimed at providing information on the location and
extent of tumor margins based on tumor biology. The
BTV should also be helpful in assessing the biological
response of tumor to therapy. In radiotherapy planning, biological imaging will also guide the radiation
oncology community in defining biologically interesting sub-volumes (field in field) of the tumor like a
target within the GTV, which could be irradiated with
a higher dose through conformal RT (Nestle et al.,
2009). Figure 15.2 shows clinical examples comparing
biologically defined tumor volumes with anatomically
defined tumor volumes. Till date, many PET tracers
have been evaluated diagnostically in different cancers.
New PET tracers are being developed for staging, RT
planning, treatment monitoring and response evaluation during and after completion of RT has gaining
more attention in biological imaging.
Techniques for PET-Guided Biological

Target Volume (BTV) Delineation
Traditionally CT and MRI have been used as primary
imaging modalities for radiation therapy treatment
134
Fig. 15.1 Example of a patient with a glioblastoma (WHO IV)

in the left temporal and frontal areas. The images shown on
the top row (temporal area) correspond to gadolinium enhanced
T2-weighted MRI (A), coregistered 18 F-FET (B) and fused
PET/MR (C) of the first study. The same is shown in the bottom
Patient#1
row for the same study in the frontal area (D, E and F). The 18 FFET PET study revealed an additional lesion missed on MRI. In
addition, the T2-weighted MRI and the 18 F-FET PET show substantially different gross tumor volume extension for radiation
therapy treatment planning
Patient#2
Patient#3
BTVPET GTVMR
Fig. 15.2 Biological (BTV, blue) and morphological gross
tumour (GTV, red) volume defining the clinical target volume
in patients with high-grade glioma. Note the common volume between the tumour volumes (yellow chicken wire). Good
BTV-GTV matching is shown (left) in 1 patient, while substantial BTV-GTV mismatch is also detailed (center and right) in 2
other patients. Adapted from Weber et al. (2008)
planning. X-ray computed tomography and MRI offer

excellent spatial resolution, and soft tissue contrast,
but both imaging modalities fail to provide functional
properties of the imaged tissues. In cases where the
true extent of the disease may extend beyond anatomically defined volumes, despite its inferior resolution
compared with CT and MRI, PET has been shown
valuable for defining the extent of target volumes.
Some of the key contributions of PET in addition to
or in combination with other imaging modalities are
delineation of tumor volumes, biological characterization of tumor, and assessment of treatment response.
With the widespread adoption of hybrid PET/CT scanners, in radiotherapy clinics, PET-based delineation of
target volumes appears to be an attractive option in RT
treatment planning. One of the most difficult issues
facing PET-based RT treatment planning is the accurate delineation of target regions from typical noisy
functional images. The major problems encountered
in functional volume quantification are image segmentation and imperfect system response function. Image
segmentation is defined as the process of classifying
the voxels of an image into a set of distinct classes. The
difficulty in image segmentation is compounded by the
low spatial resolution and high-noise characteristics of
PET images. Medical image segmentation has been
identified as the key problem of medical image analysis and remains a popular and challenging area of
research. Despite the difficulties and known limitations, several image segmentation approaches have
been proposed and used in clinical setting including
thresholding, region growing, classifiers, clustering,
edge detection, Markov random field models, artificial
neural networks, deformable models, atlas-guided, and
many other approaches (Zaidi and El Naqa, 2010).
Medical image segmentation remains an unsolved
problem that has captured the imagination of image
analysis scientists over the past three decades. Manual
segmentation methods available on most commercial
software packages to identify lesion boundaries and
to quantify GTVs in terms of standardized uptake
value are very laborious and tedious. They discourage physicians from taking advantage of the inherently
quantitative data and compel them to use qualitative
means in their diagnosis, therapy planning, and assessment of patient response to therapy. Semi- or fully
automated segmentation methods enable physicians to
easily extract maximum and mean standardized uptake
value estimates from a lesion volume. This also allows
135
the physician to track changes in lesion size and uptake

after radio/chemotherapy. At present, various methods are used in practice to delineate PET-based target
volumes.
Manual delineation of target volumes using different window level settings and look up tables is
the most common and widely used technique in
the clinic. However, the method is highly operatordependent and is subject to high variability between
operators. Rather large intra-observer variability was
reported for many localizations including HGG as
shown in Fig. 15.3 (Weber et al., 2008). In this respect,
semi- or fully-automated delineation techniques might
offer several advantages over manual techniques by
reducing operator error/subjectivity, thereby improving reproducibility.
Assessment of PET-Guided Biological

Target Volumes in High-Grade Gliomas
(HGG)
Accurate determination of the tumor boundary at the
microscopic level, assessment of the tumor sensitivity, and/or prognosis to the therapy is essential for
successful treatment planning. Glioma cells are one
of the most treatment resistant cells and their inherent heterogeneous cell populations, diffuse infiltration
into normal brain tissues are the biggest challenge in
the tumor localization and precise delineation of tumor
extent. Prognosis of cerebral gliomas has continued
to remain poor for several decades, albeit significant
advances in multimodality diagnostic and therapeutic procedures. Therefore, development of a highly
specific and sensitive non-invasive imaging modality
is required. The ability to closely correlate diagnosis
with pathology, distinguish inflammation and necrosis from tumors, differentiate tumor grades, accurately
delineate tumor volume, monitor treatment responses,
and identify residual tumor/recurrence are the desirable goals for imaging brain tumors to improve their
clinical management.
The specificity of anatomical imaging modalities
in distinguishing neoplastic disease from vascular or
inflammatory processes can be problematic in highgrade gliomas. Treatment effects including surgical
trauma, corticosteroid-induced reduction of edema
and contrast enhancement, and radionecrosis cannot
136

Interobserver Varibility in BTV delineation
80.0
70.0
60.0
Volume cm3
50.0
40.0
Observer #1
Observer #2
Observer #3
30.0
20.0
10.0
0.0
1
10
11
Case No.
12
13
14
15
16
17
18
19
Fig. 15.3 Biological tumor volume measurements by three observers for each high-grade glioma case (1 through 19)
always reliably be distinguished from tumor recurrence or response to therapy. Molecular imaging allows
a better understanding of pathology at molecular level.
This ability is especially useful in the brain tumors
where tissue sampling in vivo is associated with significant risks. Availability of a suitable imaging modality
or a multimodality combination to obtain the information of interest noninvasively is also vital for the
basic research and development of novel and effective
experimental therapeutics required to improve prognosis. Accurate quantitative information on the metabolic
state of glioma tumor cells can be achieved through
biological imaging using PET. Depending on the radiotracer used, various molecular processes can be visualized through PET imaging, most of them relating to an
increased cell proliferation, metabolic rates, and DNA
synthesis as well as abnormal microvessel density and
thereby define tumor extent better than morphologic
imaging in malignant gliomas (Tsien et al., 2009).
PET radiotracers are especially helpful (1) in
the localization, grading and finding the extent of
glioma cells; (2) in the identification of metabolically
active residual tumor after therapy; (3) monitoring of
tumor progression; and (4) most importantly in the
differentiation between recurrent tumor and radiation
necrosis. Various metabolism and biochemical pathways are exploited by PET tracers for glioma imaging.
Energy metabolism of cells is imaged by [F-18]-2fluoro-2-deoxyglucose (FDG). Amino acid transport
and incorporation of tumor cells are tracked by
L -methyl-[C-11]methionine (MET), L -[C-11]tyrosine,
L -[F-18]fluorotyrosine. DNA synthesis is imaged
by 2-[C-11]thymidine, methyl-[C-11]thymidine,
[F-18]-3 deoxy-3 -fluorothymidine (FLT). Cell membrane/lipid biosynthesis is tracked by 1-[C-11]acetate,
[C-11]choline, [F-18]fluorocholine. Hypoxia is an
important aspect to consider for assessing the aggressiveness of tumor and predicting the outcome of
therapy (Sun et al., 2011). Tumor hypoxia is imaged
by [F-18]fluoromisonidazole (FMISO) and many
other tracers.
The preferential uptake of malignant glioma cells in
comparison to normal cells are exploited by tracers like
18 F-FDG, 11 C-MET, 18 F-FET and 18 F-FLT, depending
on the tumor grade as a reflection of increased activity of membrane transporters for amino acids (11 CMET and 18 F-FET) and nucleosides (18 F-FLT) as well
as increased expression of cellular hexokinase (18 FFDG) and thymidine kinase (18 F-FLT) genes, which
phosphorylate 18 F-FDG and 18 F-FLT, respectively.
Many hypoxia tracers (18 F-FMISO, 18 F-FAZA, 64 CuATSM and 18 F-EF5) have already shown their importance in target volume delineation or patient management in RT (Grosu et al., 2005a).
The most commonly available PET tracer 18 F-FDG
has the potential to detect abnormal metabolic rate,
through increased cellular glucose metabolism in brain
tumors. However, its use in target definition is complicated by the high level of intrinsic glucose uptake
in the brain. FDG imaging is useful in distinguishing
low-grade gliomas (LGG) from HGG based on tumorto-cortex (T/C) uptake ratio and tumor-to-white matter
(T/WM) uptake ratio. Selecting the optimum site for
tumor biopsy can be done based on the maximum
uptake of FDG for sampling of the most malignant
areas of tumors. For assessing response by pre- to posttreatment comparisons, FDG appears to be limited in
clinical usefulness. The ability of 18 F-FDG PET to
differentiate recurrent tumor from radiation necrosis
is also limited. The false-positive and false-negative
FDG-PET could result in unacceptably low sensitivity,
specificity, and negative predictive values.
The goal of PET imaging with radio-labeled amino
acids is to assess the protein synthetic process of tumor
growth. Amino acid uptake in normal brain tissues
is low relative to FDG uptake so that the tumor to
normal tissue contrast is better with amino acid imaging than with FDG. Radiolabeled amino acids can
also penetrate the bloodbrain barrier independently
of its disturbance. A variety of 11 C- and 18 F-labeled
amino acids such as 11 C-methionine (11 C-MET) and
18 F-fluoro-ethyl-tyrosine (18 F-FET) have been studied for potential use in oncologic PET. Most brain
tumors show an increased uptake of amino acids compared with normal brain tissue. In particular, the uptake
of 18 F-FET by brain tumors especially by high-grade
glioma cells is intense relative to the low uptake
in normal cerebral tissue and has shown the potential in the detection of primary and recurrent brain
tumors with high sensitivity and specificity. Compared
with 11 C-MET, 18 F-FET PET findings in brain tumors
are similar. One of the advantages of 18 F-FET over
11 C-MET is that the half life, which makes it possible
to be used in clinics not having on-site cyclotron.
11 C- methionine PET and 18 F-fluorothymidine
(FLT), provide better differentiation of the tumor from
brain background signals than 18 F-FDG PET. The
11 C-methionine PET scan reflects metabolic activity
through increased transport of amino acid carriers
137
at the level of the blood-brain barrier that is highly

expressed in malignant tumors compared with low
uptake in the normal brain. On the other hand FLT,
is a thymidine analog that is incorporated exclusively
into DNA. It measures the activity of cellular thymidine kinase, which increases several-fold as cells enter
the S-phase and begin DNA synthesis. Increased FLT
uptake and therefore thymidine kinase levels provide a
direct measure of the cellular proliferation rate.
Tumor hypoxia remains the most challenging condition for treatment. Though oxygen metabolism in
gliomas differs from that of normal brain tissue, the
lack of oxygen appears to be an important factor in
determining glioma aggressiveness and response to
therapy. It has been documented in several types of
cancers that low levels of oxygen tension are associated with persistent tumor following RT and with
the subsequent development of local recurrences. In
gliomas, spontaneous necrosis suggests the presence
of hypoxic regions that are radioresistant. 18 F-FMISO
imaging of hypoxic glioma cells shows significant
promise, however larger patient population studies are
required to ascertain its clinical impact. Identifying the
regional distribution of hypoxia may improve planning of resections and allow targeting higher doses of
radiotherapy more precisely to the hypoxic areas.
Glioma cell membrane biosynthesis is imaged using
11 C-acetate and 11 C-choline. The rationale for imaging
membrane and lipid biosynthesis is that tumor growth
requires both of these processes in parallel with DNA
and protein synthesis. These will likely show retention
in tumor tissue but not by gray matter, an important advantage over FDG. A very few studies have
been done so far on this front to show the potential
advantages of this approach.
PET Imaging for Differentiating Recurrent

Brain Tumor from Radiation Necrosis
Radiation damage to vascular endothelial cells
and oligodendrocytes causes necrosis. Differentiating
tumor growth from post-treatment radiation effect
(PTRE) remains a common challenge in high-grade
glioma tumors. On MRI, appearances of radiation
necrosis and of recurrent tumor are quite similar,
as both causes areas of increased signal intensity.
Conventional MRI/MRSI are currently used for the
138
detection of early treatment-bed changes, though accurate diagnosis is challenging because tumor growth,
PTRE, and admixed lesions can all have identical
MR imaging appearances. Microscopic tissue analysis distinguishes these entities and can document
intra-lesion heterogeneity by resolving distinct subregions of tumor from pure PTRE within different
locations of the same lesion. Early work on utilizing PET in differentiating radiation induced necrosis
from recurrent brain tumor was conducted by Patronas
and coworkers using FDG (Patronas et al., 1982). The
rationale for using FDG is that radiation necrosis is
expected to show decreased uptake in comparison with
recurrent tumors. However, in many cases, distinguishing recurrent tumor from radiation necrosis is found
to be difficult based on FDG-PET alone (Hustinx
et al., 2005). Radiolabeled amino acid analogues like
11 C-MET, 18 F-FET and proliferation marker 18 F-FLT
are suggested to perform better in PTRE evaluation,
than FDG in detecting residual and recurrent tumors
after fractionated irradiation (Hustinx et al., 2005;
Reinhardt et al., 1997). Though the exact incidence
of true radiation necrosis is largely unknown, differentiating it from recurrent tumor has a larger clinical
implication in the clinical management of patients and
PET tracers seem to play a major role in it.
Current Practice of Target Volume

Delineation in HGG
Conventional target volume definitions in high grade
gliomas (HGG) have not incorporated PET. The gross
tumor volume (GTV) is mostly defined on postoperative MRI and includes the contrast-enhancing
lesion as well as the surgical cavity (Fueger et al.,
2010). Clinical target volumes (CTVs) include an MRI
defined volume with an addition of uniform margin
(about 2 cm) within the brain, which includes areas of
microscopic extension and peritumoral edema. When
therapy is planned for a second CTV (boost volume) a margin of typically 1.5 cm around the GTV
is added to account for areas of microscopic disease. These CTVs are then further expanded by a
uniform margin of 0.30.5 cm to create a PTV to
account for treatment setup uncertainties. Initial imaging for HGG is usually done with CT or MRI, which
provides accurate information about lesion anatomy
and location. However, follow-up assessment of primary HGG tumors after radiation therapy, chemotherapy and surgery, is often difficult, since the anatomical
imaging modalities are usually not able to differentiate recurrent tumor from radiation necrosis, surgical
scar or inflammation. Identifying radiation necrosis
and differentiating them from tumor recurrence pose
a potential diagnostic challenge because the accurate
diagnosis has important implications for the patient
management.
PET-Guided Biological Target Volumes

in HGG
Grosu and co-workers have investigated the use of
amino-acid PET and single-photon emission computed
tomography (SPECT) in gross tumor volume definition for radiotherapy treatment planning of gliomas
(Grosu et al., 2005a, c). They have shown that 11 CMET-PET offers significant additional information
about tumor extent in HGG, compared to CT and
MRI alone. This study suggested that integration of
amino-acid PET in target volume definition might
contribute to an improved outcome in HGG patients
treatment. It was also showed that abnormal MET
PET activity was detected beyond the area of the
contrast-enhancing lesion on MRI. In a retrospective
study on newly diagnosed glioblastoma multiforme
(GBM) who underwent MET PET before radiation, the
area of MET uptake was found to be larger than the
contrast-enhancing gadolinium volume in 29 (74%) of
39 patients (Grosu et al., 2005c). In the same study
it was showed that patients who underwent treatment
planning based on 11 C-MET PET/SPECT imaging had
improved survival compared with treatment planning
based on CT/MRI in recurrent gliomas, with a median
survival of 9 months versus 5 months, respectively.
Other investigators have also suggested that 11 C-MET
PET has the potential to improve target volume definition in the radiation treatment planning of high-grade
gliomas by identifying residual tumor after resection
and also in recurrent gliomas. Yamane et al. (2010)
have recently showed that 11 C-MET PET can provide
useful information in initial diagnosis and differentiating tumor recurrence from radiation necrosis. They
have also claimed that use of 11 C-MET PET has
changed the intended clinical management in half of
the patients.
Our group (Vees et al., 2009) has studied the

contribution of 18 F-FET PET in the delineation of
GTV in HGG patients compared with MRI alone.
In this study PET based tumor volumes were delineated in 18 patients using seven image segmentation
techniques. The PET image segmentation techniques
included manual delineation of contours (GTV(man) ),
a 2.5 standardized uptake value (SUV) cutoff
(GTV(2.5) ), a fixed threshold of 40 and 50% of the
maximum signal intensity (GTV(40%) and GTV(50%) ),
signal-to-background ratio (SBR)-based adaptive
thresholding (GTV(SBR) ), gradient find (GTV(GF) ),
and region growing (GTV(RG) ). Figure 15.4 shows
an example of image segmentation using these
techniques. Overlap analysis was also conducted
to assess geographic mismatch between the GTVs
delineated using the different techniques. Contours
defined using GTV(2.5) failed to provide successful
delineation technically in three patients (18% of cases)
as SUV(max) <2.5 and clinically in 14 patients (78%
of cases). Overall, the majority of GTVs defined
on PET-based techniques were usually found to be
smaller than GTV(MRI) (67% of cases). Yet, PET
139
detected frequently tumors that are not visible on

MRI and added substantial tumor extension outside
the GTV(MRI) in six patients (33% of cases). The
study showed that the selection of the most appropriate 18 F-FET PET-based segmentation algorithm is
crucial, as it impacts both the volume and shape of
the resulting GTV. The SBR-based PET technique
shown to be useful and suggested that it may add
considerably important information on tumor extent
to conventional MRI-guided GTV delineation. In a
recent study by (Pichler et al., 2010), it has been
shown that 18 F-FET-PET is highly sensitive for detecting high-grade glioma in patients with neurological
symptoms. It was also suggested that, in the evaluation
of new brain lesions of unknown significance via
18 F-FET-PET a negative image can encourage a wait
and see strategy-of course in accordance with the
clinical picture and morphological imaging.
There are a few good reviews on the use of functional PET and MRI imaging for tumor volume definition in high-grade gliomas (Benard et al., 2003;
Grosu et al., 2005a; Guha et al., 2008; Jacobs et al.,
2002; Lecchi et al., 2008; Pirzkall et al., 2001; Tsien
Fig. 15.4 (A) Gadolinium enhanced T1-weighted MRI, (B)

corresponding 18 F-FET PET, and fused PET/MR (C) transaxial
slices of a clinical study with a glioblastoma showing differences
in target-volume definition. Indicated are (D) the gross tumour
volume (GTV) delineated on MRI (GTVMRI ), and (E) enhanced
details of PET-based GTVs obtained by manual delineation of
contours (GTVman ; magenta), an isocontour of a standardized
uptake value (SUV) of 2.5 (GTV2.5 ; purple), a fixed threshold
of 40% (GTV40% ; green) and 50% (GTV50% ; cyan) of the maximum signal intensity, signal-to-background ratio (SBR)-based
adaptive thresholding (GTVSBR ; yellow), gradient find (GTVGF ;
blue), and region growing (GTVRG ; red) segmentation algorithms. Note that GTVMRI overestimates the tumour extension
relative to GTVman . Reprinted with permission from Vees et al.
(2009)
140
et al., 2009). The key conclusions are that the PET

images could help in detecting metabolic and functional abnormalities beyond the tumor volume seen on
conventional MRI, assess early response to treatment,
and delineate the regions of high risks for failure in
high-grade gliomas.
Concluding Remarks and Future

Perspectives
PET imaging provides the opportunity to image noninvasively many biological processes in brain tumors.
Regional biological information and pathophysiology of gliomas can be obtained by studying energy
metabolism, amino acid transport, hypoxia, proliferation and cell death. None of these has been thoroughly
studied and utilized to allow judgment of their potential benefit to the management of gliomas. Future large
clinical trials will shed some light on the potential benefits of imaging these specific biological processes.
The issue of gross tumor volume delineation using biological imaging modalities for high-grade gliomas is
not clearly resolved at present for all clinical situations. A multimodality approach of biological imaging (PET, SPECT, CT, MRI and MRSI) might prove
to be an effective way as all the imaging modalities complement each other for better target volume
delineation. Despite the limited number of published
reports, the use of amino-acid PET, SPECT and MRSI,
along with anatomical imaging are shown superior
to either CT or MRI alone in visualizing the tumor
extent in gliomas. There have been studies comparing
SPECT and PET with MRI or MRSI to see the combined effect of anato-molecular approach to identify
what volume needs to be treated with high conformity
(Grosu et al., 2005b). A number of studies showed the
existence of significant differences in the target volumes delineations that result from the use of either
PET or MRI/MRSI imaging and it is still too early
to recommend the use of biologic imaging as the sole
determinant of target volumes. However, the incorporation of biological imaging into a treatment planning
process that currently depends entirely on anatomic
imaging seems advantages in most of the clinical presentations. In essence PET guided biological target
volumes helps treatment planning in radiotherapy, in
terms of identifying biological processes. However,
the real outcome of biological imaging is yet to be

analyzed in a larger clinical setup.
At present limited clinical data is available to prove
the superior outcomes in HGG patients with PET
defined GTV for RT planning. Retrospective analysis
of the outcome of controlled dose-escalation trials that
used biological imaging alongside MRI/MRSI for target volume segmentation could provide useful insights
for better design of prospective clinical trial for accurate tumor delineation (Grosu et al., 2005a). New
studies focusing on the wide applicability of biological imaging in HGG is ongoing, and it remains to be
seen how the high precision in gross volume segmentation translates into fruitful results with conformal
radiation delivery techniques such as IMRT, IGRT
etc, in a large clinical trial settings. In order to take
maximum advantages of advances in high precision
dose deliver techniques (dose sculpting and dose
painting) most accurate available assessment of tumor
extent by biological imaging modalities should be used
for high-grade gliomas. Future developments of new
tracers for biological imaging and robust automated
image segmentation techniques will bridge the existing knowledge gap in the high-grade gliomas tumor
volume delineation.
References
Benard F, Romsa J, Hustinx R (2003) Imaging gliomas with
positron emission tomography and single-photon emission
computed tomography. Semin Nucl Med 33:148162
Boss A, Bisdas S, Kolb A, Hofmann M, Ernemann U, Claussen
CD, Pfannenberg C, Pichler BJ, Reimold M, Stegger L
(2010) Hybrid PET/MRI of intracranial masses: initial experiences and comparison to PET/CT. J Nucl Med 51:1198
1205
Fueger BJ, Czernin J, Cloughesy T, Silverman DH, Geist CL,
Walter MA, Schiepers C, Nghiemphu P, Lai A, Phelps
ME, Chen W (2010) Correlation of 6-18F-Fluoro-L-Dopa
PET uptake with proliferation and tumor grade in newly
diagnosed and recurrent gliomas. J Nucl Med 51:15321538
Gregoire V, Haustermans K, Geets X, Roels S, Lonneux M
(2007) PET-based treatment planning in radiotherapy: a new
standard?. J Nucl Med 48(Suppl 1):68S77S
Grosu AL, Piert M, Weber WA, Jeremic B, Picchio M,
Schratzenstaller U, Zimmermann FB, Schwaiger M, Molls
M (2005a) Positron emission tomography for radiation treatment planning. Strahlenther Onkol 181:483499
Grosu AL, Weber WA, Franz M, Stark S, Piert M, Thamm R,
Gumprecht H, Schwaiger M, Molls M, Nieder C (2005b)
Reirradiation of recurrent high-grade gliomas using amino
acid PET (SPECT)/CT/MRI image fusion to determine gross

tumor volume for stereotactic fractionated radiotherapy. Int J
Radiat Oncol Biol Phys 63:511519
Grosu AL, Weber WA, Riedel E, Jeremic B, Nieder C, Franz M,
Gumprecht H, Jaeger R, Schwaiger M, Molls M (2005c)
L-(methyl-11C) methionine positron emission tomography
for target delineation in resected high-grade gliomas before
radiotherapy. Int J Radiat Oncol Biol Phys 63:6474
Guha C, Alfieri A, Blaufox MD, Kalnicki S (2008) Tumor
biology-guided radiotherapy treatment planning: gross tumor
volume versus functional tumor volume. Semin Nucl Med
38:105113
Hustinx R, Pourdehnad M, Kaschten B, Alavi A (2005) PET
imaging for differentiating recurrent brain tumor from radiation necrosis. Radiol Clin N Am 43:3547
Jacobs AH, Dittmar C, Winkeler A, Garlip G, Heiss WD (2002)
Molecular imaging of gliomas. Mol Imaging 1:309335
Lecchi M, Fossati P, Elisei F, Orecchia R, Lucignani G (2008)
Current concepts on imaging in radiotherapy. Eur J Nucl Med
Ling CC, Humm J, Larson S, Amols H, Fuks Z, Leibel S,
Koutcher JA (2000) Towards multidimensional radiotherapy
(MD-CRT): biological imaging and biological conformality.
Int J Radiat Oncol Biol Phys 47:551560
Nestle U, Weber W, Hentschel M, Grosu A-L (2009) Biological
imaging in radiation therapy: role of positron emission
tomography. Phys Med Biol 54:R1R25
Patronas NJ, Di Chiro G, Brooks RA, DeLaPaz RL, Kornblith
PL, Smith BH, Rizzoli HV, Kessler RM, Manning RG,
Channing M, Wolf AP, OConnor CM (1982) Work in
progress: [18F] fluorodeoxyglucose and positron emission
tomography in the evaluation of radiation necrosis of the
brain. Radiology 144:885889
Pichler R, Dunzinger A, Wurm G, Pichler J, Weis S, Nussbaumer
K, Topakian R, Aigner RM (2010) Is there a place for
FET PET in the initial evaluation of brain lesions with
unknown significance?. Eur J Nucl Med Mol Imaging 37:
15211528
Pirzkall A, McKnight TR, Graves EE, Carol MP, Sneed PK,
Wara WW, Nelson SJ, Verhey LJ, Larson DA (2001)
141
MR-spectroscopy guided target delineation for high-grade

gliomas. Int J Radiat Oncol Biol Phys 50:915928
Reinhardt MJ, Kubota K, Yamada S, Iwata R, Yaegashi H (1997)
Assessment of cancer recurrence in residual tumors after
fractionated radiotherapy: a comparison of fluorodeoxyglucose, L-methionine and thymidine. J Nucl Med 38:280287
Rickhey M, Moravek Z, Eilles C, Koelbl O, Bogner L (2010)
18 F-FET-PET-based dose painting by numbers with protons.
Strahlenther Onkol 186:320326
Sun X, Niu G, Chan N, Shen B, Chen X (2011) Tumor hypoxia
imaging. Mol Imaging Biol 2010 Sep 14. [Epub ahead of
print] DOI: 10.1007/s11307-010-0420-z
Tsien CI, Cao Y, Lawrence TS (2009) Functional and metabolic
magnetic resonance imaging and positron emission tomography for tumor volume definition in high-grade gliomas.
Semin Radiat Oncol 19:155162
Vees H, Senthamizhchelvan S, Miralbell R, Weber DC, Ratib
O, Zaidi H (2009) Assessment of various strategies for 18 FFET PET-guided delineation of target volumes in high-grade
glioma patients. Eur J Nucl Med Mol Imaging 36:182193
Weber DC, Zilli T, Buchegger F, Casanova N, Haller G,
Rouzaud M, Nouet P, Dipasquale G, Ratib O, Zaidi H, Vees
H, Miralbell R (2008) [(18)F]Fluoroethyltyrosine- positron
emission tomography-guided radiotherapy for high-grade
glioma. Radiat Oncol 3:44
Weissleder R (2006) Molecular imaging in cancer. Science
312:11681171
Yamane T, Sakamoto S, Senda M (2010) Clinical impact of
(11)C-methionine PET on expected management of patients
with brain neoplasm. Eur J Nucl Med Mol Imaging 37:685
690
Yap JT, Carney JP, Hall NC, Townsend DW (2004) Imageguided cancer therapy using PET/CT. Cancer J 10:221233
Zaidi H, El Naqa I (2010) PET-guided delineation of radiation
therapy treatment volumes: a survey of image segmentation
techniques. Eur J Nucl Med Mol Imaging 37:21652187
Zaidi H, Vees H, Wissmeyer M (2009) Molecular PET/CT
imaging-guided radiation therapy treatment planning. Acad
Radiol 16:11081133
Chapter 16
Skin Metastases of Glioblastoma

Rebecca Senetta and Paola Cassoni
Abstract Extra-cranial metastases of glioblastoma

(GBM) are rare events, and the mechanisms involved
in their development are still controversial and need
further elucidations. Although because of their low
frequency GBM metastases do not represent a routinely therapeutic challenge in patients with GBM,
such modality of tumour progression could be of crucial interest for the comprehension of the biology of
GBM. More specifically, metastases of GBM to the
skin have been reported in literature in very few and
heterogeneous cases. We here review the state of the art
of GBM skin metastases, focusing on the hypothesized
mechanisms involved, as well as on the suggested diagnostic and therapeutic approaches to these patients,
and on the possible relevance of a glial to mesenchymal transition as a facilitating event involved in the
extra-neural spreading of GBM.
Keywords Metastases GBM Scalp Skin
Mesenchymal FNAc
and diffuse infiltration into surrounding brain matter

so that patients survival is very short. In spite of its
rapid and aggressive growth, extra-cranial metastases
from cerebral GBM are surprisingly rare, occurring
only in 0.22% of cases (Cervio et al., 2001; Hsu et al.,
1998; Mentrikoski et al., 2008) and usually develop
at the same time of intracranial tumour progression.
However, in the last years an increasing number of
extra-neural metastases has been reported, possibly
due to longer patient survivals, related to therapeutic
advances and improved diagnostic modalities.
Lung and pleura are the most frequent sites of GBM
metastases (in about 60% of cases) (Houston et al.,
2000; Laraqui et al., 2001), followed by cervical lymph
nodes (51%) (Campora et al., 1993; Wallace et al.,
1996), bone marrow and bones (30%) (Houston et al.,
2000; Rajagopalan et al., 2005), liver (22%) (Astner
et al., 2006) and skull; metastases in other organs (as
pancreas, small bowel and spleen) are extremely rare
(Mujic et al., 2006; Yasuhara et al., 2003).
Introduction
Extra-Axial Tumor Spread: Why So Rare?
Glioblastoma multiforme (GBM) represents the most

common and malignant astrocytic-derived intracranial
primitive brain tumour in adult patients and accounts
for 1925% of all primary CNS tumours (Bauchet
et al., 2010). GBM usually grows by a quick, direct
Extra-neural dissemination of intracranial GBM can

arise from a local seeding after invasive surgical procedures (Bouillot-Eimer et al., 2005; Figueroa et al.,
2002; Mentrikoski et al., 2008) or after ventricular
shunting (Newton et al., 1992). Haematogenous spread
into extra-cranial tissue and tumour transdural extension is extremely rare in absence of previous surgery;
however, spontaneous (Wallace et al., 1996) GBM
dissemination has been reported in literature.
The rarity of metastases as well as their causes have
been diffusely debated in literature by several authors,
P. Cassoni ()
Department of Biomedical Sciences and Human Oncology,
University of Turin, 10100 Turin, Italy
e-mail: Paola.cassoni@unito.it

143
144
but the mechanism of GBM cell spread outside the

SNC is not fully understood yet. Certainly, different
factors play a concurrent relevant role in preventing
GBM metastatic spread (Bernstein et al., 1995; Datta
et al., 1998; Figueroa et al., 2002; Mourad et al., 2005;
Sadik et al., 1984; Wallace et al., 1996), such as:
(1) the short patient survival;
(2) the absence of a brain lymphatic system;
(3) the presence of a dense and impassable structure represented by the dura, contrasting tumor
invasion;
(4) the presence the blood-brain barrier, as a physical entity, constituted by a dense extracellular
matrix and a tough basement membrane surrounding blood vessels;
(5) an early collapse of the thin-walled cerebral veins
due to the increased intracranial pressure consequent to tumour growth, which prevents any
penetration/diffusion of tumour cells;
(6) the immune response against malignant glial cells,
which would fail to survive in distant organs
because of their foreign antigenic specificity
which would trigger an exaggerated immunological response, preventing tumour metastatic
growth.
In addition, some authors highlighted the necessity
of neoplastic glial cells to have a specific intracranial
supporting micro-environment to survive (Figueroa
et al., 2002).
Altogether, the previously listed assumptions may
explain the rarity of extra-SNC metastases, mainly in
absence of surgery.
Skin GBM Metastases: Facts

and Hypotheses
To date, only eighteen cases of intracranial GBM
metastases to skin have been reported overall (see
Table 16.1). In one case, the cutaneous spreading
occurred in a child (13 years old boy) (Saad et al.,
2007) whereas the other seventeen cases developed
in adult patients aged 1974 years (mean age 48,5
years) without any sex prevalence. The most common
site of skin metastases was the scalp, though Miliaras
R. Senetta and P. Cassoni
et co-workers (2009) described a left scapular subcutaneous GBM spreading. In general, the reported
cutaneous lesions macroscopically appeared as fixed,
ulcerated, soft, rubbery masses or nodules, rarely
bleeding, measuring from 1 to 5 centimetres in size,
generally localised close to the surgical scar or the
craniotomy site. A hypothetical tumour dissemination
during and/or after surgery might explain the proximity to surgical suture, as a consequence of a neoplastic
cells subcutaneous implantation or a cell escape along
the tract of excision through the dura toward skin. The
involvement of the scalp occurs more commonly after
craniotomy (Allan, 2004; Figueroa et al., 2002; Hata
et al., 2001; Jain et al., 2005; Matsuyama et al., 1989;
Santos et al., 2003; Schultz et al., 2005; Wallace et al.,
1996) and less frequently after a stereotactic biopsy
(Bouillot-Eimer et al., 2005; Houston et al., 2000), further underlining the crucial role of the loss of brain
anatomical integrity in the pathogenesis of extra-neural
involvement. In addition, in the time frame close to
surgery, the post-operative neo-vascularization might
play a key role in promoting the dissemination of
neoplastic cells (Jain et al., 2005).
Skin metastases can either be isolated as single
site of extra-neural involvement (Allan, 2004; Jain
et al., 2005; Mentrikoski et al., 2008; Santos et al.,
2003; Schultz et al., 2005) or coexist with other visceral and/or lymph node tumour localisation (BouillotEimer et al., 2005; Figueroa et al., 2002; Hata et al.,
2001; Houston et al., 2000; Matsuyama et al., 1989;
Saad et al., 2007; Wallace et al., 1996). In general,
cutaneous dissemination develops concurrently and
synchronous with progression of intracranial disease;
only three cases of GBM skin metastases in absence of
a synchronous brain or extracranial progression have
been reported (Louis et al., 2007; Santos et al., 2003;
Senetta et al., 2009). In fact, in our report we described
two cases of cutaneous spreading after a 14 and 11month latency from the primary diagnosis and surgery:
unexpectedly, in both cases, at the time of skin involvement the intracranial disease was absent or responsive
to therapy, suggesting a divergent response to treatment between intracranial (responsive) and skin (resistant) GBM. In both patients, surgery was followed
by a conformational radiotherapy (total dose 59.4 and
60 Gy respectively) and adjuvant Temozolomide: in
one patient the cutaneous swelling developed after 12
cycles of chemotherapy without any recurrence of the
intracranial disease, whereas in the other patient a scalp
42, F
49, F
19, M
Santos et al. (2003)

Jain et al. (2005)
L, parietal
32, M
Houston et al. (2000)
34, F
L, temporal
NM
Hata et al. (2001)
Figueroa et al. (2002)
R, frontal
41, M
Supra-tentorial
R, temporoparietal
L, temporal
NM
R, temporal
68, M
Matsuyama et al.
(1989)
Wallace et al. (1996)
Tumor location
Age Sex
Author, year
Stereotactic biopsy; 3
months later craniotomy
with partial resection, RT
(60 Gy), CT (2 cycles of
BCNU), 125 I
Craniotomy for subtotal
excision, 125 I
brachytherapy, RT
(59.4 Gy), CT (2 cycles of
BCNU)
Craniotomy with partial
resection + Ommaya, RT
(64.8 Gy)
Craniotomy
Craniotomy, RT
Craniotomy, RT, CT
Craniotomy with total

resection, RT, CT
Craniotomy with a total
resection, RT (60 Gy), CT
(2 cycles with carmustine)
Therapy
Table 16.1 Reported cases of intracranial GBM metastases to skin
36 months
10 months
8 months
10 months
6 months
NM
3 months
NM
Timing to
cutaneous
progression
3 months later
the cutaneous
spread
No
Yes
Not mentioned
2 months later
the cutaneous
spread
NM
No
NM
Intracranial
progression
No
No
No
LN and
mediastinum
Liver, spleen,
spinal cord
Cervical LN,
extraocular
muscles,
limbus of the
right eye
Lung, LN and
heart
Skull, lung,
liver
Additional
metastatic
organs
Unknown
Excision
Excision
Biopsy, RT
(30 Gy), CT
Excision
NM
Local RT to
scalp and
neck (45 Gy)
and CT
Unknown
Diagnosis/
therapy of
cutaneous
lesions
>36 months
12 months
13 months
17 months
13 months
NM
8 months
5 months
Survival
16 Skin Metastases of Glioblastoma

145
L, fronto-parietal
L, frontal
13, M
58, F
41, M
48, F
53, F
Mentrikoski et al.
(2008)
Senetta et al. (2009)
L, frontotemporal
L, frontal
NM
Therapy
Craniotomy
Craniotomy, RT, CT
Craniotomy, RT, CT (TMZ)

Craniotomy, RT and
intratumoral seed implants
Craniotomy, RT, CT
Craniotomy, RT, Local CT

with DTI-015
Stereotactic biopsy, RT, CT
(2 cycles of BCNU)
Craniotomy, RT, CT (TMZ)
Craniotomy, RT
Craniotomy, RT (55.8 Gy)
7 months
14 months
14 months
16 months
2 months
NM
11 months
12 months
12 months
48 months
Timing to
cutaneous
progression
Yes
No
No
Yes
NM
Yes
Yes
Yes
Yes
Yes
Intracranial
progression
Parotid gland,
LN, bone
(L4)
No
No
Leptomeninges,
lung, liver
No
No
Axial skeleton
No
LN,
leptomeniges
No
Additional
metastatic
organs
FNAc
Total removal
and focal RT
Biopsy and focal
RT
Biopsy
NM
FNAc/partial
excision
CT
Excision
No treatment
CT
10 months
25 months
26 months
NM
NM
10 months
12 months
13 months
14 months
52 months
Diagnosis/therapy
of cutaneous
lesions
Survival
NM not mentioned, F female, M male, L left, R right, RT radiotherapy, CT chemotherapy, TMZ temozolomide, FTM fotemustine, PC procarbazine and CCNU, LN limph nodes,
FNAc fine-needle aspiration cytology
63, M
L, frontal
60, F
Bouillot-Eimer et al.
(2005)
Saad et al. (2007)
Miliaras et al. (2009)
L, parietal
74, F
Schultz et al. (2005)
Supra-tentorial
L, Temporooccipital
L, temporal
Allan (2004)
Moon et al. (2004)
Tumor location
Age Sex
60, M
35, F
Author, year
Table 16.1 (continued)
146
147
Fig. 16.1 Histopathology and immunophenotype shift in one

case skin metastatic GBM. The intracranial GBM showed (at
haematoxylin and eosin) a marked hypercellularity with an area
of classical pseudo-palisade necrosis and prominent glomeruloid
vascular proliferation (a). At immunohistochemistry, a diffuse
and intense GFAP reactivity was present (b). The skin metastatic
tumour was composed of small neoplastic cells with a scant
cytoplasm and hyper-chromatic nuclei (c) with only few neoplastic elements positive to GFAP (d) and an intense vimentin
immunoreactivity (e)
mass appeared after 6 cycles of chemotherapy, in presence of a partial response of the intracranial tumour
(reduction of about 90% of the enhancing area). This
observation induced us to verify in these two cases
if primary and metastatic lesions still shared a main
phenotypical consistency or if the skin localization
acquired a prevalent divergent phenotype (Fig. 16.1).
In fact, the immunophenotypical profile of cutaneous
tumours revealed a strongly reduced GFAP and EGFR
expression, paralleled by an increased vimentin and
YKL-40 (which is a marker of radio-resistance in
GBM) (Pelloski et al., 2005) staining at IHC (Senetta
et al., 2009). Other authors similarly observed an
intense vimentin-immunoreactivity in the cutaneous
GBM sites associated either to a focal (as in our
cases) or a robust GFAP expression (Jain et al., 2005;
Mentrikoski et al., 2008; Schultz et al., 2005; Senetta
et al., 2009). As previously discussed in our paper
(Senetta et al., 2009), these changes altogether may
suggest a glial-to-mesenchymal transition of the
cutaneous metastases: the newly acquired immunophenotype could account for the therapy resistance of the
skin lesions in contrast to the synchronous responsiveness of the intracranial GBM.
This hypothesis is in agreement with previous evidence of the overexpression of mesenchymal and
angiogenesis related genes in GBMs with more aggressive behaviour (Phillips et al., 2006) and could then
be extended to their metastatic spread: still, it should
be understood if such shift towards a mesenchymal
de-differentiation in skin lesions can be a spontaneous event in the pathway of tumour progression or if
radiotherapy could favour the sarcomatous transformation and/or the selection of resistant mesenchymal
cell clones (Burger et al., 1979; Schiffer et al., 1990).
Diagnostic and Therapeutic Approaches

to Skin Metastases of GBM
As summarized in Table 16.1, the diagnosis of a cutaneous dissemination of GBM is mainly performed on
a biopsy, a partial resection or a total excision of the
skin lesion. In two cases a fine-needle aspiration cytology (FNAc) approach was used, and the diagnosis
was then re-confirmed on the tumour excision histology (Schultz et al., 2005; Miliaras et al., 2009). Some
authors highlighted the importance of patient neoplastic history (whether intracranial or extra-neural) for an
accurate diagnosis (Schultz et al., 2005; Mentrikoski
et al., 2008): without a knowledge of intracranial
tumour story, the differential diagnosis must include,
above all, a melanoma, followed by a sarcomatoid squamous cell carcinoma and atypical fibroxanthoma (Mentrikoski et al., 2008). In fact, a cutaneous
lesions with atypical morphological appearance at histology (spindle-shape cells, nuclear pleomorphism,
irregular nuclear membranes, prominent nucleoli and
mitoses) associated to immunohistochemical positivity
for S-100 protein (as frequently described in GBMs)
(Bouillot-Eimer et al., 2005; Figueroa et al., 2002;
Mentrikoski et al., 2008) should first be considered as
a possible melanoma localization, even in presence of
148
a known history of GBM. Schultz and colleagues listed

the cytological features of malignant neoplasm that can
arise in the scalp and that must be distinguished from
a metastatic GBM, when an anamnesis of glioma is
present (Schultz et al., 2005).
The treatment of GBM scalp metastases can consist
in surgery alone (with a partial or total lesion excision
with surrounding subcutaneous and bone tissue), or in
surgery in combination to chemotherapy or radiotherapy, or in chemotherapy and radiotherapy alone. In the
two cases of GBM skin metastases that we reported,
total surgery+local radiotherapy (total dose: 40 Gy)
proved to be more effective in preventing further cutaneous relapses compared to biopsy+local radiotherapy (total dose: 55 Gy)+6 cycles of Temozolomide.
Therefore, in our experience, a radical surgical treatment seems to allow the best local tumour control
(Senetta et al., 2009).
In conclusion, GBM metastases are rare events,
and therefore do not represent an ordinary clinicaloncologic concern: however, comprehension of the
mechanisms involved in promoting their development
could bear relevance in elucidating GBM biology overall, and considered as a unique model to focus on the
crucial steps necessary to GBM progression.
References
Allan RS (2004) Scalp metastasis from glioblastoma. J Neurol
Neurosurg Psychiatry 75:559
Astner ST, Pihusch R, Nieder C, Rachinger W, Lohnerm H,
Tonn JC, Molls M, Grosu AL (2006) Extensive local and
systemic therapy in extraneural metastasized glioblastoma
multiforme. Anticancer Res 26:49174920
Bauchet L, Mathieu-Daud H, Fabbro-Peray P, Rigau V, Fabbro
M, Chinot O, Pallusseau L, Carnin C, Lain K, Schlama
A, Thiebaut A, Patru MC, Bauchet F, Lionnet M, Wager
M, Faillot T, Taillandier L, Figarella-Branger D, Capelle
L, Loiseau H, Frappaz D, Campello C, Kerr C, Duffau H,
Reme-Saumon M, Trtarre B, Daures JP, Henin D, Labrousse
F, Menei P, Honnorat J; with the participation of Socit
Franaise de Neurochirurgie (SFNC) and the Club de NeuroOncologie of the Socit Franaise de Neurochirurgie (CNOSFNC) and Socit Franaise de Neuropathologie (2010)
Oncological patterns of care and outcome for 952 patients
with newly diagnosed glioblastoma in 2004. Neurooncol.
12:725735
Bernstein JJ, Woodard CA (1995) Glioblastoma cells do not
intravasate into blood vessels. Neurosurgery 36:124132
Bouillot-Eimer S, Loiseau H, Vital A (2005) Subcutaneous
tumoral seeding from a glioblastoma following stereotactic biopsy: case report and review of the literature. Clin

Burger PC, Mahley MS Jr, Dudka L, Vogel FS (1979) The
morphologic effects of radiation administered therapeutically
for intracranial gliomas. A post-mortem study of 25 cases.
Cancer 44:12561272
Campora RG, Salaverri CO, Ramirez FV, Villadiego MS,
Davidson HG (1993) Metastatic glioblastoma multiforme in
cervical lymph nodes: report of a case with diagnosis by fine
needle aspiration. Acta Cytol 37:938942
Cervio A, Piedimonte F, Salaberry J, Alcorta SC, Salvat J,
Diez B, Sevlever G (2001) Bone metastases from secondary glioblastoma multiforme: a case report. J Neurooncol
52:141148
Datta CK, Weinstein JD, Bland JE, Brager PM, Stewart MA
(1998) A case of cervical lymph node metastasis resulting
from glioblastoma multiforme. W V Med J 94:276278
Figueroa P, Lupton JR, Remington T, Olding M, Jones RV,
Sekhar LN, Sulica VI (2002) Cutaneous metastasis from an
intracranial glioblastoma multiforme. J Am Acad Dermatol
46:297300
Hata N, Katsuta T, Inoue T, Arikawa K, Yano T, Takeshita M,
Iwaki T (2001) Extracranial metastasis of glioblastoma to the
lung and the heart with a histological resemblance to small
cell carcinoma of the lung: an autopsy case. No Shinkei Geka
29:433438
Houston SC, Crocker IR, Brat DJ, Olson JJ (2000) Extraneural
metastatic glioblastoma after interstitial brachytherapy. Int J
Hsu E, Keene D, Ventureyra E, Matzinger MA, Jimenez C, Wang
HS, Grimard L (1998) Bone marrow metastasis in astrocytic
gliomata. J Neurooncol 37:285293
Jain N, Mirakhur M, Flynn P, Choudhari A (2005) Cutaneous
metastasis from glioblastoma. Br J Neurosurg 19:6568
Laraqui L, Amarti A, Zouiadia F, Maher M, Kettani F, Saidi
A (2001) Pulmonary metastasis from glioblastoma: a case
report. Rev Pneumol Clin 5713:225228
Louis DN, Ogaki H, Wiestler OD, Cavanee WK (2007) WHO
classification of tumours of the central nervous system, 4th
edn. IARC Lyon Press, Lyon
Matsuyama J, Mori T, Hori S, Nakano T, Yamada A (1989)
Gliosarcoma with multiple extracranial metastases. Case
report. Neurol Med Chir 29:938943
Mentrikoski M, Johnson MD, Korones DN, Scott GA (2008)
Glioblastoma multiforme in skin: a report of 2 cases
and review of the literature. Am J Dermatopathol 30:
381384
Miliaras G, Tsitsopoulos PP, Markoula S, Kyritsis A, Polyzoidis
KS, Malamou-Mitsi V (2009) Multifocal glioblastoma with
remote cutaneous metastasis: a case report and review of the
literature. Cen Eur Neurosurg 70:3942
Moon KS, Jung S, Lee MC, Kim IY, Kim HW, LEE JK, Kim TS
(2004) Metastatic glioblastoma in cervical lymph node after
repeated craniotomies: report of a case with diagnosis by fine
needle aspiration. J Korean Med Sci 19:911914
Mourad PD, Farrell L, Stamps LD, Chicoine MR, Silbergeld
DL (2005) Why are systemic glioblastoma metastases rare?
Systemic and cerebral growth of mouse glioblastoma. Surg
Neurol 63:511519
Mujic A, Hunn A, Taylor AB, Lowenthal RM (2006)
Extracranial metastases of a glioblastoma multiforme to
the pleura, small bowel and pancreas. J Clinic Neurosci
13:677681

Newton HB, Rosenblum MK, Walker RW (1992) Extraneural
metastases of infratentorial glioblastoma multiforme to the
peritoneal cavity. Cancer 69:21492153
Pelloski CE, Mahajan A, Maor M, Chang EL, Woo S, Gilbert
M, Colman H, Yang H, Ledoux A, Blair H, Passe S,
Jenkins RB, Aldape KD (2005) YKL-40 expression is
associated with poorer response to radiation and shorter
overall survival in glioblastoma. Clin Cancer Res 11:
33263334
Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH,
Wu TD, Misra A, Nigro JM, Colman H, Soroceanu L,
Williams PM, Modrusan Z, Feuerstein BG, Aldape K (2006)
Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble
stages in neurogenesis. Cancer Cell 9:157173
Rajagopalan V, El Kamar FG, Thayaparan R, Grossbard ML
(2005) Bone marrow metastases from glioblastoma multiforme: a case report and review of the literature. J
Neurooncol 72:157161
Saad AG, Sachs J, Turner CD, Proctor M, Marcus KJ, Wang
L, Lidov H, Ullrich NJ (2007) Extracranial metastases of
glioblastoma in a child: case report and review of the literature. J Pediatr Hematol Oncol 29:190194
149
Sadik AR, Port R, Garfinkel B, Bravo J (1984) Extracranial
metastasis of cerebral glioblastoma multiforme: case report.
Santos AV, Saraiva PF, Santiago B (2003) Extracranial metastasis of glioblastoma multiforme. Acta Med Port 16:209211
Schiffer D, Chi A, Giordana MT, Mauro A, Migheli A, Soffietti
R, Vigliani MC (1990) Vascular response to irradiation in
malignant gliomas. J Neurooncol 8:7384
Schultz S, Pinsky GS, Wu NC, Chamberlain MC, Rodrigo
AS, Martin SE (2005) Fine needle aspiration diagnosis of
extracranial glioblastoma multiforme: case report and review
of the literature. CytoJournal 2:19
Senetta R, Trevisan E, Rud R, Benech F, Soffietti R, Cassoni
P (2009) Skin metastases of glioblastoma in the absence of
intracranial progression are associated with a shift towards a
mesenchymal immunophenotype: report of two cases. Acta
Wallace CJ, Forsyth PA, Edwards DR (1996) Lymph node
metastases from glioblastoma multiforme. Am J Neuroradiol
17:19291931
Yasuhara T, Tamiya T, Meguro T, Ichikawa T, Sato Y, Date I,
Nakashima H, Ohmoto T (2003) Glioblastoma with metastasis to the spleencase report. Neurol Med Chir (Tokyo)
43:452456
Part II
Therapy
Chapter 17
Diffuse Low-Grade Gliomas: What Does Complete

Resection Mean?
Johan Pallud
Abstract Supratentorial
hemispheric
diffuse
low-grade gliomas (LGG), i.e., World Health
Organization grade II gliomas, are generally revealed
by seizures in young adults with no or only mild
neurological deficits. These progressive tumors are
characterized by a continuous growth, by tumor
recurrences and by a progression into a higher grade
of malignancy. Maximal safe resection preserving
eloquent brain areas, when possible, is currently
considered as the optimal primary treatment modality
of LGG. Imaging determinations of the spatial extent
of LGG are of paramount importance in evaluating the
risk-to-benefit ratio of surgical resection. However,
it is not yet clear how accurately MRI can delineate
LGG. Indeed, LGG may recur postoperatively even
after a MRI-based complete resection and recurrences
generally occurred in the resection margins. The value
of conventional MRI in determining the spatial extent
of LGG is thus questionable. As demonstrated by
multi-scale correlative approaches with histological
and imaging data, conventional MRI underestimates
the actual spatial extent of LGG, even when they
are well delineated on T2-weigthed and FLAIR
sequences. Cycling isolated tumor cells are present
beyond MRI-defined abnormalities and permeate
surrounding normal parenchyma at sites up to at
least 15 mm outside MRI-defined tumor limits. Clear
tumor boundaries do not actually exist as LGG are diffusely infiltrative tumors with a decrease of tumor cell
density as a function of distance from MRI-defined
J. Pallud ()
Service de Neurochirurgie, Hpital Sainte-Anne, Paris
Cedex 14, France
e-mail: johanpallud@hotmail.com
abnormalities. This implies that a MRI-based complete resection of a LGG leaves isolated tumor cells
beyond the surgical field. These results should be considered when planning a surgical resection of a LGG:
(1) a maximal safe resection preserving eloquent
brain area is recommended because of the infiltrative
nature of LGG and the frequent juxtaposition close
to and/or within critically eloquent brain areas;
(2) surgical resection should be tailored according
to cortico-subcortical functional boundaries rather
than MRI-based limits; (3) an extended resection of a
margin beyond MRI-defined abnormalities, whenever
feasible in non-eloquent brain areas, might increase
the survival of patients harboring a LGG; (4) an early
surgical treatment while the LGG is smaller might
delay tumor progression by decreasing the number of
residual isolated tumor cells.
Keywords LGG Complete resection WHO Grade
II gliomas Risk-to-benefit ratio MRI
Introduction
Supratentorial hemispheric diffuse low-grade gliomas
(LGG), i.e., World Health Organization (WHO) grade
II gliomas are generally revealed by seizures in young
adults with a normal life and no or only mild neurological deficits (Duffau, 2005). However, these progressive tumors are characterized by a continuous growth
(Mandonnet et al., 2003; Pallud et al., 2006), by tumor
recurrences and by a progression into a higher grade
of malignancy that is the major cause of mortality
(Duffau, 2005; Pallud et al., 2006).

153
154
As a consequence, the better knowledge of the natural history of LGG resulted in an active therapeutic
management and maximal safe resection preserving
eloquent brain areas, when possible, is currently considered as the optimal primary treatment modality
(Berger et al., 1994; Duffau, 2005, 2009; Sanai and
Berger, 2008; Soffietti et al., 2010). Indeed, despite the
lack of Class I evidence, there is growing evidence that
the extent of resection is a significant prognostic factor for progression free and overall survivals (Sanai
and Berger, 2008; Soffietti et al., 2010). The goal of
surgery is thus to optimize the extent of resection while
preserving the quality of life. Therefore, individual
determinations of the spatial extent of LGG and of the
cortical and subcortical functional organization are of
paramount importance in evaluating the risk-to-benefit
ratio of surgery (Duffau, 2009).
The spatial extent of LGG is defined preoperatively on conventional MRI although it is not clear
how accurately MRI can delineate LGG. Indeed, they
may recur after a MRI-based complete resection and
recurrences generally occurred at the site of the initial
tumor, in the resection margins (Kelly, 1992, 2004).
The value of MRI in determining the actual spatial
extent of LGG is questionable. There is thus a need in
assessing how MRI reflects their actual tumor limits.
Such data are best provided by pathological analysis of spatially oriented surgical samples obtained
from sites outside MRI-defined abnormalities. Studies
are lacking for LGG as such samples performed are
rarely available in a current neurosurgical practice.
Therefore, the aim of the present chapter is to clarify
the value of conventional MRI in delineating LGG and
determining their actual spatial extent. The absence
of a clear tumor boundary will be stressed at the
light of histologic-imaging correlative studies focused
on LGG.
Multiscale Approach of the Spatial

Configuration of Cerebral Gliomas
Microscopic Scale
The understanding of the spatial configuration and
growth pattern of a particular glioma is important
for planning the best therapeutic strategy. The first
J. Pallud
data were acquired on postmortem analyses before

the modern imaging era (Daumas-Duport and Szikla,
1981). They showed that cerebral diffuse gliomas:
Lack of a definite tumor capsule and spread
diffusely throughout the brain in many cases.
May present without formation of a definite mass
and without pronounced destruction of the preexisting tissue.
Present a microscopic tumor extension greater than
that observed grossly with a tumor cell infiltration
beyond the gross demarcation of the tumor.
The brain invasion of diffuse gliomas was found
to follow particular anatomical structures, such as
myelinated fiber tracts, basement membranes of blood
vessels or other basement membranelike structures,
subependyma and the perivascular spaces (Giese et al.,
2003).
The study of spatially oriented pathological samples
obtained from image-based serial stereotactic biopsies of all components of the imaging abnormalities
allowed the assessment of in vivo growth pattern of
diffuse gliomas (Daumas-Duport et al., 1983, 1987;
Daumas-Duport and Szikla, 1981; Kelly et al., 1987).
Daumas-Duport et al. studied the histopathological
tridimentionnal architectural pattern of diffuse gliomas
and identified two distinct tumor components:
The solid tumor tissue component where the
tumor cells are in contact with each other, containing microvascular proliferation and only residual
brain parenchyma or none.
The isolated tumor cells component associated
with edema but no microvascular proliferation, permeating functional brain parenchyma.
This allowed proposing a spatial classification
scheme with three tridimentionnal architectural subtypes for diffuse gliomas (Fig. 17.1) (Daumas-Duport
et al., 1983, 1987; Daumas-Duport and Szikla, 1981;
Kelly et al., 1987):
Type I: solid tumor tissue only without any peripheral isolated tumor cells.
Type II: solid tumor tissue and peripheral
parenchyma infiltrated by isolated tumor cells.
Type III: brain parenchyma infiltrated by isolated
tumor cells only without solid tumor tissue.
17 Diffuse Low-Grade Gliomas: What Does Complete Resection Mean?
Fig. 17.1 (a, b) Histopathological tridimentionnal architectural

pattern of diffuse gliomas. The three different architectural subtypes comprise a solid tumor tissue component where the
tumor cells are in contact with each other, containing microvascular proliferation and only residual brain parenchyma or none
and an isolated tumor cells component associated with edema
but no microvascular proliferation, permeating functional brain
parenchyma. (Architectural subtype I: solid tumor tissue only
without any peripheral isolated tumor cells; Subtype II: solid
tumor tissue and peripheral parenchyma infiltrated by isolated
tumor cells; Subtype III: brain parenchyma infiltrated by isolated tumor cells only without solid tumor tissue). (c, d, e, f)
The WHO grade IV gliomas are architectural type II tumors.
155
A solid tumor tissue component is observed peroperatively

(c). They comprised generally a contrast-enhanced mass on
axial T1-weighted sequences after gadopentetate dimeglumine
(d) surrounded by peripheral hypointensity on T1-weighted
sequences (f) and by hyperintensity on FLAIR sequences (e).
(g, h, i, j) The WHO grade II gliomas are architectural type
III tumors. Isolated tumors cells infiltrate the brain parenchyma
and expand the involved gyri, as observed peroperatively (h).
They comprise generally an hypointensity on axial T1-weighted
sequences (i) and an hyperintensity on axial FLAIR sequences
(j) without definite contrast-enhancement on axial T1-weighted
sequences after gadopentetate dimeglumine (g)
156
The identification of different architectural subtypes

for diffuse gliomas has provided a guide for surgical
decision-making. Indeed, the spatial architectural subtype III of LGG, implying a possible tumor invasion
of functional brain areas, argues for the use of functional preoperative and peroperative mapping methods
to optimize the benefit-to-risk ratio of surgical resection. In addition, this spatial classification scheme is
generally correlated with WHO tumor grades:
Pilocytic astrocytomas are primarily type I tumors.
Grade II gliomas (LGG) are frequently type III
tumors comprised only of a volume of brain
parenchyma infiltrated by isolated tumor cells.
However, there are some LGG that are type II
because patches of tumor tissue exist within areas
of infiltrated parenchyma.
The majority of higher-grade diffuse gliomas
(WHO grade III and IV) are type II tumors; they
have a central area of solid tumor tissue surrounded
by a zone of peripheral infiltration.
Macroscopic Scale
The MRI findings are correlated with the architectural
subtypes of the spatial classification scheme described
previously (Fig. 17.1):
Type I corresponds to a contrast-enhancing mass
with no peripheral hypointensity on T1-weighted
and no hyperintensity on T2-weighted or FLAIR
sequences.
Type II corresponds to a contrast-enhanced mass
surrounded by peripheral hypointensity on T1weighted and by hyperintensity on T2-weighted or
FLAIR sequences.
Type III corresponds to an hypointense on T1weighted and hyperintense on T2-weighted or
FLAIR sequences area without definite contrastenhancement.
The architectural subtype of LGG explains the typical absence of contrast enhancement and why its
occurrence is associated with a worsened prognosis
(Pallud et al., 2009). Indeed, contrast enhancement
should be considered as a key event in the malignant
progression of LGG as it reflects macroscopically the
microscopic neoangiogenesis.
J. Pallud
The tumor extension along myelinated white matter fiber tracts is easily demonstrated on MRI. Several
reports have illustrated that glioma cells migrate along
intra-hemispheric tracts, inter-hemispheric tracts and
descending pathways and use them as preferential
ways of invasion (Mandonnet et al., 2006; Pallud et al.,
2005).
Dynamic Scale
Mandonnet et al. first showed quantitatively the spontaneous radiological growth of LGG on successive
MRIs in a subset of 27 patients that were followed
before oncological treatment (Mandonnet et al., 2003).
The study suggested a linear evolution of the mean
tumor diameter over time (velocity of diametric expansion) quantified at a 4 mm/year average rate. These
results were confirmed on a larger series of 143 LGG
that ranged individual velocities of diametric expansion from 1 to 36 mm/year and demonstrated the strong
prognostic significance of individual tumor growth
rates on overall survival (Pallud et al., 2006). As other
teams reported similar observations (Hlaihel et al.,
2010; Peyre et al., 2010; Ricard et al., 2007), it is now
accepted that LGG present a spontaneous and continuous radiological growth before any transformation into
a higher grade of malignancy.
MRI-Based Estimation of the Spatial

Extent of LGG
Static Approach
Conventional MRI is widely used in practice to study
the anatomical extent of diffuse gliomas (Laster et al.,
1984; Price, 2009). LGG appear as a hypointense area
on T1-weighted sequence and hyperintense area on
T2-weighted or FLAIR sequences. The margins of
LGG is usually indistinct on T1-weighted sequence
and is better demonstrated on T2-weighted and FLAIR
sequences that show maximal visible imaging abnormalities (Bynevelt et al., 2001; Connor et al., 2004;
Price, 2009). The hyperintense area is either with
well-circumscribed margins or with indistinct borders,
particularly on oligodendrogliomas with a loss of heterozygothy at chromosome 1p19q (Daumas-Duport

et al., 1997; Jenkinson et al., 2006).
It is suggested that Proton Magnetic Resonance
Spectroscopic Imaging (H1 -MRSI) may be more valuable than conventional MRI to assess the spatial extent
of diffuse gliomas, as abnormal metabolic areas may
exceed areas of MRI-defined abnormalities (Croteau
et al., 2001; Ganslandt et al., 2005). One previous
report studied histological data of peripheral areas of
gliomas that appeared normal on MRI and abnormal
on H1 -MRSI and identified tumor infiltration, using
proliferation markers, in four studied LGG (Ganslandt
et al., 2005). However, this study could not provide
reliable data regarding the true spatial extent of LGG
from H1 -MRSI since it included gliomas of different
histopathological subtypes and grades without imaging data, the biopsy samples were performed with
a frameless stereotactic system within one centimeter from MRI-defined abnormalities. In addition, the
MRSI techniques are currently limited by the poor resolution (Price et al., 2006). Studies have shown that
diffusion tensor imaging reveals peritumoral abnormalities in diffuse gliomas that are not apparent on
conventional MRI and that diffusion tensor imaging
may delineate tumor margins better than conventional
MRI (Price et al., 2006).
Dynamic Approach
As LGG present a continuous growth, studying their
radiological changes over time by the mean of repeated
MRI is a simple way to follow the changes in the spatial extent of LGG (Mandonnet et al., 2008). Imaging
changes are correlated with changes in the natural
history of LGG and have been shown to be associated with a worsened prognosis, as they reflect the
progression into a higher grade of malignancy:
An elevated tumor growth rate with a velocity of

diametric expansion at 8 mm/year or more (Pallud
et al., 2006).
The change over time of a contrast enhancement
pattern or the occurrence of a new contrast enhancement (Pallud et al., 2009).
157
Does MRI Reflect the Actual Spatial

Extent of LGG?
The actual spatial extent of LGG and their histological
margins of tumor infiltration are difficult to determine.
Previous studies have demonstrated that conventional
MRI underestimate the spatial extent of malignant
gliomas and that isolated tumor cells are observed well
beyond MRI-defined abnormalities on T2-weighted
sequence (Daumas-Duport et al., 1983, 1987; DaumasDuport and Szikla, 1981; Earnest et al., 1988; Kelly,
1992; Kelly et al., 1987). The accurate correlation
of MRI with histopathological findings remains the
only means of characterizing imaging abnormalities
and of establishing tumor margins. However, such data
are scarse for LGG as samples performed beyond
MRI-defined abnormalities are rarely available.
Iwama et al. reported a correlation study between
MRI and postmortem histopathological examination
in cerebral gliomas (Iwama et al., 1991). In one
studied LGG, scattered isolated tumor cells were
identified beyond MRI-defined abnormalities on T2weighted sequence. These findings might possibly be
explained by tumor growth, since a 19-week interval elapsed between MRI and post-mortem studies.
An ideal correlative study would imply correlating
MRI and histopathological specimens taken simultaneously. Watanabe et al. identified isolated tumor
cells in spatially oriented surgical samples obtained
from sites outside MRI-defined abnormalities on T2weighted sequence, from five of the eight studied LGG
(Watanabe et al., 1992). These results also question the
accuracy of histological and imaging correlation procedures as the location of specimens was performed
intraoperatively, based on the depth below the anatomically identified cortex. Accurate real-time correlative
studies between MRI and spatially oriented specimens
could be allowed by image-based serial stereotactic
biopsies.
Stereotactic spatially oriented histological specimens from LGG peripheral areas that appeared normal
on MRI have already been presented. Using T1weighted MRI-based stereotactic biopsies, Jenkinson
et al. (2006) showed an insidious change in cellularity
at the radiological margin on T1-weighted sequence
in most of 43 LGG. The spatially limited sampling
23 mm beyond the MRI-defined tumor border did
not allow a precise definition of the spatial extent
158
of LGG. Earnest et al. showed the tumor to be confined to the lesion margins as demonstrated by MRI
on T1-weighted and T2-weighted in one case of LGG
(Earnest et al., 1988). In contrast, Kelly et al. (1987)
reported a study of serial stereotactic biopsies obtained
from 10 untreated LGG and identified isolated tumor
cells in three of the five biopsy specimens obtained
beyond MRI-defined abnormalities on T2-weighted
sequence. However, these studies included a small
number of LGG, were based on histological morphological criteria alone, and did not specify the MRI
tumor delineation of the studied LGG.
MRI Underestimates the Spatial Extent

of LGG
In a recent study, Pallud et al. (2010) retrospectively
analyzed all MRI-based serial stereotactic biopsies that
were performed according to the Talairach stereotactic
method in their institution for the diagnosis of diffuse cerebral gliomas. Such stereotactic biopsies were
performed until 2002 to sample all components of
the imaging abnormalities for an accurate pathological
diagnosis, for classification into histological subtypes
and grading and to establish the real tumor extent
in a surgical context (Daumas-Duport et al., 1983;
Daumas-Duport and Szikla, 1981). They selected,
from a series of 109 procedures performed for the diagnosis of a WHO grade II glioma from January 1992 to
December 2001, 16 untreated adult patients harboring
a well-delineated and non contrast-enhanced supratentorial LGG for whom MRI-based serial stereotactic
biopsies were performed, in part, beyond MRI-defined
abnormalities before any oncological treatment. In
addition to classical morphological histological criteria, the identification of isolated tumor cells was
ascertained using multiple immunostainings and cell
cycle marker (MIB-1 immunostaining). Moreover, the
inaccuracy in the histological and imaging correlation
procedure has been limited by several features of this
study:
The short interval between preoperative MRI and
histological diagnosis should restrict tumor growth
between imaging and biopsy.
Biopsy samples were performed in a single operation before any treatment, so that artificially induced
J. Pallud
ultrastructural changes of the tumor should be limited and the use of a Sedan-Vallicioni biopsy cannula should act to reduce brain movements during
biopsy procedures.
The superimposition of reformatted preoperative
MRI and intraoperative teleradiographic X-rays
taken after each biopsy samples allowed each
biopsy site to be directly displayed on MRI on the
same reference plane. The superimposition accuracy was controlled with postoperative MRI or
CT-scan demonstrating the biopsy tract.
Only isolated tumor cells detected in biopsy samples that were at least 10 mm distant from MRIdefined abnormalities were considered.
A total of 101 biopsy samples (median 6, range
39 per patient) were performed in the 16 patients.
A total of 37 biopsy samples (median 2.2, range
15 per patient) were performed beyond MRI-defined
abnormalities on T2-weighted and FLAIR sequences.
The maximal distance of the biopsy samples sites from
MRI-defined abnormalities on T2-weighted sequence
ranged from 10 to 26 mm.
In all biopsy samples performed beyond MRIdefined abnormalities, the cortex and the white
matter had a normal appearance on routine staining.
There was no increased cell density, no edema and
no gliosis.
MIB-1 immunostaining revealed MIB-1 positive
cells (i.e., cycling cells), in all but two of the 37
samples where MIB-1 positive cells were indifferently distributed within the cortex and the white
matter and their number was often variable from
one BS to one another in a same patient.
None of the MIB-1 positive cells coexpressed glial
fibrillary acidic protein and that all MIB-1 positive cells coexpressed OLIG2, thus excluding the
possibility that MIB-1 positive cells correspond to
reactive astrocytes or activated microglia.
MIB-1 positive cells identified beyond MRI-defined
abnormalities on T2-weighted and FLAIR sequences
were cycling isolated tumor cells, since (Fig. 17.2):
Their morphological characteristics reflected those
of tumor cells.
Their number was significantly higher than that of
non-tumor controls.
159
Of note, in this study MIB-1 immunostaining was

used to detect cycling tumor cells in brain parenchyma
surrounding MRI-defined LGG. MIB-1 immunostaining only labels the fraction of the tumor cells, which
at a given time are cycling cells. However, the proliferation rates are low for LGG. Therefore, only a
low percentage of tumor cells (i.e. the cycling tumor

cells) could be identified using MIB-1 immunostaining. Moreover, glioma cells under migration present
a decreased proliferation rate (Giese et al., 2003). As
isolated tumor cells may represent invasive gliomas
cells, it is then probable that the brain parenchyma
around MRI-defined abnormalities on T2-weighted
and FLAIR sequences contains a higher proportion
of tumor cells than those demonstrated with MIB-1
immunostaining. This study demonstrated the inability of conventional MRI on T2-weighted and FLAIR
sequences in determining the actual spatial extent
of LGG.
Fig. 17.2 (a) Preoperative MRI of a patient on axial T2weighted sequences demonstrating the location of serial stereotactic biopsy samples performed in a right fronto-insular WHO
grade II glioma. Each biopsy samples are defined as inside
(black) or outside (white) areas of hypersignal. (b) A significant
decreasing gradient of the number of MIB-1 positive cells with
the distance from the MRI-defined abnormalities is observed.
The number of cycling cells is expressed as MIB-1 positive
cells/cm2 and the distance from MRI-defined abnormalities is
expressed in mm. (c, d, e, f) Comparative histological features of
biopsy samples performed within (c) and outside (d, e, f) MRIdefined abnormalities. Conventional Hemalun-Phloxin stainings
showed that samples performed within the MRI-defined tumor

limits (c) are constituted, in the white matter, of infiltrative tumor
cells associated by interstitial edema and gliosis (400). In samples performed outside MRI-defined abnormalities (d), the white
matter cell density appeared normal without any edema or gliosis (400). Double immunostainings revealing that cycling cells
do not shared astrocytic marker but correspond to Olig2 positive
cells. Double immunufluorescent labelling (e) showed that all
MIB-1 positive cells (green) co-expressed the oligodendrocyte
cell marker olig2 (red). Double chromogenic immunostaining
(f) revealed that MIB-1 positive cells (red) do not share Glial
Fibrillary Acidic Protein astrocytic marker (brown) (600)
Their number per square centimeter significantly

decreased with distance from the MRI-defined
tumor borders.
Their number was significantly correlated with the
tumor growth fraction.
160
Practical Conclusions
As demonstrated by Pallud et al. (2010) using a
multi-scale correlative approach with histological and
imaging data, conventional MRI underestimates the
actual spatial extent of LGG, even when they are well
delineated on T2-weigthed and FLAIR sequences:
Cycling isolated tumor cells are present beyond
MRI-defined abnormalities in all LGG studied and
permeate surrounding normal parenchyma.
Isolated tumor cells can be detected at sites up to at
least 15 mm beyond MRI-defined abnormalities.
It is clearly difficult for present imaging techniques
to distinguish between intact brain parenchyma and
that infiltrated by scattered isolated tumor cells. As
already explicited by Kelly (2004), tumor cells could
be found far from any MRI-defined abnormality, even
in LGG. Thus, clear tumor boundaries do not actually exists and LGG are diffusely infiltrative tumors
with a decrease of tumor cell density as a function
of distance from the tumor component associated with
abnormalities on conventional MRI. This implies that
a gross total removal of a LGG either macroscopically
or on postoperative imaging leaves isolated tumor cells
beyond the surgical field. This observation may explain
clinical events hampering the natural history of LGG:
The observation of tumor recurrences in the margin
of the resection (Kelly, 1992, 2004) were isolated
tumor cells are the more numerous.
The prognostic value of the extent of resection
for progression free survival (Sanai and Berger,
2008) as extensive resection decreases the number
of remaining isolated tumor cells.
The prognostic value of the extent of resection
for overall survival (Berger et al., 1994; Sanai
and Berger, 2008) as extensive resection theoretically decreases the cumulative odds of malignant
progression of residual tumor cells.
The prognostic value of the tumor volume (Berger
et al., 1994) as larger LGG probably have more
surrounding isolated tumor cells and more residual
tumor cells after surgical resection.
As a practical consequence, the knowledge of the
inability of MRI in determining the spatial extent of
J. Pallud
LGG is of paramount importance in evaluating the

risk-to-benefit ratio of surgery, which goal is optimizing the extent of resection while preserving the quality
of life. These results should be considered when planning a surgical resection of a LGG and several points
have to be highlighted:
A maximal safe resection preserving eloquent brain
area is recommended because of the infiltrative
nature of LGG and the frequent juxtaposition close
to and/or within critically eloquent brain areas.
Surgical resection should be tailored according
to cortico-subcortical functional boundaries rather
than MRI-based limits.
An extended resection of a margin beyond MRIdefined abnormalities, whenever feasible in noneloquent brain areas, might increase the survival of
patients harboring a LGG.
An early surgical treatment while the LGG is
smaller might delay tumor progression by decreasing the number of residual isolated tumor cells that
escape the gross total removal.
Acknowledgments Johan Pallud wants to thank FranoisXavier Roux, Edouard Dezamis and Bertrand Devaux of the
department of Neurosurgery, Catherine Daumas-Duport and
Pascale Varlet of the department of Neuropathology, JeanFranois Meder and Catherine Oppenheim of the department
of Neuroradiology of the Sainte-Anne Hospital Center in Paris,
France. Johan Pallud wants to thanks all the members of the
French Glioma Network (REG, Rseau dEtude des Gliomes)
and particularly Emmanuel Mandonnet, Laurent Capelle, Luc
Taillandier and Hugues Duffau.
References
Berger MS, Deliganis AV, Dobbins J, Keles GE (1994) The
effect of extent of resection on recurrence in patients
with low grade cerebral hemisphere gliomas. Cancer 74:
17841791
Bynevelt M, Britton J, Seymour H, MacSweeney E, Thomas N,
Sandhu K (2001) FLAIR imaging in the follow-up of
low-grade gliomas: time to dispense with the dual-echo?
Neuroradiology 43:129133
Connor SE, Gunny R, Hampton T, OGorman R (2004)
Magnetic resonance image registration and subtraction in the
assessment of minor changes in low grade glioma volume.
Eur Radiol 14:20612066
Croteau D, Scarpace L, Hearshen D, Gutierrez J, Fisher JL,
Rock JP, Mikkelsen T (2001) Correlation between magnetic
resonance spectroscopy imaging and image-guided biopsies:
semiquantitative and qualitative histopathological analyses
of patients with untreated glioma. Neurosurgery 49:823829

Daumas-Duport C, Meder JF, Monsaingeon V, Missir O, Aubin
ML, Szikla G (1983) Cerebral gliomas: malignancy, limits and spatial configuration. Comparative data from serial
stereotaxic biopsies and computed tomography (a preliminary study based on 50 cases). J Neuroradiol 10:5180
Daumas-Duport C, Monsaigneon V, Blond S, Munari C,
Musolino A, Chodkiewicz JP, Missir O (1987) Serial stereotactic biopsies and CT scan in gliomas: correlative study
in 100 astrocytomas, oligo-astrocytomas and oligodendrocytomas. J Neurooncol 4:317328
Daumas-Duport C, Szikla G (1981) Definition of limits and
3D configuration of cerebral gliomas. Histological data,
therapeutic incidences (authors transl). Neurochirurgie 27:
273284
Daumas-Duport C, Varlet P, Tucker ML, Beuvon F, Cervera P,
Chodkiewicz JP (1997) Oligodendrogliomas. Part I: Patterns
of growth, histological diagnosis, clinical and imaging correlations: a study of 153 cases. J Neurooncol 34:3759
Duffau H (2005) Lessons from brain mapping in surgery for lowgrade glioma: insights into associations between tumor and
brain plasticity. Lancet Neurol 4:476486
Duffau H (2009) Surgery of low-grade gliomas: towards a
functional neurooncology. Curr Opin Oncol 21:543549
Earnest Ft, Kelly PJ, Scheithauer BW, Kall BA, Cascino TL,
Ehman RL, Forbes GS, Axley PL (1988) Cerebral astrocytomas: histopathologic correlation of MR and CT contrast enhancement with stereotactic biopsy. Radiology 166:
823827
Ganslandt O, Stadlbauer A, Fahlbusch R, Kamada K, Buslei R,
Blumcke I, Moser E, Nimsky C (2005) Proton magnetic resonance spectroscopic imaging integrated into image-guided
surgery: correlation to standard magnetic resonance imaging
and tumor cell density. Neurosurgery 56:291298
Giese A, Bjerkvig R, Berens ME, Westphal M (2003) Cost of
migration: invasion of malignant gliomas and implications
for treatment. J Clin Oncol 21:16241636
Hlaihel C Guilloton L, Guyotat J, Streichenberger N,
Honnorat J, Cotton F (2010) Predictive value of multimodality MRI using conventional, perfusion, and spectroscopy
MR in anaplastic transformation of low-grade oligodendrogliomas. J Neurooncol 97:7380
Iwama T, Yamada H, Sakai N, Andoh T, Nakashima T, Hirata T,
Funakoshi T (1991) Correlation between magnetic resonance
imaging and histopathology of intracranial glioma. Neurol
Res 13:4854
Jenkinson MD, du Plessis DG, Smith TS, Joyce KA, Warnke PC,
Walker C (2006) Histological growth patterns and genotype
in oligodendroglial tumors: correlation with MRI features.
Brain 129:18841891
Kelly PJ (1992) Stereotactic resection and its limitations in glial
neoplasms. Stereotact Funct Neurosurg 59:8491
Kelly PJ (2004) Technology in the resection of gliomas and the
definition of madness. J Neurosurg 101:284286
Kelly PJ, Daumas-Duport C, Kispert DB, Kall BA, Scheithauer
BW, Illig JJ (1987) Imaging-based stereotaxic serial biopsies in untreated intracranial glial neoplasms. J Neurosurg
66:865874
Laster DW, Ball MR, Moody DM, Witcofski RL, Kelly DL
Jr (1984) Results of nuclear magnetic resonance with
cerebral glioma. Comparison with computed tomography.
Surg Neurol 22:113122
161
Mandonnet E, Capelle L, Duffau H (2006) Extension of paralimbic low grade gliomas: toward an anatomical classification based on white matter invasion patterns. J Neurooncol
78:179185
Mandonnet E, Delattre JY, Tanguy ML, Swanson KR, Carpentier
AF, Duffau H, Cornu P, Van Effenterre R, Alvord EC Jr,
Capelle L (2003) Continuous growth of mean tumor diameter
in a subset of grade II gliomas. Ann Neurol 53:524528
Mandonnet E, Pallud J, Clatz O, Taillandier L, Konukoglu E,
Duffau H, Capelle L (2008) Computational modeling of the
WHO grade II glioma dynamics: principles and applications
to management paradigm. Neurosurg Rev 31:263269
Pallud J, Capelle L, Taillandier L, Fontaine D, Mandonnet E,
Guillevin R, Bauchet L, Peruzzi P, Laigle-Donadey F, Kujas
M, Guyotat J, Baron MH, Mokhtari K, Duffau H (2009)
Prognostic significance of imaging contrast enhancement for
WHO grade II gliomas. Neurooncology 11:176182
Pallud J, Devaux B, Daumas-Duport C, Oppenheim C, Roux FX
(2005) Glioma dissemination along the corticospinal tract.
J Neurooncol 73:239240
value of initial magnetic resonance imaging growth rates for
World Health Organization grade II gliomas. Ann Neurol
60:380383
Pallud J, Varlet P, Devaux B, Geha S, Badoual M, Deroulers
C, Page P, Dezamis E, Daumas-Duport C, Roux FX (2010)
Diffuse low-grade oligodendrogliomas extend beyond MRIdefined abnormalities. Neurology 74:17241731
Peyre M, Cartalat-Carel S, Meyronet D, Ricard D, Jouvet A,
Pallud J, Mokhtari K, Guyotat J, Jouanneau E, Sunyach
MP, Frappaz D, Honnorat J, Ducray F (2010) Prolonged
response without prolonged chemotherapy: a lesson from
PCV chemotherapy in low-grade gliomas. Neurooncology
20:10781082
Price SJ (2009) Advances in imaging low grade gliomas. Adv
Tech Stand Neurosurg 35:134
Price SJ, Jena R, Burnet NG, Hutchinson PJ, Dean AF, Pea
A, Pickard JD, Carpenter TA, Gillard JH (2006) Improved
delineation of glioma margins and regions of infiltration with
the use of diffusion tensor imaging: an image-guided biopsy
study. AJNR Am J Neuroradiol 27:19691974
Mandonnet E, Kujas M, Mokhtari K, Taillibert S, LaigleDonadey F, Carpentier AF, Omuro A, Capelle L, Duffau
H, Cornu P, Guillevin R, Sanson M, Hoang-Xuan K,
Delattre JY (2007) Dynamic history of low-grade gliomas
before and after temozolomide treatment. Ann Neurol 61:
484490
Sanai N, Berger MS (2008) Glioma extent of resection and its
impact on patient outcome. Neurosurgery 62:753764
Soffietti R, Baumert BG, Bello L, von Deimling A, Duffau H,
Frenay M, Grisold W, Grant R, Graus F, Hoang-Xuan K,
Klein M, Melin B, Rees J, Siegal T, Smits A, Stupp R,
Wick W (2010) Guidelines on management of low-grade
gliomas: report of an EFNS-EANO Task Force. Eur J Neurol
17:11241133
Watanabe M, Tanaka R, Takeda N (1992) Magnetic resonance imaging and histopathology of cerebral gliomas.
Chapter 18
Quantitative Approach of the Natural Course of Diffuse

Low-Grade Gliomas
Johan Pallud and Emmanuel Mandonnet
Abstract Supratentorial hemispheric diffuse lowgrade gliomas (LGG), i.e., World Health Organization
grade II gliomas, are a heterogeneous group of tumors
with distinct clinical, histological and molecular characteristics. The prognosis of LGG varies between
series and reflects their heterogeneity with different
subgroups harboring specific intrinsic properties. The
natural course of LGG, as observed in clinical practice
can be summarized as a three-step process although
there is an actual continuum from low-grade to
high-grade of malignancy. The two first steps correspond to the histological low-grade of malignancy
with an initial silent period before clinical revelation
followed by a symptomatic period. The third step
corresponds to the progression to a higher grade of
malignancy leading to neurological disability and ultimately to death. It has been well demonstrated that
LGG are progressive tumors that present a systematic, spontaneous and continuous radiological growth
all along their natural course, even during the initial
silent period and during the symptomatic period before
any transformation into a higher grade of malignancy.
The Velocity of Diametric Expansion (VDE), estimated from the evolution of the mean tumor diameter
over time, can easily quantify the radiological tumor
growth. The median VDE is at about 4 mm/year for
LGG albeit with a great heterogeneity. Several intrinsic factors may influence spontaneous VDE (1p19q
codeletion and p53 overexpression) or not (histological subtype) and extrinsic factors may modify VDE
J. Pallud ()
Service de Neurochirurgie, Hpital Sainte-Anne, 14 Paris
Cedex, France
e-mail: johanpallud@hotmail.com
(hormonal changes during pregnancy). The spontaneous VDE has a strong prognostic significance on
overall and progression free survivals. As a consequence, the analysis of the spontaneous VDE, a
dynamic macroscopic parameter easily available in
clinical practice, may be a useful tool to overcome
biological diversity of LGG and could be integrated
along with the other static parameters (histological and molecular analyses) in a multi-scale approach
to understand better the individual natural course of
LGG. The VDE remains unchanged after surgical
resection, whereas it decreases markedly during and
after chemotherapy with temozolomide or PCV. Thus,
a precise assessment of the VDE obtained before and
after treatment would help guiding and analyzing the
effects of different oncological treatment modalities on
an individual basis.
Keywords LGG VDE Heterogeneity MTD
codeletion overexpression
Introduction
Supratentorial hemispheric diffuse low-grade gliomas
(LGG), i.e., World Health Organization grade II
gliomas, are a heterogeneous group of tumors with distinct clinical, histological and molecular characteristics
(Soffietti et al., 2010). The prognosis of LGG varies
between series and reflects their heterogeneity with
different subgroups harboring specific intrinsic properties. Indeed, the 5-year overall and progression-free
survival rates in randomized studies range from 58 to
72% and 37 to 55%, respectively (Soffietti et al., 2010).
In addition, the natural history of LGG is complex
and poorly understood. During their natural course,

163
164
LGG grow continuously and tend to progress to a

higher grade of malignancy, leading to neurological
disability and ultimately to death. Hence, reference
to these lesions as benign gliomas has been abandoned (Lang and Gilbert, 2006). However, a wait and
see policy has been proposed in the past years in the
initial management of LGG before their progression
towards a higher grade of malignancy. Thus, long clinical and radiological follow-ups with repeated MRI of
untreated LGG are available (Wessels et al., 2003).
They allow correlating the quantitative radiological
growth rates over time with clinical and histological
features.
Therefore, the aim of the present chapter is to
review quantitative data regarding the natural history
of LGG. A 3-step natural history scheme will be proposed. We will show that LGG present systematically a
spontaneous and continuous radiological tumor growth
before oncological treatment and before progression
into a higher grade of malignancy. The practical implications will be detailed, mainly the prognostic significance of spontaneous growth rates of LGG regarding overall survival and progression-free survival, and
the effects of oncological treatments on radiological
growth rates.
The Three-Step Natural History of LGG

A Step-By-Step Approach of a Continuum
The natural course of LGG, as observed in clinical
practice can be summarized as a three-step process.
The two first steps correspond to the histological lowgrade of malignancy with an initial silent period
before clinical revelation followed by a symptomatic
period. The third step corresponds to the progression
to a higher grade of malignancy. Beyond this artificial clustering of the LGG natural course, inspired
by WHO histological classifications, one should keep
in mind the actual continuum from low-grade to
high-grade of malignancy as the tumor acquires
progressively genotypic and phenotypic characteristics
leading to malignancy.
The Initial Silent Period

The natural course of LGG during the initial silent
period and their date of birth are unknown and
J. Pallud and E. Mandonnet
only analyses of incidental LGG could help the

understanding. However, the natural course of an incidental LGG has been very scarcely reported.
In a recent report, Pallud et al. (2010a) reviewed
1296 cases of histologically confirmed supratentorial
hemispheric LGG in adults and identified 47 incidental tumors, defined as a previously undetected and
incidental LGG at the time of radiological diagnosis that was unexpectedly discovered and unrelated to
the purpose of the MRI examination in asymptomatic
patients. They showed that incidental LGG are similarly proliferative tumours continuously growing on
sequential MRI at the same rates than the symptomatic
LGG:
The radiological tumour growth and the proliferation rates of incidental LGG were within reported
ranges of symptomatic LGG.
When not treated, incidental LGG became symptomatic with time, at a median interval of 48 months
since radiological discovery.
Incidental LGG present tumor recurrence and progression to higher grade of malignancy and led to
death in several cases despite oncological treatment.
In addition, the data suggested that incidental LGG
represent an earlier step than symptomatic LGG in
the tumor natural course as:
Age at radiological discovery was lower.
Incidental LGG presented with significantly smaller
tumour volumes. Accordingly, tumours were limited to one lobe in most of cases, corpus callosum
involvement was rare and no midline crossing was
observed.
Only rare cases of contrast enhancement were
observed, whereas contrast enhancement has been
reported in ~16% of cases for all LGG (Pallud et al.,
2009a).
As the date of birth of a LGG is not known, the
tumor genesis is still a mater of debate. They probably occur during the young adult period as suggested
by rare reports of a normal MRI examination several years before the LGG discovery. This finding is
in agreement with literature, which showed that LGG
are distinct entities in pediatric and adult populations
with different natural courses and prognoses (Bristol,
2009).
18 Quantitative Approach of the Natural Course of Diffuse Low-Grade Gliomas
The Symptomatic Period of Low-Grade of

Malignancy
Histologically, LGG comprise generally only a volume
of isolated tumor without solid tumor tissue associated
with edema but no microvascular proliferation permeating brain parenchyma preferentially along particular
anatomical structures (Daumas-Duport et al., 1997;
Giese et al., 2003). They correspond to a spatial architectural subtype III, as described by Daumas-Duport
et al. (1983, 1987) and Daumas-Duport and Szikla
(1981). However, there are some LGG that are type
II because patches of solid tumor tissue exist within
areas of infiltrated parenchyma by isolated tumor cells.
In addition, the proliferation rates of LGG are very low
(Soffietti et al., 2010).
These spatial architectural findings are correlated
with the MRI findings:
Repeated MRI examinations demonstrate a spontaneous and slow growth pattern of LGG (Chen et al.,
2010; Mandonnet et al., 2003).
Isolated tumor cells present with an area of
hypointensity on T1-weighted sequence and
hyperintensity on T2-weighted and FLAIR
sequences. MRI abnormalities are either with
well-circumscribed margins or with indistinct
borders, involve the cortex from the surface to its
deeper part, extend more or less in the adjacent
white matter and occasionally expand the involved
gyrus (Daumas-Duport et al., 1997).
Definite contrast enhancement is typically absent
in accordance with the architectural subtype III
of LGG. Indeed, contrast enhancement reflects
macroscopically the microvascular proliferation
thus explaining why its occurrence is associated
with a worsened prognosis (Pallud et al., 2009a).
The preferential brain invasion along myelinated
white matter fiber tracts is easily demonstrated
on MRI. Several reports have illustrated that
glioma cells migrate along intra-hemispheric tracts,
inter-hemispheric tracts and descending pathways
(Mandonnet et al., 2006; Pallud et al., 2005).
Of note, it has to be underlined that conventional MRI underestimates the actual spatial extent
of LGG, even when they are well delineated on T2weigthed and FLAIR sequences since cycling isolated
tumor cells can be detected at sites up to at least
165
15 mm beyond MRI-defined abnormalities (Pallud

et al., 2010c).
Taken together, histological and imaging abnormalities explain the silent initial period and the long
symptomatic period with no or mild neurological disorders:
The invasive behavior of LGG explains why functional areas may persist within brain parenchyma
permeated by isolated tumor cells (Daumas-Duport
et al., 1997).
The slow tumor growth allows functional compensation by brain plasticity mechanisms to be
efficiently implemented, explaining the lack of neurological deficits despite frequent involvement of
eloquent areas (Duffau, 2005, 2009).
The superficial tumor location and the frequent
cortical involvment explains why seizures are the
commonest presenting symptom and occur in ~90%
of LGG (Soffietti et al., 2010).
The lack of solid tumor tissue component explains
the rare occurrence of focal neurological deficit
and of increased intracranial pressure, despite clear
radiological mass effects (Soffietti et al., 2010).
Of note, LGG may remain asymptomatic until progression to a higher grade of malignancy.
The Malignant Progression

The time to progression to a higher grade of malignancy is variable from one patient to one another
since:
LGG are a heterogeneous group of tumors with
distinct clinical, histological and molecular characteristics.
LGG are diagnosed at various steps along the continuum of their natural course.
Histologically, a solid tumor tissue component
where the tumor cells are in contact with each other,
containing microvascular proliferation and only residual brain parenchyma or none, occurs within the
isolated tumor cells component. The tumor now corresponds to a spatial architectural subtype II, as
described by Daumas-Duport et al. (1983, 1987) and
Daumas-Duport and Szikla (1981). The malignant
transformation is spatially heterogeneous and areas of
166
high-grade and low-grade of malignancy may coexist. It is sometimes difficult to perform an accurate
histological grading due to the limitations of histological sampling. In addition, proliferation rates
increase.
These spatial architectural findings are correlated
with the MRI findings:
A change of the radiological growth pattern with
growth rates at ranges over those of LGG (Pallud
et al., 2006).
The occurrence of nodular-like and ring-like contrast enhancement patterns reflecting macroscopically the microvascular proliferation within the
architectural subtype II. They are known to be associated with a worsened prognosis (Pallud et al.,
2009a).
The occurrence of necrosis within the contrast
enhanced areas.
Taken together, histological and imaging abnormalities
explain functional deficits that are associated with
malignant transformation. Neurological disability
occurs as:
Brain plasticity mechanisms are overtaken by the
fast growing tumor.
The increased mass effect and intracranial pressure
injure the peripheral brain areas that were initially
recruited by brain plasticity mechanisms.
The solid tumor tissue component destroys the
remaining functional infiltrated brain parenchyma.
This malignant transformation leads ultimately to
death (Soffietti et al., 2010).
Image-Based Monitoring of the Natural

Course of LGG
Evolution Over Time of the Radiological
Growth Pattern of LGG
Only rare papers have analyzed qualitatively the evolution of the radiological growth pattern over time of
LGG. Chen et al. reported that tumors originating from

the grey matter will remain bulky while continously
growing, whereas tumors originating at the junction
of gray and white matters will grow predominantly
along the adjacent white matter fiber tracts (Chen et al.,
2010). Mandonnet et al., demonstrated the dynamic
invasion of major white matter fiber tracts surrounding
the insular lobe for paralimbic LGG (Mandonnet et al.,
2006).
Quantitative Radiological Growth Rates

of LGG
All quantitative longitudinal studies of radiological
growth patterns demonstrated a spontaneous and continuous growth of LGG (Fig. 18.1a). Mandonnet et al.
first quantified the spontaneous radiological growth
of LGG on successive MRIs in 27 histologically
proven LGG followed before oncological treatment
(Fig. 18.1b) (Mandonnet et al., 2003). These initial
results were confirmed on a larger series of 143 histologically proven LGG that ranged individual velocities of diametric expansion from 1 to 36 mm/year
(Fig. 18.1c) (Pallud et al., 2006). Several independent
groups have since reported the same results (Hlaihel
et al., 2010; Pallud et al., 2010c; Peyre et al., 2010;
Ricard et al., 2007).
The quantitative analyses demonstrated a linear
evolution of the spontaneous Mean Tumor Diameter
(MTD) over time and quantified the Velocity of
Diametric Expansion (VDE) at about 4 mm/year average rate (Fig. 18.1b) (Mandonnet et al., 2003). Further
studies, performed by distinct groups, have confirmed
this range value (Hlaihel et al., 2010; Pallud et al.,
2010c; 2006; Peyre et al., 2010; Ricard et al., 2007).
Of note, these quantitative studies never reported indolent untreated LGG or LGG alternating indolent and
growing periods. In addition, Pallud et al., recently
demonstrated that incidental LGG present a similar
radiological growth during the silent period preceding their clinical revelation with VDE in the known
ranges for symptomatic LGG (Pallud et al., 2010a).
As different independent groups have reported similar observations, it has to be accepted that LGG are
progressive tumors that present a spontaneous and
Fig. 18.1 (a) Example of the natural course of a diffuse lowgrade glioma through the evolution of its Velocity of Diametric
Expansion (VDE) over time. A right fronto-temporo-insular
glioma was discovered incidentally in a 29-year-old righthanded woman. The tumor was initially followed and consecutive MRI showed a spontaneous growth with a VDE at
5.3 mm/year (e). About 3 years after radiological discovery,
partial seizures occurred. A subtotal surgical resection (general
anaesthesia) was performed as the first oncological treatment
and confirmed a WHO grade II oligodendroglioma on pathological analysis. After surgery, the residual tumor progressed
with a VDE at 2.3 mm/year. Three years after histological diagnosis, the patient refused further clinical and radiological follow-up. One year later, epilepsy reccured and the
patient presented in emergency with a left hemiparesis. Images
167
demonstrated a radiological progression toward a higher grade

of malignancy with a VDE increase at 28.4 mm/year and the
occurrence of a nodular-like area of contrast enhancement (e).
A second subtotal surgical resection (general anaesthesia) was
performed and an external conformational radiotherapy plus
concomitant and adjuvant chemotherapy with temozolomide
was started. b. The evolution of the Mean Tumor Diameter
(MTD) is comparable among patients with a median VDE
at about 4 mm/year (Adapted from Mandonnet et al., 2003).
c. Distribution of patients by individual radiological growth
rates. The VDE vary considerably among patients but their distribution results in two groups. About 85% presented a VDE less
than 8 mm/year (blue bars) and about 15% presented a VDE at
8 mm/year or more (purple bars) (adapted from Pallud et al.,
2006)
168
continuous radiological growth all along their natural

course (i.e., during the initial silent period and during
the symptomatic period before any transformation into
a higher grade of malignancy).
Radiological Growth Rates Reflect

Different Natural Courses for LGG
Effects of Intrinsic and Extrinsic Factors on

Spontaneous Tumor Growth Rates of LGG
Pallud et al. first studied the prognostic value of spontaneous MRI growth rates on overall survival in a
retrospective series of 143 histologically proven LGG
with measurements of the evolution of the MTD
over time (Pallud et al., 2006). The low growth rates
subgroup (VDE lower than 8 mm/year) exhibited
a significantly longer overall survival than the high
growth rates subgroup (VDE at 8 mm/year or more)
(Fig. 18.2c). In multivariate analysis, tumor growth
rates and initial tumor volume were two independent
prognostic factors significantly associated with overall
survival.
The prognostic significance of spontaneous MRI
growth rates on predicting progression into a higher
grade of malignancy was addressed by Hlailel et al.
(2010). They showed that an elevated VDE higher than
3 mm/year was correlated with a greater risk of progression into a higher grade of malignancy with an
average VDE at 7.87 mm/year in transformers group
versus an average VDE at 2.14 mm/year in non transformers group. Brasil Caseiras et al. (2009) proved
that tumor growth within 6 months was better than
baseline volumes, rCBV, or ADC in predicting time to
malignant transformation in untreated LGG using the
evolution of the tumor volume over time.
Although the different LGG histological subtypes

(oligodendroglioma, astrocytoma, mixed glioma) are
associated with different prognoses (Pignatti et al.,
2002; Soffietti et al., 2010), they do not exhibit significant differences regarding the radiological tumor
growth rates (Mandonnet et al., 2009; Pallud et al.,
2006; Ricard et al., 2007). As an example, Mandonnet
et al. quantified the median VDE at 3.35 mm/year
in the astrocytoma subgroup, at 4.2 mm/year in the
oligodendroglioma subgroup, and at 3.53 mm/year
in the mixed glioma subgroup, without any statistical differences between these subgroups (Fig. 18.2a)
(Mandonnet et al., 2009).
Ricard et al. (2007) investigated the link between
LGG radiological growth rates and genetic alterations.
They showed that LGG with 1p-19q codeletion grew
significantly slower than LGG without 1p-19q codeletion and that LGG with overexpression of p53 grew
significantly faster than LGG without overexpression
of p53 (Fig. 18.2a). There is little documentation of
the effects of extrinsic factors on glioma growth properties. It has been suggested that pregnancy can have
a negative impact on the natural course of glioma, by
the mean of hormonal changes (Pallud et al., 2009b).
Only one study investigated quantitatively the impact
of pregnancy on the radiological growth rates of LGG
in 12 pregnancies in 11 adult women harbouring a
LGG prior to pregnancy (Pallud et al., 2010b). Pallud
et al. (2010b) showed that LGG accelerated their radiological growth rates during pregnancy (Fig. 18.2a).
Indeed, VDE significantly increased during pregnancy
above levels detected either before pregnancy or after
delivery in 75% of cases, changes in tumour growth
were associated with an increase in seizure frequency
in 40% of cases and radiological and clinical changes
during pregancy motivated further oncological treatment after delivery in 25% of cases.
Prognostic Significance of the

Spontaneous Growth Rates
Radiological Growth Rates as a New

Prognostic Factor for LGG?
The prognosis of LGG varies between series and
reflects the heterogeneity of the observed natural histories. The 5-year overall (OS) and progression-free
survival (PFS) rates in randomized studies range from
58 to 72% and 37 to 55%, respectively (Soffietti et al.,
2010).
Regarding clinical findings, age over 40 years
and presence of pre-operative neurological deficits
are adverse prognostic factors (Pignatti et al., 2002;
Soffietti et al., 2010). Regarding radiological findings,
Fig. 18.2 (a) Factors influencing the spontaneous Velocity of

Diametric Expansion (VDE) of WHO grade II gliomas. The different histological subtypes do not exhibit significant differences
of VDE (black bars, adapted from Mandonnet et al., 2009). The
1p-19q codeletion molecular status is associated with a significantly slower VDE (dark grey bars, adapted from Ricard et al.,
2007). The p53 overexpression molecular status is associated
with a significantly faster VDE (light grey bars, adapted from
Ricard et al., 2007). Pregnancy increases significantly VDE as
compared to pre-pregnancy and post-delivery rates (white bars,
adapted from Pallud et al., 2010b). (b) Effects of oncological
treatments on the Velocity of Diametric Expansion (VDE). VDE
169
are expressed before and after surgery (black bars, adapted from
Mandonnet et al.), before and during Temozolomide (TMZ)
chemotherapy (dark grey bars, adapted from Ricard et al., 2007),
before and during PCV chemotherapy (light grey bars, adapted
from Peyre et al., 2010). (c) Kaplan-Meier estimates of overall
survival by individual Velocity of Diametric Expansion (VDE).
The subgroup with a VDE less than 8 mm/year (blue line)
presents a significant longer overall survival (median survival
more than 15 years) than the subgroup with a VDE at 8 mm/year
or more (purple line) where the overall survival is closer to
that of more malignant gliomas (median survival at 5.16 years)
(adapted from Pallud et al., 2006)
170
larger tumors and tumors crossing the midline correlate

with a short OS and PFS (Pignatti et al., 2002; Soffietti
et al., 2010) and growth rates are inversely correlated
with OS (Pallud et al., 2006). There are conflicting
reports as to whether contrast enhancement is associated with a worsened prognosis (Soffietti et al., 2010).
However, Pallud et al. recently showed that the presence of contrast enhancement alone, regardless of its
pattern, had no prognostic value and that only the presence of a nodular-like pattern or of progressive contrast
enhancement over time were associated with a worsened prognosis (Pallud et al., 2009a). A low Cerebral
Blood Volume and a low uptake of 11C-methionine
correlate with longer PFS and OS (Soffietti et al.,
2010). Regarding pathological and molecular findings,
the astrocytoma histological subtype is associated with
a worsened prognosis (Pignatti et al., 2002; Soffietti
et al., 2010). 1p loss (with or without 19q loss) is
a favorable prognostic factor (Soffietti et al., 2010).
IDH1 codon 132 mutation has been recently suggested
as an independent favorable prognostic factor (Soffietti
et al., 2010). When risk factors are cumulated, the survival of LGG is dramatically decreased with survival
times closer to those of higher grade gliomas.
The limitation of the histological diagnosis
explains, in part, this prognostic heterogeneity.
Analyses on microscopic and macroscopic, static and
dynamic, scales are required to refine the prognostic
evaluation. A multi-scale approach is a way to circumvent the diagnostic limitations of the histological
examination. Thus, analysis of VDE, a dynamic
macroscopic parameter easily available in clinical
practice by the mean of repeated measurements of
MTD over time, may be a useful tool to overcome
biological diversity of LGG (Mandonnet et al., 2008;
Pallud et al., 2006). As a practical consequence,
VDE could be integrated along with the other static
parameters (histological and molecular analyses)
in a multi-scale approach to understand better the
individual natural course of LGG.
as promising tool in the follow-up of LGG and in the

monitoring of the different oncological treatments.
Mandonnet et al. have recently shown that VDE
remain unchanged after surgical resection in a retrospective study of 54 resected LGG (Fig. 18.2b)
(Mandonnet et al., 2009). This suggests that the survival benefit of surgery is mediated by a cytoreductive
effect. Of note, 2 patients exhibited a decrease of their
tumor growth rates greater than 3 mm/year whereas
in 2 other patients, surgery failed to stop an ongoing malignant transformation, resulting in an increase
of 3 mm/year on the post-operative tumor growth
rates. Thus, the precise quantitative assessment of the
VDE obtained pre and post-operatively would help
analyzing the effects of surgical resection on an individual basis and guiding the postoperative oncological
treatment.
Similarly, the tumor response to chemotherapy
can be quantified by the VDE evolution over time.
Ricard et al. first quantified the tumor response after
temozolomide and almost all LGG exhibited an initial decrease of the VDE after temozolomide onset
(Fig. 18.2b) (Ricard et al., 2007). They evidenced different patterns of response, depending on the 1p-19q
codeletion status as tumor relapse occured more frequently and earlier in tumors without 1p-19q codeletion. Peyre at al. first quantified the tumor response
after PCV chemotherapy and all LGG presented an initial VDE decrease after PCV onset (Fig. 18.2b) (Peyre
et al., 2010). They demonstrated that the median VDE
decrease after PCV onset was very close to the tumor
growth decrease after temozolomide onset found by
Ricard et al. (2007). In addition, they showed an ongoing VDE decrease after PCV continuation that was
prolonged more than 2 years in 60% of the LGG under
study. These results raise the issue of a chemotherapy
monitoring based on the quantitative changes of the
tumor VDE. Of note, at that time, no study has quantified the radiological response, using VDE evolution
over time, after radiation therapy.
Quantitative Assessment of Treatment

Efficacy
Conclusions
Along with clinical response, the quantitative assessment of the individual VDE changes by the mean of the
MTD evolution over time on consecutive MRI, appears
Quantitative analyses on MRI have demonstrated that

LGG are progressive tumors that present a spontaneous and continuous radiological growth all along
their natural course during the initial silent period and
during the symptomatic period before any transformation into a higher grade of malignancy. LGG are a
heterogeneous group of tumors with distinct prognoses
and a multi-scale approach may help overpassing the
diagnostic limitations of the sole histological examination. Thus, analysis of VDE, a dynamic macroscopic
parameter easily available in clinical practice by the
mean of repeated measurements of MTD over time
(for technical details, see (Mandonnet et al., 2008)),
may be a useful tool to understand the biological diversity of LGG. At the light of its strong prognostic
significance, VDE could be integrated along with the
other static parameters (histological and molecular
analyses) in a multi-scale approach to understand better the individual natural course of LGG. In addition, the precise quantitative assessment of the VDE
obtained before and after treatment would help guiding and analyzing the effects of different oncological
treatment modalities on an individual basis.
Acknowledgments Johan Pallud and Emmanuel Mandonnet
want to thanks all the members of the French Glioma Network
(REG, Rseau dEtude des Gliomes) and particularly Laurent
Capelle, Luc Taillandier and Hugues Duffau. Johan Pallud wants
to thank Franois-Xavier Roux, Edouard Dezamis and Bertrand
Devaux of the department of Neurosurgery, Catherine DaumasDuport and Pascale Varlet of the department of Neuropathology,
Catherine Oppenheim and Jean-Franois Meder of the department of Neuroradiology of the Sainte-Anne Hospital Center in
Paris, France.
References
Brasil Caseiras G, Ciccarelli O, Altmann DR, Benton CE, Tozer
DJ, Tofts PS, Yousry TA, Rees J, Waldman AD, Jager HR
(2009) Low-grade gliomas: six-month tumor growth predicts
patient outcome better than admission tumor volume, relative
cerebral blood volume, and apparent diffusion coefficient.
Bristol RE (2009) Low-grade glial tumors: are they all the same?
Semin Pediatr Neurol 16:2326
Chen X, Dai J, Jiang T (2010) Supratentorial WHO grade
II glioma invasion: a morphologic study using sequential
conventional MRI. Br J Neurosurg 24:196201
Daumas-Duport C, Meder JF, Monsaingeon V, Missir O, Aubin
ML, Szikla G (1983) Cerebral gliomas: malignancy, limits and spatial configuration. Comparative data from serial
stereotaxic biopsies and computed tomography (a preliminary study based on 50 cases). J Neuroradiol 10:5180
Daumas-Duport C, Monsaigneon V, Blond S, Munari C,
Musolino A, Chodkiewicz JP, Missir O (1987) Serial stereotactic biopsies and CT scan in gliomas: correlative study
171
in 100 astrocytomas, oligo-astrocytomas and oligodendrocytomas. J Neurooncol 4:317328

Daumas-Duport C, Szikla G (1981) Definition of limits and 3D
configuration of cerebral gliomas. Histological data, therapeutic incidences (authors transl). Neurochirurgie 27:273
284
Daumas-Duport C, Varlet P, Tucker ML, Beuvon F, Cervera
P, Chodkiewicz JP (1997) Oligodendrogliomas. Part I: patterns of growth, histological diagnosis, clinical and imaging
correlations: a study of 153 cases. J Neurooncol 34:3759
Duffau H (2005) Lessons from brain mapping in surgery for lowgrade glioma: insights into associations between tumor and
Giese A, Bjerkvig R, Berens ME, Westphal M (2003) Cost of
migration: invasion of malignant gliomas and implications
for treatment. J Clin Oncol 21:16241636
Hlaihel C, Guilloton L, Guyotat J, Streichenberger N, Honnorat
J, Cotton F (2010) Predictive value of multimodality MRI
using conventional, perfusion, and spectroscopy MR in
anaplastic transformation of low-grade oligodendrogliomas.
Lang FF, Gilbert MR (2006) Diffusely infiltrative low-grade
gliomas in adults. J Clin Oncol 24:12361245
78:179185
Mandonnet E, Delattre JY, Tanguy ML, Swanson KR, Carpentier
AF, Duffau H, Cornu P, Van Effenterre R, Alvord EC
Jr, Capelle L (2003) Continuous growth of mean tumor
diameter in a subset of grade II gliomas. Ann Neurol 53:
524528
Mandonnet E, Pallud J, Clatz O, Taillandier L, Konukoglu E,
Duffau H, Capelle L (2008) Computational modeling of the
WHO grade II glioma dynamics: principles and applications
to management paradigm. Neurosurg Rev 31:263269
Mandonnet E, Pallud J, Fontaine D, Taillandier L, Bauchet
L, Peruzzi P, Guyotat J, Bernier V, Baron MH, Duffau
H, Capelle L (2009) Inter- and intrapatients comparison of
WHO grade II glioma kinetics before and after surgical
resection. Neurosurg Rev 33:9196
Pallud J, Capelle L, Taillandier L, Fontaine D, Mandonnet E,
Guillevin R, Bauchet L, Peruzzi P, Laigle-Donadey F, Kujas
M, Guyotat J, Baron MH, Mokhtari K, Duffau H (2009a)
Prognostic significance of imaging contrast enhancement for
WHO grade II gliomas. Neuro-Oncology 11:176182
Pallud J, Devaux B, Daumas-Duport C, Oppenheim C, Roux FX
(2005) Glioma dissemination along the corticospinal tract. J
Pallud J, Duffau H, Razak RA, Barbarino-Monnier P, Capelle
L, Fontaine D, Frenay M, Guillet-May F, Mandonnet E,
Taillandier L (2009b) Influence of pregnancy in the behavior
of diffuse gliomas: clinical cases of a French glioma study
group. J Neurol 256:20142020
Pallud J, Fontaine D, Duffau H, Mandonnet E, Sanai N,
Taillandier L, Peruzzi P, Guillevin R, Bauchet L, Bernier V,
Baron MH, Guyotat J, Capelle L (2010a) Natural history of
incidental world health organization grade II gliomas. Ann
Neurol 68:727733
172
Pallud J, Mandonnet E, Deroulers C, Fontaine D, Badoual M,
Capelle L, Guillet-May F, Page P, Peruzzi P, Jouanneau E,
Frenay M, Cartalat-Carel S, Duffau H, Taillandier L Club de
Neuro-Oncologie de la Socit Franaise de Neurochirurgie
(SFNC), Association des Neuro-Oncologues dExpression
Franaise (ANOCEF) (2010b) Pregnancy increases the
growth rates of World Health Organization grade II gliomas.
Ann Neurol 67:398404
60:380383
Pallud J, Varlet P, Devaux B, Geha S, Badoual M, Deroulers
C, Page P, Dezamis E, Daumas-Duport C, Roux FX (2010c)
Diffuse low-grade oligodendrogliomas extend beyond MRIdefined abnormalities. Neurology 74:17241731
Peyre M, Cartalat-Carel S, Meyronet D, Ricard D, Jouvet A,
Pallud J, Mokhtari K, Guyotat J, Jouanneau E, Sunyach
MP, Frappaz D, Honnorat J, Ducray F (2010) Prolonged
response without prolonged chemotherapy: a lesson from
PCV chemotherapy in low-grade gliomas. Neuro-Oncology
10:10781082
Pignatti F, van den Bent M, Curran D, Debruyne C, Sylvester
R, Therasse P, Afra D, Cornu P, Bolla M, Vecht C, Karim

AB European organization for research and treatment of cancer brain tumor cooperative group, European organization
for research and treatment of cancer radiotherapy cooperative group (2002) Prognostic factors for survival in adult
patients with cerebral low-grade glioma. J Clin Oncol 20:
20762084
Mandonnet E, Kujas M, Mokhtari K, Taillibert S, LaigleDonadey F, Carpentier AF, Omuro A, Capelle L, Duffau
H, Cornu P, Guillevin R, Sanson M, Hoang-Xuan K,
Delattre JY (2007) Dynamic history of low-grade gliomas
before and after temozolomide treatment. Ann Neurol 61:
484490
Soffietti R, Baumert BG, Bello L, Von Deimling A, Duffau H,
Frnay M, Grisold W, Grant R, Graus F, Hoang-Xuan K,
Klein M, Melin B, Rees J, Siegal T, Smits A, Stupp R,
Wick W (2010) Guidelines on management of low-grade
gliomas: report of an EFNS-EANO Task Force. Eur J
Neurol 17:11241133
Wessels PH, Weber WE, Raven G, Ramaekers FC, Hopman
AH, Twijnstra A (2003) Supratentorial grade II astrocytoma: biological features and clinical course. Lancet Neurol
2:395403
Chapter 19
Impact of Extent of Resection on Outcomes in Patients

with High-Grade Gliomas
Debraj Mukherjee and Alfredo Quinones-Hinojosa
Abstract Throughout the course of human history,

surgical therapy has remained a viable option for the
treatment of certain, well-defined lesions. However,
particularly given the technological advancements over
the last half-century, the fields of neurosurgery and
neurooncology have been able to embrace multimodal
forms of therapy, including maximal safe surgical
resection, radiotherapy, and both local and systemic
chemotherapy, in an effort to improve survival and
decrease the odds of developing new post-operative
deficits in patients with highly malignant gliomas.
Some studies do appear to indicate that extent of resection plays a favorable role in survival of patients with
high grade gliomas, although the exact relationship
between extent of resection and survival or poor outcomes is not completely defined. Future prospective,
randomized trials may need to be developed to assess
the effect of multimodal forms of imaging, motor mapping, and combination forms of therapy upon shortand long-term patient outcomes.
Keywords High-grade astrocytomas Mapping
Central nervous system Temozolomide
Chemotherapy Radiotherapy
Introduction
High-grade astrocytomas are the most common malignant primary central nervous system tumors in adults,
D. Mukherjee ()
Maxine Dunitz Neurosurgical Institute, Cedars-Sinai Medical
Center, Los Angeles, California 90048
e-mail: debraj.mukherjee@cshs.org
encompassing both anaplastastic astrocytomas (also

known as WHO Grade III tumors) and glioblastoma
multiforme (also known as WHO Grade IV tumors).
Despite advances in medical and surgical therapy, the
median survival remains less than 2 years. While mean
survival for patients with high-grade gliomas have
remained relatively short, individual patient survival
is heterogeneous. Thus, there has been an emphasis
on studying intraoperative factors that hold prognostic significance in predicting survival in patients with
malignant astrocytomas. For many solid organ malignant tumors, gross total resection with clear margins
has been associated with significantly extended survival in well-designed studies. However, the effects
of such extensive resection on prolonging survival in
patients with malignant astrocytomas is less clear as no
prospectively, randomized trials have been conducted
in this areas of study. Extensive resection of highgrade gliomas is made increasingly difficult because
these tumors are frequently invasive, widely infiltrative, and often involve eloquent areas. Improvements
in surgical adjuncts including functional MR imaging,
cortical mapping, and intraoperative MR imaging have
made it possible to achieve extensive resection of many
gliomas. It has remained incompletely understood
whether more extensive resection of malignant astrocytomas is associated with significantly prolonged
survival in well-designed studies.
Historical Considerations
Dating back to the early attempts by Neolithic
man, surgery of the brain, meninges, or skull have
slowly evolved over 12,000 years. Through most of
this history, interventions were technically crude by

173
174
modern measures and aimed almost exclusively toward

the removal of extracranial, overtly visualized lesions
of the skull. Within the modern era, this work progressed toward surgical intervention of the coverings
of the brain by Zanobi Pecchioli, who resected a fungus of the dura mater in a patient presenting with a
large extracranial mass in 1835. Yet, there continued to
be no widely reported surgical interventions to remove
primary intraparenchymal lesions of the brain through
the late nineteenth century, as only a limited amount
was known of localized areas of cortical functioning
until the work of Paul Broca and Hughlings Jackson in
the 1860s. With modest improvements in neurosurgical
techniques and in the collective understanding of cortical functioning, the first widely recognized resection
of a primary brain tumor was performed by Rickman
J. Godlee in 1884 on a 25-year old who presented
with a nearly 3 year history of intractable seizures.
Although removal of this seemingly low-grade glioma
was deemed an operative success, the patient died of
complications from meningitis 3-weeks after surgical
intervention. Yet, at autopsy, the patient was noted
to have no residual tumor, likely alluding, in part, to
the implicit importance Godlee may have placed in
achieving a gross total resection.
Attempts at intracranial surgery continued to some
degree through the early part of the twentieth century and reached a milestone with Walter Dandys
1928 case series of patients undergoing hemispherectomy for removal of invasive high-grade gliomas.
This series included one patient surviving three and a
half years following surgical resection. Nevertheless,
most patients died within 3 months of surgery, often
of infectious complications, including meningitis and
pneumonia.
Over the following decades, advancements in surgical science, including the appropriate use of antibiotics and sterile technique, decreased immediate postoperative morbidity and mortality. Furthermore, the
advent of cerebral angiograms in the 1930s, computerized topography (CT) scans in the 1970s, and magnetic
resonance imaging (MRI) technology in the 1980s further facilitated in operative planning for these complex
operations. With the last two decades, advancements
in microsurgery have further increased the possibility
of achieving operative gross total resection in patients
with both low and high grade gliomas.
However, the lack of well-designed, prospective
studies have made the role of extensive resection of
D. Mukherjee and A. Quinones-Hinojosa
high-grade astrocytomas incompletely defined, despite

the great advances made in developing adjuvant radiotherapy and chemotherapy protocols that significant
improve survival among these patients (Stupp et al.,
2005). For instance, many past studies used varying cut-offs for defining extent of resection thresholds
for defining gross total, near total, or subtotal resection, including varying methods of assessing degree
of resection (particularly among the non-volumetric
studies). Additionally, a majority of these studies did
not use standard WHO tumor grading criteria, thus
making it difficult to fully generalize the findings of
these studies to the general neuro-oncological literature. Furthermore, although most studies has comparable levels of adjuvant radiotherapy or chemotherapy
between patients of varying degrees of extent of resection, all studies did not adjust likely predictors of overall survival, including patient comorbidities, location
of tumor (for instance near eloquent cortical areas), or
known predictors of survival, including 1p19q status
among oligodendrogliomas and MGMT status among
patient with glioblastoma multiforme. Furthermore,
the use of particular pre-operative imaging modalities possibly helpful for operative planning, including
DTI or fMRI for patients with lesions involving motor
tracts or eloquent cortical areas, respectively, were not
accounted for in most studies.
The Impact of Extent of Resection:

Lessons Learned from Low-Grade
Gliomas
Over approximately 20 years, at least eleven studies have independently and retrospectively reviewed
the impact of extent of resection on outcomes in
patients with low-grade astrocytomas, as reviewed
by Sanai and Berger (2008a). Of these, most have
used non-volumetric means, including for instance
intraoperative surgeon report or gross radiographic
evidence of residual disease, to categorize extent of
resection. A majority also presented evidence either
in univariate or multivariate analysis of supporting
greater 5-year progression-free survival or overall
survival with greater extent of resection. Only the
Johannesen et al. (2003) report from Norway studied
5-year overall survival and found no statistically significant difference in survival between biopsy, subtotal
19 Impact of Extent of Resection on Outcomes in Patients with High-Grade Gliomas
resection, or gross total resection, although there was

a trend toward greater survival in those with greater
extent of resection. Additionally, a few studies have
retrospectively reviewed this association using more
precise radiographic volumetrics, as opposed to selfreport or two-dimensional impressions of extent or
resection from post-operative radiographic imaging, to
assess degree of resection. All retrospective studies
demonstrated significantly improved 5-year in those
with greater extent of resection, although the threshold for definitely greatest degree of resection ranged
from 75 to 100% in these studies. Taken in total, all
major retrospective studies analyzing the association
between extent of resection and 5-year survival have
demonstrated at least a trend toward better outcomes
in those with greater resection. These findings have
been supported in two reviews and one meta-analyses
on this area of study (Pouratian et al., 2007; Sanai and
Berger, 2008a; Smith et al., 2008). However, findings
in those with low-grade gliomas may not necessarily
translate directly to patient with more infiltrative highgrade gliomas with often inherent much worse overall
prognoses.
The Impact of Extent of Resection on

Survival in Patients with High-Grade
Gliomas: A Review of the Literature
In the modern era, at least 29 studies have assessed
the impact of extent of resection on progression-free
or overall survival in patients with high-grade astrocytomas, as reviewed by Sawaya (1999) as well as
Sanai and Berger (2008a). Of these, a majority used
non-volumetric means of assessing extent or resection
while just four studies used the more precise measure
of volumetrics (Keles et al., 1999, 2006; Lacroix et al.,
2001; Pope et al., 2005). Among the non-volumetric
studies, an equal number demonstrated a statistically
significant improvement in overall survival in multivariate analysis as those that showed no significant
association. Among all studies of high-grade glioma
patients, only three have assessed progression-free
survival as an outcome (Jeremic et al., 1994; SandbergWollheim et al., 1991; Ushio et al., 2005). Within these
studies, two reports demonstrated a statistically significant association between greater extent of resection
and longer progression-free survival in multivariate
175
analysis (Jeremic et al., 1994; Ushio et al., 2005), while

one reports showed no significant association in multivariate analysis (Sandberg-Wollheim et al., 1991).
Four volumetric studies have thus far been widely published on the topic (Keles et al., 1999, 2006; Lacroix
et al., 2001; Pope et al., 2005). Among these studies, two reports demonstrated significantly improved
overall survival in multivariate analysis in those with
greater extent of resection (Keles et al., 1999; Lacroix
et al., 2001), while the remaining two studies showed
no statistically significant relationship in multivariate analysis (Keles et al., 2006; Pope et al., 2005).
Only two volumetric studies have thus far assessed the
impact of degree of resection on progression-free survival, with one study showing significant association
(Keles et al., 1999) and the other demonstrating no
significance (Keles et al., 2006), both in multivariate
analyses.
The Importance of Multimodal Therapy

in Improving Survival
Although some progress has been made in understanding the impact of extensive resection on survival,
the treatment of high-grade astrocytomas remains
multimodal in approach, combining surgical resection with both adjuvant chemotherapy and radiation. The efficacy of combination local chemotherapy following extensive tumor resection was found to
improve survival in patient with recurrent malignant
gliomas from 23 to 31 weeks after revision resection
(Brem et al., 1995). Amongst all high-grade glioma
patients, a major advance in the treatment of highgrade gliomas occurred when Valtonen et al. (1997)
found that median time from surgery to death in primary, high-grade glioma patients was 58.1 weeks for
patients who received carmustine-loaded biodegradable polymers delivering localized chemotherapy in
surgically resection tumor beds versus 39.9 weeks for
those who only underwent surgical resection. More
specifically patients with glioblastoma multiforme, the
most deadly form of high-grade gliomas, survived
53.3 weeks compared to 39.9 weeks in those given
carmustine-loaded wafer with resection, as compared
to those undergoing resection alone.
Truly multimodal therapy, incorporating extensive surgical resection along with both adjuvant
176
temozolamide chemotherapy and adjuvant radiotherapy demonstrated a further significant improvement

in glioblastoma survival through the Stupp et al.
(2005) report, which has now become standard
of care for all patients with glioblastoma multiforme. Additionally, recent retrospective work has
demonstrated that combination surgical resection and
both local and systemic chemotherapy may significantly improvement outcomes among these patients
as well (McGirt et al., 2009b). In a recent analysis,
glioblastoma multiforme patients receiving extensive
surgical resection along with both local carmustineimpregnated wafer chemotherapy and systemic temozolamide chemotherapy had 9 months of greater
survival relative to counterparts only receiving surgical resection and localized wafer chemotherapy;
the two groups did not have significantly different rates of complication (Fig. 19.1). Thus, surgical resection coupled with carmustine-impregnanted,
biodegradable wafers, systemic temozolamide, and
radiation therapy are currently considered the most
effective treatment modalities available for treating
high-grade gliomas. Additionally approaches to modifying multimodal therapy, such as using temozolomide
in a dose-dense relationship in patient with MGMT
silencing, or using temozolomide in combination with
MGMT inhibitors (such as O6 -benzylguanine), are
currently under further investigation (Quinn et al.,
2009).
The Impact of New Post-operative

Deficits on Survival
Fig. 19.1 Kaplan-Meier plot of survival after primary resection of glioblastoma multiforme and radiotherapy in patients
70 years old. Patient receiving concomitant temozolamide
according to the Stupp protocol in addition to Gliadel wafer
implantation demonstrated improved survival relative to those

receiving Gliadel wafters and radiotherapy alone (n = 78). Mean
survival was 21.3 versus 12.4 months, respectively (p = 0.005).
Source: McGirt (2009b)
While studies have suggested an association between

greater degree of resection among malignant astrocytomas and improved survival, the negative effects
that new post-operative deficits, potentially induced by
greater extensive resection, may have upon survival
is less clear. A single retrospective analysis of 306
patients undergoing primary resection for glioblastoma
multiforme was conducted at a single, high-volume
institution to assess the relationship between new
post-operative motor or language deficits on survival
(McGirt et al., 2009a). Even after adjusting for patient
age, preoperative KPS score, adjuvant therapy, and
extent of resection, those patients with newly acquired
motor and language deficits had reductions in median
survival of 3.9- and 3.3-months, respectively, relative
to their deficit-free counterparts (Fig. 19.2). Those
patients who only demonstrated transient deficits in
the immediately post-operative period had a smaller
decrease in median survival relatively to those who
developed permanent deficits. Whether this association
was one of causation or simple statistical association currently remains unknown given the retrospective
nature of the study (McGirt et al., 2009a).
Although a definitive mechanism to explain the
observed association is not yet known, past work
177
Fig. 19.2 Kaplan-Meier plot demonstrating survival after

resection of glioblastoma multiforme in patients without a new
post-operative neurological deficit (mean survival 12.8 months),
with a surgically acquired language deficit (mean survival
9.6 months), or with surgically acquired motor deficits (mean
survival 9.0 months). Those with new deficits experienced significantly worse survival (p<0.05 for both types of new deficits)
relative to their non-deficit counterparts. Source: McGirt
(2009a)
has alluded to poor quality of life possibly serving

as a conduit between new post-operative deficits and
worse survival. Prospective analyses of patients with
high-grade gliomas have demonstrated that those with
worse baseline quality of life scores, including measures of verbal communication, have worse survival
compared to those with normal quality of life scores
(Brown et al., 2006; Mauer et al., 2007). Additionally,
retrospective studies have shown significant associations between post-operative depression in glioma
patients and worse long-term survival (Gathinji et al.,
2009; Litofsky et al., 2004). Furthermore, in the study
by Mauer et al. (2007), baseline deficits in motor or
speech among glioma patients was associated a significant prognostic factor affecting survival. Additionally,
in theory, decreased performance status and functionality may contribute to the accelerated development
of terminal comorbidities such as pneumonia or cardiovascular deconditioning in post-operative patients.
This is consistent with the well-known fact that the
KPS score carries prognostic significance. Thus, while
the full importance of extensive surgical resection
in high-grade glioma patients remains incompletely
described, further studies should also likely study
the negative effects that extensive resection with new
deficits may play balancing possible long-term survival
benefits.
The Importance of Neuro-Imaging

and Cortical Mapping
Over the past 20 years, technological advances have
developed with the aim of optimizing the balance between maximizing extent of surgical resection while minimizing poor neurological outcomes.
Such advances have included preoperative magnetic
source imaging, positron emission tomographic imaging, functional MRI, intraoperative MRI, diffuse tensor
imaging, electrical stimulation mapping, and negative language mapping (Sanai et al., 2008b). The use
of multimodal forms of imaging has been used in
more recent years in an attempt to further optimize
this balance. Although reports have demonstrated the
validity of these multimodal approaches, few studies have clarified the impact true impact of such
approaches upon survival or minimizing the odds of
post-operative neurological sequela (Jack et al., 1994;
Quiones-Hinojosa et al., 2003; Simos et al., 1999).
New forms of mapping have also been used
with greater frequency in recent years to achieve
similar goals. For instance, the work of Ojemann
et al. (1989) demonstrated the usefulness of electrical stimulation mapping to identify cortical language
areas, as patients studied demonstrated significant
178
inter-personal variability in the localization of common language areas. Schiffbauer et al. demonstrated
that preoperative magnetic source (MS) imaging correlated well with intraoperative electrophysiological
stimulation mapping, thus making it possible to neurosurgeons to more appropriately plan neurosurgical
intervention for tumors near eloquent areas (2002).
Furthermore, Signorelli et al. (2007) demonstrated
through a retrospective study that the use of such techniques significantly reduced the proportion of patients
who developed new post-operative deficits. Similarly,
a recent prospective, randomized study involving diffuse tensor imaging in glioma patients found that
those patients who underwent pre-operative imaging
were significantly less likely to develop post-operative
motor deficits (Wu et al., 2007).
References
Brem H, Piantadosi S, Burger PC, Walker M, Selker R, Vick
NA, Black K, Sisti M, Brem S, Mohr G, Muller P, Morawetz
R, Schold SC (1995) Placebo-controlled trial of safety and
efficacy of intraoperative controlled delivery by biodegradable polymers of chemotherapy for recurrent gliomas.
The Polymer-brain Tumor Treatment Group. Lancet 345:
10081012
Brown PD, Ballman KV, Rummans TA, Maurer MJ, Sloan
JA, Boeve BF, Gupta L, Tang-Wai DF, Arusell RM, Clark
MM, Buckner JC (2006) Prospective study of quality of
life in adults with newly diagnosed high-grade gliomas.
Gathinji M, McGirt MJ, Attenello FJ, Chaichana KL, Than
K, Olivi A, Weingart JD, Brem H, Quinones-Hinojosa A
(2009) Association of preoperative depression and survival
after resection of malignant brain astrocytoma. Surg Neurol
71:299303
Jack CR, Thompson RM, Butts RK, Sharbrough FW, Kelly
PJ, Hanson DP, Riederer SJ, Ehman RL, Hangiandreou
NJ, Cascino GD (1994) Sensory motor cortex: correlation
of presurgical mapping with functional MR imaging and
invasive cortical mapping. Radiology 190:8592
Jeremic B, Grujicic D, Antunovic V, Djuric L, Stojanovic M,
Shibamoto Y (1994) Influence of extent of surgery and tumor
location on treatment outcome of patients with glioblastoma multiforme treated with combined modality approach.
Johannesen TB, Langmark F, Lote K (2003) Progress in longterm survival in adult patients with supratentorial low-grade
gliomas: a population-based study of 993 patients in whom
tumors were diagnosed between 1970 and 1993. J Neurosurg
99:854862
Keles GE, Anderson B, Berger MS (1999) The effect of extent
of resection on time to tumor progression and survival in
patients with glioblastoma multiforme of the cerebral hemisphere. Surg Neurol 52:371379

Keles GE, Chang EF, Lamborn KR, Tihan T, Chang CJ, Chang
SM, Berger MS (2006) Volumetric extent of resection and
residual contrast enhancement on initial surgery as predictors
of outcome in adult patients with hemispheric anaplastic
astrocytoma. J Neurosurg 105:3440
Lacroix M, Abi-Said D, Fourney DR, Gokaslan ZL, Shi W,
DeMonte F, Lang FF, McCutcheon IE, Hassenbusch SJ,
Holland E, Hess K, Michael C, Miller D, Sawaya R (2001)
A multivariate analysis of 416 patients with glioblastoma
multiforme: prognosis, extent of resection, and survival.
Litofsky NS, Farace E, Anderson F, Meyers CA, Huang
W, Laws ER, Glioma Outcomes Project Investigators
(2004) Depression in patients with high-grade glioma:
results of the glioma outcomes project. Neurosurgery 54:
358366
Mauer M, Stupp R, Taphoorn MJ, Coens C, Osoba D, Marosi C,
Wong R, de Witte O, Cairncross JG, Efficace F, Mirimanoff
RO, Forsyth P, van den Bent MJ, Weller M, Bottomley A
(2007) The prognostic value of health-related quality-of-life
data in predicting survival in glioblastoma cancer patients:
results from an international randomised phase III EORTC
Brain tumour and Radiation Oncology Groups, and NCIC
Clinical Trials Group study. Br J Cancer 97:302307
McGirt MJ, Mukherjee D, Chaichana KL, Than KD, Weingart
JD, Quinones-Hinojosa A (2009a) Association of surgically
acquired motor and language deficits on overall survival
after resection of glioblastoma multiforme. Neurosurgery 65:
463469
McGirt MJ, Than KD, Weingart JD, Chaichana KL, Attenello
FJ, Olivi A, Laterra J, Kleinberg LR, Grossman SA, Brem H,
Quiones-Hinojosa A (2009b) Gliadel (BCNU) wafer plus
concomitant temozolomide therapy after primary resection
of glioblastoma multiforme. J Neurosurg 110:583588
Ojemann G, Ojemann J, Lettich E, Berger M (1989) Cortical
language localization in left, dominant hemisphere. An electrical stimulation mapping investigation in 117 patients.
Pope WB, Sayre J, Perlina A, Villablanca JP, Mischel PS,
Cloughesy TF (2005) MR imaging correlates of survival in
patients with high-grade gliomas. AJNR Am J Neuroradiol
26:24662474
Pouratian N, Asthagiri A, Jagannathan J, Shaffrey ME, Schiff
D (2007) Surgery Insight: the role of surgery in the management of low-grade gliomas. Nat Clin Pract Neurol 11:
628639
Quinn JA, Jiang SX, Reardon DA, Desjardins A, Vredenburgh
JJ, Gururangan S, Sampson JH, McLendon RE, Herndon
JE, Friedman HS (2009) Phase 1 trial of temozolomide plus
irinotecan plus O6-benzylguanine in adults with recurrent
malignant glioma. Cancer 115:29642970
Quiones-Hinojosa A, Ojemann SG, Sanai N, Dillon WP, Berger
MS (2003) Preoperative correlation of intraoperative cortical mapping with magnetic resonance imaging landmarks
to predict localization of the Broca area. J Neurosurg 99:
311318
Sanai N, Berger MS (2008a) Glioma extent of resection and its
impact on patient outcome. Neurosurgery 4:753764
Sanai N, Mirzadeh Z, Berger MS (2008b) Functional outcome
after language mapping for glioma resection. N Engl J Med
358:1827

Sandberg-Wollheim M, Malmstrm P, Strmblad LG, Anderson
H, Borgstrm S, Brun A, Cronqvist S, Hougaard K, Salford
LG (1991) A randomized study of chemotherapy with
procarbazine, vincristine, and lomustine with and without
radiation therapy for astrocytoma grades 3 and/or 4. Cancer
68:2229
Sawaya R (1999) Extent of resection in malignant gliomas:
a critical summary. J Neurooncol 3:303305
Schiffbauer H, Berger MS, Ferrari P, Freudenstein D, Rowley
HA, Roberts TP (2002) Preoperative magnetic source imaging for brain tumor surgery. A quantitative comparison with
intraoperative sensory and motor mapping. J Neurosurg
97:13331342
Signorelli F, Ruggeri F, Iofrida G, Isnard J, Chirchiglia D,
Lavano A, Volpentesta G, Signorelli CD, Guyotat J (2007)
Indications and limits of intraoperative cortico-subcortical
mapping in brain tumor surgery: an analysis of 101 consecutive cases. J Neurosurg Sci 51:113127
Simos PG, Papanicolaou AC, Breier JI, Wheless JW,
Constantinou JE, Gormley WB, Maggio WW (1999)
Localization of language-specific cortex by using magnetic source imaging and electrical stimulation mapping.
Smith JS, Chang EF, Lamborn KR, Chang SM, Prados M,
Cha S, Tihan T, VandenBerg S, McDermott MW, Berger
MS (2008) Role of extent of resection in the long-term
179
outcome of low-grade hemispheric gliomas. J Clin Oncol

26:13381345
Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B,
Taphoorn MJ, Belanger K, Brandes AA, Marosi C, Bogdahn
U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeier
A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff
RO, European Organisation for Research and Treatment of
Cancer Brain Tumor and Radiotherapy Groups, and National
Cancer Institute of Canada Clinical Trials Group (2005)
Radiotherapy plus concomitant and adjuvant temozolomide
for glioblastoma. N Engl J Med 352:987996
Ushio Y, Kochi M, Hamada J, Kai Y, Nakamura H (2005) Effect
of surgical removal on survival and quality of life in patients
with supratentorial glioblastoma. Neurol Med Chir (Tokyo)
45:454461
Valtonen S, Timonen U, Toivanen P, Kalimo H, Kivipelto. L,
Heiskanen O, Unsgaard G, Kuurne T (1997) Interstitial
chemotherapy with carmustine-loaded polymers for
high-grade gliomas: a randomized double-blind study.
Wu JS, Zhou LF, Tang WJ, Mao Y, Hu J, Song YY, Hong
XN, Du GH (2007) Clinical evaluation and follow-up outcome of diffusion tensor imaging-based functional neuronavigation: a prospective, controlled study in patients
with gliomas involving pyramidal tracts. Neurosurgery
61:935948
Chapter 20
Glioma Surgery: Intraoperative Low Field Magnetic

Resonance Imaging
Christian Senft
Abstract For decades, surgical resection has been

marking the first step in the treatment of gliomas.
The benefit of radical yet safe tumor resection persists
even after the introduction of multimodal treatment
concepts. While the development of imaging techniques has enhanced the visualization of brain tumors,
the intraoperative discrimination between tumor tissue and healthy parenchyma remains problematic. The
incorporation of computed tomography and magnetic
resonance imaging scans into neurosurgical navigation systems has helped to improve the intraoperative
delineation of tumor boundaries. However, navigation
systems do not account for intraoperative alterations
of the anatomy. To that end, intraoperative imaging
technologies such as ultrasound or magnetic resonance
imaging have been introduced into modern neurosurgery. Low field units were the first intraoperative
magnetic resonance imaging devices that entered neurosurgical operating rooms. They were specifically
developed to facilitate brain tumor surgery. Although
an increasing number of high field units are being
installed today, low field units still represent the most
frequently used ones today. This chapter deals with the
development and setup of intraoperative low field magnetic resonance imaging as well as the implications and
consequences of its use in modern glioma surgery.
Keywords Gliomas iMRI Surgery Karnofsky
score Neuronavigation Brain-shift
Introduction
Rationale of Glioma Surgery
Ever since the first brain tumor surgery was performed
in the late nineteenth century by Rickham John Godlee
(Bennett and Godlee, 1884), radical surgical resection
has been the first step in the treatment of gliomas.
Aside from obtaining tissue samples for histopathology, resection of brain tumors may relieve mass effect
and reduce symptoms caused by the compression of
neural structures. There has long been a scientific
debate regarding the value of cytoreductive surgery
in gliomas after the introduction of radiation therapy
and chemotherapy (Curran et al., 1992; Tortosa et al.,
2003), but there has been increasing evidence recently
that the extent of tumor resection is an independent
prognostic factor in both high- and low-grade gliomas
(Sanai and Berger, 2008; McGirt et al., 2009).
Yet, due to the infiltrative nature of these tumors,
glioma patients cannot be cured by surgery, whatever its extensiveness may be. The clinical patient
status as determined by the Karnofsky performance
scale has often been shown to be a strong prognostic
factor in glioma patients (Laws et al., 2003; Tortosa
et al., 2003). Therefore, tumor resections must not be
undertaken at the cost of neurological deterioration.
Neuronavigation and Neurophysiological

Monitoring
C. Senft ()
Department of Neurosurgery, Johann Wolfgang
Goethe-University, 60528 Frankfurt, Germany
e-mail: c.senft@med.uni-frankfurt.de
To avoid surgically induced neurological deficits and

clinical deterioration, eloquent structures, especially
the motor cortex and the corticospinal tract, need to be

181
182
identified and spared during tumor resection. Eloquent

structures may be identified either anatomically or with
the help of intraoperative neurophysiological monitoring techniques (Berger et al., 1990).
The introduction of neuronavigation systems in the
1990s has dramatically influenced brain tumor surgery.
Being a technology similar to GPS-systems, these systems allow neurosurgeons to transfer preoperative neuroimaging data, e.g. anatomical as well as functional
MRI, into a computer workstation in the operating
room. After matching points on the patients head with
corresponding points on the images, a navigational tool
is depicted on a screen in its corresponding anatomic
location along with the preoperative data. As a result,
neuronavigation assists in cranial neurosurgery to confirm the exact localization of anatomic structures or
lesions in an online fashion. Neuronavigation has
become a standard part of modern neurosurgical operating rooms, helping to plan skin incisions, tailor
craniotomies, and make approaches to intraparenchymal lesions less invasive (Wirtz et al., 2000). The
usefulness of neuronavigation, however, is limited during the dissection of brain tissue due to anatomical
alterations caused by leakage of cerebrospinal fluid,
brain edema, or removal of tissue itself. These alterations, called brain-shift, cause significant inaccuracy
of the neuronavigation because it relies on preoperative
images (Roberts et al., 1998).
Rationale of Intraoperative Imaging

Low-grade gliomas in particular, but to lesser extent
also high-grade gliomas, are often difficult to distinguish from healthy brain parenchyma, which needs to
be spared during surgery. This renders tumor resections guided by (even microscopic) vision incomplete in many cases, even if the surgeon believes to
have achieved a complete resection (Albert et al.,
1994). Further, the phenomenon of brain-shift mentioned above necessitates an intraoperative update of
the imaging data used for neuronavigation. Thus, intraoperative imaging was developed in order to overcome the limitations of neuronavigation based on
pre-operative images and to visualize tumor remnants
that might leave otherwise unrecognized by the neurosurgeon (Nimsky et al., 2003; Ram and Hadani,
2003).
C. Senft
Magnetic resonance imaging represents the best

means of imaging the brain, and obviously it would
also represent the best means of intraoperative imaging. Consequently, MRI devices were developed for
intraoperative neurosurgical use. The first intraoperative MRI (iMRI) unit was installed in 1994 at the
Brigham and Womens Hospital in Boston (Black
et al., 1997). For technical reasons, it was designed as
a low magnetic field strength device with 0.5 T (Signa
SP, General Electric Medical Systems, Milwaukee,
WI).
Later, several iMRI systems have been set up at
different neurosurgical centers. These systems vary
by field strength of the magnet and require different
adjustments of OR infrastructure, e.g. ferromagnetic
shielding or the need for special instruments. Also, the
costs for the installation of an iMRI vary significantly
between the currently available systems, ranging from
$ 1.5 million to well above $ 5 million, especially for
high field (1.5 or 3.0 T) devices. Today, low field iMRI
units are the ones predominantly used.
Low Field iMRI: Setup

Intraoperative magnetic resonance imaging can be performed with a scanner installed in a room adjacent to
the OR (two-room concept) (Tronnier et al., 1997), or
with a scanner installed in the OR itself (one-room
concept) (Black et al., 1997; Hadani et al., 2001). The
two-room concept may be advantageous from a reimbursement perspective, since the scanner can be used
for outpatients while it is not needed for intraoperative
imaging.
Example of iMRI Setup

With more than 50 installations worldwide, one of
the most frequently employed iMRI devices today is
R
the PoleStar
iMRI navigation suite (Medtronic Inc.,
Louisville, CO). It was specifically designed for cranial
neurosurgical procedures and follows the one-room
concept of intraoperative imaging and offers an integrated neuronavigation system (Hadani et al., 2001;
R
Schulder et al., 2006). The PoleStar
N20 consists
of two permanent magnets employing a field strength
of only 0.15 T. In contrast to the Signa SP, its design
20 Glioma Surgery: Intraoperative Low Field Magnetic Resonance Imaging
183
Fig. 20.1 Schematic drawing

R
of the PoleStar
intraoperative OR Setup (with
permission from Medtronic
Navigation, Inc.)
and lower field strength obviate the need for special

surgical instruments. Also, continuous neurophysiological monitoring in this setting is feasible (Szelenyi
et al., 2008). The two magnets are placed perpendicular to each other with an aperture of 27 cm, in
which the patients head is positioned during scanning.
The scanner depicts an ellipsoid field of view measuring approximately 20 cm by 15 cm. Figure 20.1
gives an example of the OR-setup. Solely the neurosurgeon and OR personnel, without the need of a
radiologist or MR technician, operate the scanner. The
R
use of the PoleStar
system in glioma surgery has
been described in detail (Hadani et al., 2001; Schulder
et al., 2006; Ntoukas et al., 2008; Senft et al., 2008).
Example of Surgical Workflow

In a typical case, a glioma patient is brought into the
OR in general anesthesia. Using a non-ferromagnetic
head-holder system, the patient is positioned supine,
prone, or lateral as needed, and the head is fixated
to the head holder using sterile titanium skull pins.
Additionally, a patient reference frame is attached to
the head holder to allow for neuronavigation. The scanner is then wheeled to its position beneath the patients
head, and it is elevated for the pre-surgical navigation
scan (Fig. 20.2a). After image acquisition, the magnet is again lowered to its parking position beneath the
patients head. The acquired images are automatically
transferred to the neuronavigation system, and the skin
incision and craniotomy are planned. Figure 20.2b,
shows the setup after disinfection and sterile draping.
Surgery commences with neuronavigation-guidance

and is performed using regular microsurgical instruments (Fig. 20.2c). When the surgeon believes to have
resected the tumor, the scanner is brought upwards
to its scan position (Fig. 20.2d). Intraoperative imaging is performed, and the neuronavigation data are
updated. Pre-surgical and intraoperative images can
be compared directly to identify any residual tumor
and, if present, to exactly localize it and to determine
its extension. Figure 20.3 gives an example of image
quality and continued resection after intraoperative
scanning.
In comparison with conventional microsurgical
resection of a brain tumor, patient positioning is a little more cumbersome in the iMRI-setting, and image
acqisition is also time consuming, demanding approximately 1020 min per scan session depending on the
imaging sequences. These time demands sum up to
roughly 1 h on average.
Safety Concerns in iMRI

Aside from any potential benefits of applying iMRI
technology in neurosurgery, there are also safety concerns associated with its use.
First, intraoperative image acquisition prolongs surgical time with the skull opened, raising the question
of an increased risk of infection. This is even more the
case in the two-room concept of iMRI, as the patient
needs to be transferred out of the OR into an adjacent room for image acquisition. There are, however,
no reports of increased rates of infections (Tronnier
184
C. Senft
R
Fig. 20.2 Example of glioma surgery with the PoleStar
iMRI
Fig. 20.3 Example images of a patient with a left temporal

glioblastoma. (a) preoperative T1-weighted scan obtained at 1.5
T; (b/c) postoperative T1-weighted scans obtained at 3 T with
(b) and without (c) contrast agent confirming complete resection
of contrast enhancing tumor tissue; (df) intraoperative images
obtained at 0.15 T before (d) and during (e, f), tumor resection.
Note that in residual contrast enhancement along the resection
cavity is visible in (e), leading to further tumor resection. Finally,
a complete tumor resection is documented (f)
et al., 1997; Nimsky et al., 2002), even if patients are

taken out of the OR to perform scanning (Ramina et al.,
2010). In an own unpublished series of more than 200
patients, we only observed one case of a wound infection, occurring about half a year after surgery followed
by combined radio-/chemotherapy. On the contrary,
iMRI has been described as helpful in the surgical
treatment of brain infections (Bernays et al., 2002;
Senft et al., 2009).
Second, the magnetic field exerted by the scanner might attract surgical instruments, causing patient
harm. In some iMRI setups, such as the Signa SP,
special non-ferromagnetic instruments are therefore
essential in the PoleStar-setup, regular instruments
can be used. So far, intraoperative magnetic resonance
imaging has been shown safe with respect to effects of
magnetic force (Seifert et al., 1999; Hall et al., 2000).
Finally, there are arguments that the intraoperative
detection of tumor tissue might lead to more aggressive and extensive resections, leading to higher rates of
neurological complications. Yet, especially when used
in combination with neurophysiological monitoring,
complication rates in iMRI-guided brain tumor surgery
are comparable to those in conventional microsurgery
(Hall et al., 2000; Bohinski et al., 2001; Senft et al.,
2008). In another report, continued resection after
intraoperative scanning was not associated with the
occurrence of new neurological deficits (Senft et al.,
2010a).
Results of iMRI-Guided Glioma Surgery

Influence on the Course of Surgery
Low field iMRI has been used to facilitate the surgical resection of both low- and high-grade gliomas.
Unanimously, reports on the use of (low field) iMRI
devices have been positive (Schulder and Carmel,
2003), regardless of the system used (Tronnier et al.,
1997; Seifert et al., 1999; Muragaki et al., 2006;
Schulder, 2009). The ability to detect residual tumor
by intraoperative imaging has led to continued resections in many cases, ranging from 10% (Nimsky et al.,
2002) to up to 71% (Schneider et al., 2005), but
mostly in approximately 3040% of the cases (Nimsky
et al., 2003; Senft et al., 2010a). Not surprisingly,
the use of iMRI was reported to be superior than the
use of sole neuronavigation in terms of the extent
185
of tumor resection (Bergsneider et al., 2005). Using

iMRI-guidance has improved resection rates in both
low- and high-grade gliomas alike, although tumor
remnants resulting in continued are encountered more
often in low-grade than in high-grade gliomas (Senft
et al., 2008; Senft et al., 2010b).
Influence on Survival
Intraoperative MRI and, consequently, enhanced rates
of complete tumor resection have also benefitted
patients. Several groups have reported improved survival for glioma patients undergoing complete vs.
subtotal resections in iMRI-guided surgery. In a large
series of 156 low-grade glioma patients undergoing
iMRI-guided surgery using the Signa SP, a trend
towards prolonged survival after total resection was
observed (Claus et al., 2005). When analyzing patients
with glioblastoma, two groups independently reported
significantly improved survival in patients undergoing
complete resections. In a series of 31 patients undergoing image-guided surgery using the Signa SP, median
survival was 1.5 years for patients with complete and
only 0.6 years for patients with incomplete resection
in a univariate analysis (Schneider et al., 2005). This
large difference in may be in part explained by differences in patient age and preoperative Karnofsky score,
which both are known prognostic factors. In an own
R
series of 58 patients using the PoleStar
, respective
median survival was 1.1 and 0.7 years, respectively
(Senft et al., 2010a).
In summary, theses reports add to the increasing
evidence that extensive surgical resection translates
into better outcome (Stummer et al., 2008). From
an evidence-based medicine perspective, however, the
value of iMRI, regardless of the field strength of the
system, has not been proven, as all studies mentioned
above have been retrospective series without a control
group. There is only one matched-group analysis in a
series of 32 patients, which failed to show a statistically
significant benefit of iMRI (Hirschberg et al., 2005).
Discussion
Neurosurgery has long been using and relying on
the most advanced technologies, more than any other
subspecialty in medicine. The transportation of technology into the neurosurgical treatment of brain tumors
186
has benefitted patients starting with diagnostic imaging

of the nervous system using CT or MRI. It continued
with the introduction of microsurgical concepts, and
did not stop with the invention of image guided surgery
or neuronavigation. Surgeries have become safer, and
patients today recuperate from brain surgery faster than
just a few decades ago.
The development of iMRI represents a major technological challenge and also advance in brain tumor
surgery. Intraoperative MRI systems bear a tremendous
advantage over frameless neuronavigational systems
in that they allow for the correction of intraoperative brain shift, which can lead to incomplete tumor
resections. In the past 15 years, iMRI systems have
come from unique application to every-day routine.
Simplicity of operation, minimal interference with the
neurosurgical workflow, as well as improvements in
surgical results have been critical issues in their introduction.
Neurosurgeons and hospital administrators are
under pressure to justify the enormous costs associated with the installation of an iMRI unit, regardless
of manufacturer or specific design. Currently, arguments in favor of iMRI-guidance primarily rely on
so-called expert-opinion, and the possibility of performing iMRI-guided brain tumor surgery is limited to
a few dedicated centers.
There are many advocates of intraoperative imaging in brain tumor surgery, claiming that residual
tumor, which would otherwise remain unnoticed, can
be detected and resected due to intraoperative image
acquisition. Critics of intraoperative imaging argue
that imaging is only performed because the technology is at hand, so that the intraoperative detection of
residual tumor may only be a self-fulfilling prophecy.
Certainly, extremely high rates of continued resection must be doubted, but a large multi-center study
reported unintentional residual tumor in more than
60% of the cases after conventional microsurgical
resection of glioblastoma (Stummer et al., 2006), indicating that there is a need for the ability to visualize
tumor tissue intraoperatively by other means than mere
microscopic vision.
Still, a scientifically sound proof of the usefulness
of intraoperative image guidance in terms of improved
rates of tumor resection or improved clinical outcome is missing. Glioma patients succumb to their
disease despite neurosurgeons utilizing state-of-the-art
techniques. Today, iMRI is a helpful tool in glioma
C. Senft
surgery, but randomized trials governing its use have

not been conducted. The true value of iMRI as part of
a skilled neurosurgeons armamentarium is yet to be
determined.
References
Albert FK, Forsting M, Sartor K, Adams HP, Kunze S (1994)
Early postoperative magnetic resonance imaging after resection of malignant glioma: objective evaluation of residual tumor and its influence on regrowth and prognosis.
Neurosurgery 34:4560, discussion 6061
Bennett A, Godlee R (1884) Excision of a tumour from the brain.
Lancet 124:10901091
Berger MS, Ojemann GA, Lettich E (1990) Neurophysiological
monitoring during astrocytoma surgery. Neurosurg Clin N
Am 1:6580
Bergsneider M, Sehati N, Villablanca P, McArthur DL, Becker
DP, Liau LM (2005) Mahaley clinical research award: extent
of glioma resection using low-field (0.2 T) versus highfield (1.5 T) intraoperative MRI and image-guided frameless
neuronavigation. Clin Neurosurg 52:389399
Bernays RL, Kollias SS, Yonekawa Y (2002) Dynamic changes
during evacuation of a left temporal abscess in open MRI:
technical case report. Neuroradiology 44:438442
Black PM, Moriarty T, Alexander E 3rd, Stieg P, Woodard EJ,
Gleason PL, Martin CH, Kikinis R, Schwartz RB, Jolesz
FA (1997) Development and implementation of intraoperative magnetic resonance imaging and its neurosurgical
applications. Neurosurgery 41:831842, discussion 842845
Bohinski RJ, Kokkino AK, Warnick RE, Gaskill-Shipley MF,
Kormos DW, Lukin RR, Tew JM Jr (2001) Glioma resection in a shared-resource magnetic resonance operating room
after optimal image-guided frameless stereotactic resection.
Claus EB, Horlacher A, Hsu L, Schwartz RB, Dello-Iacono
D, Talos F, Jolesz FA, Black PM (2005) Survival rates in
patients with low-grade glioma after intraoperative magnetic
resonance image guidance. Cancer 103:12271233
Curran WJ Jr, Scott CB, Horton J, Nelson JS, Weinstein AS,
Nelson DF, Fischbach AJ, Chang CH, Rotman M, Asbell SO
et al. (1992) Does extent of surgery influence outcome for
astrocytoma with atypical or anaplastic foci (AAF)? A report
from three Radiation Therapy Oncology Group (RTOG)
trials. J Neurooncol 12:219227
Hadani M, Spiegelman R, Feldman Z, Berkenstadt H, Ram Z
(2001) Novel, compact, intraoperative magnetic resonance
imaging-guided system for conventional neurosurgical operating rooms. Neurosurgery 48:799807, discussion 807809
Hall WA, Liu H, Martin AJ, Pozza CH, Maxwell RE, Truwit
CL (2000) Safety, efficacy, and functionality of high-field
strength interventional magnetic resonance imaging for neurosurgery. Neurosurgery 46:632641, discussion 641642
Hirschberg H, Samset E, Hol PK, Tillung T, Lote K (2005)
Impact of intraoperative MRI on the surgical results for
high-grade gliomas. Minim Invasive Neurosurg 48:7784
Laws ER, Parney IF, Huang W, Anderson F, Morris AM, Asher
A, Lillehei KO, Bernstein M, Brem H, Sloan A, Berger MS,

Chang S (2003) Survival following surgery and prognostic
factors for recently diagnosed malignant glioma: data from
the glioma outcomes project. J Neurosurg 99:467473
McGirt MJ, Chaichana KL, Gathinji M, Attenello FJ, Than K,
Olivi A, Weingart JD, Brem H, Quinones-Hinojosa A (2009)
Independent association of extent of resection with survival
in patients with malignant brain astrocytoma. J Neurosurg
110:156162
Muragaki Y, Iseki H, Maruyama T, Kawamata T, Yamane F,
Nakamura R, Kubo O, Takakura K, Hori T (2006) Usefulness
of intraoperative magnetic resonance imaging for glioma
surgery. Acta Neurochir Suppl 98:6775
Nimsky C, Ganslandt O, Buchfelder M, Fahlbusch R (2003)
Glioma surgery evaluated by intraoperative low-field magnetic resonance imaging. Acta Neurochir Suppl 85:5563
Nimsky C, Ganslandt O, Tomandl B, Buchfelder M, Fahlbusch
R (2002) Low-field magnetic resonance imaging for intraoperative use in neurosurgery: a 5-year experience. Eur Radiol
12:26902703
Ntoukas V, Krishnan R, Seifert V (2008) The new generation
polestar n20 for conventional neurosurgical operating rooms:
a preliminary report. Neurosurgery 62:8289, discussion
8990
Ram Z, Hadani M (2003) Intraoperative imagingMRI. Acta
Neurochir Suppl 88:14
Ramina R, Coelho Neto M, Giacomelli A, Barros E Jr, Vosgerau
R, Nascimento A, Coelho G (2010) Optimizing costs of
intraoperative magnetic resonance imaging. A series of 29
glioma cases. Acta Neurochir (Wien) 152:2733
Roberts DW, Hartov A, Kennedy FE, Miga MI, Paulsen
KD (1998) Intraoperative brain shift and deformation: a
quantitative analysis of cortical displacement in 28 cases.
Sanai N, Berger MS (2008) Glioma extent of resection and
its impact on patient outcome. Neurosurgery 62:753764,
discussion 764766
Schneider JP, Trantakis C, Rubach M, Schulz T, Dietrich J,
Winkler D, Renner C, Schober R, Geiger K, Brosteanu
O, Zimmer C, Kahn T (2005) Intraoperative MRI to
guide the resection of primary supratentorial glioblastoma multiformea quantitative radiological analysis.
Schulder M (2009) Intracranial surgery with a compact, lowfield-strength magnetic resonance imager. Top Magn Reson
Imaging 19:179189
Schulder M, Carmel PW (2003) Intraoperative magnetic resonance imaging: impact on brain tumor surgery. Cancer
Control 10:115124
Schulder M, Salas S, Brimacombe M, Fine P, Catrambone
J, Maniker AH, Carmel PW (2006) Cranial surgery with
187
an expanded compact intraoperative magnetic resonance

imager. Technical note. J Neurosurg 104:611617
Seifert V, Zimmermann M, Trantakis C, Vitzthum HE, Kuhnel
K, Raabe A, Bootz F, Schneider JP, Schmidt F, Dietrich
J (1999) Open MRI-guided neurosurgery. Acta Neurochir
(Wien) 141:455464
Senft C, Franz K, Ulrich CT, Bink A, Szelenyi A, Gasser
T, Seifert V (2010a) Low field intraoperative MRI-guided
surgery of gliomas: a single center experience. Clin Neurol
Senft C, Seifert V, Hermann E, Franz K, Gasser T (2008)
Usefulness of intraoperative ultralow-field magnetic resonance imaging in glioma surgery. Neurosurgery 63:257266,
discussion 266267
Senft C, Seifert V, Hermann E, Gasser T (2009) Surgical
treatment of cerebral abscess with the use of a mobile
ultralow-field MRI. Neurosurg Rev 32:7785
Senft C, Ulrich CT, Seifert V, Gasser T (2010b) Intraoperative
magnetic resonance imaging in the surgical treatment of
cerebral metastases. J Surg Oncol 101:436441
Stummer W, Pichlmeier U, Meinel T, Wiestler OD, Zanella
F, Reulen HJ (2006) Fluorescence-guided surgery with
5-aminolevulinic acid for resection of malignant glioma:
a randomised controlled multicentre phase III trial. Lancet
Oncol 7:392401
Stummer W, Reulen HJ, Meinel T, Pichlmeier U, Schumacher
W, Tonn JC, Rohde V, Oppel F, Turowski B, Woiciechowsky
C, Franz K, Pietsch T (2008) Extent of resection and survival
in glioblastoma multiforme:identification of and adjustment
for bias. Neurosurgery 62:564576, discussion 564576
Szelenyi A, Gasser T, Seifert V (2008) Intraoperative neurophysiological monitoring in an open low-field magnetic resonance
imaging system:clinical experience and technical considerations. Neurosurgery 63:268275, discussion 275276
Tortosa A, Vinolas N, Villa S, Verger E, Gil JM, Brell M, Caral
L, Pujol T, Acebes JJ, Ribalta T, Ferrer I, Graus F (2003)
Prognostic implication of clinical, radiologic, and pathologic features in patients with anaplastic gliomas. Cancer 97:
10631071
Tronnier VM, Wirtz CR, Knauth M, Lenz G, Pastyr O, Bonsanto
MM, Albert FK, Kuth R, Staubert A, Schlegel W, Sartor K,
Kunze S (1997) Intraoperative diagnostic and interventional
magnetic resonance imaging in neurosurgery. Neurosurgery
40:891900, discussion 900902
Wirtz CR, Albert FK, Schwaderer M, Heuer C, Staubert A,
Tronnier VM, Knauth M, Kunze S (2000) The benefit of
neuronavigation for neurosurgery analyzed by its impact on
glioblastoma surgery. Neurol Res 22:354360
Chapter 21
Low-Grade Gliomas: Intraoperative Electrical Stimulations

Hugues Duffau
Abstract The goal of low-grade gliomas (LGG)

surgery is to optimize the extent of resection, while
minimizing the risk of permanent neurological deficit.
Because LGG often invade eloquent areas and
because of major interindividual anatomofunctional
variability, the cortical functional organization, the
subcortical connectivity and the brain plastic potential should be studied at the individual scale. Indeed,
cortical and subcortical structures essential for brain
functions must be detected and preserved. Presurgical
functional neuroimaging and tractography can show
the relationships between eloquent regions and the
tumor, but they have several methodological limitations. Consequently, intraoperative electro-stimulation
mapping (IES) is more and more used by neurosurgeons, to tailor the resection according to individual
functional boundaries. IES can be performed under
general anesthesia for motor mapping, or on awake
patient for somatosensory (and even motor), visuospatial, language and cognitive mapping. This is an
easy, accurate, reliable and safe technique of detection of cortical areas, white matter pathways and deep
gray nuclei crucial for the function. Thus, IES enables:
(i) to study the individual cortical functional organization before any resection; (ii) to understand the
pathophysiology of cerebral eloquent areas; (iii) to
map the subcortical structures throughout the resection, allowing a study of the anatomo-functional
H. Duffau ()
Department of Neurosurgery and INSERM U1051, Institute for
Neurosciences of Montpellier, Montpellier University Medical
Center, Hpital Gui de Chauliac, CHU Montpellier 34295
Montpellier Cedex 5, France
e-mail: h-duffau@chu-montpellier.fr
connectivity; (iv) to analyze the mechanisms of

on-line short-term plasticity, using repeated IES;
(v) to tailor the resection according to individual
cortico-subcortical functional boundaries. Moreover,
IES can be combined with repeated pre-and postoperative functional neuroimaging, before and after
surgery, both to validate these non-invasive techniques
and to study the reorganization of the functional maps
over time at the individual scale. Such plastic potential can open the door to multiple surgeries spaced
by several months or years, with the goal to optimize the benefit/risk ratio of surgery for LGG, i.e. to
increase the extent of resection while preserving and
even improving the quality of life. IES also represents
a unique tool to improve our knowledge of brain processing and to revisit the classical model of cognitive
neurosciences, by switching from a localisationnist to
a hodotopic view of cerebral organization.
Keywords LGG Neuroimaging Tractography
IES EOR Mapping
Introduction
Hemispheric WHO grade II gliomas, that is, diffuse
low-grade gliomas (LGG) are usually revealed by
seizures in young adults with no or only mild neurological deficit. However, these precancerous tumors
will inescapably continue to grow (Pallud et al., 2006),
to migrate along white matter pathways (Mandonnet
et al., 2006) and to evolve into anaplasia. As a consequence, in the past decade, this better knowledge of
the natural history of LGG resulted in a switch from a
classical wait and see attitude to an early therapeutic
strategy.

189
190
In this context, the dilemma of surgery for LGG is

to optimize the extent of resection (EOR) while preserving the quality of life. Indeed, although matter of
debate for a long time, maximal resection of glioma,
when possible, represents currently the first therapeutic option (Duffau, 2009a; Smith et al., 2008). In the
recent series measuring objectively the extent of resection (EOR) on repeated postoperative MRI, all of them
supported EOR as a statistically significant predictor
of overall survival (OS). When no signal abnormality
was visible on control MRI, especially on FLAIRweighted imaging (i.e. the so-called complete resection), patients had a significantly longer OS compared
with patients having any residual abnormality (Claus
et al., 2005; Duffau et al., 2005a; McGirt et al., 2008;
Smith et al., 2008; Rezvan et al., 2009). Interestingly,
even in cases of incomplete tumor removal, patients
with a greater percentage of resection had a significantly longer OS. In addition to the percentage of
resection, the postoperative tumor volume is also a predictor of survival, with a significantly longer OS when
the residue is less than 10 ml (subtotal resection)
compared with more than 10 ml (partial resection)
(Duffau et al., 2005a).
On the other hand, due to the frequent location of
LGG near or within the so-called eloquent areas
(Duffau and Capelle, 2004), and due to their infiltrative
feature (poorly delineated), it was for a long time considered that the chances to perform an extensive glioma
removal were low, whereas the risk to induce postoperative sequelae was high. Indeed, many surgical series
have reported a rate of permanent and severe deficit
between 13 and 27.5% following LGG removal (for a
review, see Duffau et al., 2005a).
Therefore, to optimize the benefit-to-risk ratio of
surgery, an increase number of authors proposed the
use of functional mapping methods. Indeed, considerable interindividual anatomofunctional variability was
demonstrated in healthy volunteers (Vigneau et al.,
2006). Moreover, this variability is increased in LGG,
due to brain plasticity, explaining why many patients
have no or only slight symptoms except epilepsy
(Duffau, 2005). It is thus mandatory to study for each
patient (1) the cortical functional organization, (2)
the effective connectivity and (3) the cerebral plastic
potential, in order to tailor the resection according to
both oncological as well as cortico-subcortical functional boundaries.
The goal of this article is to review how, in addition to functional neuroimaging, the technique of
H. Duffau
intraoperative cortical and subcortical electrical stimulations mapping has contributed to improve the results
in LGG surgery, both concerning the impact on the
natural course of the glioma as well as regarding the
preservation of the quality of life. The fundamental
applications of IES will also be detailed, particularly
in the field of cognitive neurosciences.
Presurgical Neurocognitive Evaluation

In spite of a neurological examination often normal, recent extensive neuropsychological assessments
showed that most of patients with LGG had cognitive disturbances, especially concerning working memory and executive functions (Teixidor et al., 2007).
Therefore, a systematic presurgical neurocognitive
assessment is highly recommended (Mauer et al.,
2008), with four goals: (i) to search the possible neurocognitive deficit not detected by a standard neurological examination (ii) to adapt the surgical methodology
(for instance to select an awake procedure) to the
results of this assessment (iii) to obtain a presurgical
baseline enabling a comparison with the postsurgical evaluation (iv) and to plan an adapted functional
rehabilitation.
Interestingly, although LGG regularly involves eloquent areas, neurological deficits are unfrequent or
mild, at least in the first period of the clinical history.
This is due to brain reshaping, which allows functional compensation in slow-growing lesions. Indeed,
cerebral remapping is possible, thanks to recruitment
of perilesional or remote areas within the ipsilesional hemisphere and/or recruitment of contralesional
homologous areas. The recent integration of these
concepts into the therapeutic strategy has resulted in
dramatic changes in the surgical management of LGG
patients, with an increase of surgical indications within
eloquent regions classically considered as inoperable
(Duffau, 2005).
Presurgical Functional Neuroimaging

Advances in functional neuroimaging, especially functional MRI (fMRI), now make possible to perform
a non-invasive cortical mapping of the whole brain.
fMRI provides data about the location of the eloquent
21 Low-Grade Gliomas: Intraoperative Electrical Stimulations
191
regions (namely, the areas involved in sensorimotor,

language, visual and even higher-cognitive function)
in relation to LGG, and helps to determine the hemispheric language lateralization. Therefore, this method
is currently a standard before resection of LGG, since
it participates in the selection of the surgical indications, partly depending on the location of the tumor and
its relationships with eloquent areas detected by fMRI
(allowing an estimation of the tumor resectability).
Neurofunctional imaging is also useful for the surgical planning, i.e. the selection of the surgical approach
and the delineation of the limits of resection. Finally,
fMRI may contribute to the selection of the surgical technique, notably to the decision to perform an
awake craniotomy if the LGG is closed to language or
cognitive regions.
Nonetheless, fMRI is not yet enough reliable at the
individual scale, mainly because the results depend
on the bio-mathematical models used for the reconstruction. A recent review based on correlations with
intraoperative electrophysiology showed that the sensitivity of fMRI was currently only around 71% for
movement, and from 59 to 100% for language (specificity from 0 to 97%) (Giussiani et al., 2010). Such
discrepancies can be due to a neurovascular decoupling in LGG, by inadequate tasks (not adapted to the
location of the glioma or to the neurological status),
or to methodological problems such as the selection
of the threshold. Thus, there is a risk of false negative and then to operate a patient without intraoperative
mapping although the glioma is actually located in crucial areas (even if not detected by presurgical fMRI).
Moreover, an erroneous interpretation of brain plasticity (pseudo-reorganization) can be made. Finally,
these methods are not able to differentiate the structures essential for the function, which should be surgically preserved, from those which can be functionally
compensated and so potentially resected without permanent deficit. In summary, fMRI can lead not to select
a patient for surgery while the tumor was operable or
to stop prematurely the resection.
Recent advances in diffusion tensor imaging (DTI)
also enable the detection of the main white matter
pathways and their relationships to the LGG. Yet, this
new techniques needs to be validated before it can
be used routinely for surgical planning, particularly
because results of DTI, as fMRI, strongly depend on
the bio-mathematical models used for the tractography. Indeed, comparison between distinct fiber tracking software tools found different results. Moreover,
correlations between DTI and intrasurgical subcortical

stimulation showed that, despite a good correspondence, DTI is not yet optimal to map language tracts.
Negative tractography does not rule out the persistence
of a fiber tract, especially when invaded by a LGG
(Leclercq et al., in press). Furthermore, DTI studies
only the anatomy of the subcortical pathways, but not
their function.
To overcome these pitfalls, one can currently consider to perform longitudinal studies based on pre-,
intra- and post-surgical mapping rather than to
content itself with static information based on a
unique preoperative functional neuroimaging analysis. Consequently, the additional use of invasive electrophysiological investigations remains highly recommended for LGG surgery in eloquent structures.
Principles of Intraoperative Functional

Mapping
Intrasurgical, the incorporation of multimodal imaging into frameless stereotactic surgery was extensively
used and referred to as functional neuronavigation.
Nonetheless, a randomized trial failed to demonstrate
the significant impact of navigation on postoperative
results (Willems et al., 2006). This is due to the limitations of the preoperative fMRI and DTI previously
detailed, as well as to the high risk of intraoperative
brain shift explained by surgical retraction, mass
effect, gravity, EOR and cerebrospinal fluid leakage.
Despite methodological improvements to reduce the
effects of the shift (e.g. combination with intraoperative ultrasound, producing real-time imaging; use of
mathematical models based on data from ultrasonography or digital images that track cortical displacement; intraoperative MRI), their reliability has to be
improved.
Therefore, invasive electrophysiological investigations remain the gold-standard when removing LGG
in eloquent brain structures. To this end, intrasurgical somatosensory and motor evoked potentials were
proposed to detect the central region. However, its reliability to identify the rolandic sulcus is not optimal,
since accurate localization of the central sulcus was
reported only between 91 and 94%. Estimation of the
overall sensitivity and negative predictive value of this
method is 79 and 96%, respectively. Moreover, phase
192
reversal recording identifies the central sulcus itself,

but offers no direct information on the particular distribution of motor function on the adjacent exposed cerebral structures. In addition, when recording compound
muscle action potentials, only the monitored muscles
can be controlled, that is, there is an unability to detect
motor deficits in nonmonitored muscles. Above all,
monitoring of muscle action potentials does not mean
monitoring of complex movements and action adapted
to the environment, which is the ultimate goal for the
patient. Finally, intraoperative evoked potentials cannot map language, memory or other higher functions
crucial for the quality of life (for a review, see Duffau
et al., 2003).
The use of extraoperative electrophysiological
recordings and stimulations via the implantation of
subdural grids was also suggested (Kral et al., 2007).
The patient is in optimal conditions, in his room, to
perform the tasks a point very important for children. Furthermore, advances in the interpretation of
the electrophysiological signal, such as electrocorticographic spectral analysis evaluating the event-related
synchronization in specific bands of frequency, now
enable a better understanding of the organization of
the functional cortex, and a study of the connectivity notably via the recording of cortico-cortical
evoked potential. Extraoperative electrophysiological
mapping nonetheless usually used grids with 1 cmspaced electrodes, thus with a limited accuracy. Two
surgical procedures are also necessary, one to implant
grids and a second to remove the lesion. In addition,
there is a risk of infectious complications due to the
presence of subdural grids during several days. Above
all, this technique can map only the cortex, but not the
subcortical structures, i.e. it provides no information
about the axonal connectivity.
The Method of Intraoperative Electrical

Stimulations
Due to the limitations of these different mapping
techniques, intraoperative electrical stimulations (IES)
were increasingly used in the past decade, under general or local anesthesia, during LGG surgery in eloquent areas (Duffau et al., 2008; Sanai et al., 2008).
Indeed, except for gliomas involving the motor structures, the mapping is performed in awake patients.
H. Duffau
Nonetheless, as mentioned, since movements and

action are more complex than single muscle contractions, it is now proposed to map the motor function
under local anesthesia with an active collaboration of
the patient (Duffau, in press). The principle is to use
IES as a focal and transitory virtual lesion, to obtain
an individual functional mapping both at cortical and
subcortical levels, and to test if a structure involved
by a LGG is still essential for the function what
is observed in 1520% of the cases. Stimulation of
a crucial area elicites a transitory disruption of the
task performed by the patient, and this region must
be preserved. An individual cortical mapping is thus
obtained before the resection, which will be tailored
according to functional limits (Fig. 21.1). In practice, a bipolar electrode tips spaced 5 mm apart and
delivering a biphasic current (pulse frequency 60 Hz,
single-pulse phase duration 1 msec) is applied to the
brain. The current intensity adapted to each patient is
determined by progressively increasing the amplitude
in 1 mA increments from a baseline of 2 mA until a
functional response is elicited, with 6 mA as the upper
limit under local anesthesia, and with 16 mA as the
upper limit under general anesthesia in order to avoid
seizures. The patient is never informed when the brain
is stimulated. At least 1 picture presentation without
stimulation should separate each stimulation, and no
site is stimulated twice in succession. Each cortical site
(size 5 5 mm, due to the spatial resolution of the
probe) of the entire cortex exposed by the bone flap is
tested 3 times. Indeed, it is currently admitted that 3 trials are sufficient to assure the eloquence of an area, by
generating disorders during its 3 stimulations, and with
normalization of the function as soon as the stimulation
is stopped. This limitation of trials and tasks is required
by the timing of the surgical procedure, because the
patient is awake and has a risk to be tired at the end of
the resection.
Interestingly, recent data demonstrated that the surgical procedure could be simplified, by avoiding the
use of intraoperative electrocorticography despite an
equivalent reliability of IES, and without increasing the
rate of seizures (Duffau et al., 2008). Nonetheless, in
cases of stimulation-induced seizures, irrigation cold
Ringers lactate is able to abrogate the seizure activity. In addition, some authors emphasized the value
of negative mapping (no identification of eloquent
sites) in the setting of a tailored cortical exposure
(Sanai et al., 2008). Although such recommendation
193
Fig. 21.1 (a) Axial T2-weighted MRI in a right-handed young

pharmacist with inaugural seizures. No oncological treatment
was administrated (b) Axial FLAIR-weighted MRI 4 years later.
The patient met several neurosurgeons, who said that the tumor
was stable, and that it was impossible to operate due to the invasion of Brocas area. In fact, volumetric measurement showed
that the tumor doubled, with now involvement of the corpus
callosum. An awake surgery was thus proposed in our institution (b) Intraoperative views before (left) and after (right)
glioma resection, delineated by letter tags. IES shows a reshaping of the eloquent maps, with a recruitment of perilesional
language sites located behind the glioma, within the precentral
gyrus (1, 2, 3). There was no crucial site within the left inferior frontal gyrus. Thus, an extensive resection of Brocas area
was possible, by preserving the subcortical connectivity in the
depth of the cavity (49 and 50, corresponding to language pathways). (c) Postoperative axial FLAIR- and coronal T2-weighted
MRI, demonstrating a near-complete resection of the glioma,
with removal of the corpus collosum, in a patient with neither
neurological nor neurocognitive deficit, leading a normal socioprofessional life. It is worth-noting that a FLAIR-hyperintensity
is visible in the deep of the cavity, i.e. within the deep gray nuclei
and white matter tracts, still functional
is acceptable for high-grade gliomas, since the surgical goal is mainly to remove the enhanced part of the
tumor, a negative mapping can be dangerous in surgery
of diffuse LGG, especially in non expert hands. Indeed,
because LGG is poorly demarcated, the boundaries
of the resection will be essentially guided according

to functional criteria. Since negative mapping can be
due to false negative for methodological reasons, it
does not guarantee the absence of eloquent sites. In
the experience reported by Sanai et al., all 4 of the
194
patients with permanent postoperative deficits had no

positive sites detected prior to their resections (Sanai
et al., 2008). Consequently, other authors continue to
prone a wider boneflap, to obtain a systematic positive
mapping before to perform the resection (Duffau et al.,
2003, 2008). Moreover, a positive mapping also allow
to increase the EOR (de Benedictis et al., in press),
since the resection can be pursued until eloquent areas
are encountered, i.e. with no margin around the functional structures (Gil Robles and Duffau, in press).
A recent study demonstrated that, in a consecutive and
homogeneous series of 115 LGG in the left dominant
hemisphere, the rate of permanent deficit remained
lower than 2% despite the absence of margin around
the language sites (Duffau et al., 2008).
IES allow the mapping of motor function (possibly under general anesthesia, by inducing unvoluntary
motor response, but also in awake patient by eliciting
a disturbance of the movement), somatosensory function (by eliciting dysesthesias described by the patient
himself intraoperatively), visual function (by eliciting
phosphenes and/or visual field deficit described by the
patient), auditivo-vestibular function (by inducing vertigo), language (spontaneous speech, counting, object
naming, comprehension, writing, reading, bilingualism, language switching from one language to another
. . .), and also the mapping of higher-order functions
such as calculation, memory, spatial cognition, crossmodal judgement or even emotional processing
by generating transient disturbances if the electrical
stimulation is applied at the level of a functional
epicenter (Duffau, in press). A speech therapist or
neuropsychologist or neurologist must be present in
the operative theater, to interpret accurately the kind of
disorders induced by IES, for instance speech arrest,
anarthria, speech apraxia, phonological disturbances,
semantic paraphasia, perseveration, anomia, dyscalculia, and so on. Thus, IES allow the on-line detection
of the cortical sites essential for the function before
the beginning of the resection, thus the selection of the
best surgical approach as well as the definition of the
cortical boundaries of the glioma removal.
In addition to the cortical mapping, another major
issue is the use of subcortical IES throughout the
resection (Duffau et al., 2003, 2008). Brain lesion
studies showed that lesion of the white matter pathways induced more severe deficit than cortical damage. Thus, the subcortical tracts sub-serving motor,
somatosensory, visual, auditivo-vestibular, language
and cognitive functions should be identified during
H. Duffau
the glioma removal. The goal is to preserve cerebral connectivity while optimizing the EOR, namely,
to pursue the resection until eloquent pathways are
detected. Interestingly, according to the same principle as that described at the cortical level, IES can also
identify eloquent subcortical structures. IES enable
the study of the anatomo-functional connectivity by
directly and regularly stimulating the white matter
tracts and deep gray nuclei throughout the resection,
and by eliciting functional response when in contact
with deep crucial areas (Fig. 21.1). Moreover, IES
allow a better understanding of the brain connectivity,
showing that dynamic cerebral processing is underlain
by parallel distributed and interactive networks, the socalled hodology (Duffau, 2008a). This connectionist
view also opens the door to the concept of cerebral
plasticity, crucial in LGG surgery.
One of the major advantages of IES for brain mapping in adult patients is that they intrinsically do not
cause any false negatives if nonetheless the methodology is rigorously applied, as detailed above. Indeed,
IES are highly sensitive for detecting the cortical and
axonal eloquent structures, and, as mentioned, they
also provide a unique opportunity to study brain connectivity, since each area responsive to stimulation is in
fact an input gate into large-scale network rather than
an isolated discrete functional site. IES, however, have
a limitation: the specificity is suboptimal. Indeed, IES
may lead to interpretation that a structure is crucial,
due to the induction of a transient functional response
when stimulated, whereas (i) this effect is caused by
the backward spreading of the stimulation along the
network to an essential area, and (ii) the stimulated
region can be functionally compensated thanks to longterm brain plasticity mechanisms. Thus, although IES
are still the gold standard for brain mapping, due to
the risk of false positives, their combination with
new methods such as fMRI and biomathematical modeling is now mandatory, to clearly differentiate those
networks that are actually essential to function from
those that can be compensated (Mandonnet et al., in
press).
IES: A New Door to Cerebral Functional

Anatomy
In addition to fMRI, IES participate in a better understand the pathophysiology of functional areas, with
195
the goal to improve the surgical planning in eloquent

regions.
was not basically involved in speech production,

but in high-level language processing, namely: its
posterior part (pars opercularis) more involved in
phonological processing, its superior part (pars triangularis) implied in syntatic processing, while its
anterior part (pars orbitaris) is more involved in
a large semantic network underlied by the inferior fronto-occipital fasciculus (see below) (Duffau
et al., 2005b). Interestingly, these data provided by
IES are in agreement with those obtained using
fMRI, as shown in a recent meta-analysis of the
literature (Vigneau et al., 2006).
Premotor cortex and language: In addition to its
implication in motor function, IES showed that
stimulation of the dominant dorsal premotor area
(i.e. the structure lateral to the SMA, in front of the
primary motor area of the hand), induced anomia
when stimulated. Furthermore, stimulation of the
dominant ventral premotor cortex regularly elicited
anarthria (Duffau et al., 2008). Therefore, IES support the involvement (i) of the dorsal premotor
cortex in the naming network, in accordance with
fMRI studies which showed its involvement to lexical retrieval and to conceptual category; (ii) of the
ventral premotor cortex in the planification of articulation, explaining why lesion studies reported that
damage of the lower motor cortex induced speech
apraxia (i.e. aphemia).
Functional organization of Wernickes area: For
LGG located in the dominant temporal posterior
areas, comprehension tasks have been developed.
For instance, a triad of pictures is shown, and the
patient is asked to pair them by naming two pictures
with conceptual links. Interestingly, during IES,
several sites within the posterior part of the superior temporal gyrus specifically elicited an anomia
without comprehension disorders, although other
sites within the same gyrus elicited only comprehension disorders with preservation of the ability to
name, and other areas generated only phonological
disturbances. These results support the complexity
of the functional organization of Wernickes area
(in accordance with fMRI results), with its participation to, but also with possible dissociation,
between comprehension, naming and phonological
processing (Vigneau et al., 2006).
Left angular gyrus and calculation: The supramarginal and angular gyri, in the dominant hemisphere, participate to complex cognitive functions,
such as calculation. In patients with a left posterior
Supplementary Motor Area (SMA) syndrome: The

SMA, i.e. the frontomesial area in front of the primary motor cortex, involved in the planning of the
movement, is often invaded by LGG. Its resection
induces the classical SMA syndrome, with a complete akinesia and even mutism (if damage of the
left dominant SMA), which occurs approximately
30 min after its resection. Then, this syndrome suddenly and spontaneously resolves around the tenth
day following surgery, even if some rehabilitation is
often needed during 13 months. Preoperative fMRI
showed that the occurrence of this syndrome was
not related to the volume of the frontal resection, but
directly to the removal of a specific structure called
the SMA-proper. Thus, fMRI enables to predict,
before surgery, if a SMA syndrome will occur or not
postoperatively, and to inform the patient and his
family (Krainik et al., 2004). Moreover, by combining preoperative fMRI, the pattern of clinical deficit
after surgery, and the extent of resection on the
postoperative MRI, the existence of a somatotopy
within the SMA-proper has been demonstrated
with, from anterior to posterior: the representation
of language (at least in the dominant hemisphere),
of the face, then the upper limb, then the lower
limb. This is important for the planning of a specific
rehabilitation.
Involvement of the insular lobe in language and
swallowing: Although the insula is also frequently
involved by LGG, this structure was poorly studied
during a long time. Recent fMRI studies showed
the role of this multimodal lobe in many functions, in particular in language. These results were
confirmed by IES, which induced language disorders, and more specifically articulatory disturbances
when applied on the insular cortex, supporting the
role of this structure in the complex planning of
speech (Duffau, 2009b). It means that in left dominant (fronto-temporo)-insular LGG, resection carries a high risk to be incomplete. Furthermore, after
resection of right insulo-opercular LGG, a transient Foix-Chavany-Marie syndrome, i.e. a bilateral facio-linguo-pharyngo-laryngal palsy, may be
elicited, with a reversible unability for the patient to
speak and swallow.
Functional organization of the left inferior frontal
gyrus: IES showed that the classical Brocas area
196
parietal LGG, both multiplication and subtraction

can be tested using IES. Interestingly, functional
epicenters more involved in arithmetic facts such
as rote multiplication, with tables learned by heart,
were found to be located immediately above the
posterior end of the sylvian fissure, thus very close
to the language sites. On the other hand, actual
calculation such as subtraction recruited functional
sites located in the superior part of the angular
gyrus, immediately below the intraparietal sulcus,
namely close to the areas involved in working
memory. These results support the existence of a
calculotopy within the angular gyrus. Despite a
transient dyscalculia following surgery, the patients
recovered.
Involvement of frontal eye field and cingulate
eye field in oculomotor behavior: The functional
anatomy of the frontal eye field was studied by
preoperative fMRI and IES. This region, located
laterally and in front of the primary motor area of
the face, is implied in the regulation of the voluntary and unvoluntary ocular saccades. IES over
this area evoked contraversive smooth eye movements recorded electro-oculographically. In addition, stimulation of an anterior sub-region of this
electrically determined frontal eye field disclosed
both smooth eye movement and interfered with oculomotor behavior, supressing self-paced saccades in
awake patient. The posterior part of the anterior
cingulum, namely the cingulate eye field, also
plays a role in suppression of unwanted saccades
(antisaccades), thus in attentional processing.
Role of the right supramarginal gyrus and posterior temporal areas in spatial awareness: The use
of a line bisection task during awake surgery in
patients with a LGG in the right parieto-temporal
junction enables the mapping the areas involved
in the spatial awareness. A significant rightward
deviation is usually observed during the stimulation of the antero-inferior part of the supra-marginal
gyrus and the caudal part of the superior temporal gyrus. A transient and reproducible left neglect
is thus induced by electrical inactivation of cortical sites essential for the visuo-spatial integration,
in the right parieto-temporal junction. If these eloquent areas are preserved, the patients show no
signs of neglect a few days after surgery. These findings demonstrate that the supramarginal gyrus and
the caudal part of the superior temporal gyrus, at
H. Duffau
least in the right hemisphere, are critical to the symmetrical processing of the visual scene in humans
(Thiebaut et al., 2005).
Involvement of the left dorsolateral cortex in judgment: For LGG in the left dominant prefrontal
cortex, task of cross-modal (visual-verbal) congruent and incongruent judgment can be performed
in awake patient. Visual and auditory stimuli were
presently simultaneously, both referring to the same
item (congruence condition), or not (semantic or
phonemic incongruent condition). Brain areas not
involved in naming processing elicited reproducible
deficit of incongruent judgment when stimulated,
especially at the level of the left dorso-lateral prefrontal region even if an interindividual variability
was observed, as for other functions (Plaza et al.,
2008). Preservation of such executive functions is
essential for the daily life, in particular regarding the decision-making and planning of complex
strategy.
Interestingly, other anatomo-functional correlations
can also be made using IES in patients operated on
for LGG, in particular with regard to writing, reading, bilingualism and language switching, memory,
emotional processing or even control of micturition.
Therefore, the neurosurgeon must adapt his strategy,
particularly the surgical technique (e.g. the selection of
the functional tasks to optimize the reliability of the
intrasurgical mapping) to the better knowledge of the
functional anatomy applied to each patient.
IES: New Insights into the Subcortical

Connectivity
As mentioned, the study of individual anatomofunctional connectivity underlying the eloquent networks is highly recommended in LGG surgery, to
avoid postoperative permanent neurological deficit.
Subcortical motor tracts: In precentral LGG, after
detection and preservation of the primary motor
cortical areas using IES, it is also important
to detect the corresponding descending motor
pathways using subcortical stimulation, and their
somatotopy i.e. the different fibers of the corona
radiata, with the pyramidal bundles of the lower
197
limb medially, of the upper limb and of the face

more laterally. These subcortical motor fibers constitute the posterior and deep functional limits of
the resection, until the opening of the ventricle. The
pyramidal pathways may also be identified within
the posterior limb of the internal capsule, particularly in (fronto-temporo-)insular LGG, in which the
deep boundaries of the resection are given when
subcortical stimulation induce motor responses in
the inferior part of the corona radiata up to the superior part of the mesencephalic peduncles (Duffau
et al., 2003).
Subcortical somatosensory pathways: In the same
way, the thalamo-cortical somatosensory pathways
and their somatotopy can be identified by IES,
which induce dysesthesias in awake patients, in
cases of retrocentral LGG (Duffau et al., 2003).
Subcortical visual pathways: Optic radiations can
be mapped in patients who undergo awake surgery
for temporo-occipito-parietal LGG, with induction
of a transient shadow or phosphenes in the controlateral visual field during stimulation of the posterosuperior and deep part of the surgical cavity,
sometimes with also metamorphopsia (i.e. visual
illusion). Thus, if resection is stopped at this level,
patients are left with only a residual quadrantanopsia without consequence on the quality of life,
especially for the driving.
Subcortical language pathways: In left dominant
precentral LGG, after identification of the motor
and language cortical sites within the pre-rolandic
and inferior frontal gyri, IES also enable the detection of the language pathways (Duffau et al., 2008).
Medially, IES can identify the fasciculus subcallosal medialis (running from the SMA and cingulate
gyrus to the head of the caudate nucleus) which
elicites transient transcortical motor aphasia during
its stimulation: this tract is involved in the initiation
of language. Posteriorly, the fibers coming from the
premotor ventral cortex must be detected, inducing
anarthria when stimulated: this pathway is crucial
for speech production. More laterally, the operculoinsular connections should also to be detected, by
generating a complete speech arrest during stimulation: these connections are involved in speech
planning.
pathways, with first of all, the deep part of the superior

longitudinal fasciculus namely the arcuate fasciculus (AF). In patients with a LGG involving the left
insula or left inferior frontal gyrus, IES can identify
the anterior part of AF, located within the anterior floor
of the external capsule (under the superior part of the
insula) and also under the posterior part of the so-called
Brocas area (namely the pars opercularis and pars
triangularis of the inferior frontal gyrus). Stimulation
induces transient symptoms observed in conduction
aphasia, i.e. phonemic paraphasia and repetition disorders. AF must also be detected at the level of its
postero-superior loop, located under the supramarginal
gyrus, in patients operated on for a left parietal LGG.
The same symptoms associating phonemic paraphasias
and repetition disorders are induced during stimulation. Again, AF is detected in posterior temporal LGG,
its posterior part corresponding to the anterior functional limit of the resection. Finally, the anterior part
of the lower AF must also been used as the posterior functional boundary of left dominant anterior
and mid-temporal lobectomy (Duffau et al., 2008).
AF seems also to subserve a wide network involved
in language schwitching (from a native language to
another language or vive versa): IES can disrupt such
function, crucial to preserve in bilingual patients. In
addition to the AF, there is a lateral part of the superior longitudinal fasciculus. In patients harboring a left
retrocentral supra-sylvian LGG, IES detect not only
the language cortical sites at the level both of the ventral premotor cortex in front of the tumor and of the
supramarginal gyrus and/or angular gyrus behind it,
but also a fronto-parietal subcortical network inducing speech apraxia when stimulated. This operculoopercular loop might underly the anatomo-functional
connectivity of the working memory circuit. Indeed,
this loop corresponds to the anterior segment of an
indirect pathway of the classical superior longitudinal
fasciculus, which runs parallel and lateral to the AF, by
connecting Brocas territory with the inferior parietal lobe as recently showed by DTI. This tract might
be involved in the vocalization of semantic content.
Therefore, this example illustrates well that IES and
DTI can be combined in order to better understand the
anatomo-functional connectivity of the brain (Duffau,
2008a).
In addition to this dorsal phonological root, IES
supported the likely role of the inferior fronto-occipital
fasciculus (IFOF) in the semantic system the ventral
In addition to these loco-regional language pathways, IES also detect the long-distance association
198
semantic root. In patients with a frontal LGG in front

and above the Brocas area i.e. the pars orbitaris of
the left inferior frontal gyrus and the dorsolateral prefrontal area the anterior part of the IFOF has been
identified under these regions, by eliciting semantic
paraphasias during subcortical stimulation. IFOF was
also detected in surgery for left insular LGG, by inducing the same symptoms (semantic paraphasias) when
stimulated, in its intermediate part located in the anterior floor of the internal capsule (in front and inferiorly
to the AF, and behind and superiorly to the uncinate
fasciculus). Again, IFOF was detected in left temporal
LGG, by eliciting semantic disorders when stimulated:
it constituted the deep limit of the resection (above
the roof of the temporal horn of the ventricle) (Duffau
et al., 2005b).
Interestingly, stimulation of the anterior part of the
inferior longitudinal fasciculus, in front of the visual
word form area (i.e. the basal part of the temporooccipital junction, involved in high-level visual processing such as reading), and stimulation of the
uncinate fasciculus, never generated language disturbances. Thus, these fasciculi can be removed without
aphasia. This indirect pathway from the temporooccipital areas to the prefrontal region, with a relay
in the temporal pole (temporo-occipital area, inferior
longitudinal fasciculus, temporal pole, uncinate fasciculus, orbito-frontal and prefrontal areas) might be
compensated by the direct pathway constituted by the
IFOF (Duffau et al., 2008).
IES also allow the mapping of the deep gray nuclei,
sometimes invaded by LGG. Indeed, stimulation of
the head of the dominant caudate in patients with
a frontomesial LGG coming in contact of the striatum in the depth, generally generates perseverations,
namely the repetition of the previous item while the
next item is presented to the patient. These results
support an inhibitor role of the caudate in the control of cognition. Equally, it is important to map the
lateral part of the dominant lentiform nucleus, at the
end of the resection of insular LGG. Lentiform stimulation induces anarthria, supporting the likely role of
this structure in the planning of articulation, in association with the insula and ventral premotor cortex
(Duffau et al., 2008). Finally, it is also important to
use IES for language mapping, both at cortical and
subcortical levels, for LGG involving the right hemisphere in right-handed or even ambidextrous patients,
due to the bilateral distribution of language. In all
H. Duffau
cases, these language bundles should constitute the

subcortical functional limits of the resection.
Subcortical pathway involved in spatial awareness: Using the task of line bisection previously
described during awake surgery in patients with a
right parieto-temporal LGG, IES can also detect
the white matter tracts implied in spatial processing, to avoid postoperative left neglect. During the
stimulation of the lateral part of the superior longitudinal fasciculus, a significant rightward deviation
is regularly observed (Thiebaut et al., 2005). This
parieto-frontal pathway seems to subserve spatial
awareness, and a lesion at its level may generate
a permanent left neglect. Stimulation of the right
superior longitudinal fasciculus may also induce
vertigo, by disrupting a large network between the
parieto-insular vestibular cortex, the visual and the
sensory-motor areas.
These results suggest that damage to restricted
regions of white matter can cause the dysfunction of
large-scale cognitive networks. Also, they show that it
is possible to adapt the intraoperative testing to each
patient with the goal to map the subcortical pathway
underlying other cognitive functions than language.
Interestingly, although IES of the inter-hemispheric
white matter pathways has been performed, no functional responses were elicited by the stimulation of the
corpus callosum. Thus, it is possible to remove LGG
within this structure without consequence on the quality of life, whatever the location of the callosectomy.
In summary, the vision of the neural basis of cognition begins to shift from a localisationist then an
associationist view towards a hodological organization (i.e. dynamic parallel large-scale networks able
to compensate themselves). Indeed, from Lichtheim
to Geschwind, cognitive functions such as language
were conceived in associationist terms of centers and
pathways, the general assumption being that visual
and auditory linguistic information were processed
in localized cortical regions with the serial passage
of information between regions through white matter
tracts. Currently, an alternative hodological account is
proposed, in which language is conceived as resulting from parallel distributed processing performed by
distributed groups of connected neurons rather that
individual centers (Duffau, 2008a). Consequently, it is
crucial for the neurosurgeon to improve his knowledge
199
of the anatomo-functional connectivity, thus to integrate more easily and more systematically the concept
of subcortical mapping in his surgical strategy. First
because the gliomas, in essence, involve both cortical and subcortical structures, and thus they may alter
the connectivity. Second, because lesions of the white
matter may elicite more severe permanent deficits than
cortical damages. In addition, such hodological view
may explain why some epicentres considered as essential for language in a localisationist model, for instance
Brocas area, can be in certain conditions involved by
a LGG (or even surgically removed) with no aphasia due to a functional compensation within a large
distributed network, i.e. the so-called brain plasticity
(Desmurget et al., 2007).
the use of IES allows a better study of the dynamic

reorganization of the eloquent maps induced by LGG
at the individual scale, and thus opens new surgical
indications in classically not operable areas.
IES: A Study of Cerebral Plastic Potential

Brain plasticity can be defined as the continuous processings allowing short, middle and long-term remodelling of the neurono-synaptic organization, in order
to optimize the functioning of the networks of the
brain during phylogenesis, ontogeny, physiological
learning and following lesions involving the peripheral as well as the central nervous system. Several
hypotheses about the pathophysiological mechanisms
underlying plasticity have been considered. At a microscopic scale, these mechanisms seem to be represented by: synaptic efficacy modulations, unmasking
of latent connections, phenotypic modifications, synchrony changes and neurogenesis. At a macroscopic
scale, diaschisis, functional redundancies, cross-modal
plasticity with sensory substitution and morphological
changes are suggested to be involved. The behavioral
consequences of such cerebral phenomena have been
analyzed in humans in the last decade, both in physiology ontogeny and learning and in pathology.
In particular, the ability to recover after a lesion of
the nervous system, and the patterns of functional
reorganization within eloquent area and/or within distributed networks, allowing such compensation, have
been extensively studied (Duffau, 2008a).
Interestingly, the field of slow-growing brain tumors
such as LGG, has demonstrated that large amounts
of cerebral tissue could be removed, inside or outside
the so-called eloquent areas, with impressive recovery
allowing no permanent detectable functional consequences (Duffau, 2005). Such knowledge combined to
Preoperative plasticity: As mentioned, it could

seem surprising that patients with a LGG have
usually only mild functional deficit, despite the
frequent invasion of eloquent structures. These
slow-growing lesions have induced progressive
functional brain reshaping, as suggested by preoperative fMRI. Interestingly, the patterns of reorganization may differ between patients, a very important
notion to know by the neurosurgeon with the goal
to optimize both surgical indications and planning
(Duffau, 2005). Indeed, fMRI showed that three
kinds of preoperative functional redistribution are
possible. In the first one, due to the infiltrative feature of LGG, function still persists within the tumor,
thus with a very limited chance to perform a fair
resection without inducing postoperative sequelae.
In the second one, eloquent areas are redistributed
around the tumor, thus with a reasonable chance
to perform at least a near-total resection despite
an immediate transient deficit but with secondary
recovery within some weeks. In the third one, there
is already a preoperative compensation by remote
areas within the lesional hemisphere and/or by
the controlateral homologuous: consequently, the
chances to perform a real total resection of this kind
of gliomas are very high, with only a slight and very
transient deficit. Therefore, in cases of LGG involving eloquent areas, plasticity mechanisms seem to
be based on an hierarchically organized model,
i.e.: first with intrinsic reorganization within injured
areas (indice of favorable outcome); second, when
this reshaping is not sufficient, other regions implicated in the functional network are recruited, in the
ipsilateral hemisphere (close and even remote to the
damaged area) then in the controlateral hemisphere
if necessary (Duffau, 2008b).
Intraoperative plasticity: IES before any resection
has allowed the confirmation of the existence of a
functional reshaping induced by LGG, notably with
a possible remapping of the sensorimotor homonculus and also a reorganization of the language
sites. Moreover, acute reorganization of functional
maps was observed throughout the resection, likely
due to the surgical act itself which can generate a
200
loco-regional hyper-excitability as in head injury

(Duffau, 2005). For instance, in several patients
with a frontal LGG, although stimulation of the
precentral gyrus induced motor responses only at
the level of a limited number of cortical sites
before resection, an acute unmasking of redundant motor sites located within the same precentral gyrus and eliciting the same movements than
the previous adjacent sites when stimulated, was
observed immediately following lesion removal.
Acute unmasking of redundant somatosensory sites
was also regularly observed within the retrocentral
gyrus in patients operated on for a parietal glioma.
Furthermore, a redistribution within a more larger
network involving the whole rolandic region was
detected, i.e. with unmasking of functional homologuous located in the precentral gyrus for the
first cortical representation and in the retrocentral
gyrus for its redundancy (or vice versa). Finally,
intraoperative mapping has also a prognostic value
concerning the postoperative recovery: a positive
response means that the patient will recover.
Postoperative plasticity: The mechanisms of such
a plasticity induced by surgical resection within
eloquent areas were also studied, by performing
postoperative fMRI once the patient has recovered
his preoperative functional status. For instance, several patients were examined following the resection
of LGG involving the SMA, which elicited a transient postsurgical SMA syndrome. fMRI showed,
in comparison to the preoperative fMRI, the occurrence of activations of the SMA and premotor
cortex contralateral to the lesion: the contrahemispheric homologuous thus participated to the postsurgical functional compensation (Krainik et al.,
2004).
IES: Applications to LGG Surgery

It was recently proposed to incorporate such a better
understanding of the individual plastic potential in the
surgical strategy in LGG, with the aims (i) to extent
the indications of resection in eloquent structures so
far considered as inoperable (ii) to maximize the
EOR, by performing the resection according to (not
fixed) functional boundaries (iii) while minimizing the
risk of postoperative permanent deficit or even while
H. Duffau
improving the quality of life (Duffau, 2005, 2008b).

Consequently, several surgical series showed that it
was possible to remove LGG invading the following
eloquent brain structures (Fig. 21.2):
SMA resection: it induces the occurrence of an
SMA syndrome. As mentioned, all the patients
recover, and postoperative fMRI supported functional compensation by the controlateral SMA and
premotor cortex as well as by the ipsilesional primary motor cortex;
Insular resection: despite a transient hemiparesis
after right insula removal, likely because this region
is a non-primary motor area, and transient speech
disturbances following left dominant insula resection, all patients recover except in rare cases
of deep stroke (Duffau, 2009b). Moreover, it was
possible in right non-dominant fronto-temporoinsular LGG involving the deep grey nuclei, to
remove the clautrum without any cognitive disorders (despite its role suggested in consciousness), and to remove the invaded striatum without
inducing neither motor deficit nor movement disorders. This compensation can be explained by a
recruitment of parallel subcortical circuits such as
pallido-luyso-pallidal, strio-nigro-striate, corticostrio-nigro-thalamo-cortical and cortico-luysal networks.
Resection of the primary somatosensory area: the
first results using pre- and post-operative magnetoencephalography suggested the possible recruitment of redundant eloquent sites around the
cavity, within the postcentral gyrus. It is in accordance with the IES data, showing unmasking of
redundant somatosensory sites during resection,
likely explained by the decrease of the corticocortical inhibition. The recruitment of the second
somatosensory area or posterior parietal cortex, primary motor area (due to strong anatomo-functional
connections between the pre- and retro-central
gyri), and controlateral primary somatosensory area
are also possible to explain the recovery;
The resection of the (dominant) parietal posterior
lobe can be performed without permanent deficit,
and even with a possible improvement in comparison to the preoperative status, especially using
pointing task (Desmurget et al., 2007);
Resection of non-dominant sensorimotor area of
the face: the recovery of the usual transient central
201
Fig. 21.2 Examples of

extensive glioma resection
performed within eloquent
areas using IES, with
preservation the quality of life
thanks to brain plasticity:
(a) left SMA (b) entire left
frontal lobe including Brocas
area (c) right primary
sensorimotor area of the face
(d) left primary sensorimotor
area of the face (e) primary
motor area of the hand (f) left
primary somatosensory area
and parietal lobe (g) right
paralimbic system (h) left
insula (i) left dominant
temporal lobe
facial palsy, with a potential Foix-Chavany-Marie

syndrome when the insula is also involved, is likely
explained by the disinhibition of the controlateral
homologous sites, via the transcallosal pathways
(Duffau, 2005);
Resection of the knob of the hand: on the basis
of the existence of multiple cortical motor representations showed in humans using fMRI and IES,
the compensation of the motor function could be
explained by the recruitment of parallel networks
within the primary motor region allowing the
superior limb area removal, eventually using two
consecutive surgeries in order to induce durable
remapping following the first one (see below);
Brocas area resection: the language compensation may be underlain by the recruitment of adjacent regions, in particular the pars orbitaris of the
inferior frontal gyrus, the dorsolateral prefrontal

cortex and the insula (Duffau, 2008b);
Temporal language area resection: language compensation following a left dominant temporal resection could be explained by the fact that this function
seems to be organized with multiple parallel networks (Vigneau et al., 2006). Consequently, beyond
the recruitment of areas adjacent to the surgical
cavity, the long term reshaping could be related
to progressive involvement of (i) remote regions
within the left dominant hemisphere such as
the posterior part of the superior temporal gyrus,
the pars orbitaris of inferior frontal gyrus or other
left frontolateral regions (ii) the controlateral
right non-dominant hemisphere due to a transcallosal disinhibition phenomenon (Desmurget et al.,
2007).
202
IES: Functional and Oncological Results

in LGG Surgery
A dramatic improvement of the surgical results was
provided by advances in IES. First, the use of IES
has allowed to significantly increase the surgical indications in eloquent areas which were classically considered as not operable (see above) (Duffau et al.,
2005a). In addition, despite a frequent transitory neurological worsening in the immediate postoperative
period (due to the attempt to perform a maximal
glioma removal according to cortico-subcortical functional limits using IES), leading to a specific functional
rehabilitation, more than 98% of patients recovered
the same status than before surgery after LGG resection within eloquent areas guided by functional mapping, and returned to a normal socio-professional life
(Duffau et al., 2008; Sanai et al., 2008). Even more,
1520% of patients can improve in comparison to
their preoperative neurological and neuropsychological assessment, and 80% of patients with presurgical
intractable epilepsy can benefit from a relief of their
seizures (Duffau, 2009b). Therefore, LGG surgery is
currently not only able to preserve brain functions
but may also improve the quality of life of patients,
as demonstrated by extensive neurocognitive assessment performed after surgical resection (Teixidor et al.,
2007). These data support the existence of additional
brain plasticity mechanisms occurring after the operation, likely facilitated by a systematic and adapted
rehabilitation (Ghering et al., 2009). Interestingly, this
rate of less than 2% of sequelae is very reproducible
among the teams using IEM worldwide. In comparison, in series which did not use IEMS, the rate
of sequelae ranged from 13 to 27.5%, with a mean
around 19% (for a review, see Duffau et al., 2005a).
Finally, comparative studies between LGG resection
performed without or with IES showed that the EOR
was significantly increased thanks to IES, despite better functional results following resection within eloquent areas (Duffau et al., 2005a; de Benedictis et al.,
in press). Indeed, since IES allows identification of the
cortical and subcortical eloquent structures individually, it seems logical to perform a resection according
to functional boundaries. The resection is continued
until the functional structures are detected by IES, and
not before, to optimize the EOR without increasing
the risk of permanent deficit. Moreover, as mentioned
H. Duffau
in the introduction, a more extensive resection is now

demonstrated as inducing an increased impact on the
natural course of LGG.
Towards a Multiple-Stages Surgical

Approach in LGG
The price to pay to obtain such favorable functional
results is sometimes to perform incomplete LGG
resection, when the tumor invades sites still crucial for
the function. A new concept recently proposed is to
use postoperative fMRI, since it can be easily repeated
due to its non-invasive feature, when the patient has
totally recovered, in order to compare the new maps
to those obtained before surgery. Indeed, even if this
method has some limitations detailed above, subtraction between a pre- and post-surgical acquisition may
nonetheless shows a possible additional functional
reshaping, due to (i) the resection itself (ii) the rehabilitation (iii) the re-growth of the residual tumor (as
before surgery).
Interestingly, recent series demonstrated that such
remapping was not a theoretical concept, but a concrete
reality. Postoperative fMRI performed some months
or years following the surgery clearly showed a new
recruitment of perilesional areas and/or remote regions
within the ipsilesional hemisphere and/or a recruitment of controlateral structures (Krainik et al., 2004).
Based on these data, a second surgery was proposed in
patients who continued to lead a normal life, before
the occurrence of new symptoms (except possible
seizures), only because of an increase of the volume of LGG. The second surgery was also conducted
using intraoperative cortical and subcortical mapping,
in order to validate the mechanisms of brain reshaping supposed but not proven by preoperative fMRI,
before to perform the additional resection (Gil Robles
et al., 2008). The preliminary results have supported
the efficacy and the safety of such re-operation for
LGG not totally removed during a first surgery, due
to their location within eloquent areas. Indeed, in this
recent experience, 74% resections were complete or
subtotal (less than 10 ml of residue) following the
second operation, despite no additional serious neurological deficit at the contrary, with an improvement
of the neurological status in 16% of cases. Again,
the seizures were reduced or disappeared in 82% of
203
patients with epilepsy before the second operation.

The median time between the two operations was 4.1
years, and all patients were still alive with a median
follow-up of 6.6 years despite an initial incomplete
resection. Therefore, these original data demonstrated
that, thanks to mechanisms of cerebral plasticity, it
is possible to re-operate patients with a LGG involving eloquent areas with a minimal morbidity and an
increase of the EOR. However, 58% of tumors had
already progressed to high grade glioma during the
second surgery, raising the problem of the timing of reoperation. It was thus suggested to over indicate an
early re-intervention, in order to anticipate the second
surgery before the anaplastic transformation (Martino
et al., 2009). Interestingly, one can currently consider
to perform postoperative fMRI after rehabilitation and
recovery following a second surgery, in order to open
the door to a possible third or even fourth resection
several years after the previous operations. The goal is
both to allow the patient to enjoy a normal life as well
as to increase the OS. It is also possible to integrate
surgeries within a dynamic therapeutic strategy including chemotherapy and radiotherapy, especially when
a wide removal is not possible for functional reasons
(Duffau, 2009a). To this end, neoadjuvant chemotherapy was recently advocated in LGG, with the goal to
induce a shrink of the tumor before an operation or a
re-operation (Duffau et al., 2006).
(accurately analyzed by a speech therapist along the

surgical procedure). Combination of these intraoperative anatomo-functional data with those provided
by DTI (subcortical anatomical informations), magnetoencephalography (temporal data), and fMRI (perioperative functional data) could enable one to elaborate
individual and predictive models of the functioning of
neurono-synaptic circuits, i.e. to open a new door to
hodology (Duffau, 2008b). Such correlations with IES,
which remains the gold-standard for functional brain
mapping, can also enable to validate the non-invasive
method of neuroimaging, especially the new technique
of DTI (Leclercq et al., 2010). Anatomic dissections of
the brain, in particular the white matter pathways, are
now to be revisited in the lights of functional mapping
(Martino et al., in press).
Moreover, such connectionist models may lead to a
better knowledge of the dynamic potential of spatiotemporal reorganization of the parallel and interactive
networks, namely the mechanisms of brain plasticity,
thought to play a major role of functional compensation in slow-growing tumors and in their surgical resection. In this way, individual plastic potential could be
better understood using repeated intraoperative mappings combined to post-surgical fMRI, then possibly
guiding specific post-operative rehabilitation program
in order to optimize the quality of functional recovery
(Duffau, 2005, 2009b; Gil Robles et al., 2008).
In addition to its fundamental implications, IES also
allow to perform tumor resection according to functional boundaries, therefore, leading to the optimization of the benefit-to-risk ratio of surgical removal in
cerebral glioma. Indeed, these new techniques (serial
brain mappings) and concepts (hodology and plasticity) have allowed (i) to significantly extend the indications of resection in eloquent areas classically considered as inoperable, e.g. Brocas area, the insula
even in the left dominant hemisphere, the central
area or the left posterior temporal regions (ii) to significantly decrease the rate of permanent deficit at less
than 2%, instead of 1327% (mean 19%) without mapping (iii) and even to improve the quality of life, thanks
to seizures control (in around 80% of cases, especially
in insular and/or temporal LGG) and thanks to cognitive rehabilitation (iv) and to significantly increase
the EOR compared with the series without brain mapping, thus with an increased impact on the median
survival. In practice, in order to evolve towards a multistage surgical approach (i.e. second or third surgery
Conclusions and Perspectives

LGG surgery now benefits from important technical developments in the field of functional mapping,
using complementary non-invasive methods of fMRI
and invasive IES. Such recent advances have enabled
to better understand the organization of the eloquent
brain for each patient, in order to integrate the concept of inter-individual anatomo-functional variability in the surgical strategy. Furthermore, intraoperative real-time subcortical stimulation, in association
with cortical mapping, gives an unique opportunity
to study the so-called effective connectivity, since
it allows on-line correlations between discrete and
transient virtual lesions which can be performed at
each place of a distributed network (each cortical and
subcortical sites being perfectly identified anatomically using 3D MRI) and their functional consequences
204
more extensive than the first one in cases of initial

incomplete resection within crucial areas), a dynamic
strategy has to be envisaged for functional neuroimaging (Gil Robles et al., 2008; Martino et al., 2009).
The goal is to switch from a static use of a unique
preoperative fMRI assessment (limited technique with
lack of reliability), to longitudinal studies based on
the repetition of the fMRI before and after surgical
resection(s), with the goal to analyze a possible brain
reshaping at the individual scale, and to select the candidates to re-operation(s). The next step is now to
use bio-mathematical models able to examine brain
functional interaction through effective connectivity,
in order to attempt to predict before surgery the patterns of postsurgical remapping at the individual scale
on the basis of the data provided by the preoperative
fMRI.
References
Claus EB, Horlacher A, Hsu L, Schwartz RB, Dello-Iacono
D, Talos F, Jolesz FA, Black PM (2005) Survival rates in
patients with low-grade glioma after intraoperative magnetic
resonance image guidance. Cancer 103:12271233
de Benedictis A, Moritz-Gasser S, Duffau H (2010) Awake mapping optimizes the extent of resection for low grade gliomas
in eloquent areas. Neurosurgery 66:10741084
Desmurget M, Bonnetblanc F, Duffau H (2007) Contrasting
acute and slow growing lesions: a new door to brain plasticity. Brain 130:898914
Duffau H (2005) Lessons from brain mapping in surgery for lowgrade glioma: insights into associations between tumour and
Duffau H (2008a) The anatomo-functional connectivity of language revisited: new insights provided by electrostimulation
and tractography. Neuropsychologia 4:927934
Duffau H (2008b) Brain plasticity and tumors. Adv Tech Stand
Neurosurg 33:333
Duffau H (2009a) Surgery of low-grade gliomas: towards a
Duffau H (2009b) A personal consecutive series of surgically
treated 51 cases of insular WHO grade II glioma: advances
and limitations. J Neurosurg 110:696708
Duffau H (2010) Awake surgery for non-language mapping.
Duffau H, Capelle L (2004) Preferential brain locations of lowgrade gliomas. Cancer 100:26222626
Duffau H, Capelle L, Denvil D, Sichez N, Gatignol P, Taillandier
L, Lopes M, Mitchell MC, Roche S, Muller JC, Bitar A,
Sichez JP, van Effenterre R (2003) Usefulness of intraoperative electrical subcortical mapping during surgery for
low-grade gliomas located within eloquent brain regions:
functional results in a consecutive series of 103 patients. J
Neurosurg 98:764778
H. Duffau
Duffau H, Gatignol P, Mandonnet E, Capelle L, Taillandier L
(2008) Contribution of intraoperative subcortical stimulation
mapping of language pathways: a consecutive series of 115
patients operated on for a WHO grade II glioma in the left
dominant hemisphere. J Neurosurg 109:461471
Duffau H, Gatignol P, Mandonnet E, Peruzzi P, TzourioMazoyer N, Capelle L (2005b) New insights into the
anatomo-functional connectivity of the semantic system:
a study using cortico-subcortical electrostimulations. Brain
128:797810
Duffau H, Lopes M, Arthuis F, Bitar A, Sichez JP, Van Effenterre
R, Capelle L (2005a) Contribution of intraoperative electrical
stimulations in surgery of low grade gliomas: a comparative study between two series without (19851996) and
with (19962003) functional mapping in the same institution.
J Neurol Neurosurg Psychiatry 76:845851
Duffau H, Taillandier L, Capelle L (2006) Radical surgery after
chemotherapy: a new therapeutic strategy to envision in
grade II glioma. J Neurooncol 80:171176
Ghering K, Sitskoorn MM, Gundy CM, Sikkes SA, Klein
M, Postma TJ, van den Bent MJ, Beute GN, Enting RH,
Kappelle AC, Boogerd W, Veninga T, Twijnstra A, Boerman
DH, Taphoorn MJ, Aaronson NK (2009) Cognitive rehabilitation in patients with gliomas: a randomized, controlled
trial. J Clin Oncol 27:37123722
Gil Robles S, Duffau H (2010) Surgical management of WHO
grade II gliomas in eloquent areas: is the preservation of a
margin necessary around functional structures? Neurosurg
Focus 28(2):Intro
Gil Robles S, Gatignol P, Lehricy S, Duffau H (2008)
Long-term brain plasticity allowing multiple-stages surgical approach for WHO grade II gliomas in eloquent
areas: a combined study using longitudinal functional MRI
and intraoperative electrical stimulation. J Neurosurg 109:
615624
Giussani C, Roux FE, Ojemman J, Sganzerla EP, Pirillo D,
Papagno C (2010) Is preoperative functional magnetic resonance imaging reliable for language areas mapping in brain
tumor surgery? Review of language functional magnetic resonance imaging and direct cortical stimulation correlation
studies. Neurosurgery 66:113120
Krainik A, Duffau H, Capelle L, Cornu P, Boch AL, Mangin JF,
Le Bihan D, Marsault C, Chiras J, Lehricy S (2004) Role
of the healthy hemisphere in recovery after resection of the
supplementary motor area. Neurology 62:13231332
Kral T, Kurthen M, Schramm J, Urbach H, Meyer B (2007)
Stimulation mapping via implanted grid electrodes prior to
surgery for gliomas in highly eloquent cortex. Neurosurgery
61:319325
Leclercq D, Duffau H, Delmaire C, Capelle L, Gatignol P,
Ducros M, Chiras J, Lehricy S (2010) Comparison of diffusion tensor imaging tractography of language tracts and intraoperative subcortical stimulations. J Neurosurg 112:503511
78:179185
Mandonnet E, Winkler PA, Duffau H (2010) Direct electrical
stimulation as an input gate into brain functional networks:
principles, advantages and limitations. Acta Neurochir
(Wien) 152:185193
205
Martino J, Brogna C, Gil Robles S, Vergani F, Duffau H (2010)

Anatomic dissection of the inferior fronto-occipital fasciculus revisited in the lights of brain stimulation data. Cortex
46:691699
Martino J, Taillandier L, Moritz-Gasser S, Gatignol P,
Duffau H (2009) Re-operation is a safe and effective therapeutic strategy in recurrent WHO grade II gliomas within
eloquent areas. Acta Neurochir (Wien) 151:427436
Mauer ME, Bottomley A, Taphoorn MJ (2008) Evaluating
health-related quality of life and symptom burden in brain
tumour patients: instruments for use in experimental trials
and clinical practice. Curr Opin Neurol 21:745753
McGirt MJ, Chaichana KL, Attenello FJ, Weingart JD, Than K,
Burger PC, Olivi A, Brem H, Quinones-Hinojosa A (2008)
Extent of resection is independently associated with survival
in patients with hemispheric infiltrating low-grade gliomas.
60:380383
Plaza M, Gatignol P, Cohen H, Berger B, Duffau H (2008)
A discrete area within the left dorsolateral prefrontal cortex involved in visual-verbal incongruence judgment. Cereb
Cortex 18:12531259
Rezvan A, Christine D, Christian H, Olga Z, Lutz E, Marius H,
Stephanie C, Christal HM, Rainer WC, Andreas U (2009)
Long-term outcome and survival of surgically treated supra-
tentorial low-grade glioma in adult patients. Acta Neurochir

(Wien) 151:13591365
Sanai N, Mirzadeh Z, Berger MS (2008) Functional outcome
after language mapping for glioma resection. N Engl J Med
358:1827
Smith JS, Chang EF, Lamborn KR, Chang SM, Prados MD,
Cha S, Tihan T, Vandenberg S, McDermott MW, Berger
MS (2008) Role of extent of resection in the long-term
outcome of low-grade hemispheric gliomas. J Clin Oncol
26:13381345
Teixidor P, Gatignol P, Leroy M, Masuet-Aumatell C, Capelle
L, Duffau H (2007) Assessment of verbal working memory
before and after surgery for low-grade glioma. J Neurooncol
81:305313
Thiebaut de Schotten M, Urbanski M, Duffau H, Volle E, Lvy
R, Dubois B, Bartolomeo P (2005) Direct evidence for
a parietal-frontal pathway subserving spatial awareness in
humans. Science 309:22262228
Vigneau M, Beaucousin V, Herve PY, Duffau H, Crivello
F, Houd O, Mazoyer B, Tzourio-Mazoyer N (2006)
Meta-analyzing left hemisphere language areas: phonology, semantics, and sentence processing. NeuroImage 30:
14141432
Willems PW, Taphoorn MJ, Burger H, Berkelbach van der
Sprenkel JW, Tullekan CA (2006) Effectiveness of neuronavigation in resecting solitary intracerebral contrast-enhancing
tumors: a randomized controlled trial. J Neurosurg 104:
360368
Chapter 22
Malignant Gliomas: Present and Future Therapeutic Drugs

Linda Coate and Warren Mason
Abstract Malignant gliomas are the most common

primary brain tumors in adults. Surgery, radiation therapy and chemotherapy are the principal components
of therapy for this heterogenous group of tumors.
Concurrent and adjuvant chemotherapy with temozolomide in addition to radiotherapy is now a standard
of care for the treatment of glioblastoma multiforme
(GBM). The role of adjuvant chemotherapy for other
types of malignant glioma is less clear. Temozolomide
has become the most commonly used agent for GBM
and even when used adjuvantly, rechallenge with this
agent is also effective at the time of relapse as salvage therapy. Molecular classification of gliomas is
of increasing importance and several prognostic and
predictive biomarkers are currently under investigation. Anti-angiogenic drugs, such as bevacizumab have
shown efficacy in relapsed GBM and this agent has
recently been approved by several regulatory agencies for recurrent tumors. Other targeted therapies,
such as those directed against the epidermal growth
factor receptor (EGFR), have been less successful.
Multiple trials are underway investigating new agents
and attempting to validate prospective biomarkers that
have shown promise in retrospective series.
Keywords Temozolomide
Astrocytoma
Oligodendroglioma Glioblastoma multiforme

Molecular biomarkers PCV
L. Coate ()
Department of Medical Oncology, Princess Margaret Hospital,
Toronto, ON, Canada M5G 2M9
e-mail: Linda.Coate@uhn.on.ca
Introduction
Malignant gliomas are the most frequently occurring primary brain tumors in adults. They are rapidly
progressive brain tumors further classified as anaplastic astrocytoma (AA) and oligodendroglioma (AO),
and glioblastoma multiforme (GBM). The classification of primary brain tumors is based on histopathological features, although molecular identification is
gaining increasing importance in the evaluation of
these malignancies. The most commonly occurring
glioma is GBM, accounting for more than half of all
gliomas. GBM tends to be a cancer of older individuals, with peak incidence occurring at the age of 65.
Unfortunately, the prognosis for GBM is dire, with
a median survival of 1218 months with aggressive
multimodal therapy. The survival for patients with
less aggressive gliomas is estimated at 25 years for
patients with AA, and 515 years for patients with
oligodendroglial tumors.
Conventional treatment of malignant brain tumors
with existing agents and drug development of experimental agents presents unique challenges. Many
patients who present with these tumors are prescribed
drugs (such as anti-epileptic medications and corticosteroids) that may potentially interfere with the
bioavailability of anticancer agents by induction of the
cytochrome p450 system. Many targeted agents are
metabolized by the cyp2A4 enzyme, creating unique
challenges for the evaluation of novel therapies in neurooncology. In addition, drug delivery may present a
challenge as a result of the blood-brain barrier. These
challenges may in part explain the generally negative
results of clinical trials of targeted therapies and the
relative lack of active agents available to combat these
tumors.

207
208
Currently, the drug temozolomide forms the cornerstone of medical treatment for malignant gliomas
(Athanassiou et al., 2005; Stupp et al., 2005, 2009).
It is the gold standard of chemotherapy for gliomas,
having demonstrable efficacy in the adjuvant setting
and at relapse (Perry et al., 2010). However, for most
patients with progressive malignant gliomas, participation in a clinical trial is the ideal approach, particularly
if the patient has been treated aggressively with a
temozolomide containing regimen at first presentation.
For patients who have been treated with temozolomide and are not candidates for a clinical study,
bevacizumab, either alone or in combination with a
cytotoxic chemotherapy agent, has emerged as an
effective treatment option. More conventional options
include cytotoxic chemotherapies such as lomustine, procarbazine, etoposide and platinum analogues.
Recent advances in the understanding of the molecular changes that underpin the genesis of malignant
gliomas have led to some progress in the treatment
of these tumors and have identified novel molecular
targets against which to evaluate emerging therapies.
However, despite intensive research and extensive clinical research endeavors, there remains a paucity of
active drugs for the treatment of malignant gliomas.
In this chapter we review currently utilized treatment paradigms and discuss current drug development
and future directions for systemic treatment of this
challenging group of malignant tumors.
Current Treatment Paradigms for

Malignant Gliomas-Initial Therapy
For patients with newly diagnosed malignant gliomas,
maximal feasible surgical excision is the preferred
initial treatment strategy. This approach provides adequate tissue for accurate pathological diagnosis, effectively reduces tumor burden and potentially improves
tumor-related symptoms and signs. Following surgery,
most patients with malignant gliomas will receive radiation therapy, a treatment that has been a standard of
care for over 30 years (Walker et al., 1980). Cranial
irradiation is usually delivered to the tumor with a
23 cm margin to a total dose of 5860 Gy in 1.8
2.0 Gy fractions. For low grade gliomas, radiotherapy
delivered at lower doses provides local control, but
increasingly radiotherapy is deferred for patients with
L. Coate and W. Mason
these tumors because survival is independent of the

time at which radiation therapy is delivered. While the
role of chemotherapy in the management of low grade
gliomas remains one of the most controversial issues
in neurooncology, early chemotherapy in the treatment
of GBM is now the standard of care for most patients
with this disease.
Temozolomide is an oral alkylating agent and is the
treatment of choice for adjuvant therapy of malignant glioma. A pivotal phase III trial conducted by the
European Organization for the Research and Treatment
of Cancer and National Cancer Institute of Canada
Clinical Trials Group (EORTC/NCIC CTG) established the role of temozolamide for GBM (Stupp et al.,
2005). A 5 year follow-up of this study was recently
published, providing convincing evidence that longterm survival is now possible for patients with this
disease (Stupp et al., 2009). This study randomized
573 patients with glioblastoma multiforme to postoperative involved-field radiation therapy (30 fractions of
radiation to a total dose of 60 Gy) to either adjuvant
radiation alone or radiation in combination with concurrent temozolomide followed by six further cycles of
temozolomide. This demonstrated an absolute overall
survival benefit of 8% (10% versus 2%) with a hazard ratio for death of 0.63 (0.530.75) (Stupp et al.,
2005) (Fig. 22.1). In a follow-up publication, the overall survival advantage remained at long-term follow up
(Stupp et al., 2009).
A retrospective study of patients from the
EORTC/NCIC CTG trial demonstrated that hypermethylation of the promoter for methyl guanine methyl
transferase (MGMT) was a major prognostic factor
for improved survival and was predictive of significant
benefit from chemotherapy. Remarkably, for patients
receiving radiotherapy and concurrent and adjuvant
temozolomide chemotherapy, the 5-year survival rates
were 15% for those with hypermethylated tumors,
versus 2% for those patients without (Hegi et al.,
2005).
This seminal study showed no adverse impact on
health-related quality of life related to the addition
of temozolomide, and benefit from temozolomide
was seen across all subsets of patients, including
22 Malignant Gliomas: Present and Future Therapeutic Drugs
209
Fig. 22.1 Overall survival (reproduced from Stupp et al., 2005)
those patients with poor baseline prognostic features.

Important questions regarding the application of temozolomide for newly-diagnosed GBM remain unanswered, particularly its role in treatment of the elderly,
and for those with poor performance status who were
excluded from the EORTC/NCIC CTG trial.
Although temozolomide is the only approved drug
for GBM in the adjuvant setting, a number of promising agents including the antiangiogenic agent bevacizumab are being evaluated in the adjuvant setting. A phase II study of 70 patients treated with
bevacizumab in addition to standard chemoradiation
demonstrated a nonsignificant improvement in overall survival compared to a matched group treated
with radiotherapy and temozolomide alone (25 versus 21 months). The role of bevacizumab in newly
diagnosed GBM is currently being addressed by two
large phase III trials. Additionally, the CENTRIC trial
is a molecularly directed study of adjuvant administration of cilengitide, an integrin inhibitor that has
demonstrated promising activity in newly diagnosed
and recurrent GBM, with temozolomide and radiation therapy for newly diagnosed GBM patients
whose tumors have methylated promoter of MGMT
(http://clinicaltrials.gov/ct2/show/NCT00689221).
Anaplastic Astrocytoma
Patients with AA frequently receive chemotherapy in
the adjuvant setting. Historically the combination of
procarbazine, lomustine and vincristine chemotherapy
(PCV) was administered to these patients following
surgery and radiotherapy for a maximal duration of
12 months, although this regimen has recently been
replaced by the more tolerable agent temozolomide.
However, there is no class I evidence to support
the administration of any adjuvant chemotherapy for
patients with AA, as several large phase III trials have
consistently failed to show significant improvement in
survival when chemotherapy was used immediately
following surgery and radiation therapy. However,
the practice of adjuvant chemotherapy for AA is
widespread because several meta-analyses have shown
that there may be as much as a 15% reduction in the
risk of death when patients with AA are treated with
adjuvant chemotherapy (Stenning et al., 1987; Fine
et al., 1993; Stewart, 2002).
While temozolomide appears to have activity in
patients with this AA, analysis of retrospective data
from two trials do not provide convincing evidence
for the use of this drug in the adjuvant setting for
210
this disease. In this analysis (Brandes et al., 2006),

109 patients with AA were identified, 60 of which
were treated with temozolomide. No differences were
observed in the 2-year or median survivals between
the two cohorts. Moreover, adjuvant chemotherapy
has been compared with adjuvant radiotherapy in a
trial that allowed patients with anaplastic astrocytoma
to participate (Wick et al., 2009). The chemotherapy in this study was either PCV or temozolomide.
Patients were randomly assigned to either treatment
and patients were crossed over to the alternative regimen upon progression. There was no difference in
overall time to treatment failure depending upon the
treatment sequence. There were also no differences
in outcome seen between patients treated with PCV
or temozolomide. Therefore data to support the role
of early chemotherapy in AA is not convincing, and
the optimal chemotherapy regimen and its timing for
patients with AA remain areas of urgent study.
Low Grade and Anaplastic Astrocytoma

and Oligodendroglioma
Radiotherapy is indicated for some patients with
low-grade astrocytic tumors, however, early postoperative radiation does not improve overall survival,
rather it only impacts time to tumor progression.
Chemotherapy has no firm role in the initial management of patients with low-grade astrocytic neoplasms.
Enhanced efficacy is observed in patients with oligodendroglial neoplasms with 1p19q co-deletion, and
so patients with this deletion may benefit from early
chemotherapy, but this has yet to be prospectively validated. The prognosis for low grade oligodendroglial
tumors is more favourable than that for low grade
astrocytomas.
Adjuvant chemoradiation in the postoperative management of patients with anaplastic oligodendroglial
tumors has been addressed in two large phase III
trials that compared two strategies for postoperative
adjuvant chemotherapy plus radiotherapy with radiotherapy alone.
Although oligodendroglial tumors are more sensitive to chemotherapy than other glial tumors, these
trials did not provide evidence of a survival advantage
when adjuvant PCV chemotherapy was added to radiotherapy, rather than delayed until there was evidence of
disease progression. Patients treated with radiotherapy
and PCV chemotherapy had increased progressionfree survival (PFS), but overall survival in both arms
were similar because PCV chemotherapy was effective
as salvage therapy at the time of radiotherapy failure
for patients initially treated with radiotherapy alone.
Management of Recurrence Current

Paradigms
Although for patients with GBM there is a definite
survival advantage from treatment with upfront temozolomide chemotherapy, the majority of patients with
this disease will experience tumor progression and
death. Prior to temozolomide becoming the standard
of care in the adjuvant setting for GBM, this agent
had been in widespread use for progressive disease,
having been shown in randomized phase II trials to
extend progression-free survival at 6 months. Because
most patients with GBM receive temozolomide in the
first-line setting as part of an adjuvant chemoradiation,
at recurrence patients with adequate performance status should be considered for entry into clinical trials
whenever possible. If a trial is not available, use of the
antiangiogenic agent bevacizumab either alone or with
cytotoxic chemotherapy has become an increasingly
popular option. This recent change in practice is a consequence of promising results of a randomized, noncomparative phase II trial that allocated patients with
recurrent or progressive GBM to treatment with bevacizumab alone or bevacizumab (Friedman et al., 2009)
in conjunction with irinotecan. This study demonstrated remarkable radiographic response rates and
progression-free survivals at 6 months in both arms,
and similar results have been observed in another albeit
smaller study (Kreisl et al., 2009). Bevacizumab, however, is not a drug devoid of risk, as its use in this
patient population has been associated with hypertension, arteriovenous thrombosis, and low incidence
of intracranial hemorrhage and craniotomy dehiscence
(Nghiemphu et al., 2008). In patients initially treated
with temozolomide who are not candidates for clinical
study or bevacizumab, nitrosurea-based chemotherapy
might be considered either as single agent therapy
(lomustine or carmustine) or in recognised combinations (PCV). (Brandes et al., 2004; Schmidt et al.,
2006; Fabrini et al., 2009).
An interesting approach to the management of
progressive GBM involves temozolomide rechallenge
using alternate extended dosing schedules (Perry et al.,
2010). The principal rationale for this strategy was the
notion that continuous exposure of tumor cells to temozolomide potentially depletes tumor MGMT thereby
overcoming resistance to this agent. Additionally,
extended temozolomide schedules result in dose intensification and potentially represent a form of metronomic chemotherapy that may have antiangiogenic
effects (Tuettenberg et al., 2005). The continuous
dose temozolomide regimen explored by Perry et al.
(2010) was very well-tolerated and produced an overall
progression-free survival (PFS) at 6 months of 23.9%
for patients with GBM and 35.7% for patients with
AA. Among patients with GBM, those who recurred
early had a PFS of 27.3% at 6 months, while those
who experienced disease progression after completion
of adjuvant therapy, and never truly experienced treatment failure with temozolomide, had a PFS6 of 35.7%,
suggesting that the regimen may be most beneficial in
these two subgroups.
Future Therapeutic Drugs

New cytotoxic agents to treat malignant gliomas are
generally evaluated initially in patients with recurrent or progressive disease. Procarbazine, etoposide,
tamoxifen and pegylated liposomal doxorubicin, carboplatin, cyclophosphamide, paclitaxel, and irinotecan
and topotecan have all been investigated and have
shown some limited activity (Warnick et al., 1994;
Fulton et al., 1996; Chamberlain and Kormanik, 1999a,
b; Cloughesy et al., 2003; Macdonald, 2003; Hau et al.,
2004).
An increased understanding in the molecular aberrations underpinning the development and progression
of central nervous tumors has revealed putative targets
for therapy, particularly relevant are inhibitors targeting receptor tyrosine kinases such as epidermal growth
factor receptor (EGFR),the platelet derived growth
factor receptor (PDGFR) and the vascular endothelial growth factor receptors (VEGFR) and related
211
targets, as well as on signal-transduction inhibitors

targeting mammalian target of rapacycin (mTOR), farnesyltransferase and phosphatidyl inositol 3 kinase
PI3K (Wen and Kesari, 2008). However, despite the
valiant efforts of researchers, many targeted agents
tested in malignant gliomas to date have been tested
as monotherapy at the time of recurrence and have
demonstrated only modest activity.
As malignant gliomas are highly vascular tumors,
key molecular targets include those involved in angiogenesis, namely the vascular endothelial growth factor
(VEGF) and its receptor. A humanized monoclonal
antibody against VEGF, bevacizumab has shown efficacy in recurrent GBM and is currently under investigation in the adjuvant setting. A novel pan VEGFR
tyrosine kinase inhibitor (cedirinib) has shown efficacy as documented by functional imaging in patients
with recurrent GBM (Batchelor et al., 2007, 2010).
The epidermal growth factor receptor (EGFR) has been
extensively studied in malignant glioma as its pathway is dysregulated in the majority of patients with
malignant glioma. Studies involving agents that inhibit
this pathway have uniformly revealed disappointing
results. The tyrosine kinase inhibitor erlotinib has been
investigated both as a single agent (van den Bent
et al., 2009) and in combination with chemotherapy
(Brown et al., 2008) and has failed to demonstrate an
improvement in survival in either context. Monoclonal
antibodies directed against the EGFR such as panitumumab are currently under investigation in GBM.
Vorinostat, a histone deacetylase inhibitor is currently
under investigation for patients with recurrent GBM.
This follows a phase II study which met the predetermined criteria to further evaluate vorinostat in this
setting (Galanis et al., 2009). Drugs targeting other
pathways such as mTOR and PI3K are also under
investigation. In addition other novel approaches such
as oncolytic viral therapy are currently being trialled.
Molecular Biomarkers
The impact of unravelling the molecular underpinning
of gliomagenesis is perhaps more evident in terms
of the predictive and prognostic markers which have
been identified, have the potential to change the face
of current management of glioma and have shaped
212
ongoing and future trials. Perhaps most importantly

MGMT promotor hypermethylation has been found to
be a positive molecular genetic prognostic feature in
patients with GBM. It is also possibly predictive of a
favourable response to temozolomide-containing therapy, a key question that is addressed in a large phase
III trialrecently completed by the Radiation Therapy
Oncology Group (RTOG 0525). Until the results of this
trial are available, the clinical utility of this biomarker
remains limited, as MGMT promotor hypermethylation status should not be used to guide therapeutic
decisions.
In anaplastic oligodendroglioma the identification
of 1p19q co-deletion has uncovered a subpopulation
of tumors that grow slowly and are more exquisitely
sensitive to both chemotherapy and radiation than
those tumors not sharing this molecular characteristic
(Walker et al., 2006). Increasingly anaplastic gliomas
are being classified by the presence or absence of this
molecular signature, and 1p19q codeletion is being
used to assist in the histologic diagnosis of these
tumors. Studies are underway to establish therapies
based on 1p19q classification of malignant gliomas.
Mutations of p53 and overexpression, mutation and
amplification of EGFR are common molecular aberrations observed in astrocytic tumors. Unravelling their
prognostic and predictive impact has not been achieved
however and their potential clinical utility remains
unclear (Mason and Cairncross, 2008). Novel, emerging biomarkers may include mutations in IDH1 and
IDH2 which seem to be early molecular events in
the gliomagenesis of particular tumors, and seem to
be associated with a more favourable prognosis, but
this finding requires prospective evaluation (Yan et al.,
2009).
Conclusion
The last decade has revealed dramatic changes in the
treatment of malignant gliomas. Temozolomide is now
the agent of choice in the adjuvant setting and may
also yield favourable results upon rechallenge for most
patients with GBM. Bevacizumab is under investigation in the adjuvant setting for GBM and may be
useful in the patient with relapsed disease. Targeting
the VEGF pathway appears to be a particularly promising strategy, as preliminary trials using the pan-VEGF
receptor tyrosine kinase inhibitor cedirinib have also

demonstrated signs of efficacy in patients with progressive GBM.
As our understanding of the molecular biology of
malignant glioma advances, and as biomarkers such
as MGMT promotor hypermethylation and 1p/19q codeletion undergo prospective validation in large phase
III trials, the possibility of tailoring therapy for patients
with malignant glioma based on molecular genetic
factors approaches reality. Clearly neurooncology has
emerged as an area of intense clinical research for the
evaluation of novel molecularly targeted therapies, and
it is hoped that these efforts will result in improved
therapies and better outcomes for patients afflicted
with these intractable and devastating malignancies.
References
Athanassiou H, Synodinou M, Maragoudakis E, Paraskevaidis
M, Verigos C, Misailidou D, Antonadou D, Saris G,
Beroukas K, Karageorgis P (2005) Randomized phase II
study of temozolomide and radiotherapy compared with
radiotherapy alone in newly diagnosed glioblastoma multiforme. J Clin Oncol 23:23722377
Batchelor TT, Duda DG, di Tomaso E, Ancukiewicz M, Plotkin
SR, Gerstner E, Eichler AF, Drappatz J, Hochberg FH,
Benner T, Louis DN, Cohen KS, Chea H, Exarhopoulos
A, Loeffler JS, Moses MA, Ivy P, Sorensen AG, Wen PY,
Jain RK (2010) Phase II study of cediranib, an oral panvascular endothelial growth factor receptor tyrosine kinase
inhibitor, in patients with recurrent glioblastoma. J Clin
Oncol 28:28172823
Batchelor T, Sorensen AG, Ancukiewicz M, Duda DG, Louis
DN, Plotkin SR, Ivy P, Wen PY, Loeffler JS, Jain R (2007)
A phase II trial of AZD2171 (cediranib), an oral pan-VEGF
receptor tyrosine kinase inhibitor, in patients with recurrent
glioblastoma. J Clin Oncol (Meeting Abstracts) 25:2001
Brandes AA, Tosoni A, Amista P, Nicolardi L, Grosso D, Berti
F, Ermani M (2004) How effective is BCNU in recurrent
glioblastoma in the modern era? A phase II trial. Neurology
63:12811284
Brandes AA, Tosoni A, Cavallo G, Bertorelle R, Gioia V,
Franceschi E, Biscuola M, Blatt V, Crino L, Ermani M
(2006) Temozolomide 3 weeks on and 1 week off as firstline therapy for recurrent glioblastoma: phase II study from
gruppo italiano cooperativo di neuro-oncologia (GICNO). Br
J Cancer 95:11551160
Brown PD, Krishnan S, Sarkaria JN, Wu W, Jaeckle KA, Uhm
JH, Geoffroy FJ, Arusell R, Kitange G, Jenkins RB, Kugler
JW, Morton RF, Rowland KM Jr, Mischel P, Yong WH,
Scheithauer BW, Schiff D, Giannini C, Buckner JC (2008)
Phase I/II trial of erlotinib and temozolomide with radiation
therapy in the treatment of newly diagnosed glioblastoma
multiforme: North Central Cancer Treatment Group Study
N0177. J Clin Oncol 26:56035609

Chamberlain MC, Kormanik P (1999a) Salvage chemotherapy with taxol for recurrent anaplastic astrocytomas. J
Neurooncol 43:7178
Chamberlain MC, Kormanik PA (1999b) Salvage chemotherapy
with tamoxifen for recurrent anaplastic astrocytomas. Arch
Neurol 56:703708
Cloughesy TF, Filka E, Kuhn J, Nelson G, Kabbinavar F,
Friedman H, Miller LL, Elfring GL (2003) Two studies evaluating irinotecan treatment for recurrent malignant glioma
using an every-3-week regimen. Cancer 97:23812386
Fabrini MG, Silvano G, Lolli I, Perrone F, Marsella A, Scotti
V, Cionini L (2009) A multi-institutional phase II study on
second-line Fotemustine chemotherapy in recurrent glioblastoma. J Neurooncol 92:7986
Fine HA, Dear KB, Loeffler JS, Black PM, Canellos GP (1993)
Meta-analysis of radiation therapy with and without adjuvant chemotherapy for malignant gliomas in adults. Cancer
71:25852597
Friedman HS, Prados MD, Wen PY, Mikkelsen T, Schiff D,
Abrey LE, Yung WK, Paleologos N, Nicholas MK, Jensen
R, Vredenburgh J, Huang J, Zheng M, Cloughesy T (2009)
Bevacizumab alone and in combination with irinotecan in
recurrent glioblastoma. J Clin Oncol 27:47334740
Fulton D, Urtasun R, Forsyth P (1996) Phase II study of prolonged oral therapy with etoposide (VP16) for patients with
recurrent malignant glioma. J Neurooncol 27:149155
Galanis E, Jaeckle KA, Maurer MJ, Reid JM, Ames MM,
Hardwick JS, Reilly JF, Loboda A, Nebozhyn M, Fantin VR,
Richon VM, Scheithauer B, Giannini C, Flynn PJ, Moore DF
Jr, Zwiebel J, Buckner JC (2009) Phase II trial of vorinostat
in recurrent glioblastoma multiforme: a north central cancer
treatment group study. J Clin Oncol 27:20522058
Hau P, Fabel K, Baumgart U, Rummele P, Grauer O, Bock
A, Dietmaier C, Dietmaier W, Dietrich J, Dudel C, Hubner
F, Jauch T, Drechsel E, Kleiter I, Wismeth C, Zellner A,
Brawanski A, Steinbrecher A, Marienhagen J, Bogdahn U
(2004) Pegylated liposomal doxorubicin-efficacy in patients
with recurrent high-grade glioma. Cancer 100:11991207
RC, Stupp R (2005) MGMT gene silencing and benefit from
temozolomide in glioblastoma. N Engl J Med 352:9971003.
http://clinicaltrials.gov/ct2/show/NCT00689221
27:740745
Macdonald DR (2003) New frontiers in the treatment of malignant glioma. Semin Oncol 30:7276
Mason WP, Cairncross JG (2008) Invited article: the expanding
impact of molecular biology on the diagnosis and treatment
of gliomas. Neurology 71:365373
Nghiemphu PL, Green RM, Pope WB, Lai A, Cloughesy TF
(2008) Safety of anticoagulation use and bevacizumab in
patients with glioma. Neuro Oncol 10:355360
Perry JR, Belanger K, Mason WP, Fulton D, Kavan P, Easaw J,
Shields C, Kirby S, Macdonald DR, Eisenstat DD, Thiessen
B, Forsyth P, Pouliot JF (2010) Phase II trial of continuous
213
dose-intense temozolomide in recurrent malignant glioma:
RESCUE study. J Clin Oncol 28:20512057
Schmidt F, Fischer J, Herrlinger U, Dietz K, Dichgans J, Weller
M (2006) PCV chemotherapy for recurrent glioblastoma.
Neurology 66:587589
Stenning SP, Freedman LS, Bleehen NM (1987) An overview
of published results from randomized studies of nitrosoureas
in primary high grade malignant glioma. Br J Cancer 56:
8990
Stewart LA (2002) Chemotherapy in adult high-grade glioma:
a systematic review and meta-analysis of individual
patient data from 12 randomised trials. Lancet 359:
10111018
(2009) Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in
glioblastoma in a randomised phase III study: 5-year analysis
of the EORTC-NCIC trial. Lancet Oncol 10:459466
adjuvant temozolomide for glioblastoma. N Engl J Med 352:
987996
Tuettenberg J, Grobholz R, Korn T, Wenz F, Erber R, Vajkoczy P
(2005) Continuous low-dose chemotherapy plus inhibition of
cyclooxygenase-2 as an antiangiogenic therapy of glioblastoma multiforme. J Cancer Res Clin Oncol 131:3140
van den Bent MJ, Brandes AA, Rampling R, Kouwenhoven MC,
Kros JM, Carpentier AF, Clement PM, Frenay M, Campone
M, Baurain JF, Armand JP, Taphoorn MJ, Tosoni A, Kletzl H,
Klughammer B, Lacombe D, Gorlia T (2009) Randomized
phase II trial of erlotinib versus temozolomide or carmustine
in recurrent glioblastoma: EORTC brain tumor group study
26034. J Clin Oncol 27:12681274
Walker MD, Green SB, Byar DP, Alexander E Jr, Batzdorf
U, Brooks WH, Hunt WE, MacCarty CS, Mahaley MS Jr,
Mealey J Jr, Owens G, Ransohoff J 2nd, Robertson JT,
Shapiro WR, Smith KR Jr, Wilson CB, Strike TA (1980)
Randomized comparisons of radiotherapy and nitrosoureas
for the treatment of malignant glioma after surgery. N Engl J
Med 303:13231329
Walker C, Haylock B, Husband D, Joyce KA, Fildes D,
Jenkinson MD, Smith T, Broome J, Kopitzki K, du Plessis
DG, Prosser J, Vinjamuri S, Warnke PC (2006) Genetic and
metabolic predictors of chemosensitivity in oligodendroglial
neoplasms. Br J Cancer 95:14241431
Warnick RE, Prados MD, Mack EE, Chandler KL, Doz F,
Rabbitt JE, Malec MK (1994) A phase II study of intravenous carboplatin for the treatment of recurrent gliomas. J
Neurooncol 19:6974
Wen PY, Kesari S (2008) Malignant gliomas in adults. N Engl J
Med 359:492507
214

Chapter 23
Recurrent Malignant Glioma Patients: Treatment

with Conformal Radiotherapy and Systemic Therapy
Abhirami Hallock and Lauren VanderSpek
Abstract Malignant gliomas have a propensity to

recur despite initial radical treatment. The diagnosis,
therapy, and prognosis of recurrent malignant glioma
will be discussed with an emphasis on fractionated
stereotactic radiotherapy (FSRT) for the reirradiation
of recurrent malignant glioma alone or in combination
with systemic therapy. The challenges of re-irradiation
with a focus on central nervous system radiation tolerances and late effects will be addressed. Studies
conducted between 2000 and 2010 for reirradiation
with fractionated stereotactic radiotherapy (FSRT) +/
systemic therapy are provided in table format as well
as a diagrammatic flowchart to help decision making
in patients with recurrent malignant gliomas.
Keywords Malignant gliomas Fractionated
stereotactic radiotherapy Systemic therapy
Conformal radiotherapy Reirradiation CTV
Introduction
Malignant gliomas are a therapeutic challenge due
to their infiltrative growth and biologic behavior
with a propensity to recur close to their original
location. Therapeutic interventions at the time of
relapse include surgery, systemic therapy, reirradiation,
supportive care, or a combination of interventions,
the choice of which should involve multidisciplinary
A. Hallock ()
Department of Radiation Oncology, London Regional Cancer
Program, University of Western Ontario, E. London, ON,
Canada
e-mail: Abhirami.hallock@londonhospitals.ca
discussion taking into account the clinical status of the

patient and his or her goals of therapy. Previous treatments and response must also be taken into account
in deciding on management at relapse so as to provide patients with a therapy where potential benefits
outweigh potential toxicity.
Surgery may be considered for malignant glioma
at the time of relapse; however the infiltrative nature
of gliomas, proximity to eloquent areas of the brain,
and the risk for surgery-associated side effects should
be weighed against the benefit of surgical intervention (Combs et al., 2007). Systemic therapies have
shown modest benefit when used as a single agent or
in combination; however, the use of systemic therapies alone at relapse may be limited by prior exposure
to chemotherapy during primary treatment or due to
decreased bone marrow reserve (Combs et al., 2007).
In regard to targeted agents, there are some promising results from Friedman et al. (2009) in their study
of 167 patients with recurrent glioblastoma randomized to either Avastin alone or in combination with
Irinotecan. In the Avastin alone and the Avastin plus
Irinotecan groups, the 6-month progression-free survival rates were 42.6 and 50.3%, respectively. The
median overall survival times were 9.2 and 8.7 months,
respectively.
The use of radiotherapy at relapse has advanced
from whole brain techniques, to 3D conformal techniques and subsequently to the use of stereotactic
localization either in single fraction or fractionated regimens with the goal of improving target coverage while
sparing surrounding normal structures. The use of reirradiation in combination with systemic agents seeks
to improve the therapeutic ratio (through either additive or synergistic effects, i.e. radiosensitization) using
chemotherapy or targeted agents.

215
216
Patterns of Failure
Initial radiation volumes for GBM are variable
depending on the amount of peritumoral edema
covered. The Radiation Therapy Oncology Group
(RTOG) has historically included peritumoral edema
within the clinical target volume (CTV) (Chang
et al., 2007).
Arguing in favor of including areas of peritumoral
edema are histopathologic correlative studies demonstrating infiltrating tumor cells within this region
(Halperin et al., 1989). However, studies suggest most
recurrences are confined to the immediate contrast
enhancing area on imaging. For instance, using whole
brain radiotherapy, Hochberg and Pruitt (1980) demonstrated that 80% of GBM recurrences occur within
2 cm of the margin of the primary tumor bed. Oppitz
et al. (1999) studied 79 patients with GBM, with the
macroscopic preoperative tumor radiated with a 2 cm
safety margin, and showed that in 33 out of 34 patients
for which the CT-study showing tumor-recurrence
was available, the recurrence was completely situated
within the original 90%-isodose. Based on these findings, Chang et al. (2007) and Oppitz et al. (1999)
advocate that proximity to the gross tumor volume
(GTV) rather than the presence of peritumoral edema
is the more important factor in predicting the potential site of recurrence. Chang et al. (2007) recommend
an initial CTV defined as the GTV + 2 cm (with
no attempt to include all of the edema volume) and
a planning target volume (PTV) as CTV + 0.5 cm.
They demonstrated in a study of 48 patients treated in
this fashion that all relapses where contained within
this more limited volume and that irradiated volumes
were smaller than comparison plans where edematous
regions were deliberately included.
Patterns of tumor failure localized to regions close
to the original have driven clinical investigations of
radiation dose escalation in an attempt to improve
local control. Dose escalation attempts to doses up to
90 Gy have not improved local control and this remains
the predominant failure pattern (Chan et al., 2002).
External beam conformal radiotherapy plus an implant
has not resulted in improvements in local control or
survival compared to external beam therapy alone
(Laperriere et al., 1998). GBM is therefore known
to display an aggressive local recurrence pattern, but
current imaging modalities are insufficient to determine where a recurrence is most likely to occur within
A. Hallock and L. VanderSpek
the local region (Chang et al., 2007). The contrast

enhancing volume on CT or MRI has been used as the
target for radiation dose escalation and there is emerging evidence that this may not represent the whole
biologically important tumor target. Improved visualization of the microscopic extension of malignant
glioma cells is the goal of new imaging techniques;
biologic imaging may provide complementary information to that of conventional data from CT and MRI
and thus improve the efficacy of treatment while minimizing toxicity by better distinguishing the target from
uninvolved brain (Dhermain et al., 2005).
Biologic imaging that has been evaluated in
previous studies for planning radiation treatment
includes functional magnetic resonance imaging
(fMRI), positron emission tomography (PET), 11Cmethionine-PET (MET-PET), magnetic resonance
spectroscopy (MRSI), and single photon computed
emission tomography (SPECT) combined with
CT/MRI. Preliminary studies have demonstrated discordance between anatomically defined (T1 contrast
enhanced or T2 MRI) and biologically defined (MRSI
or PET/SPECT) volumes, however, it is not clear that
patterns of failure correlate between with biologic
versus anatomically defined targets. For this reason,
currently the use of biologic imaging for radiation
planning in malignant glioma is investigational, but
highlights the need for more accurate determination of
the target to enable more efficacious treatment design.
For example, in the setting of recurrent malignant
glioma, Grosu et al. (2005) demonstrated a statistically
significant longer survival time in the univariate analysis for patients reirradiated using (11)C-methionine
positron emission tomography (MET-PET) or (123)Ialpha-methyl-tyrosine (IMT)-SPECT/CT/MRI image
fusion in the treatment planning compared to patients
treated based on MRI/CT alone. The study was
small, including 44 patients (1 patient anaplastic
oligodendroglioma (AO), 8 with AA, 33 with GBM,
and 2 with gliosarcoma) after previous surgery and
post-operative conventional radiotherapy with or
without chemotherapy. Median survival times were
11 months for patients who received fractionated
stereotactic radiotherapy (SFRT) and temozolomide
compared to 6 months for patients treated with SFRT
without biologic imaging, without temozolomide, or
without both (p = 0.008) (Grosu et al., 2005).
Imaging advances are also important is in the
diagnosis of recurrent malignant glioma as MRI
interpretation can be challenging in distinguishing
23 Recurrent Malignant Glioma Patients
between recurrent tumor and treatment effects, both of

which may appear as new or worsened areas of contrast
enhancement (Chang et al., 2006). Magnetic resonance spectroscopy determines the biology of regions
of abnormality on MRI images by measuring the
ratio of tissue metabolites such as N-acetylaspartate,
creatine, choline, and lactate (Chang et al., 2006).
An increase of the choline peak and decrease of
N-acetylaspartate indicates tumor recurrence, however,
there have been some reports of transient increased
choline and decreased N-acetylaspartate after radiotherapy especially in areas exposed to lower doses
(Peca et al., 2009). Cerebral perfusion imaging can
demonstrate increased cerebral blood volume in recurrent tumor to distinguish it from normal brain tissue
(Chang et al., 2006). Diffusion-weighted MR imaging may also be helpful in distinguishing treatment
effects from recurrent tumors by measuring differences in apparent diffusion coefficient (Chang et al.,
2006). Similarly, PET and SPECT imaging may provide complementary information to help distinguish
necrosis from active tumor (Gmez-Rio et al., 2008).
If the diagnosis of recurrent malignant glioma remains
uncertain then surgical biopsy (or serial imaging) may
be required prior to treatment (Chang et al., 2006).
Radiation Response of the Brain to Initial

Radiation
The brain is a slowly proliferating and late reacting
tissue. The /ratio is thought to be approximately
22.9 Gy suggesting a low repair capacity (Mayer
and Sminia, 2009; Yaacov et al., 2010). Concerns of
late effects like radiation necrosis, diffuse white matter injury and/or leukoencephalopathy have been the
limiting factor for radiation treatments. The significant
reluctance to undertake reirradiation stems from the
potential for cumulative late toxicity and lack of clinical data to base estimates of recovery of tolerance after
initial treatment.
Much of our understanding of late radiation effects
on the brain come from limited retrospective clinical
studies and extrapolated data from animal studies of
radiation myelopathy. Such data often reflects older
radiation techniques (larger treatment volumes such as
whole brain radiotherapy, lack of conformal radiotherapy techniques, non-conventional treatment schedules
such as treating one field per day, larger dose per
217
fraction, etc). Both clinical and laboratory studies have

shown that the CNS toxicity is a function of total
dose delivered, dose per fractionation, volume treated
and interval between fractions. The pathophysiology
of CNS radiation late injury is poorly understood.
There is consensus that it is likely secondary to both
neuro-glial damage and secondary vascular changes.
The abnormal production and release of cytokines by
astrocytes or microglia as normal responses to injury,
contribute to CNS damage.
Late Effects
Focal radiation necrosis is often limited to the white
matter, with a latency period of 6 months to 2
years, its presentation mimicking tumor recurrence,
and symptoms dependant on the neuro-anatomic location. Histopathological analyses of these areas show
focal coagulative necrosis and demyelination likely
secondary to fibrinoid necrosis of small arterial vessels
(Schultheiss et al., 1995). Imaging changes associated
with treatment necrosis make it difficult to distinguish
it from tumor recurrence, or tumor necrosis. After radiation, irregular enhancement with diffuse edema and
low density white matter changes is seen within the
high dose volume regions of treated brain on CT scans,
and greater extent of similar white matter changes and
edema can be seen on MRI. MR spectroscopy and
PET imaging may be useful in deciphering the changes
seen on conventional CT and MRI imaging. There is
evidence that there is little or no radiation necrosis
below 60 Gy delivered in 2 Gy per fractions. For fraction sizes larger than 2.22.5 Gy or doses >60 Gy the
risk of radiation related necrosis increases (Schultheiss
et al., 1995). Yaacov et al. (2010) found a 5% risk of
necrosis for at BED of 120 Gy (100140 Gy) and a
10% risk at BED of 150 Gy (140170 Gy) for fraction
sizes < 2.5 Gy (equivalent to 72 and 89 Gy in 2 Gy
per day fractions respectively, /= 3). This risk rises
significantly for twice daily fractionation and becomes
unpredictable for fractionation >2.5 Gy. Based on their
analysis, they conclude dose, fraction size, and treatment volume are major players in the risk of necrosis,
while location, use of chemotherapy, and, diabetes
mellitus are less well established risk factors. They
estimate a median time of 12 years between initial
radiation and development of radiation necrosis.
218
The development of diffuse white matter injury is

also a potential part of the late effect syndrome and
has been documented after both whole brain radiation and focal radiotherapy. For the latter, the changes
can extend beyond the peripheral of the high dose
volume. They found resulting neurological sequelae
ranging from mild lethargy to progressive memory loss
leading to dementia. Histopathologically the damaged
white matter shows reactive astrocytosis and white
matter pallor. White matter injury resulting in cognitive changes seems to be related to volume, dose and
fraction size.
Leukoencephalopathy describes diffuse white matter injury in the setting of chemotherapy with or
without radiation. It is noted in adults beyond 1218
months and has a spectrum of symptoms with lethargy,
dysarthia to progressive severe CNS damage resulting in ataxia, memory loss, and dementia. Imaging
changes are similar to those described above and differ
with late development of calcifications mostly limited
to the basal ganglia and grey-white matter interface.
Histopathology analysis show diffuse gliosis and multiple non-inflammatory foci of necrosis (Schultheiss
et al., 1995).
Late cerebrovascular effect causing stroke like
deficits as a result of obliteration of vessels is of less
concern in patients with malignant gliomas as these
side effects are experienced by those who experience long survivorship, more commonly documented
in children and for those receiving treatments for base
of skull tumors.
Recovery of Occult Injury and

Re-irradiation
The recovery of occult radiation injury is important
in understanding the safety of retreatments and determination of re-irradiation dose. Rodent studies have
shown that the spinal cord is able to recover significant and that the extent of repair is dependant on the
initial dose received and time interval to retreatment
(Bauman et al., 1996). Studies on monkeys suggest that
most of the occult injury induced by an initial treatment is recovered within 2 years; however, there is
variability in the rate of recovery (Schultheiss et al.,
1995). In this study, the time to late toxicity was similar in both the retreatment and single treatment group.
However, a retrospective clinical experience of 300

cases of radiation myelopathy from 76 papers showed
that latency period for development of toxicity was
reduced for the lumbar spine and with increasing doses
of radiation. For the brain, given the lack of similar
studies, an estimate of 4050% of the previous radiation dose after 12 years maybe accounted for in
retreatments (Bauman et al., 1996). It is important to
recognize the inadequate understanding of impact of
high dose per fractions in setting of retreatments on the
brain.
Both a recent literature review by Mayer et al.
(2009) and Yaacov et al. (2010) highlight the limitations of the present evidence evaluating the sequelae of
reirradiation. Most clinical data is derived from single
institution studies that often lack details in target volumes, spatial factors or time between radiation treatments. Reporting toxicity as only a crude rate (ratios
to number of patients treated) instead of an actuarial
risk (ratios to number of survivors at risk) can result
in underestimation of true risks (Yaacov et al., 2010).
These limitations acknowledged, Mayer et al. (2009)
evaluated 21 studies using techniques of radiosurgery
and FSRT and conventionally fractionated radiotherapy. They concluded that radiation can be given
safely using different techniques with necrosis becoming evident at (normalized total doses) NTDcumulative
>100 Gy. The NTDcumulative is the total dose delivered
in 2 Gy per fraction using the linear quadratic formulation (accounting for differences in fraction size and
total dose) to sum radiation courses. In their review, the
major factor contributing to risk of necrosis was total
dose and the time interval to retreatment (minimum
interval of 3 months) did not significantly correlate
to risk of necrosis. Previous published reports on salvage therapies including those for radiosurgery and
brachytherapy estimate a necrosis of 1020% (Bauman
et al., 1996) and ranging from 4 to 40% for SRT
or 3DCRT (Butowki et al., 2006). The timelines of
progression to these symptoms in the setting of reirradiation is not well established particularly for those
patients retreated for malignant glioma where competing risk of tumor progression may underestimate the
potential for long term toxicity. As systemic therapies
combined with focal therapies become a more common component of salvage regimes; it is important to
gain an understanding of the impact of both cytotoxic
chemotherapy and targeted agents on both acute and
late toxicities in the setting of retreatments.
Re-irradiation of Malignant Glioma

There is no consensus on the general management
of recurrent malignant gliomas or re-irradiation for
recurrence. No comparative studies of the salvage
options of surgery, radiation, or systemic therapies
are available in the literature. Feasibility of salvage
option (i.e. resectability and operability when considering reoperation), expected therapeutic response,
and potential for cumulative toxicity should be guiding principles in the management of patients with
recurrent malignant gliomas. Figure 23.1 provides a
suggested approach. Conventional, fractionated stereotactic radiotherapy, brachytherapy, radiosurgery (Linac
based, GammaKnife, and Cyberknife) have been utilized in the salvage setting; the highly focal techniques
allow the possibility of maximizing therapeutic efficacy, tolerance, and convenience. Radiosurgery for
progressive malignant glioma is the subject of another
chapter.
219
For this section of the chapter, a literature search for

re-irradiation studies in recurrent malignant gliomas
between 2000 and 2010 was conducted. Studies with
less than ten patients, involving brachytherapy, radiosurgery, or inadequate information about treatment
regimes were excluded. Tables 23.2, 23.3 and 23.4
lists these retrospective, prospective and phase I trials and are provided as a guide to the doses, time
intervals and regimes used and for an appreciation
of tolerances of normal brain. It is important to reemphasize the challenges in making comparison and
caution against making overarching general conclusions based on these studies. These studies span eras
that have seen significant advancements in technology
such as improved neuro-imaging and neurosurgical
techniques, have inherent issues of selection biases,
and differing institution treatment practices. Also,
there is no gold standard for imaging changes to diagnose true normal tissue necrosis; many of the studies
did not grade their necrosis or report on re-operation
rates or confirm suspected necrosis with biopsy.
Fig. 23.1 A suggested approach to the management of recurrent malignant glioma
220
Figure 23.2 shows these studies in a temporal relationship, with median overall survival from time of
retreatment.
calculation for specific patient scenarios, the following

formulas have been provided (Table 23.1):
NTDcumulative = BEDcumulative /2;
3D Conformal/Stereotactic Radiotherapy
As stated earlier in this chapter, the customization of
salvage radiation treatment regime including radiation
technique, total dose, and dose per fraction for individual patients should account for initial dose with first
treatment, time interval to retreatment, new treatment
volume, and when appropriate consider placement in
a clinical trial. As suggested by Mayer and Sminia
(2009), attention should be paid to the NTDcumulative .
They found that both NTDcumulative and total retreatment dose can be increased when using conformal
techniques or radiosurgery, without added risk of
necrosis.
Based on their review, the range of NTDcumulative
for the conventional technique re-irradiation was 81.6
101.9 Gy; for FSRT was 90133.9 Gy. To aid in
and
BEDcumulative = BEDinitial + BEDreirradiation
BED = nd(1 + d/[/])
D = dose/fraction (Gy); n = number of fractions;
/ = repair capacity of tissue.
Table 23.2 includes studies that have evaluated
treatments of recurrence with different radiation techniques and suggest that re-irradiation can be delivered
safely by various techniques. There is significant heterogeneity in treatments that patients received within
institutions in terms of prior chemotherapy regimes,
resections, etc. The increased exploration of FSRT
started after some early radiosurgery studies suggested
high rates of reoperation for necrosis. FSRT allows for
22
Fig. 23.2 Scatter plot of

re-irradiation studies in
recurrent malignant gliomas.
Median survival is from time
of retreatment. XRT
(conventional radiation),
FSRT (fractionated
stereotactic radiotherapy),
CTX (chemotherapy), TG
(targeted therapy)
Median Overall Survival
20
18
XRT
XRT+CTX
FSRT
FSRT+CTX
FSRT+TG
16
14
12
10
8
6
2000
Table 23.1 BED for examples of common radiation regimes

Dose/fraction BED/=10
BED/=2
Dose (Gy)
Fractions
(Gy)
(Gy)
(Gy)
2005
Year of Study
NTD (Gy)
2010
Comments
60
30
2
72
120
60
Standard initial treatment
35
7
5
53
122.5
61
Hypofractionated Re-XRT
20
1
20
60
220
110
Radiosurgerya
a The linear quadratic model was derived for fractionated radiotherapy. Its reliability for small number of fractions or
single fractions is uncertain. There are models suggested for radiosurgery but these do not account for all possible
variables, and are not well validated.
4561 Gy/1.82 Gy
daily OR 45 Gy/3 Gy
daily OR
54.8 Gy/1.8 Gy
bid+CT (94%)
60 Gy/2 Gy per day/5
days per week +CT
(16 patients)
59.4 Gy/2 Gy per day/5
days per week +CT
(16 patients)
4560 Gy/Dose/fx: NA
5060 Gy/2 Gy per

day/5 days per week

days per week

days per week
19
33
11
15
NA
12
7.9
15.4 (AA)
6.9
Acute edema (1)/necrosis

(2): clinical necr +
cognitive decline (1) at
62 months/all late
toxicity at BED>
204 Gy. Dose >40 Gy
major risk factor.
Necrosis(2)
Mixed (tumor +radiation)
necrosis (1)
GBM(51)
10
49
8
Necrosis(0)
GS(1)
days per week
GCG(1)
Combs et al. (2005a)
AA(22)
34
56.2
16
Necrosis (0)
AO(10)
days per week
AOA(8)
Ernst-Stecken et al.
GBM(11)
10
22.4
12
Necrosis (0)
(2007)
AA(4)
days per week
AA (anaplastic astrocytoma), AO (anaplastic oligodendroglioma), AOA (Anaplastic oligoastrocytoma), AE (Anaplastic Ependymoma), GBM (glioblastoma multiforme),
GS (gliosarcoma), GCG (giant cell glioblastoma), LG (low grade), DE (Dose escalation), NA (not available), CT (chemotherapy), fx (fraction).
a Median interval is from the time from the first radiation treatment.
b Median survival is from the time of retreatment.
Vordermark et al.
(2005)
Combs et al. (2005b)
GBM(14)
AA(5)
Veninga et al. (2001)
20-36 Gy/46 Gy per

day/5 days per week
Necrosis(0)

days per week
Selch et al. (2000)
GBM(15)
AOA(3)
LG(3)
Unk(1)
AA(29)
Graded
toxicity(n)/necrosis
(n)/comments
Table 23.2 Re-irradiation with conventional radiotherapy, fractionated stereotactic radiotherapy, or LINAC based stereotactic radiosurgery
First treatment radiation
Median
Median
Median
(median or range)/dose/
Median re-radiation/
volume
intervala
survivalb
Author
Histology (n)
(months)
(months)
schedule
dose/schedule
treated (cc)

221
222
delivery for high dose per fraction treatment while taking radiobiological advantage for normal tissue repair
and maintaining short overall treatment times. Cho
et al. (1999) compared FSRT to SRS and found lower
complication rates with FSRT without impact on survival rates. Other studies of FSRT given alone or
in conjunction with systemic therapies (Table 23.3)
support the results of Cho et al. (1999). Given high progression post retreatments, newer trials are exploring
the role of dose escalation and enhancing radiosensitization.
Stereotactic Radiotherapy with Systemic

Therapies
Brain tumors provide unique challenges to treatment
with systemic therapies alone with issues of diffuse
infiltrative disease, blood brain barrier, and rapid evolution of resistance.
There is only modest benefit to conventional
chemotherapy in treatment for recurrent malignant
gliomas, with greater benefit observed in patients with
grade three gliomas.
Chemotherapeutic agents used for recurrent gliomas
include temozolomide, nitrosourea, carboplatin, procarbazine, irinotecan, etoposide and carmustine wafers
(Wen and Kesari, 2008).
Concurrent chemoradiation is becoming increasing
common as evidenced by its use in radical treatments
of gynecologic, gastrointestinal, head & neck, and
thoracic malignancies. The addition of chemotherapy
to repeat radiation treatment of recurrent malignant
glioma is another treatment strategy and has been
described in the literature for several chemotherapeutic
agents (Table 23.3).
Chemotherapy may interact with radiation through
several mechanisms: spatial cooperation, additive or
synergistic cell killing or non-overlapping toxicity. In
the retreatment of glioma, chemotherapy may potentially address microscopic tumor outside the radiated volume. Improved tumor cell kill within the
treated volume may result from the additive effects of
chemotherapy and radiation and allow for treatment
intensification without the increased risk of necrosis
seen when higher dose re-irradiation is used as
a single modality due to non-overlapping toxicity.
Synergistic effects can include alteration of cell cycle
by chemotherapy resulting in greater number of cells

within the radiosensitive G2/M phase, inhibition of
radiation repair and conversion of sub-lethal single
strand radiation injury to lethal double strand breaks.
The role of Cisplatin (CDDP), Paclitaxel,
Lomustine, Topotecan, Temozolomide and hyperbaric oxygen with radiation in recurrent gliomas have
been reported (Table 23.3).
Cisplatin (CDDP) is used as a radiosensitizer in several tumor types (anal, cervical, head and neck, bladder), but has shown only limited benefit in the setting
of re-irradiation for malignant glioma (VanderSpek
et al., 2008; Glass et al., 1997). Topotecan has been
shown to be efficacious in combination with radiation,
both because of its synergistic effect and substantial
CNS penetration.
Temozolomide is an oral alkylating agent that readily crosses the blood brain barrier that has had the
greatest impact on treatment of GBM (Stupp et al.,
2005), and its efficacy was first found in recurrent
gliomas. Continuous administration of this temozolomide appears to "deplete" the DNA repair enzyme
MGMT (O6 methylguanine-DNA methyltransferase),
which is increased during radiotherapy (Stupp et al.,
2005).
Despite these biologic correlations, clinical studies
of combined modality treatment for recurrent glioma
suggest limited benefit. While the morbidity experienced with combined use of systemic therapy and
re-irradiation seems acceptable there do not appear
to be any apparent gains in disease control or outcomes compared to re-irradiation alone (Table 23.3,
Fig. 23.2).
Given the limited benefit of conventional systemic
therapies, there is great interest in the role of targeted molecular agents and antiangiogenic agents as
our understanding of the molecular pathogenesis of
gliomas increases. For malignant gliomas, agents that
target Endothelial Growth Factor Receptor (EGFR),
Platelet Growth Factor Receptor (PDGFR), Vascular
Growth Factor Receptor (VEGFR), mammalian target
of rapamycin (mTOR), farnesyltransferase, and P13k
are of interest (Wen and Kesari, 2008).
Malignant gliomas are known to be highly vascular
tumors, expressing VEGF as an important angiogenesis factor and have some response to older, likely
less potent antiangiogenic agents like Thalidomide. It
is hypothesized that the normalization of abnormal
blood vessels with anti-angiogenic therapy targeting
GBM (5)
AA (7)
AO(2)
HG (25)
60 Gy/1.83 Gy 30 Gy/5 Gy per

per day/5 days
day/5 days per
per week +
week
chemo (20 pts)
NA
20-30 Gy (2 1.2
Gy per day)
60 Gy/2 Gy per
30 Gy/2 Gy per
day/5 days per
day/5 days per
week
week
Median radiation/
dose/schedule
TMZ
(66%)
TMZ
TMZ
Chemotherapy
16
14
NA
Median
intervala
(months)
11 (received
functional
imaging)
163 (PTV)
NA
Median volume
treated (cc)
20
7.5
9.3
Median
survivalb
(months)
No G 3 or 4
toxicities/necrosis(0)
Grade 3 hematologic (1)
Cephalgia(1)
Mental degradation
(1,query tumor
progression)
Late toxicity: NA
No acute >grade 3 neuro
toxicity; hematological
toxicity:NA
Mixed necrosis (3)
Improved survival with
PET (SPECT /CT/
MRI compared to CT/
MRI alone; 9 vs.
5 months)
84% previous
chemotherapy for
recurrence
Graded toxicity/
necrosis/comments
60 Gy/2 Gy per
2530 Gy/56 Gy
Topotecan
NA
14.5
16.5
day/5 days per
per day
week OR
54.4 Gy/1.6 Gy
twice daily/3.5
weeks
Combs et al.
GBM(8)
60 Gy/2 Gy per
36 Gy/2 Gy per
TMZ
36
50
8
No severe toxicities
(2008)
HG(10)
day/5 days per
day/5 days per
(PTV)
Necrosis(0)
week
week
GBM (glioblastoma multiforme), AA (anaplastic astrocytoma), AO (anaplastic oligodendroglioma), AA (anaplastic astrocytoma), AO (anaplastic oligodendroglioma),
AOA (Anaplastic oligoastrocytoma), GS(gliosarcoma), GCG (giant cell glioblastoma), HG (high grade), LG (low grade), TMZ (Temozolomide), NA (not available)
a Median Interval is from time from the first radiation treatment.
b Median Survival is from the time of retreatment.
Wurm et al. (2006) GBM(18)

GCG(2)
AA(3)
AOA(2)
Grosu et al. (2005) GBM(33)

GS(2)
AA(8)
AO(1)
Schonekaes et al.
(2002)
Schafer et al.
(2004)
Table 23.3 Reirradiation and chemotherapy

First treatment r
radiation
(median or
range)/Dose/
Author
Histology (n)
schedule

223
224
either VEGF or VEGFR will increase the therapeutic benefit of radiation and chemotherapies. While it
is thought that glioma stem cells produce VEGF, use
of these single agent VEGF inhibiters have shown
have limited benefit (Wen and Kesari, 2008). This
has prompted evaluation of these agents in combination with chemotherapy. Work from Friedman et al.
(2009) has resulted in the approval of bevacizumab
in combination with irinotecan for use in recurrent
GBM.
Given these findings, potential role of targeted therapies with radiation is exciting especially for brain
tumors, but we are still in the preliminary stages of our
understanding. There is a suggestion that promotion of
angiogenesis may occur during radiation, implying the
potential benefit of combining anti-angiogenic agents
during radiotherapy. In vivo and vitro studies show that
exposure to radiation induces VEGF expression in several tumor cell lines and that an anti-VEGF treatment
potentiates radiation induced tumor kill (Gorski et al.,
1999).
Blocking of VEGF-receptor 2 potentiates radiation
induced tumor control in vivo human tumor xenograft
of glioblastoma multiforme (U87) (Gutin et al., 2009).
Gutin et al. (2009) studied the benefits of combination
of 7 cycles of bevacizumab with FSRT in recurrent
gliomas and reported a 6 months PFS of 65%, with
50% overall response rate. However, three out of 25
patients discontinued treatments due to grade 3 toxicities of intratumoral hemorrhage, wound dehiscence
and bowel perforation (but there was no noted necrosis). Evaluation of responses of anti-angiogenic therapies may be complicated by the decreased enhancement on imaging seen with these drugs, independent
of their anti-tumor effect. Thus trials evaluating these
agents should include survival endpoints (such as 6
month survival) as a complement to radiographic endpoints. Another interesting advantage of combining
anti-angiogenic agents is the potential reduction in
necrosis suggesting that a potential protective role for
this agent in combination with re-irradiation, perhaps
sequentially administered given the toxicity seen with
concurrent administration in the series by Gutin et al.
(2009).
Endothelial Growth Factor Receptors (EGFR) are
involved in pathways known to cell differentiation,
angiogenesis, metastatic spread, resistance to apoptotic death and thought to be expressed highly in cells
resistant to radiation and chemotherapy. The majority
of glioblastomas show a deregulation as a result of

mutations, overexpressions and gene amplifications
in the molecular pathway involving EGFR. Small
molecule inhibitors targeted against the EGFR tyrosine kinase pathways or the EGFR receptor has been
evaluated in glioblastoma. For example, the epidermal
growth factor receptor tyrosine kinase inhibitor (EGFR
TKI), Gefitinib (ZD1839 or Iressa), has demonstrated
a median survival for patients with recurrent GBM
or 39.4 weeks when used as a single agent after first
relapse (Schwer et al., 2008). Combination of the
EGFR TKI with radiation can have potentiating effects
through greater number of cells in the radiosensitive
phase of the cell cycle; greater radiation induced apoptosis and decreased DNA repair mechanisms within the
cancer cells.
The addition of molecular targeted therapies with
re-irradiation provides an exciting new therapeutic
option for patients with recurrent malignant gliomas.
There are too few studies to make any meaningful conclusions and even these studies report outcomes comparable other studies for salvage therapies (Table 23.4).
At present, there is no clear evidence for the standard use of targeted therapies with re-irradiation. This
suggests the need for further exploration into their
use and identifying patients who may benefit such
therapies. For example, EGFR TKI agents are small
molecules and should have a theoretical benefit in
terms of intracranial delivery, however, only 10-20%
of GBM respond to these targeted agents. Mellinghoff
et al. (2005) showed that the sensitivity of GBMs to
these EFGR inhibitors was based on co-expression of
EGFRvIII and PTEN, implying that therapeutic benefit
may not be for all patients. There are currently phase
2 trials evaluating multi tyrosine kinase inhibitors
underway to determine if therapeutic outcomes can
be improved by targeting multiple molecular abnormalities. Consideration of enrolment of patients in
trials evaluating molecular targeted therapies with
radiation is recommended to help establish clear
benefits.
Fractionated Stereotactic Reirradiation

Techniques
The radiation planning volumes for FSRT used in
reirradiation of malignant glioma need to take into
account the proximity of the lesion to eloquent areas
GBM(11)
AA(4)
60 Gy/1.82 Gy
per day/5 days
per week + TMZ
1836 Gy/612 Gy
per day/
consecutive days
Median
radiation/dose/
schedule
Gefitinib
(Iressa)
Target therapy
12
Median
intervala
(months)
41 (PTV)
Median
volume
treated (cc)
GBM(20)
59.4 Gy
30 Gy/6 Gy per
Bevacizumab
15
34
AA(4)
Fractionation: NA
day/5 days per
(Avastin)
AO(1)
week
GBM (glioblastoma multiforme), AA (anaplastic astrocytoma), AO (anaplastic oligodendroglioma), AO (anaplastic oligodendroglioma)
a Median Interval is from the time of first radiation treatment.
b Median Survival is from the time of retreatment.
Gutin et al.
(2009)
Schwer et al.
(2008)
Table 23.4 Re-irradiation with molecular targeted agents

First treatment
radiation (median
or range)/dose/
Author
Histology (n)
schedule
12.5
10
Median
survivalb
(months)
Necrosis (2): Grade 1;

BCNU wafers
given postoperative
at recurrence for 5
patients
Grade 3 CNS toxicity
Necrosis(0)
Graded toxicity/
radiation(n)/comments

225
226
of the brain and location of the recurrence in relation to the previously radiated volume, as well as the
method of immobilization, treatment planning system,
and delivery techniques available at each institution
which influence the accuracy of treatment delivery
(Bauman et al., 2006).
The typical planning process involves fabrication
of any immobilization devices followed by acquisition of a volumetric computed tomotherapy (CT) scan
with the patient in the immobilization device for radiation treatment planning purposes (called CT simulation). CT simulation scans are typically acquired using
3 mm slices thickness and with intravenous contrast
administration for optimal delineation of the target.
Typically a stereotactic image fusion is performed
between the planning CT and diagnostic imaging (MRI
or PET/SPECT). Most commonly the diagnostic highresolution T1-weighted MRI with Gadolinium contrast
is used to assist in target volume delineation. The
tumor visible on imaging is defined as the Gross Tumor
Volume (GTV) and serves as the target for radiation
treatment planning. For reirradiation, delineation of a
region of risk (clinical target volume, CTV) surrounding the GTV (as usually performed for initial treatment
to cover regions of microscopic infiltration) is omitted
in order to minimize the retreated volume.
The planning target volume (PTV) is intended to
account for geometric errors in setup and is dependent
on the immobilization system used. For example for
traditional radiosurgery systems, invasive stereotactic
frame systems with submillimeter localization accuracy are used, justifying the omission of a PTV margin.
For relocatable frame systems, selection of a PTV margin is dependent on the system used and institutional
practice. For instance, Hudes et al. (1999) defined the
target volume as the contrast-enhanced tumor edge
(GTV) without a PTV margin for patients immobilized
in the GTC (Gill-Thomas-Cosman) frame. Others utilize a margin around the GTV, such as Schwer et al.
(2008) where the PTV was defined as the GTV + 2 mm
as they used a technique with patients immobilized
using a custom-molded removable head frame and had
documented reproducibility within 2 mm. Overall the
choice of a PTV margin depends on the measured
accuracy of the overall radiation delivery at each institution as well as the proximity to organs at risk and
clinical judgment (Bauman et al., 2006).
Acknowledgements The authors would like to thank Dr. Glenn
Bauman, London Regional Cancer Program, University of

Western Ontario, for his constructive feedback during the development of the chapter.
References
Bauman GS, Sneed PK, Wara WM, Staplers LJA, Chang SM,
McDermott MW, Gutin PH, Larson DA (1996) Reirradiation
of primary CNS tumors. Int J Radiat Oncol Biol Phys
36:433441
Bauman G, Wong E, McDermott MW (2006) Fractionated
radiotherapy techniques. Neurosurg Clin N Am 17:99110,
v, 20
Butowski NA, Sneed PK, Chang SM (2006) Diagnosis and treatment of recurrent high-grade astrocytoma. J Clin Oncol 24:
12731280
Chan JL, Lee SW, Fraass BA, Normolle DP, Greenberg HS,
Junck LR, Gebarski SS, Sandler HM (2002) Survival
and failure patterns of high-grade gliomas after threedimensional conformal radiotherapy. J Clin Oncol 20:
16351642
Chang EL, Akyurek S, Avalos T, Rebueno N, Spicer C,
Garcia J, Famiglietti R, Allen PK, Chao C, Mahajan A, Woo
SY, Maor MH (2007) Evaluation of peritumoral edema in
the delineation of radiotherapy clinical target volumes for
glioblastoma. Int J Radiat Oncol Biol Phys 68:144250
Chang SM, Butowski NA, Sneed PK, Garner IV (2006) Standard
treatment and experimental targeted drug therapy for recurrent glioblastoma multiforme. Neurosurg Focus 20:E4
Cho KH, Hall WA, Gerbi BJ, Higgins PD, McGuire WA, Clark
HB (1999) Single dose versus fractionated stereotactic radiotherapy for recurrent high-grade gliomas. Int J Radiat Oncol
Biol Phys 45:11331141
Combs SE, Bischof M, Welzel T, Oertel S, Debus J, SchulzErtner D (2008) Radiochemotherapy with temozolomide as
re-irradiation using high precision fractionated stereotactic
radiotherapy (FSRT) in patients with recurrent gliomas. J
Combs SE, Debus J, Schulz-Ertner D (2007) Radiotherapeutic
alternatives for previously irradiated recurrent gliomas. BMC
Cancer 7:167177
Combs SE, Gutwein S, Thilmann C, Debus J, Schulz-Ertner D
(2005a) Reirradiation of recurrent WHO grade III astrocytomas using fractionated stereotactic radiotherapy (FSRT).
Strahlenther Onkol 181:768773
Combs SE, Gutwein S, Thilmann C, Huber P, Debus J,
Schulz-Ertner D (2005b) Stereotactically guided fractionated re-irradiation in recurrent glioblastoma multiforme. J
Dhermain F, Ducreux D, Bidault F, Bruna A, Parker F,
Roujeau T, Beaudre A, Armand JP, Haie-Meder C (2005)
Use of functional imaging modalities in radiation therapy
treatment planning in patients with glioblastoma. Bull
Cancer 92:333342
Ernst-Stecken A, Ganslandt O, Lambrecht U, Sauer R,
Grabenbauer G (2007) Survival and quality of life after
hypofractionated stereotactic radiotherapy for recurrent
malignant glioma. J Neurooncol 81:287294
Friedman HS, Prados MD, Wen PY, Mikkelsen T, Schiff
D, Abrey LE, Yung WK, Paleologos N, Nicholas MK,

Jensen R, Vredenburgh J, Huang J, Zheng M, Cloughesy
T (2009) Bevacizumab alone and in combination with
irinotecan in recurrent glioblastoma. J Clin Oncol 27:
47334740
Glass J, Silverman C, Axelrod R, Corn B, Andrews D (1997)
Fractionated stereotactic radiotherapy with cis-platinum
radiosensitization in the treatment of recurrent, progressive, or persistent malignant astrocytoma. Am J Clin Oncol
20:226229
Gmez-Rio M, Rodriguez-Fernez A, Ramos-Font C, LpezRamirez E, Llamas-Elvira JM (2008) Diagnostic accuracy
of 201 Thallium-SPECT and 18 F-FDG-PET in the clinical
assessment of glioma recurrence. Eur J Nucl Med Mol
Imaging 35:966975
Gorski DH, Beckett MA, Jaskowiak. NT, Calvin DP, Mauceri
HJ, Salloum RM, Seetharam S, Koons A, Hari DM, Kufe
DW et al (1999) Blockage of the vascular endothelial growth
factor stress response increases the antitumor effects of
ionizing radiation. Cancer Res 59:33743378
Grosu AL, Weber WA, Franz M, Stark S, Piert M, Thamm
R, Gumprecht H, Schwaiger M, Molls M, Nieder C (2005)
Reirradiation of recurrent high-grade gliomas using amino
acid PET (SPECT)/CT/MRI image fusion to determine gross
tumor volume for stereotactic fractionated radiotherapy. Int J
Gutin PH, Iwamoto FM, Beal K, Mohile NA, Karimi S, Hou BL,
Lymberis S, Yamada Y, Chang J, Abrey LE (2009) Safety
and efficacy of bevacizumab with hypofractionated stereotactic irradiation for recurrent malignant gliomas. Int J Radiat
Oncol Biol Phys 75:156163
Halperin EC, Bentel G, Heinz ER, Burger PC (1989) Radiation
therapy treatment planning in supratentorial glioblastoma
multiforme: an analysis based on post mortem topographic
anatomy with CT correlations. Int J Radiat Oncol Biol Phys
17:13471350
Hochberg FH, Pruitt A (1980) Assumptions in the radiotherapy
of glioblastoma. Neurology 30:907911
Hudes RS, Corn BW, Werner-Wasik M, Andrews D, Rosenstock
J, Thoron L, Downes B, Curran WJ Jr (1999) A phase I dose
escalation study of hypofractionated stereotactic radiotherapy as salvage therapy for persistent or recurrent malignant
glioma. Int J Radiat Oncol Biol Phys 43:293298
Laperriere NJ, Leung PMK, McKenzie S (1998) Randomized
study of brachytherapy in the initial management of patients
with malignant astrocytoma. Int J Radiat Oncol Biol Phys
41:10051011
Mayer R, Sminia P (2009) Reirradiation tolerance of the human
brain. Int J Radiat Oncol Biol Phys 70:1350-1360
Mellinghoff IK, Wang MY, Vivanco I, Haas-Kogan DA, Zhu
S, Dia EQ, Lu KV, Yoshimoto K, Huang JH, Chute DJ,
Riggs BL, Horvath S, Liau LM, Cavenee WK, Rao PN,
Beroukhim R, Peck TC, Lee JC, Sellers WR, Stokoe
D, Prados M, Cloughesy TF, Sawyers CL, Mischel PS
(2005) Molecular determinants of the response of glioblastoma to EGFR kinase inhibitors. N Engl J Med 353:
20122024
227
Oppitz U, Maessen D, Zunterer H, Richter S, Flentje M (1999)
3D-recurrence-patterns of glioblastomas after CT-planned
postoperative irradiation. Radiother Oncol 53:5357
Peca C, Pacelli R, Elefante A, Del Basso De Caro ML, Vergara
P, Mariniello G, Giamundo A, Maiuri F (2009) Early clinical and neuroradiological worsening after radiotherapy and
concomitant temozolomide in patients with glioblastoma;
tumor progression or radionecrosis? Clin Neurol Neurosurg
111:331334
Schafer U, Micke O, Schuller P, Schuck A, Willich N (2004)
The effect of sequential radiochemotherapy in preirradiated
malignant gliomas in a phase II study. J Neurooncol 67:233
239
Schonekaes K, Mucke R, Panke J, Rama B, Wagner W (2002)
Combined radiotherapy and temozolomide in patients with
recurrent high grade glioma. Tumori 88:2831
Schultheiss TE, Kun LE, Ang KK, Stephens LC (1995)
Radiation response of the central nervous system. Int J Radiat
Schwer AL, Damek DM, Kavanagh BD, Gaspar LE, Lillehei
K, Stuhr K, Chen C (2008) A phase I dose-escalation study
of fractionated stereotactic radiosurgery in combination with
gefitinib in patients with recurrent malignant gliomas. Int J
Selch MD, DeSalles AA, Solberg TD, Wallace RE, Do TM, Ford
J, Cabatan-Wong C, Withers HR (2000) Hypofractionated
stereotactic radiotherapy for recurrent malignant gliomas.
J Radiosurgery 3:312
Stupp R, Mason WP, Van den Bent MJ, Weller M, Fisher B,
U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeier
A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff
RO (2005) Radiotherapy plus concomitant and adjuvant
temozolomide for glioblastoma. N Engl J Med 352:987996
VanderSpek L, Fisher B, Bauman G, Macdonald D (2008) 3D
conformal radiotherapy and cisplatin for recurrent malignant
glioma. Can J Neurol Sci 35:5764
Veninga T, Langendijk HA, Slotman BJ, Rutten EH, van der
Kogel AJ, Prick MJ, Keyser A, van der Maazen RW (2001)
Reirradiation of primary brain tumours: survival, clinical
response and prognostic factors. Radiother Oncol 59:127
137
Vordermark D, Kolbl O, Ruprecht K, Vince GH, Bratengeier
K, Flentje M (2005) Hypofractionated stereotactic reirradiation: treatment option in recurrent malignant glioma.
BMC Cancer 5:5559
Wen PY, Kesari S (2008) Malignant gliomas in adults. N Engl J
Med 359:492507
Wurm RE, Kuczer DA, Schlenger L, Matnjani G, Scheffle D,
Cosgrove VP, Ahswede J, Woiciechowsky C, Budach V
(2006) Hypofractionated stereotactic radiotherapy combined
with topotecan in recurrent malignant glioma. Int J Radiat
Oncol Biol Phys 66:S26S32
Yaacov RL, Li AX, Naqa IE, Hahn C, Marks LB, Merchant TE,
Dicker AP (2010) Radiation dose-volume effects in the brain.
Int J Radiat Oncol Biol Phys 76:S20S27
Chapter 24
Glioblastoma: Boron Neutron Capture Therapy

Tetsuya Yamamoto, Kei Nakai, and Hiroaki Kumada
Abstract Boron neutron capture therapy (BNCT) is

a unique method that can deliver tumor-cell-selective
high-linear energy transfer (LET) particle radiotherapy
to an extended target area encompassing a microscopic invasion while avoiding radiation damage to
the surrounding normal brain tissue. The process of
BNCT is based on the nuclear interaction of 10 B
with thermal neutrons with the release of high LET
and 7 Li particles through the boron neutron capture reaction, 10 B(n, ) 7 Li. The very short path length
(<9 m) of -particles and 7 Li enables high-LET irradiation of tumor cells without undesirable damage
to 10 B-unloaded normal cells. Eight non-randomized
prospective external beam BNCT trials for glioblastoma (GBM) have been performed over 15 years using
the two currently available boron drugs and a neutron beam at a nuclear reactor. Four of the 8 studies
suggested that the external beam BNCT may improve
survival in newly diagnosed GBM. In addition, 4 of the
8 studies were primary phase I trials that demonstrated
only modest toxicity. The median time to progression
and the median survival time (MST) were 612 months
and 1227 months, respectively. The optimization of
dosage and boron delivery agents, the combined use
of different boron agents, the combination of BNCT
with other therapeutic modalities and the development
of a hospital-based neutron source have been studied
for the improvement of BNCT. The need for more
evidence-based data, either through randomized trials
T. Yamamoto ()
Department of Neurosurgery and Radiation Oncology, Institute
of Clinical Medicine, University of Tsukuba, Tsukuba City,
Ibaraki 305-8575, Japan
e-mail: tetsu_tsukuba@yahoo.co.jp
or using a prospective case control methodology, is

clear from a review of the clinical reports.
Keywords Glioblastoma BNCT Boron MST
Radiation
Introduction
Despite recent improvements in multimodal therapies, GBM, a radio-chemo-resistant malignant brain
tumor, recurs within several months and progresses
with a median overall survival time (MST) of less
than 11.5 years after the initial treatment. Survival
prolongation by the conventional fractionated photon radiotherapy has been demonstrated in several
randomized trials; these trials used total doses of
4560 Gy and achieved MSTs of 5.815.5 months
(Buatti et al., 2008). Recently, the clinical trial EORTC
26981/22981-NCIC CE3 showed that postoperative
conventional radiotherapy alone at a dose of 60 Gy
in 30 fractions resulted in an MST of 12.1 months.
Another arm, the concomitant and adjuvant use of
temozolomide with conventional radiotherapy, demonstrated a significant survival advantage (MST=14.6
months) compared to conventional radiotherapy alone,
and approximately 25% of the patients survived longer
than 24 months (Stupp et al., 2005).
The dose of radiation required to control GBM completely has not been clearly determined. In dose escalation studies using an additional stereotactic radiosurgery, fractionated proton beam radiation or other
conformal radiotherapies, the MST varied from 9.5
to 26 months. Some of those studies appeared to be

229
230
encouraging; however, they were designed as case

series of small numbers of selected patients (Buatti
et al., 2008). These dose-escalating studies and their
failure analyses imply that at least 90 Gy must be delivered to achieve local control of GBM (Fitzek et al.,
1990; Tanaka et al., 2005). Such high-dose radiation
exceeds the accepted tolerance dose for conventional
radiotherapy of normal brain tissue. Thus, high-dose
radiation should be delivered more selectively for
the main tumor mass as well as for wide-spreading
microscopic invasion, avoiding radiation damage to
the surrounding normal brain. This requirement would
seem to be beyond the abilities of current 3-D conformal radiation methods, such as stereotactic radiation
therapy (SRT), intensity modulated radiation therapy
(IMRT) and image-guided radiation therapy (IMGT),
because tumor-cell-selectivity at the microscopic level
is needed.
Boron Neutron Capture Therapy (BNCT)

Boron neutron capture therapy (BNCT) is a unique
method that can deliver tumor-cell-selective highlinear energy transfer (LET) particle radiotherapy to
an extended target area encompassing the microscopic
invasion while avoiding radiation damage to the surrounding normal brain tissue. Theoretically, BNCT
allows the preferable destruction of boron-10 (10 B)loaded tumor cells while sparing normal tissue without
10 B. The process is based on the nuclear reaction
between 10 B and thermal neutrons, which releases
high LET and 7 Li particles through the boron neutron capture reaction, 10 B(n, ) 7 Li (Fig. 24.1). The
very short path length (<9 m) of -particles and 7 Li
enables high-LET irradiation of tumor cells without
undesirable damage to 10 B-unloaded normal cells.
Although thermal neutrons can be captured more
efficiently by 10 B nuclei, epithermal (high-energy)
neutrons are essential for the external beam BNCT,
because thermal neutrons have limited penetrability
in tissue. Epithermal neutrons are moderated in the
scalp and cranial tissues and become thermal neutrons,
losing their energy. To deliver 10 B, two boron drugs,
p-dihydroxyboryl-phenylalanine (BPA) and sulfohydryl borane Na2 B12 H11 SH (BSH), are currently
available for BNCT clinical studies (Fig. 24.1).
T. Yamamoto et al.
Fig. 24.1 Neutron capture reaction of 10 B (upper) and currently available boron delivery agents: boronophenylalanine
p-dihydroxyboryl-phenylalanine (BPA, lower left) and borocaptate sodium sulofhydryl borane Na2 B12 H11 SH (BSH, lower
right). BNCT is based on the nuclear reaction between 10 B
and thermal neutrons, which releases high linear energy transfer (LET) and 7 Li particles through the boron neutron capture
reaction, 10 B(n, ) 7 Li. The effectiveness of BNCT is highly
dependent on selective 10 B accumulation in tumor cells, as well
as sufficient thermal neutrons even at depth
To determine the eligibility of a patient for BNCT,

the 10 B concentration in the tumor and normal brain
tissue has to be estimated. 18 F-labeled positron emission tomography (PET) can be used to calculate
the lesion-to-normal (L/N) ratio of BPA to estimate
the dose necessary for the tumor. The uptake in
11 C-Methionine-PET is shown to have a linear correlation with that of 18 F-BPA-PET (Fig. 24.2), indicating
the potential application of 11 C-methionine-PET for
the dose planning and candidate selection for BNCT
(Nariai et al., 2009). Boron delivery agents such
as BSH and/or BPA are administered intravenously
before neutron irradiation. Blood samples are drawn
serially after intravenous injection of boron agent to
measure the blood boron level by inductively coupled plasma-atomic emission spectroscopy (ICP-AES)
or by prompt gamma ray analysis (PGA). In the typical BNCT, epithermal neutron beam irradiation is
performed at a research reactor, normally in a single fraction, following the boron drug administration
(Fig. 24.3).
24 Glioblastoma: Boron Neutron Capture Therapy

Fig. 24.2
11 C-Methionine-PET (left)
and 18 F-BPA-PET (right) of
the left frontal GBM. Similar
uptake is shown at the
posteromedial wall of the
surgical cavity
Fig. 24.3 Neutron irradiation

room and head positioning at
JRR-4. The patients head
position is fixed under a
laser-guided positioning
device in the neutron
irradiation room. The figure
also shows the relation
between the beam direction
and the patients head
231
232
Clinical Studies of BNCT

The previous clinical data suggest that BNCT is effective as a post-operative adjuvant treatment for GBM.
The clinical trials of BNCT for newly diagnosed GBM,
which consisted of several small, non-randomized
series, revealed that the median time to progression
varies from 6 to 12 months and the MST varies from 12
to 27 months (Yamamoto et al., 2009). The most favorable survival data of the external beam BNCT appear
to be comparable with the best results of the recent
studies of high-dose radiotherapy, which show an MST
of 1926 months (Buatti et al., 2008). The results of
4 of the 8 prospective studies, even within subgroups
of the patient population, suggest that BNCT may
improve survival in newly diagnosed GBM. The other
4 primary phase I trials demonstrate only modest toxicity, but failed to demonstrate survival prolongation
(Table 24.1). The need for more clinical data, either
through randomized trials or using a prospective case
control methodology, is obvious after reviewing these
clinical reports.
Because of the limited penetrability of neutron
beams, the early clinical trials of BNCT were performed intraoperatively under craniotomy. External
beam BNCT using epithermal neutrons with greater
penetrability was initiated in 1994. In this clinical
trial, 53 GBMs were irradiated using 1, 2 or 3 irradiation fields to evaluate the safety and effectiveness
of external beam BNCT at the Brookhaven National
Laboratory (BNL) (Chanana et al., 1999; Diaz, 2003).
The best MST of 14.8 months was found in the onefield BNCT cohort. The MST of 13 months for all
53 subjects was unfavorable, but this initial applications of a single-fraction, BPA-mediated BNCT were
well tolerated by patients with GBM. Two-hour intravenous injection of BPA-F at a dose of 250330 mg/kg
caused no major adverse events; however, BPA precipitates in the urine were found in patients receiving
330 mg/kg. One of the 17 subjects treated using 2
fields and 4 of the 10 subjects treated using 3 fields
had grade 3 radiation toxicity, such as somnolence
with or without motor weakness, expressive aphasia,
or ototoxicity (Diaz, 2003; Coddere et al., 2004). A
similar BPA-mediated BNCT clinical trial for GBM
was carried out at the Harvard/Massachusetts Institute
of Technology (MIT) between 1996 and 1999 using
T. Yamamoto et al.
1 fraction or 2 fractions on consecutive days and 13

fields depending on the size and location of the tumor
(Busse et al., 2003). No adverse event was found in
relation to the intravenous injection of 250 mg/kg over
1 h, 300 mg/kg over 1.5 h, or 350 mg/kg over 1.5 h.
The tumors with volumes >60 cm3 and <60 cm3 were
associated with a 67 and 19% incidence of developing
grade 3 or higher toxicity, respectively. The MST in
that clinical trial was 12 months. The Finnish phase I/II
trial conducted by Helsinki University Central Hospital
and the VTT (Technical Research Centre of Finland)
showed that a BPA dose level of 450 mg/kg was optimal for further BNCT studies of newly diagnosed
GBM (Joensuu et al., 2003; Kankaanranta et al., 2007,
2008). BPA of 290 mg/kg was infused over 2 h in the
first 12 patients suffering from GBM using 2 fields,
and the BPA dose to subsequent patients was escalated from 330 mg/kg (n = 1) to 360 mg/kg (n = 3),
400 mg/kg (n = 3), 450 mg/kg (n = 3), or 500 mg/kg
(n = 8). The maximum tolerated dose was reached at a
BPA dose level of 500 mg/kg, where grade 3 (n = 2)
and grade 4 (n = 1) CNS toxicity was found. The
median OS values for the dose groups of 290, 330/360,
400, 450, and 500 mg BPA/kg were 13.4, 11.0, 16.9,
21.9 and 14.7 months, respectively.
A BSH-mediated BNCT clinical trial was performed at the Petten Irradiation Facility, Netherlands
(EORTC trial 11961). The BSH was applied 4 times
using a dose suitable for reaching 30 ppm 10 B in
blood over the total irradiation time, and the epithermal neutron beam was delivered in 4 fractions on 4
consecutive days. On the first day all patients received
100 mg BSH/kg bodyweight. No dose-related side
effects were observed and the limiting radiation dose
was not reached in that dose-escalation study. The
mean survival for the dose groups varied from 10.4 to
13.2 months (Wittig et al., 2002). The study showed
the feasibility and safety of a fractionated BNCT application using BSH at a dose of 100 mg/kg infused
at a dose rate of 1 mg/kg/min. BSH-mediated BNCT
was also studied at the NRI in Rez with an infusion
of 100 mg BSH/kg and was well tolerated with only
modest toxicity (Burian et al. 2002).
Preclinical data suggest that a longer infusion
time may improve the homogeneity of boron accumulation in tumors in BPA-mediated BNCT. This
method was applied to the phase II clinical trial
26 (19972002),
58 years
30 (19992005),
55.5 years for
18 pts
29 (20012003),
53 years
5 (2001),
10 (Protocol 1:
20022004), 59
years
11 (Protocol 2:
20042006),
47.5 years
EORTC 11961,
Phase I
University of
Helsinki and
VTT, Phase I/II
Studsvik, Phase II
Osaka Medical
College, Phase II
7.015.5/3.3
6.1 Gy
14.2 >/2 Gy
8.611.4 Gya
(physical boron
dose)/ND
8.113.5/36 Gy
8.716.4/2.77.4
8.414.8/1.88.5
BPA:250 mg/kg, 1 h
Prescribed peak
BSH:100 mg/kg,
dose 13 Gy>
1h
BPA:700 mg/kg, 6 h
Prescribed peak
BSH:100 mg/kg,1 h
dose 15 Gy >
BSH:100 mg/kg, 1 h
BPA:
290500 mg/kg,
2h
BPA: 900 mg/kg, 6 h
BPA:
250330 mg/kg,
2h
BPA:
250350 mg/kg,
12 h
BSHc :
100 mg/kg/min
14.1
23.5
26.965.4
ND
21.9 for
450 mg/kg
cohort
14.2e
13.2 for 10.4 Gy

cohortd
13 (1 field: 14.8, 2
fields: 12.1, 3
fields: 11.9)
12
16.363.0
ND
15.554.3
ND
ND
1855 (data from

38 of 53
subjects)
7.324.8
Studsvik AB
Sweden
LVR-15, NRI Rez,
Czech Republic
KUR, KURRI and
JRR-4, JAEA,
Japan
Single fraction
(No)
Single fraction
(No)
Single fraction
(No)
Single fraction
(2030 Gy/10
20
fraction)
Single fraction
(30 Gy/15 fr or
30.6 Gy/17
fraction)
Fir1, Helsinki,
Finland
HFR, Petten, the

Netherlands
MITR-II, M-67,
MIT, USA
BMRR, BNL,
USA
Reactor, institute,
country
Single fraction
(No)
4 fractions (No)
1 or 2 fractions
(No)
Single fraction
(No)
Neutron
irradiation,
(photon
radiation)
University of
8 (19982007), 65 BPA:250 mg/kg,1 h
8.414.1/2.53.4
15.542.5
27.1/11.9
JRR-4, JAEA,
Tsukuba and
years
BSH:5 g/body,1 h
Japan
JAEA, Phase
I/II
a Weighted dose D
w
b Includes 2 melanomas. Patients underwent BNCT at MITR-II
c Four fractions on consecutive days, 100 mg/kg was administered the first day, and this dose was sufficient to keep the average blood concentration at 30 g/g during treatment
days 24
d Mean survival time for the 3rd dose group
e Survival time calculated from the BNCT treatment
NRI Rez, Phase I
Harvard/MIT,
Phase I
53 (19941999),
(56.5 years for 1
field)
20 (19971999)b ,
56 years
BNL, Phase I/II
Table 24.1 Summary of currently underway or recently completed external beam BNCT clinical trials for newly diagnosed GBM
Number of
evaluated
Minimum tumor
patients (year),
Normal brain
Median survival
dose in GTVa
median age of
Boron drug: dose,
dosea peak/ave.
Trial (reference)
(Gy)
(months)
(Gy)
patients
infusion time

233
234
at the Studsvik BNCT facility for 29 patients suffering from GBM, who received 900 mg/kg BPA-F
in a 6-h infusion. The minimum dose to the tumor
volume and to the target volume (defined as tumor
plus edema plus a 2 cm margin) ranged from 15.4
to 54.3 Gy and 8.8 to 30.5 Gy, respectively. Four
patients developed grade 34 toxic events including epileptic seizures, hematuria, thrombosis and erythema. The median time from BNCT treatment to
tumor progression was 5.8 months, and the MST
after BNCT was 14.2 months, suggesting the survival
benefit of long-time infusion compared to conventional 2-h infusion in BPA administration (Henriksson
et al., 2008; Skld et al., 2009).
The combined use of BPA and BSH was based
on experimental data that showed different uptake
characteristics of the cell-cycle dependency of tumor
cells (Yoshida et al., 2002). Other experimental data
have also suggested that the combination of BNCT
and photon radiation leads to significant gains in survival (Barth et al., 2005). In the trial conducted by
Osaka Medical College, the first 10 patients suffering
from GBM were administered 100 mg/kg of BSH and
250 mg/kg of BPA in a 1-h infusion (protocol 1), and
the latter 11 patients were administered 100 mg/kg of
BSH and 700 mg/kg of BPA in a 6-h infusion (protocol 2). X-ray irradiation was added in protocol 2 for
a total dose of 2030 Gy. The MSTs for all patients
and for protocol 2 patients were 15.6 and 23.5 months,
respectively (Kawabata et al., 2008).
In the trial at the University of Tsukuba and
Tokushima University at Japan Research Reactor No.4
(JRR-4) of the Japan Atomic Energy Agency (JAEA),
a low dose (250 mg/kg) of BPA was administered over
1 h, and 5 g BSH/kg bodyweight was infused over
1 h in 8 patients with a single irradiation field. These
patients received additional photon radiation to the volume which was defined as a peritumoral high intensity
in the T2-weighted MRI after completion of BNCT.
The median OS and the time to progression were 27.1
and 11.9 months, respectively. The 1-year and 2-year
survival rates were 87.5 and 62.5%, respectively. The
small group of patients showed good tolerance to the
treatment and there were no severe acute or subacute adverse events. Although it is not certain whether
additional photon irradiation could play a role in the
clinical response to BNCT, the survival of this small
cohort seems to be favorable evidence (Yamamoto
et al., 2009).
T. Yamamoto et al.
Boron Compounds for Clinical BNCT

Successful BNCT is dependent on the selective accumulation and absolute level of 10 B atoms in tumor
cells. There are two boron delivery agents available for
clinical BNCT trials for high-grade glioma, BPA and
BSH (Fig. 24.1).
Boron is an essential plant nutrient, and necessary for the optimal health of mammals as a rare
element, that may be beneficial for bone growth and
maintenance, central nervous system function, and
the inflammatory response, though its physiological
role in mammals is not yet fully understood. Natural
boron consists of 80% 11 B and 20% 10 B, and the
latter 10 B isotope has high efficiency in capturing thermal neutrons to generate the boron neutron capture
reaction,10 B(n, ) 7 Li (Nielsen, 2009).
BPA, the amino acid analogue, is remarkably nontoxic. Large doses have been administered to experimental animals without perturbing hematopoietic,
hepatic or renal function (FDA-IND #43,317). BPA
has structural characteristics similar to those of a
melanin precursor, and promising clinical results were
shown in the pilot study of BNCT for skin melanoma
(Coderre et al., 2003). BPA is usually administrated
intravenously as a soluble fructose complex, BPA-F,
at doses ranging from 250 mg BPA/kg to 900 mg
BPA/kg. BPA can penetrate across the blood-brain
barrier into the normal brain, and is actively transported through the tumor cell membrane due to the
elevated rate of amino acid transport in proliferating cells. The average macroscopic concentration of
boron in tumor is 24 times higher than that in blood.
However, non-proliferating cells escape the radiation
damage following BNCT.
The primary mode of BSH distribution is passive
diffusion from blood to tumor tissue via the disrupted
blood brain barrier. The boron concentration in the normal brain with an intact blood brain barrier remains
minimal, while the tumor 10 B concentration is related
to both the tumor vessel density and blood 10 B level.
A tumor-to-blood boron concentration ratio ranging
from 0.5 to 1.0 has been reported in human patients
treated with BSH-mediated BNCT. The recent clinical
trials have used a combination of BPA and BSH with
the intention that these different compounds would
accumulate in different subpopulations of tumor cells
(Yamamoto et al., 2008).
New Boron Compound and Delivery

System
Previous studies showed that a boron concentration
of approximately 2040 g10 B/g tissue, or 109 10 B
atoms/cell, could control the tumor progression. The
coefficient rate of BSH is low, thus only 2% or less of
the total amount administered participated in the neutron capture reaction in the irradiation field. Roughly
translated, these results mean that these agents must
be as safe as H2 O or glucose, because the required
intracellular boron level could be achieved by drug
administration on the order of grams. This is one of
the reasons why the development of new boron compounds is difficult. The required characteristics for
BNCT agents are as follows: (a) non-toxic at the clinically effective dose; (b) the achievement of at least
1030 g10 B/g of tumor; (c) high tumor/brain and
tumor/blood concentration ratios; (d) rapid clearance
from blood circulation and normal tissues but persistence in tumor; (e) water solubility; and (f) chemical
stability (Coderre et al., 2004; Barth et al., 2005).
To date, a variety of boron delivery agents have been
investigated, including porphyrins, amino acids, polyhedral borane, carbohydrates, polyamines, biochemical precursors, DNA-binding agents, antisense agents,
peptides, liposomes and monoclonal antibodies.
Low molecular weight boron agents are peptides,
porphyrins, amino acids such as BPA or carborane derivatives such as BSH. The high molecular weight group contains boronated antibodies and
proteins. Liposomes are categorized independently
of boron delivery systems. BOPP, a boronated porphyrin, and GB-10, polyhedral borane Na2 B10 H10 ,
have been evaluated for their pharmacokinetics in
humans. A phase I trial for potential application
of BOPP to photodynamic therapy was interrupted
because of its general toxicity. Recently, improvement of the synthesis pathway and novel agents
such as porphyrin-cobaltacarborane conjugates or
10 B-enriched carboranyl-containing phthalocyanine
have been under examination. GB-10 has been proposed as a boron neutron capture agent for an additional dose in fast neutron therapy. In a trial in which
GB-10 was used in 15 lung cancer patients, no major
toxicity was recorded.
Using the principle of antibody-antigen reaction
for active tumor targeting, numerous molecular target
235
drugs were developed and are clinically used at

present. Nevertheless, with respect to boron delivery,
this method does not achieve a high cellular concentration of boron, even with the high selectivity and
specificity.
Most drug delivery systems (DDS) applied to
BNCT are liposomal delivery systems. Such liposoms
may include a solution containing a high concentration
of boron, or boron-containing lipids. Active targeting is achieved by modifying the surface molecules.
The current marketing of several liposomal anticancer
agents and photodynamic therapy agents for macular
degeneration shows the steady growth of the technical
improvements in this field.
Direct intratumoral injection and convectionenhanced delivery require the use of a pump to achieve
a pressure gradient and establish the bulk flow of
boron-containing agents during the interstitial infusion. Blood brain barrier disruption with hyperosmotic
mannitol or receptor-mediated permeabilizer-7 (CMP7), have helped to achieve high tumor boron concentrations in experimental animals (Yang et al., 2009).
Recent reports have revealed that preloading with LDOPA increases the uptake of BPA, that buthionine
sulfoximine may similarly enhance the uptake of BSH,
and that tumor growth is delayed following administration of either agents and neutron irradiation (Capuani
et al., 2008; Yoshida et al., 2008).
Neutron Beam Facilities for BNCT

The first clinical trials of BNCT for malignant brain
tumors were performed at the Brookhaven Medical
Research Reactor (BMRR) of BNL from 1959 to 1961.
Another group of patients with malignant gliomas was
treated using the reactor at MIT over the same period.
Several research reactors, such as the Hitachi Training
Reactor, Musashi Institute of Technology Research
Reactor (MuITR), Kyoto University Research Reactor
(KUR) and Japan Research Reactor No. 2 (JRR-2)
in JAEA, have been applied to clinical studies in
Japan. In these early BNCT trials, a low energy thermal neutron beam was used for the irradiation. To
deliver thermal neutrons to deeper regions of the brain,
the BNCT procedure included craniotomy, such as
skin flap reopening and bone removal. Since 1994,
external beam BNCT using epithermal neutron beams
was initiated using the BMRR and the High Flux
236
T. Yamamoto et al.
Fig. 24.4 Cross-section of the neutron irradiation facility of Japan Research Reactor No. 4 (JRR-4) (upper) and process flow of the
BNCT dose planning system JCDS and the setting simulation (lower)
Reactor (HFR) in Petten, the Netherlands. By using

epithermal neutron beams the therapeutic range was
expanded more deeply into the brain, compared to
that of the thermal neutron beam irradiation, and with
the application of the epithermal neutron beam it
became possible to perform non-operative irradiation
(Yamamoto et al., 2008).
At present, neutron beams exclusively for use in
BNCT are available at JRR-4 (JAEA, Japan), KUR
(Kyoto University, Japan), LVR-15 (NRI Rez, Czech
Republic), Studsvik AB (Sweden), Fir1 (Finland),
HFR (Petten, the Netherlands), MITR-II and M-67
(MIT, USA), and the TRIGA reactor (Palva, Italy).
JRR-4 in JAEA can generate a thermal neutron beam

as well as an epithermal neutron beam (Fig. 24.4).
Neutron beam facilities consist of several neutron
spectrum control devices, such as a heavy water tank,
cadmium shutter, bismuth filter or collimator. The
heavy water tank has four separated heavy water layers. The heavy water thickness in the tank can be
changed by filling the corresponding layers with heavy
water. And the cadmium shutter can be put in or out at
the beam line. When the cadmium shutter is put at the
beam line, the beam facility generates an epithermal
neutron beam. The bismuth filter is installed at the
beam line to reduce gamma-rays from the reactor
core. The utilization of the heavy water tank and the

cadmium shutter enables the generation of neutron
beams appropriate for the patients disease condition.
The BNCT clinical trial for malignant brain tumors
using the facility of JRR-4 was initiated in 1999
using a thermal neutron beam, including craniotomy,
and beginning in 2003 this treatment was switched
to external beam BNCT with an epithermal neutron
beam (Yamamoto et al., 2009). The epithermal neutron beam allowed the performance of clinical trials for
other cancers, including head-and-neck and pulmonary
tumors (Kankaanranta et al., 2007; Haginomori et al.,
2008). Now the thermal neutron beam is used only for
shallow-seated tumors such as skin melanomas.
Treatment Planning System

The treatment planning for BNCT differs markedly
from the planning for photon or electron irradiation
in conventional radiotherapy and, in some ways, is
significantly more complex. The boron neutron capture reaction provides a tumor-selective boron dose,
10 B(n, ) 7 Li, while other non-selective dose components are present and consist of proton recoils due to
fast neutrons, 1 H(n, n )p, 0.54 MeV protons from the
nitrogen capture reaction, 14 N(n, p)14 C, -rays arising from the contamination in the primary beam and
2.2 MeV prompt -rays from the hydrogen capture.
Treatment planning for BNCT involves computation
and analysis of these different components of radiation
doses and their distribution in a patient to determine
the neutron beam orientation and radiation fluence
enabling the delivery of an optimized radiation dose
and distribution. The optimized combination will comply with the dose prescription, optimize the dose to the
target volume, and respect the dose limits to normal
tissues and organs at risk.
The therapeutic advantage in BNCT is obtained
principally through the tumor-selectivity of a neutron
capture agent (e.g., 10 B) rather than through precise
geometric targeting of multiple well-collimated radiation fields onto the target volume. In conventional
photon radiotherapy, only a single low-LET dose component must be computed, the dose from primary
photons or electrons, both of which are ultimately
delivered by electrons. In contrast, in BNCT the
radiation field is a complex mixture of high- and
low-LET dose components with varying biological
237
effectiveness that depends on both the tissue and the

chemical form of the neutron capture agent. The spatial distribution of each dose component is different
and depends on the tissue composition, as well as the
neutron and photon fluence spectra.
Three BNCT treatment planning systems (TPSs)
have been reported and are being applied to clinical trials NCTPlan, SERA and JCDS (Nigg, 2003;
Kumada et al., 2007). All three TPSs are noncommercial and have been developed by small research teams,
usually with expertise in nuclear engineering. The
NCTPlan was developed at MIT and is being applied
for the treatment planning of BNCT at MITR and
the RA-6 reactor in Argentina. SERA was developed
by the Idaho National Engineering and Environmental
Laboratory (INEEL) and Montana State University
(MSU) in the USA. The system has been used for the
planning of BNCT studies that are being performed at
the FiR-1 reactor in Finland and KUR in Japan. JCDS
was developed by JAEA to perform BNCT at JRR-4,
and the system is still being developed in order to
improve the dose calculation accuracy (Kumada et al.,
2007; 2009).
The treatment planning procedure using BNCT
TPSs is almost the same as that of the conventional
TPSs for photon radiotherapy, and the user interface
of TPSs is also similar to that of photon radiotherapy.
The algorithms generally used for dose calculations in
clinical photon radiotherapy are computationally efficient, simple, and empirical. On the other hand, BNCT
TPSs exclusively rely on a Monte Carlo simulation
for dose calculations because of the complex, scatterdominated nature of the radiation transport processes
involved. Figure 24.4 shows the outline of the procedure for BNCT treatment planning with JCDS. In the
process of computational dosimetry, a patients CT and
MRI images are first used to create a 3-dimensional
model. The CT images are used to automatically differentiate compositions of the human body into bone,
soft tissues and air according to their CT values. On
the other hand, the MRI data are used to define several regions of interest (ROIs), such as the tumor region
and organs at risk. By superimposing the MRI images
onto the CT images, a detailed 3D model incorporating the information on body composition and ROIs is
created. The latest version of JCDS can deal with PET
images in addition to CT and MRI images to pick out
tumor regions in accordance with the T/N ratio on PET
images. To compute the distributions of several dose
238
components and neutron fluxes in the body effectively,

the detailed 3D model is converted into a voxel calculation model. The initial version of the JCDS made
the voxel calculation model by dividing the space into
10 10 10 mm3 voxel cells that contain the proper
material data in each. JCDS has been improved continuously to perform more accurate dosimetry, and the
current version of JCDS can produce a voxel calculation model of 2 2 2 mm3 voxel cells. The distributions of the dose components and fluxes in the voxel
model are determined by Mote Carlo transport calculation. JCDS evaluates the detailed distributions of the
dose components based on the calculation results, and
outputs a two-dimensional distribution superimposed
on CT or MRI. Finally, physicians decide the irradiation conditions for the patient based on the JCDS
results.
In conclusion, BNCT is the most complicated radiotherapy; however, it may one day be a valuable alternative treatment in cases of GBM that invades the
surrounding healthy normal brain. BNCT seems to be
preferable to currently existing high-dose radiotherapies, since in theory it allows tumor-cell-selective
high-dose particle radiation while sparing normal
cells. It remains uncertain whether BNCT can control GBM sufficiently and lead to distinct survival
benefits because of the limitations of the present clinical trials, which include the lack of a contemporary control arm. Randomized trials of comparably
selected patients are required to demonstrate the efficiency of this emerging modality more conclusively.
To improve the effectiveness of BNCT, several critical issues must be addressed. Single or combined use
of currently available boron agents must be optimized,
and more tumor-selective boron agents and/or delivery or targeting systems are needed. The combined
use of photon irradiation, chemotherapy, immunotherapy, and/or molecular targeting therapy with BNCT
should also be evaluated. Finally, the development of
an accelerator-based neutron source could lead to an
in-hospital BNCT, which would greatly enhance both
basic and clinical research on BNCT.
Acknowledgments This study was supported in part by a
Grant-in-Aid for Society Collaboration from the Ministry of
Education, Science and Culture of Japan (20390379). We
thank Professor Akira Matsumura, Dr. Tadashi Nariai and Dr.
Alexander Zaboronok for their kind support in the manuscript
preparation.
T. Yamamoto et al.
References
Barth RF, Coderre JA, Vicente GH, Blue TE (2005) Boron
neutron capture therapy of cancer: current status and future
prospects. Clin Cancer Res 11:39874002
Buatti J, Ryken TC, Smith MC, Sneed P, Suh JH, Mehta M,
Olson JJ (2008) Radiation therapy of pathologically confirmed newly diagnosed glioblastoma in adults. J Neurooncol
89:313337
Burian J, Marek M, Rataj J, Flibor S, Rejchrt J, Viererbl L, Sus
F, Honova H, Petruzelka L, Prokes K, Tovarys F, Dbaly V,
Benes V, Kozler P, Honzatko J, Tomandl I, Mares V, Marek
J, Syrucek M (2002) Report on the first patient group of the
phase I BNCT trial at the LVR-15 reactor. In: Surewein W,
Moss R, Wittig A (eds) Research and development in neutron
capture therapy. Monduzzi Editore, Bologna, pp 11071112
Busse PM, Harling OK, Palmer MR, Kiger WS III, Kaplan J,
Kaplan I, Chuang CF, Goorley JT, Riley KJ, Newton TH,
Santa Cruz GA, Lu XQ, Zamenhof RG (2003) A critical
examination of the results from the Harvard-MIT NCT program phase I clinical trial of neutron capture therapy for
intracranial disease. J Neurooncol 62:111121
Capuani S, Gili T, Bozzali M, Russo S, Porcari P, Cametti
C, DAmore E, Colasanti M, Venturini G, Maraviglia
B, Lazzarino G, Pastore FS (2008) L-DOPA preloading
increases the uptake of borophenylalanine in C6 glioma rat
model: a new strategy to improve BNCT efficacy. Int J Radiat
Chanana AD, Capala J, Chadha M, Coderre JA, Diaz AZ,
Elowitz EH, Iwai J, Joel DD, Liu HB, Ma R, Pendzick N,
Peress NS, Shady MS, Slatkin DN, Tyson GW, Wielopolski
L (1999) Boron neutron capture therapy for glioblastoma multiforme: interim results from the phase I/II doseescalation studies. Neurosurgery 44:11821193
Coderre JA, Hopewell JW, Turcottea JC, Rileyc KJ, Binnsc
PJ, Kiger WS III, Harling OK (2004) Tolerance of normal
human brain to boron neutron capture therapy. Appl Radiat
Isot 61:10831087
Coderre JA, Turcotte JC, Riley KJ, Binns PJ, Harling OK,
Kiger WS III (2003) Boron neutron capture therapy: Cellular
targeting of high linear energy transfer radiation. Technol
Cancer Res Treat 5:355375
Diaz AZ (2003) Assessment of the results from the phase
I/II boron neutron capture therapy trials at the Brookhaven
National Laboratory from a clinicians point of view. J
Fitzek MM, Thornton AF, Rabinov JD, Lev MH, Pardo FS,
Munzenrider JE, Okunieff P, Bussiere M, Braun I, Hochberg
FH, Hedley-Whyte ET, Liebsch NJ, Harsh GR 4th (1990)
Accelerated fractionated proton/photon irradiation to 90
cobalt gray equivalent for glioblastoma multiforme: results
of a phase II prospective trial. J Neurosurg 91:251260
Haginomori S, Miyatake S, Inui T, Kawabata S, Takamai A,
Lee K, Takenaka H, Kuroiwa T, Uesugi Y, Kumada H,
Ono K (2008) Planned fractionated boron neutron capture
therapy using epithermal neutrons for a patient with recurrent
squamous cell carcinoma in the temporal bone: a case report.
Head Neck 31:412418
Henriksson R, Capala J, Michanek A, Lindahl SA, Salford
LG, Franzn L, Blomquist E, Westlin JE, Bergenheim T

(2008) Boron neutron capture therapy (BNCT) for glioblastoma multiforme: A phase II study evaluating a prolonged
high-dose of boronophenylalanine (BPA). Radiother Oncol
88:183191
Joensuu H, Kankaanranta L, Seppl T, Auterinen I, Kallio M,
Kulvik M, Laakso J, Vhtalo J, Kortesniemi M, Kotiluoto
P, Seren T, Karila J, Brander A, Jrviluoma E, Ryynnen P,
Paetau A, Ruokonen I, Minn H, Tenhunen M, Jskelinen
J, Frkkil M, Savolainen S (2003) Boron neutron capture
therapy of brain tumors: clinical trials at the Finnish facility
using boronophenylalanine. J Neurooncol 62:123134
Kankaanranta L, Koivunoro H, Kortesniemi M, Vlimki P,
Seppl T, Kotiluoto P, Auterinen I, Kouri M, Savolainen
S, Joensuu H (2008) BPA-based BNCT in the treatment
of glioblastoma multiforme In: Zonta A, Altieri S, Roveda
L, Barth R (eds) A dose escaration study. Proceedings of
the 13th international congress of neutron capture therapy.
ENEA, p 30
Kankaanranta L, Seppl T, Koivunoro H, Saarilahti K, Atula T,
Collan J, Salli E, Kortesniemi M, Uusi-Simola J, Mkitie A,
Seppnen M, Minn H, Kotiluoto P, Auterinen I, Savolainen
S, Kouri M, Joensuu H (2007) Boron neutron capture therapy
in the treatment of locally recurred head and neck cancer. Int
J Radiat Oncol Biol Phys 69:475482
Kawabata S, Miyatake S, Kuroiwa T, Yokoyama K, Doi A,
Iida K, Miyata S, Nonoguchi N, Michiue H, Takahashi M,
Inomata T, Imahori Y, Kirihata M, Sakurai Y, Maruhashi
A, Kumada H, Ono K (2008) Boron neutron capture therapy for newly diagnosed glioblastoma. J Radiat Res (Tokyo)
50:5160
Kumada H, Nakamura T, Komeda M, Matsumura A (2009)
Development of a new multimodal Monte-Carlo radiotherapy planning system. Appl Radiat Isot 67:118121
Kumada H, Yamamoto K, Matsumura A, Yamamoto T,
Nakagawa Y (2007) Development of JCDS, a computational
dosimetry system at JAEA for boron neutron capture therapy.
J Phy Conf Ser 74:17
Nariai T, Ishiwata K, Kimura Y, Inaji M, Momose T, Yamamoto
T, Matsumura A, Ishii K, Ohno K (2009) PET pharmacokinetic analysis to estimate boron concentration in tumor and
brain as a guide to plan BNCT for malignant cerebral glioma.
Appl Radiat Isot 67:S348S350
Nielsen FH (2009) Micronutrients in parenteral nutrition: boron,
silicon, and fluoride. Gastroenterology 137:S55S60
Nigg DW (2003) Computational dosimetry and treatment
planning considerations for neutron capture therapy. J
Neurooncol 62:7586
239
Skld K-H, Stenstam B, Diaz AZ, Giusti V, Pellettieri L,
Hopewell JW (2009) Boron neutron capture therapy for
glioblastoma multiforme: advantage of prolonged infusion of
BPA-f. Acta Neurol Scand. [Epub ahead of print]
B, Taphoorn MJB, Belanger K, Brandes AA, Marosi C,
adjuvant temozolomide for glioblastoma. N Engl J Med
352:987996
Tanaka M, Ino Y, Nakagawa K, Tago M, Todo T (2005)
High-dose conformal radiotherapy for supratentorial malignant glioma: a historical comparison. Lancet Oncol 6:
953960
Wittig A, Hideghety K, Paquis P, Heimans J, Vos M, Goetz C,
Haselsberger K, Grochulla F, Moss R, Morrissey J, BourhisMartin E, Rassow J, Stecher-Rasmussen F, Turowski B,
Westler M, deVris MJ, Fankhauser H, Gabel D, Sauerwein W
(2002) Current clinical results of the EORTC-study 11961.
In: Surewein W, Moss R, Wittig A (eds) Research and
development in neutron capture therapy. Monduzzi Editore,
Bologna, pp 11171122
Yamamoto T, Nakai K, Kageji T, Kumada H, Endo K, Matsuda
M, Shibata Y, Matsumura A (2009) Boron neutron capture
therapy for newly diagnosed glioblastoma. Radiother Oncol
91:8084
Yamamoto T, Nakai K, Matsumura A (2008) Boron neutron capture therapy for glioblastoma. Cancer Lett 262:
143152
Yang W, Barth RF, Wu G, Huo T, Tjarks W, Ciesielski M,
Fenstermaker RA, Ross BD, Wikstrand CJ, Riley KJ, Binns
PJ (2009) Convection enhanced delivery of boronated EGF
as a molecular targeting agent for neutron capture therapy of
brain tumors. J Neurooncol 95:355365
Yoshida F, Matsumura A, Shibata Y, Yamamoto T, Nakauchi H,
Okumura M, Nose T (2002) Cell cycle dependence of boron
uptake from two boron compounds used for clinical neutron
capture therapy. Cancer Lett 187:135141
Yoshida F, Yamamoto T, Nakai K, Kumada H, Shibata
Y, Tsuruta W, Endo K, Tsurubuchi T, Matsumura
A (2008) Combined use of sodium borocaptate and
buthionine sulfoximine in boron neutron capture therapy
enhanced tissue boron uptake and delayed tumor growth
in a rat subcutaneous tumor model. Cancer Lett 263:
253258
Chapter 25
Glioblastoma: Anti-tumor Action of Cyclosporin A

and Functionally Related Drugs
Bozena Kaminska, Magdalena Tyburczy, Konrad Gabrusiewicz,
and Malgorzata Sielska
Abstract Human malignant gliomas are highly

resistant to current therapeutic approaches. Major
signaling pathways that have been identified as
playing important roles in glioblastomas are: the
PTEN/PI3K/Akt/mTOR and the Ras/Raf/MEK/ERK
signaling cascades, which support cell invasion, survival and prevent apoptosis. In the face of tumor
resistance to apoptosis, novel agents which can overcome resistance or/and affect cell survival by nonapoptotic mechanisms such as necrosis, senescence,
autophagy and mitotic catastrophe, are highly desirable. The present chapter focuses on anti-tumor action
of cyclosporin A (CsA) and rapamycin that besides
their well known immunosuppressive abilities appear
to be multitarget kinase inhibitors and moderately
effective anti-tumor agents in glioblastomas in vitro,
in vivo and in clinical trials. A compelling evidence
shows that cyclosporin A induces growth arrest and
programmed cell death in cultured rat and human
glioblastoma cells. The molecular mechanism involves
accumulation of a cell cycle inhibitor p21Cip1/Waf1,
even in the absence of functional p53 tumor suppressor. In C6 glioma cells with functional TP53 and
PTEN tumor suppressors CsA treatment up-regulates
fasL expression, activates p53 and intrinsic mitochondrial death pathway, while in human glioblastoma
cells with defects in either TP53 or PTEN, none
of those effects were observed. Molecular analysis

revealed that CsA, trough yet unknown mechanisms,
down-regulates PI3K/Akt and mTOR signaling pathways in glioblastoma cells, and interferes with proinvasive activity of tumor-infiltrating microglia. A systemically applied CsA significantly reduced growth
of intracranial gliomas, tumor invasion and angiogenesis. Pharmacological inhibitors of the mTOR
pathway: rapamycin, temsirolimus, everolimus and
AP23573 were tested as potential targeted drugs in
human glioblastoma cultures and in animal models. However, rapamycin and derivatives show moderate efficacy in patients with recurrent glioblastoma multiforme, they deserve further clinical studies,
particularly in combination with PI-3K pathway
inhibitors. Defects of innate and adaptive immunity
are common in glioblastoma patients contributing
to a lack of effective anti-tumor responses. Thus,
immunosuppressants such as CsA, rapamycin and
its derivatives may be an effective novel strategy to
treat drug-resistant gliomas or complement apoptosis
based-therapies.
Keywords Glioblastomas Anti-tumor drugs
Immunosuppressants Apoptosis Rapamycin
Inhibitors
Introduction
B. Kaminska ()
Laboratory of Transcription Regulation, Department of Cell
Biology, Nencki Institute of Experimental Biology, 02-093
Warsaw, Poland
e-mail: bozenakk@nencki.gov.pl
Glioblastomas are the most frequent and devastating primary malignant brain tumor in adults.
Glioblastomas are highly resistant to current
therapeutic approaches, in which surgery is followed

241
242
by radiotherapy with concomitant and adjuvant

chemotherapy. The prognosis remains poor with
a median survival in the range of 1215 months
(Clarke et al., 2010). Common genetic abnormalities
in glioblastoma are associated with multiple molecular
mechanisms involved in drug resistance, including
drug detoxification, aberrant activation or suppression
of cellular signal transduction pathways, deficiencies
in tumor suppressors, apoptosis mediators and death
ligand/receptor signaling. Significantly high frequency
of alterations in cell cycle regulators such as the TP53
tumor suppressor and the p16INK4A cyclin dependent
kinase inhibitor results in reduced sensitivity to a
majority of anti-cancer drugs (Ohgaki and Kleihues,
2009).
Major signaling pathways that have been identified
as playing important roles in glioblastomas are: the
PTEN/PI3K/Akt/mTOR and the Ras/Raf/MEK/ERK
signaling cascades, which support cell proliferation,
survival, invasion, and prevent apoptosis (McCubrey
et al., 2006). Components of these pathways are
frequently mutated, aberrantly expressed or constitutively activated in glioblastomas. Monoclonal antibodies or small molecular-weight kinase inhibitors
targeting specific pathways are the most common
classes of agents in cancer treatment. However,
highly selective or specific blocking of only one
of the kinases has been associated with limited or
sporadic responses. Therefore, multitargeted kinase
inhibitors and combinations of single-target kinase
inhibitors should be more effective to overcome
therapeutic resistance. Some agents could be used
together with radiation, chemotherapy, or immunotherapy to enhance treatment efficacy. Furthermore, in
the face of tumor resistance to apoptosis, novel
agents which can overcome resistance or/and induce
cell death by non-apoptotic mechanisms such as
necrosis, senescence, autophagy (type II programmed
cell death) and mitotic catastrophe, are highly
desirable.
The present chapter focuses on anti-tumor action
of drugs such as cyclosporin A and rapamycin
that besides their well known immunosuppressive
abilities appear to be kinase inhibitors and potential anti-tumor agents in glioblastomas. The present
chapter summarizes a rationale and results of preclinical/clinical studies of these agents in therapy of
glioblastomas.
B. Kaminska et al.
Mechanisms of Cyclosporin A Induced

Cell Death in Rat C6 Glioma Cells
Mode of Immunosuppressant Action
Cyclosporin A (CsA), FK506 (tacrolimus, Prograf)
and rapamycin (sirolimus) are short polypeptides
which have revolutionized transplantology due to ability to block the activation of lymphocytes and other
immune system cells. CsA, FK506 and rapamycin bind
to specific intracellular proteins called immunophilins:
CsA binds to cyclophilin, FK506 and rapamycin
bind to FKBP (FK506-binding protein). Drugimmunophilin complexes bind to a regulatory subunit
of calcineurin and inhibit its activity. Calcineurin is
a calcium- and calmodulin-dependent threonine/serine
phosphatase. CsA and FK506 exert immunosuppressive effects by inhibiting of calcineurin-mediated
dephosphorylation of NFAT (nuclear factor of activated T cells), thus preventing transcriptional induction of several cytokines and their receptors genes
(Fig. 25.1). NFAT family proteins are transcription
factors that regulate the expression of a variety of
target genes and are implicated in many functions,
including cell growth, survival, invasion and angiogenesis (Mancini and Toker, 2009). Several members of
NFAT family were detected in C6 glioma cells pointing to a new mechanism of transcription regulation in
glioma cells. A transient receptor potential 6 (TRPC6)
which is required for the development of the aggressive glioblastoma phenotype and causes a sustained
elevation of intracellular calcium, is coupled to activation of the calcineurin-NFAT pathway. Pharmacologic
inhibition of this pathway reduced the development of the aggressive glioblastoma phenotype under
hypoxia.
Molecular Mechanisms of Pro-apoptotic

Action of Cyclosporin A in C6 Glioma Cells
It was reported by Mosieniak et al. (1997) that
cyclosporin A at concentrations at the range of 30
60 M inhibits proliferation of rat C6 glioma cells
and induces cell death. CsA-induced cell death was an
25 Glioblastoma: Anti-tumor Action of Cyclosporin A and Functionally Related Drugs
243
Fig. 25.1 Mechanism of immunosuppressant action (a) Immunosuppressants bind to immunophilines: CsA to cyclophilin
(cyc) and FK506 to FKBP (FK506-binding protein); subsequently complexes bind to the calcineurin and inhibit its activity.
Calcineurin is a calcium- and calmodulin-dependent phosphatase which dephosphorylates NFAT transcription factors
allowing them to translocate to the nucleus, where NFAT proteins in cooperation with other transcription factors (f.e. AP-1)
regulate the expression of genes coding for chemokines,
cytokines and their receptors. The immunosuppressantimmunophilin complex may interfere with MAPK signaling
pathways. (b) Summary of pro-apoptotic mechanisms induced

by CsA in C6 glioma cells. CsA induces activation of JNK
(c-Jun amino-terminal kinase) and MKK3 (MAP kinaseactivated protein kinase)-p38 MAPK signaling pathways. It
leads to activation of AP-1 and transcriptional up-regulation of
FasL (ligand Fas) which binds receptor Fas expressed on glioma
cells and induces cell death. Activation of p38 MAPK signaling
leads to accumulation of p53 and induction of p53-dependent
expression of pro-apoptotic genes involved in mitochondrial
death pathway
active process requiring expression of new genes and

proteins, with typical features of apoptosis: oligonucleosomal DNA fragmentation and caspase 3 activation (Mosieniak et al., 1997; Pyrzynska et al., 2000).
Two major mechanisms responsible for induction of
apoptotic death by CsA were identified. Apoptotic cell

death induced by CsA was associated with a persistent activation of mitogen activated protein kinases
(MAPK), in particular cJun N-terminal kinase (JNK)
and p38 MAPK. Prolonged activation of JNK led
244
to accumulation of phosphorylated c-Jun and ATF2 (main substrates of JNK) and formation of the
AP-1 transcription factor followed by transcriptional
activation of the Fas Ligand expression (Pyrzynska
et al., 2000). Further studies with promoter constructs
depleted of DNA binding sites for particular transcription factors revealed that activation of the FasL
gene promoter was only partially AP-1-dependent
and collaborative action of other transcription factors
was required for promoter activation. It was demonstrated that CsA down-regulates Akt signaling to facilitate activation of Forkhead family members resulting
in transcriptional activation of the FasL expression
(Ciechomska et al., 2003). Down-regulation of Akt signaling was necessary to permit Forkhead transcription
factor translocation to the nucleus and pre-requisite
to transcriptional activation of the FasL expression.
Treatment of glioma cells with lower doses of CsA
(110 M) was sufficient to reduce Akt phosphorylation and sensitized cells to doxorubicin, and UVC
treatments (unpublished).
Another event resulting from prolonged activation of MAP kinases, in particular p38 MAPK, was
accumulation of the tumor suppressor p53 in glioma
cells. The p53-family of transcription factors consists
of three genes p53, p63, and p73 sharing significant structural and functional similarities. p53 is a
potent inducer of apoptosis and tumor suppression.
Many anti-cancer agents, from traditional chemo- and
radiation therapies to more recently developed small
molecules, exert their effects by enhancing the antiproliferative effects of p53 and transactivating p63/p73
proteins. In normal cells the p53 is expressed at
low, constitutive level and localized predominantly in
cytoplasm. The latent form of p53 is stabilized and
activated by posttranslational modifications. The activation of p53 occurs in response to DNA damage or
stress such as hypoxia, nutrients or nucleotide deprivation. p53-mediated cell cycle arrest is largely brought
about by induction of p21Waf1, an inhibitor of cyclindependent kinases. Activation of p53 may also result
in apoptosis via transcriptional activation of a number of pro-apoptotic proteins including Bax, Fas, p85,
IGF-BP3, and PIG3 and apoptotic protease activating
factor-1 (apaf-1).
It was reported by Pyrzynska et al. (2002) that CsA
treatment results in up-regulation of p53 protein level
and its accumulation in cell nuclei. Concomitantly,
the levels of p21Waf1 and Bax proteins increased,
B. Kaminska et al.
and Bcl-xL decreased in CsA-treated glioma cells.

Bax protein translocated to mitochondria, as revealed
by immunofluorescence and double staining with a
mitochondrial marker, and likely induced mitochondrial apoptotic pathway. Contribution of p53 to CsAinduced cell death was further confirmed in experiments, in which glioma cells stably transfected with a
mutant p53 (p53Val135) failed to increase p21 and Bax
protein levels and were less sensitive to CsA-induced
apoptosis. Also primary fibroblasts from p53-/knockout mice were significantly more resistant to
CsA-induced apoptosis compared to their corresponding counterparts containing functional p53 (Pyrzynska
et al., 2002).
Accumulation and activity of p53 can be regulated by phosphorylation that occurs at several Ser and
Thr residues, and a number of cellular kinases have
been proposed to directly phosphorylate p53, including
casein kinase I, casein kinase II, double-strandedRNA-dependent protein kinase, ATM, CDK7, DNAactivated protein kinase, Jun-NH2 kinase and p38
MAP kinase. The induction of cell death by CsA was
associated with a persistent activation of MKK3-p38
MAP kinase signaling pathway. Overexpression of a
dominant negative form of MKK3, an upstream activator of p38, abrogated phosphorylation of p38 MAPK
and p53 accumulation (Pyrzynska et al., unpublished).
Together, the compelling evidence demonstrate that the
apoptotic program activated by CsA can be mediated
by multiple pathways: via activation of p53 transcription factor and p53-mediated apoptosis, and through
up-regulation of FasL expression by JNK-AP-1 and
Forkhead (Fig. 25.1).
Cyclosporin A Induces Cell Death

or Growth Arrest and Senescence
of Human Glioblastoma Cells
In rat C6 glioma cells apoptotic cell death induced by
CsA was associated with induction of intrinsic death
pathway: p53- and mitochondria-dependent, as well as
extrinsic, death ligand-dependent pathway. However,
the experiments performed on glioma cells lacking a
functional p53, namely the cell line stably expressing
the temperature-sensitive p53 mutant, indicated that up
to 20% of cells died even in the absence of functional
p53, suggesting also a p53-independent mode of CsA

action.
Therefore, we tested efficacy of CsA towards
three human glioblastoma cell lines that are radioand chemotherapy resistant, in part due to mutations
in the tumor suppressor PTEN or/and TP53 genes:
T98G (mutated PTEN and TP53), U373-MG (mutated
TP53 and PTEN) and U87-MG cells (wild type
TP53/mutated PTEN). Cultured cells were exposed
to increasing concentrations of CsA (Fig. 25.2) and
30 g/ml CsA affected growth of all studied glioma
cells, produced dramatic changes in cell number and
morphology in 24 h. The inhibitory effect of CsA on
cell growth and survival was dose-dependent, and progressed with the time of exposure. CsA at the concentration range of 520 g/ml had no visible cytotoxic
effects (Zupanska et al., 2005). Morphological alterations induced by 30 g/ml CsA were characterized
by shrinkage of the cells, rounding up of the cell body
and detachment from the bottom of the plate at 24
48 h after the treatment (Fig. 25.2). Cell death induced
by CsA was blocked by cycloheximide (a protein synthesis inhibitor) and showed no signs of necrosis, such
as swelling or disruption of cells. The appearance
of numerous large vacuoles was observed in CsAtreated T98G cells and to less extent in U373-MG cells
(Zupanska et al., 2005).
Features of CsA-triggered cell death in human
glioblastoma cells do not fulfill all criteria of apoptosis.
For example, biochemical hallmarks of apoptosis such
as phosphatidylserine exposure in the external leaflet
of the plasma membrane bilayer, the ladder-like
oligonucleosomal DNA fragmentation or appearance
of the subG1 population were not detected, although
condensation of chromatin, deformation of nuclei
and sparse TUNEL labeling were observed. Cells in
CsA-treated cultures showed bean shaped nuclei
with condensed chromatin or nuclei with irregular
clumps of dense chromatin. Such nuclear alterations
as bean shaped nuclei with condensed chromatin
usually preceded oligonucleosomal DNA fragmentation in CsA-treated rat glioma cells (Mosieniak et al.,
1997). In T98G glioblastoma cells nuclear alterations
were completely abolished by cycloheximide (CHX)
treatment, indicating dependency of cell death on de
novo protein synthesis. Measurement of changes in the
mitochondrial membrane potential in response to CsA
treatment with the fluorescent probe JC-1 did not show
alterations in mitochondrial potential, which excludes
245
a possibility of initiation of mitochondrial pathway

along with apoptosome formation (Zupanska et al.,
2005).
Western blot analysis using specific antibodies recognizing intact and cleaved caspases 7, 3 and PARP
(poly (ADP-ribose) polymerase-1) was employed to
further characterize cell death induced by CsA. The
treatment of human glioblastoma cells with 30 M
CsA resulted in activation of the caspase cascade (as
evidenced by the appearance of cleaved caspase 7, 3
and cleaved PARP) in T98G cells, and to lesser extent
in U373-MG cells. No caspase activation was detected
in U87 glioma cells (Fig. 25.2).
Growth arrest may lead to replicative senescence
or differentiation. Senescent cells in contrast to presenescent, proliferating or quiescent cells express
a beta-galactosidase activity detectable at pH 6.0.
Since flatten morphology of CsA-treated U87 cells
and reduction of the number of proliferating cells
suggested the induction of differentiation process, a
senescence-associated -galactosidase (SA-beta-Gal)
staining was performed (Fig. 25.2). Untreated U87MG and T98G cells demonstrated very rare cells
positive for the SA-beta-Gal staining. After 24 h of
CsA treatment the percentage of cells positive for
SA-beta-Gal was doubled in U87-MG cells compared
to control, while only few T98G cells were positively stained. SA-beta-Gal positive U87-MG cells
increased in size and flattened out, thereby attaining
morphology of senescent-like cells (Zupanska et al.,
2005).
Cell proliferation is controlled by cell cycle
regulatory factors which include cyclins and cyclindependent kinases. p21WAF1/Cip1 protein is an
universal inhibitor of cyclin kinases and plays an
important role in inhibiting cell proliferation. The
levels of p21WAF1/Cip1 protein were increased after
exposure to 30 M CsA in all examined glioblastoma
cell lines. The highest up-regulation of p21WAF1/Cip1
protein was observed in T98G cells and the moderate
accumulation was detected in U373-MG and U87MG cells. Further studies revealed transcriptional
activation of p21WAF1/Cip1 expression preceded in
time by a long-term activation of ERK1/2 signaling,
subsequent c-Jun phosphorylation and accumulation of AP-1 complex in CsA-treated glioblastoma
cells. Pre-treatment with ERK pathway inhibitors or
overexpression of dominant negative mutants MKK1,
ERK2 and c-Jun reduced the p21WAF1/Cip1
246
Fig. 25.2 CsA induces programmed cell death or growth

arrest/senescence of malignant glioblastoma cells. (a)
Morphological alterations induced by CsA treatment in human
glioma cells. T98G cells cultured in DMEM with 10% fetal
bovine serum, were exposed to 30 M CsA alone or with
1 g/ml cycloheximide (CHX). Upper panel shows that a majority of cells lost processes and became round in CsA-treated
cultures. Lower panel shows Hoechst 33258 staining revealing deformations of cell nuclei and condensation of chromatin.
CHX inhibits morphological alterations in CsA-treated T98G
cells. (b) Caspase cascade activation in CsA-triggered cell death.
B. Kaminska et al.
A representative immunoblot shows caspase activation in total

protein extracts of human glioma cells treated with 30 M CsA.
Specific antibodies recognizing intact and cleaved caspases 7,
3 and cleaved PARP (Cell Signaling, USA) were employed.
(c) Detection of the Senescence-Associated beta-galactosidase
(SA-beta-Gal) staining in CsA-treated U87-MG glioma cells.
The substantial increase in cellular volume and the blue staining
reflecting SA-beta-Gal activity was observed in U87-MG cells
treated with CsA. Original magnification for larger photos is
10; for insets 20
expression, confirming involvement of this pathway.

Transcriptional activation of p21WAF1/Cip1 expression by CsA was independent of p53 and preceded
CsA-induced growth arrest in glioblastoma cells
(Zupanska et al., 2007).
In conclusion, the findings above presented demonstrate an ability of CsA to induce growth arrest or
programmed, but non-apoptotic cell death in human
glioblastoma cells at the concentration of 30 M or
higher. Many forms of programmed cell death different
from classical apoptosis have been described by the
criteria of morphology, biochemistry, and response to
apoptosis inhibitors, particularly in transformed cells
which contain endogenous inhibitors preventing a particular pathway.
Anti-tumor Effects of CsA in Glioma

Models
Good cytostatic and cytotoxic efficacy of CsA in rat
and human glioblastoma cultures encouraged studies
in more complex models: in organotypic brain slice
cultures and in murine glioma model. Organotypic
brain slice cultures injected with glioblastoma cells
recapitulate many features of glioblastoma and are
very useful for investigating the cellular and molecular mechanisms of glioma invasion under conditions
most analogous to those of normal brains in vivo. It
was reported by Markovic et al. (2005) that microglial
cells contribute significantly to invasion of glioma cells
in cultured brain slices. Microglia are the intrinsic
immune cells of the brain, serving principally to control the innate and the adaptive immune responses in
the central nervous system, to initiate host-defense
and tissue repair mechanisms. Microglial cells are
attracted towards glioma (glioma tissue consists of up
to 30% of microglial cells) and microglia density in
gliomas positively correlates with malignancy, invasiveness and grading of the tumors (Watters et al.,
2005). When glioma cells were injected into brain
slices depleted of endogenous microglia (by incubation with clodronate-filled liposomes for 96 h), the
invasiveness of the tumors was significantly decreased.
Inoculation of exogenous microglia together with
glioma cells into cultured brain slices increased the
infiltrative behavior of glioma cells. Experiments with
247
co-culture of microglia with glioma cells revealed that

soluble factors released from glioma cells strongly
stimulate metalloprotease-2 (MMP-2) activity increasing breakdown of extracellular matrix and thereby
promoting tumor invasiveness (Markovic et al., 2005).
Further studies showed that membrane type 1 metalloproteinase (MT1-MMP) is up-regulated in gliomaassociated microglia. Microglial MT1-MMP in turn
activates glioma-derived pro-MMP-2 and promotes
glioma expansion. Tumor growth and invasion in ex
vivo model using MT1-MMP deficient brain tissue and
in microglia-depleted animals were strongly reduced
(Markovic et al., 2009).
It was reported by Sliwa et al. (2007) that migration/invasion of fluorescently labeled GL261 glioma
cells in murine brain slices significantly decreased
upon treatment with CsA, even at low concentrations: 1, 10 and 30 M. Anti-invasive effects of
lower, non-cytotoxic doses of CsA suggested an existence of the additional mechanism of CsA action
on tumor invasion. When glioma cells were injected
into slices devoid of endogenous microglia, the
inhibitory effect of 1 M CsA on glioma invasiveness mostly vanished. It indicated that CsA abolishes
microglia-promoting effects on glioma malignancy.
The inhibitory action of CsA on microglia function
has been directly demonstrated in microglia-glioma
co-cultures. Glioma-derived factors in co-cultures
or glioma conditioned medium induce morphological transformation of microglial cells into amoeboid
phagocytes, activate MAPK signaling pathways and
production of some cytokines. CsA at low doses of
0.1. and 1 M blocked transformation of microglial
cells stimulated by glioma-conditioned medium and
abolished pro-invasive effects of microglia (Sliwa
et al., 2007). A recent study points to a crucial
role of microglia-derived transforming growth factor beta 1 (TGF-1) in regulation of glioma invasion
(Wesolowska et al., 2008). Blockade of TGF-1 signaling by silencing of TGF-1 type II receptor with
shRNA or neutralizing anti-TGF-1 antibody abolished the promoting effect of microglia on glioma
invasion.
In vivo tumor models developed by intracranial
or subcutaneous implantation of glioma cell lines
in rodents are used to test novel therapies. The
advantages of these glioma models are their highly
efficient gliomagenesis, reproducible growth rates,
and knowledge of tumor location. However, these
248
models have been criticized for not recapitulating

all pathological features of human glioblastoma multiforme (GBM), they allow to target different features of GBM, including location in the brain, invasion of brain parenchyma, angiogenesis, and secretion of immune suppressive molecules. In studies
by Sliwa et al. (2007) fluorescently labeled GL261
glioma cells were implanted into the striatum of
DBA/2 J mice and developed gliomas of considerable size in 14 days. CsA (Sandimmun, Novartis)
was administrated intraperitoneally (i.p.) every second days at doses of 2 or 10 mg/kg of body weight
starting from the second day after cell inoculation.
Systemically applied 2 or 10 mg/kg CsA significantly decreased tumor volumes (by 70%) with similar
efficacy.
Further studies in syngenic murine models, such
as GL261 mouse glioma cells implanted to C57BL6
mice, have confirmed anti-tumor action of CsA administrated intraperitoneally every second day at doses
of 2 and 10 mg/kg. Immunofluorescence studies
with Iba1 antibody demonstrated that CsA blocks
accumulation and activation of glioma-associated
microglia confirming the in vitro data (unpublished).
It was reported that macrophagecolony stimulating factor (M-CSF) and other cytokines secreted
by high-grade glioma cells stimulate differentiation of tumor-infiltrating microglia/macrophages into
cells acquiring the anti-inflammatory (M2) phenotype. M-CSF was significantly correlated with histological malignancy and with the proportion of
M2 microglia/macrophages in vivo. Such M2 cells
instead of initiating immune responses support tumor
growth and invasion. CsA with its strong effect
on pro-tumorigenic activity of microglia could be
effective drug in modulation of anti-inflammatory
M2 phenotype of microglia/macrophages in glioma
therapy.
CsA has been used to block the immune reaction towards human glioblastoma cells transplanted
to rodent brains. It was reported by Mathiesen et al.
(1989) that human glioblastoma cells implanted to
cerebrum of rat hosts survive longer in immunosuppressed animals but aggressive growth of glioblastoma cells and tumor proliferation was not observed.
Human glioblastoma cells implanted into the brain
of Wistar rats immunosuppressed with CsA applied
at 12 mg/kg/daily survived under such conditions but
formed dense and poorly diffused tumors (Strojnik
B. Kaminska et al.
et al., 2006). Up to now CsA has not been

tested as anti-invasive/cytotoxic agent in clinical
trials.
Anti-tumor Action of Rapamycin

and Derivatives
Mechanism of Action of Rapamycin
Rapamycin (Sirolimus) is a macrolide discovered
1970 as a product of the bacterium Streptomyces
hygroscopicus in a soil sample from the Easter Island
(Rapa Nui). Rapamycin was originally developed as
an antifungal agent, but turned out to be a potent
immunosuppressive drug and since 1997 has been
used to prevent host-rejection in kidney transplantation. An immunosuppressive effect of rapamycin is
due to the inhibition of interleukin 2 (IL-2)-mediated
T-cell proliferation and activation. The mode of action
of rapamycin is to bind the cytosolic immunophilin
FK-binding protein 12 (FKBP12). This complex
inhibits the mammalian target of rapamycin (mTOR)
pathway by directly binding the mTOR Complex
1 (mTORC1). mTOR is a kinase playing a key
role in the regulation of cell growth and proliferation by regulating ribosomal biogenesis and protein translation. Among others, mTOR can be activated by growth factors and hormones, which is
mediated by the induction of PI-3-kinase (PI-3K).
Active PI-3K generates phosphatidylinositol (3,4,5)trisphosphate [PtdIns(3,4,5)P3 ] which activates kinase
Akt phosphorylating and inactivating the tuberous
sclerosis complex 2 (TSC2). TSC2 acts as a GTPaseactivating protein for the small GTPase RHEB (Ras
homolog enriched in brain). Inhibition of TSC2 activity leads to elevated RHEB-GTP levels and activation
of mTORC1 which by its downstream effectors, such
as ribosomal protein S6 kinases (S6K) and 4E-binding
protein (4E-BP1), further up-regulates ribosome biogenesis and protein translation (Fig. 25.3). This results
in an increase in cell size and mass, and enhanced
proliferation. On the other hand, inhibition of mTOR
promotes autophagy in some cell types. Autophagy
is a process of degradation of organelles and macromolecules to supply cells with nutrients and can function as a cytoprotective mechanism or alternative cell
death process.
249
Fig. 25.3 Rapamycin inhibits proliferation or induces cell death

of malignant glioblastoma cells. (a) mTOR signaling network.
mTORC1 stimulates cell growth upon activation by growth factors or insulin (for details see the text). (b) Rapamycin inhibits
the kinase activity of mTOR measured by reduction of S6K
phosphorylation in T98G cells. Western blot analysis of the
phospho-S6K (T389) level in T98G glioma cells after treatment
with 10 nM rapamycin for 24 h. S6K served as a protein loading
control. (c) Rapamycin affects proliferation of T98G cells. Cell
proliferation was determined 24 h after exposure to 1, 10, 30
or 60 M rapamycin using BrdU test. Bars represent the ratio

of proliferating cells treated with the inhibitor related to control cells (means SEM of three independent experiments, each
in triplicate). Statistical analysis was done by one-way ANOVA
followed by Newman-Keuls test, p < 0.001. (d) Rapamycin
induces morphological alterations in T98G cells. Phase-contrast
images of T98G glioma cells performed 24 h after treatment with
10, 30 or 60 M rapamycin. Cell death was observed after using
the highest drug concentration. Original magnification is 10
Rapamycin and Its Analogs in Pre-clinical

and Clinical Trials in Glioblastomas
mTOR is a crucial downstream component of the

PTEN/Akt signaling, pre-clinical studies with pharmacological inhibitors of the mTOR pathway were
initiated in glioblastomas. Apart from rapamycin, its
analogs with improved pharmacokinetic properties
were synthesized. CCI779 (temsirolimus; Wyeth),
RAD001 (everolimus; Novartis) and AP23573 (Ariad)
are currently being tested as potential targeted drugs
in glioblastomas. Pharmacological inactivation of
mTOR decreased tumor cells proliferation and tumor
size in PTEN-deficient mice (Rajasekhar et al., 2003).
Moreover, xenografts generated from PTEN null
U87 glioma cells in mice had inhibited growth and
reduced proliferation rate after rapamycin treatment.
Enhanced or constitutive activation of the

PI3K/Akt/mTOR pathway is an important factor in
gliomagenesis (Guertin and Sabatini, 2007). In human
glioblastomas Akt is activated in approximately 70%
of tumors due to loss of PTEN (phosphatase and
tensin homolog deleted on Chromosome 10), the
tumor suppressor and negative regulator of PI-3 K/Akt
signaling, and/or up-regulation of epidermal growth
factor receptor (EGFR) and platelet-derived growth
factor receptor (PDGFR) tyrosine kinases. PTEN
mutations are present in 2040% of GBM. Since
250
Conversely, rapamycin did not decrease size of PTEN

wild-type LN229 tumors (Wei et al., 2008). These
findings provided grounds to initiate clinical trials of
mTOR inhibitors in glioblastomas.
The results of six studies using sirolimus or
CCI-779 (a dihydroxymethyl propionic acid ester of
sirolimus) as a single agent, or sirolimus and RAD001
in combination with inhibitors of EGFR in patients
with recurrent glioblastoma multiforme (GBM) are
shown in Table 25.1. Treatment with CCI-779 or
sirolimus alone had only a limited activity in recurrent
GBM, particularly in unselected patients. In the phase
II study by Chang et al. (2005) CCI-779 was administered weekly to forty-three patients with recurrent
GBM (14 patients were on enzyme-inducing antiepileptic drug). Initially CCI-779 was administered at a
dose of 250 mg intravenously but the dose was reduced
to 170 mg because of side effects. CCI-779 was well
tolerated; however, failed to demonstrate any efficacy
as a single agent in patients with recurrent GBM.
Despite initial disease stabilization in approximately
50% of patients, the durability of response was short.
In another phase II study CCI-779 was administered in
a 250-mg intravenous dose weekly to sixty-five recurrent GBM patients with 1 chemotherapy regimen
(Galanis et al., 2005). The study reported that 20 of 65
patients with recurrent GBM (36%) had radiographic
improvement. Progression-free survival at 6 months
was 7.8% and median overall survival was 4.4 months.
Median time to progression for all patients was 2.3
months and was significantly longer for responders
B. Kaminska et al.
(5.4 months) versus non-responders (1.9 months). The

authors assessed activation of the PI3K pathway by
examining tumor specimens for total/phosphorylated
Akt and p70s6 kinase, and found p70s6 kinase staining
indices being significantly more frequent in responders
versus non-responders (Galanis et al., 2005).
It was reported by Cloughesy et al. (2008) that
rapamycin treatment leads to substantial inhibition of
tumor cell proliferation in seven of 14 patients, as
demonstrated by reduction of Ki-67 staining in a subset of patients with PTEN loss. Inhibition of mTOR
signaling was observed in tumor tissue. However,
rapamycin led to the activation of Akt in some patients,
which correlated with faster tumor progression the
authors concluded that rapamycin has anti-cancer
activity in PTEN-deficient glioblastoma and deserves
further clinical study alone or in combination with
PI3K pathway inhibitors.
More recently, based on evidence of the synergism
between inhibitors of mTOR and EGFR (Rao et al.,
2005), several studies investigated such a combination.
In two trials a partial radiological response (50%
decrease in the product of perpendicular diameters
of contrast enhancing mass without new lesions) was
seen. Doherty et al. (2006) reported a trial on 22 GBMs
and six anaplastic gliomas patients treated with either
gefitinib 500 mg or erlotinib 150 mg orally/daily in
combination with sirolimus administered at a dose of
6 mg orally the first day followed by 4 mg orally once
daily thereafter. Both medications were given daily
in 28-day cycles. Out of this cohort, 19% of patients
Table 25.1 Results of targeted therapy trials of mTOR inhibitors used alone or in combination with inhibitors of EGFR in recurrent
glioblastomas
Radiological
Median PFS
6-month
Median OS
Study
Treatment
Patients (n) response (%) (months)
PFS (%)
(months)
Chang et al. (2005)
CCI-779, 170250 mg
43
5
weekly
Galanis et al. (2005) CCI-779, 250 mg weekly
65
0
Cloughesy et al.
Sirolimus, 2, 5, or 10 mg
15
14
(2008)
daily
Doherty et al.
Sirolimus, 4 mg daily;
22
18
(2006)
gefitinib, 500 mg, or
erlotinib, 150 mg daily
Kreisl et al. (2009)
RAD001, 70 mg weekly;
22
14
gefitinib, 250 mg daily
Reardon et al.
Sirolimus, 510 mg daily; 32
0
(2010)
erlotinib, 150450 mg
daily
PFS, progression-free survival; OS, overall survival; NA, not available
2.3
NA
2.3
NA
8
NA
4.4
NA
25
NA
2.6
5.8
1.8
8.5
experienced a partial response and 50% had stable

disease; 6-month progression-free survival (6 M-PFS)
was 25%. In a second study on 22 patients (all had
received prior radiation and chemotherapy) receiving gefitinib (250 mg daily) and everolimus (70 mg
weekly), 14% of patients had a partial response and
36% of patients stable disease; median overall survival was 5.8 months (Kreisl et al., 2009). Overall
response rate was 19% and a 6 M-PFS of 25% in
GBM patients. No molecular markers of response were
described. A retrospective review of eight consecutive
negative phase II trials in recurrent malignant gliomas
from the M.D. Anderson Cancer Center found a 6 MPFS for GBM of 15%. For comparison, among GBM
patients treated with temozolomide at 200 mg/m2 /day
orally for the first 5 days of a 28-day cycle, PFS
at 6 months was 18%; median progression-free survival and median overall survival were 2.1 and 5.4
months, respectively. The 6-month survival rate was
46% (Brada et al., 2001).
Although the trials demonstrate moderate clinical
benefits from the combination of two drugs, ongoing phase I/II studies are investigating the efficacy of
using mTOR inhibitors with other targeted agents and
chemoradiation, namely: sirolimus plus vandetanib,
CCI-779 or RAD001 plus temozolomide, CCI-779
plus radiation or sorafenib or bevacizumab or perifosine, RAD001 plus AEE788, and RAD001 plus Gleevec
and hydroxyurea (http://clinicaltrials.gov).
General Considerations
A use of CsA may raise reservations due to employment of drug blocking the immune system in tumor
patients, poor drug accessibility due to blood-brain
barrier or general toxicity. However, a growing
evidence shows both innate and adaptive immunity impaired in glioblastomas (Yang et al., 2010).
Glioblastomas are poorly immunogenic and do not
express specific tumor antigens (Watters et al., 2005).
Microglia infiltrating glioblastomas are converted into
tumor supportive cells and contribute to tumor growth
(Markovic et al., 2005, 2009; Sliwa et al., 2007).
Microglial cells operate as the first line innate and
adaptive immunity of the central nervous system
(CNS). In the normal CNS, microglia express low
levels of major histocompatibility complex (MHC)
251
class I and class II molecules and co-stimulatory

molecules such as CD86 and CD40. Upon activation, microglia convert to an active phenotype, upregulate MHC class I and class II and co-stimulatory
molecules and take part in CD4-and CD8-specific
T cell responses. Furthermore, malignant glioblastoma cells secrete chemoattractants and growth factors, including monocyte chemoattractant protein
1 (MCP-1), colony stimulating factor 1 (CSF-1),
granulocyte-macrophage colony stimulating factor
(GM-CSF), hepatocyte growth factor (HGF) and
recruit microglia. Even though glioblastoma accumulate many microglia, macrophages and a small population of lymphocytes, the defense mechanisms are
down-regulated. It was reported by Hussain et al.
(2006) that glioblastoma-infiltrating microglia isolated from human tumor tissue samples do not produce inflammatory cytokines: interleukin 1 (IL-1),
interleukin 6 and tumor necrosis factor (TNF-),
cytokines critical for developing effective innate
immune responses.
On the contrary, glioma-associated microglia/
macrophages might promote tumor growth by inducing immunosuppression in the tumor microenvironment. Glioblastoma cells and accumulating microglia
release several cytokines, such as interleukin 10 (IL10) and TGF- which may inhibit T-cell activation
and contribute to the local glioma immunosuppressive milieu (Watters et al., 2005). An effective antitumor T-cell response is decreased in glioblastomas,
because expression of MHC class II and co-stimulatory
B7 molecules is reduced and microglia appear deficient in proper antigen presentation. Thus, potential
immunossuppressive effects of CsA can be neglected
because the immune system in glioblastoma patients
is already paralyzed and anti-tumor responses are nonfunctioning. Although, tumors incidence in transplant
recipients is a recognized consequence, a recent examination of reported cases of tumor in the transplant
population of the Israel Penn International Transplant
Tumor Registry shows that primary brain tumors,
including gliomas, do not appear to be overrepresented
in the Registry, indicating that they may not arise with
increased frequency in transplant recipients (Schiff,
2004).
As it has been shown in animal glioma model by
Sliwa et al. (2007), CsA has a systemic effect, can
affect cell interactions in tumor microenvironment and
block pro-tumorigenic activity of tumor infiltrating
252
microglia/macrophages. A recent study demonstrates

that mTOR signaling controls microglial activation in
response to cytokines and appears to play a crucial role
in regulating of microglial viability (Dello Russo et al.,
2009). Thus anti-tumor action of rapamycin in vivo
may be partly due to interference with cell interactions
in tumor microenvironment.
The blood-brain barrier (BBB) possesses a significant impediment for the delivery of therapeutic drugs
into the brain. In the normal adult brain, the BBB
mainly consists of vascular endothelial cells, astrocytes
and pericytes. In malignant glioblastomas, the blood
vessel network is broken down. Increased permeability of tumor blood vessels is induced by angiogenic
factors released by malignant glioma cells, such as
vascular endothelial growth factor (VEGF) that is associated with intravascular migration of glioma cells into
blood vessels and their transfer to distant areas in the
brain via blood flow. The mechanism of vascular permeability is used by glioma cells to facilitate migration
and invasiveness, but can be also employed to facilitate
delivery of anti-tumor medicines into the brain (Tate
and Aghi, 2009).
Previous studies demonstrated that a systemic injection of 100 mg/kg/i.p. CsA results in a blood concentration of 10 M maintained for at least 18 h
(Ciechomska et al., 2005). Thus, concentrations of
CsA (2 and 10 mg/kg) tested in animal glioma model
correspond to blood concentrations of 0.2 or 1 M,
respectively. A permeable blood-brain barrier in brain
tumors may result in similar CsA concentrations in the
brain as detected in the blood. In contrast to widely
accepted notion that CsA does not cross or poorly cross
the blood-brain barrier, studies on immunosuppressed
patients demonstrated the presence of CsA in cerebrospinal fluids suggesting that the drug is able to cross
the blood-brain barrier. Furthermore, CsA may impair
the brain endothelial barrier function by accelerating
NO production in the brain endothelial and astroglial
cells.
Notably, CsA at lower micromolar concentrations
does not affect survival of non-transformed cells.
Neurons from mixed neuronal-glial cultures developed from hippocampal dentate gyrus are affected
by CsA at the concentrations higher than 810 M
(Kaminska et al., 2001). Astrocytes are more resistant
and CsA at concentration of 40 M or higher affects
the survival of astrocytes from neonatal brain cultures.
Altogether, these results suggest that CsA, particularly
B. Kaminska et al.
at lower concentrations affecting pro-invasive activity of microglia, could be effective anti-tumor drug
without inducing neurotoxicity. CsA induces cell death
via multiple mechanisms and some of them are
able to prevail alterations of growth regulatory and
apoptotic pathways, caused by common mutations in
human glioblastomas. Rapamycin and its derivatives,
besides the well described inhibition of mTOR pathway in glioblastoma cells, may additionally affect proinvasive activity of glioblastoma infiltrating microglia.
The unique mechanism of action of CsA and pharmacological inhibitors of the mTOR pathway justifies
further research on their anti-tumoral properties either
alone or in combined therapies.
Acknowledgements This work was supported by grant
P-N/024/2006.
References
Brada M, Hoang-Xuan K, Rampling R, Dietrich PY, Dirix LY,
Macdonald D, Heimans JJ, Zonnenberg BA, Bravo-Marques
JM, Henriksson R, Stupp R, Yue N, Bruner J, Dugan M,
Rao S, Zaknoen S (2001) Multicenter phase II trial of temozolomide in patients with glioblastoma multiforme at first
relapse. Ann Oncol 12:259266
Chang SM, Wen P, Cloughesy T, Greenberg H, Schiff D,
Conrad C, Fink K, Robins HI, De Angelis L, Raizer J,
Hess K, Aldape K, Lamborn KR, Kuhn J, Dancey J, Prados
MD (2005) Phase II study of CCI-779 in patients with
recurrent glioblastoma multiforme. Invest New Drugs 23:
357361
Ciechomska I, Legat M, Golab J, Wesolowska A, Kurzaj Z,
Mackiewicz A, Kaminska B (2005) Cyclosporine A and its
non-immunosuppressive derivative NIM811 induce apoptosis of malignant melanoma cells- in vitro and in vivo studies.
Int J Cancer 117:5967
Ciechomska I, Pyrzynska B, Kazmierczak P, Kaminska B
(2003) Inhibition of Akt kinase signalling and activation
of Forkhead are indispensable for up-regulation of FasL
expression in apoptosis of glioma cells. Oncogene 22:
76177627
Clarke J, Butowski N, Chang S (2010) Recent advances in
therapy for glioblastoma. Arch Neurol 67:279283
Cloughesy TF, Yoshimoto K, Nghiemphu P, Brown K, Dang
J, Zhu S, Hsueh T, Chen Y, Wang W, Youngkin D, Liau
L, Martin N, Becker D, Bergsneider M, Lai A, Green R,
Oglesby T, Koleto M, Trent J, Horvath S, Mischel PS,
Mellinghoff IK, Sawyers CL (2008) Antitumor activity of
rapamycin in a phase I trial for patients with recurrent
PTEN-deficient glioblastoma. PLoS Med 5:e8
Dello Russo C, Lisi L, Tringali G, Navarra P (2009) Involvement
of mTOR kinase in cytokine-dependent microglial activation
and cell proliferation. Biochem Pharmacol 78:12421251

Doherty L, Gigas DC, Kesari S, Drappatz J, Kim R, Zimmerman
J, Ostrowsky L, Wen PY (2006) Pilot study of the combination of EGFR and mTOR inhibitors in recurrent malignant
gliomas. Neurology 67:156158
Galanis E, Buckner JC, Maurer MJ, Kreisberg JI, Ballman
K, Boni J, Peralba JM, Jenkins RB, Dakhil SR, Morton
RF, Jaeckle KA, Scheithauer BW, Dancey J, Hidalgo M,
Walsh DJ (2005) Phase II trial of temsirolimus (CCI-779) in
recurrent glioblastoma multiforme: a North Central Cancer
Treatment Group study. J Clin Oncol 23:52945304
Guertin DA, Sabatini DM (2007) Defining the role of mTOR in
cancer. Cancer Cell 12:922
Hussain SF, Yang D, Suki D, Grimm E, Heimberger AB (2006)
Innate immune functions of microglia isolated from human
glioma patients. J Transl Med 30:1524
Kaminska B, Figiel I, Pyrzynska B, Czajkowski R, Mosieniak G
(2001) Treatment of hippocampal neurons with cyclosporin
a results in calcium overload and apoptosis which are independent on NMDA receptor activation. Br J Pharmacol
133:9971004
Kreisl TN, Lassman AB, Mischel PS, Rosen N, Scher HI,
Teruya-Feldstein J, Shaffer D, Lis E, Abrey LE (2009) A
pilot study of everolimus and gefitinib in the treatment of
recurrent glioblastoma (GBM). J Neurooncol 92:99105
Mancini M, Toker A (2009) NFAT proteins: emerging roles in
cancer progression. Nat Rev Cancer 9:810820
Markovic DS, Glass R, Synowitz M, Rooijen N, Kettenmann
H (2005) Microglia stimulate the invasiveness of glioma
cells by increasing the activity of metalloprotease-2.
J Neuropathol Exp Neurol 64:754762
Markovic DS, Vinnakota K, Chirasani S, Synowitz M, Raguet
H, Stock K, Sliwa M, Lehmann S, Klin R, van Rooijen N,
Holmbeck K, Heppner FL, Kiwit J, Matyash V, Lehnardt S,
Kaminska B, Glass R, Kettenmann H (2009) Gliomas induce
and exploit microglial MT1-MMP expression for tumor
expansion. Proc Natl Acad Sci USA 106:1253012535
Mathiesen T, Collins VP, Olson L, Granholm L (1989)
Prolonged survival and vascularization of xenografted
human glioblastoma cells in the central nervous system of
cyclosporine a treated rats. Cancer Lett 44:151156
McCubrey JA, Steelman LS, Abrams SL, Lee JT, Chang
F, Bertrand FE, Navolanic PM, Terrian DM, Franklin
RA, DAssoro AB, Salisbury JL, Mazzarino MC, Stivala
F, Libra M (2006) Roles of the RAF/MEK/ERK and
PI3K/PTEN/AKT pathways in malignant transformation and
drug resistance. Adv Enzyme Regul 46:249279
Mosieniak G, Figiel I, Kaminska B (1997) Cyclosporin A, an
immunosuppressive drug, induces programmed cell death in
rat C6 glioma cells by a mechanism that involves the AP-1
transcription factor. J Neurochem 68:11421149
Ohgaki H, Kleihues P (2009) Genetic alterations and signaling pathways in the evolution of gliomas. Cancer Sci 100:
22352241
Pyrzynska B, Mosieniak G, Kaminska B (2000) Changes of the
trans-activating potential of AP-1 transcription factor during
cyclosporin a-induced apoptosis of glioma cells are mediated
by phosphorylation and alterations of AP-1 composition.
J Neurochem 74:4251
253
Pyrzynska B, Serrano M, Martinez-A C, Kaminska B

(2002) Tumor suppressor p53 mediates apoptotic cell
death triggered by cyclosporine A. J Biol Chem 277:
1410214108
Rajasekhar VK, Viale A, Socci ND, Wiedmann M, Hu X,
Holland EC (2003) Oncogenic Ras and Akt signaling contribute to glioblastoma formation by differential recruitment of existing mRNAs to polysomes. Mol Cell 12:
889901
Rao RD, Mladek AC, Lamont JD, Goble JM, Erlichman C,
James CD, Sarkaria JN (2005) Disruption of parallel and
converging signaling pathways contributes to the synergistic antitumor effects of simultaneous mTOR and EGFR
inhibition in GBM cells. Neoplasia 7:921929
Reardon DA, Desjardins A, Vredenburgh JJ, Gururangan S,
Friedman AH, Herndon JE 2nd, Marcello J, Norfleet JA,
McLendon RE, Sampson JH, Friedman HS (2010) Phase
2 trial of erlotinib plus sirolimus in adults with recurrent
glioblastoma. J Neurooncol 96:219230
Schiff D (2004) Gliomas following organ transplantation: analysis of the contents of a tumor registry. J Neurosurg 101:
932934
Sliwa M, Markovic D, Gabrusiewicz K, Synowitz M, Glass
R, Zawadzka M, Wesolowska A, Kettenmann H, Kaminska
B (2007) The invasion promoting effect of microglia on
glioblastoma cells is inhibited by cyclosporin A. Brain
130:476489
Strojnik T, Kavalar R, Lah TT (2006) Experimental model and
immunohistochemical analyses of U87 human glioblastoma
cell xenografts in immunosuppressed rat brains. Anticancer
Res 26:28872900
Tate MC, Aghi MK (2009) Biology of angiogenesis and invasion
in glioma. Neurotherapeutics 6:447457
Watters JJ, Schartner JM, Badie B (2005) Microglia function in
brain tumors. J Neurosci Res 81:447455
Wei LH, Su H, Hildebrandt IJ, Phelps ME, Czernin J, Weber
WA (2008) Changes in Tumor Metabolism as Readout
for Mammalian Target of Rapamycin Kinase Inhibition
by Rapamycin in Glioblastoma. Clin Cancer Res 14:
34163426
Wesolowska A, Kwiatkowska A, Slomnicki L, Dembinski
M, Master A, Sliwa M, Franciszkiewicz K, Chouaib S,
Kaminska B (2008) Microglia-derived TGF-beta as an
important regulator of glioblastoma invasion-an inhibition
of TGF-beta-dependent effects by shRNA against human
TGF-beta type II receptor. Oncogene 27:918930
Yang I, Han SJ, Kaur G, Crane C, Parsa AT (2010) The role
of microglia in central nervous system immunity and glioma
immunology. J Clin Neurosc 17:610
Zupanska A, Adach A, Dziembowska M, Kaminska
B (2007) Alternative pathway of transcriptional
induction of p21WAF1/Cip1 by cyclosporine A in p53deficient human glioblastoma cells. Cell Signal 19:
12681278
Zupanska A, Dziembowska M, Ellert-Miklaszewska A,
Gaweda-Walerych K, Kaminska B (2005) Cyclosporine A
induces growth arrest or programmed cell death of human
glioma cells. Neurochem Int 47:430441
Chapter 26
Glioblastoma Patients: Chemotherapy with Cisplatin,

Temozolomide and Thalidomide
Fable Zustovich and Giuseppe Lombardi
Abstract When relapse occurs the treatment of

glioblastoma patients remains a challenge for oncologists, the prognosis is poor and no therapy has proven
benefit on survival. There are preclinical evidences
of a synergism between cisplatin and temozolomide
and between cisplatin and thalidomide. In the contest of a phase I and a subsequent phase II study we
treated 17 patients with relapsed high-grade malignant glioma with cisplatin plus temozolomide and then
with the same combination plus thalidomide. Toxicity
was generally manageable but an incidence of 36.4%
of vascular thrombo-embolic events was observed in
patients treated with thalidomide. Sixteen patients
were evaluable for response: 1 patient had a partial
response, 8 patients had stable disease and 5 patients
had progression. The median time to progression was 4
months (range 0.914.9), the progression-free survival
at 6 month was 28% and the median overall survival
was 7.8+ months (range 2.624.2+). Due to the high
incidence of thalidomide-associated thrombo-embolic
events we continued the phase II study treating patients
only with cisplatin and temozolomide. The availability
of new antiangiogenetic agents as bevacizumab could
in the future improve the efficacy and tolerability of
such combinations.
Keywords Cisplatin Temozolomide Thalidomide
Toxicity AGAT Glioblastoma
F. Zustovich ()
Oncologia Medica 1, I.O.V. IRCCS, Ospedale Busonera,
35128 Padova, Italy
e-mail: fable.zustovich@ioveneto.it
Introduction
Malignant gliomas are relatively rare diseases.
Although prognosis is poor, recent improvement in the
multi-disciplinary approach, such as earlier diagnosis,
better and repeated surgery plus local chemotherapy,
improved radiotherapy and more effective systemic
chemotherapy, has prolonged the survival. Long-term
survivors are more frequent in the clinical practice, the
median overall survival of patients with glioblastoma
has reached the 15 months in phase III studies and ~20
months for selected patients in phase II studies.
In a phase III study, temozolomide at a daily dose of
75 mg/m2 when associated with standard radiotherapy,
and then administered at a dose of 200 mg/m2 day for 5
days every 28 days for at least 6 cycles, has proven to
provide a statistically significant advantage in overall
2-year survival (26.5 vs. 10.4%) in comparison to standard radiotherapy alone in patients with glioblastoma
after surgical resection (Stupp et al., 2005). The further
analysis of the O6 -methylguanine-DNA methyltransferase (MGMT) methylation status in the tumour cells
gave more details on survival results and in particular how patients with methylated-MGMT (45% of the
total) had a median survival of 18.2 months compared
to 12.2 months of the non-methylated counterpart
(Hegi et al., 2005).
There are no phase III trials investigating second line therapy for patients progressing during or
after temozolomide treatment. Surgery or further
chemotherapy is always considered whenever possible. Some data are available using more temozolomide
treatment with modified schedules, the combination
of procarbazine, carmustine and vincristine (PCV regimen), irinotecan and carmustine, fotemustine and

255
256
carmustine, among the other drugs. Unfortunately,

all these attempts of treatment are characterized by
low response rates and the lack of durability when
responses occur. In contrast, the use of inhibitors of
the angiogenetic pathways appeared very promising.
In a phase II study bevacizumab, a monoclonal antibody, targeted against the vascular endothelial growth
factor (VEGF), in combination with irinotecan, gave
an overall response rate (ORR) of 57% in 35 patients
with progressive refractory glioblastoma, the 6-month
progression-free survival (PFS-6) was 46% and the
6-month overall survival was 77% (Goli et al., 2007).
Rationale
Temozolomide is an alkylating agent which belongs
to the family of imidazotetrazine. It can be orally
used, and permeates through the brain-blood barrier.
Temozolomide has an excellent tolerability profile at
the usual dose of 200 mg/m2 a day for five consecutive days every 4 weeks. The dose limiting toxicity is
thrombocytopenia. Initial phase II studies in patients
with relapsed glioblastomas showed that temozolomide given in monotherapy was able to obtain an ORR
of 19% and a PFS-6 of 24% (Brandes et al., 2000).
Temozolamide acts through methylation of the O6
position of guanylic acid in DNA with intrachain links.
Consequent mismatch-repair system activation leads
to apoptotic cell death. Resistance to temozolomide is
due to high intracellular levels of O6 -alkyl-guaninealkyltransferase (AGAT) which is involved in DNArepair mechanisms, transferring an alkyl group from
DNA to another cysteine residue. Therefore, cells with
increased concentrations of AGAT or with a deficiency
in mismatch repair before drug administration, may be
resistant to temozolomide (DIncalci et al., 1988).
Cisplatin is able to reduce in vitro intracellular
AGAT levels in proportion to the drug concentration
and the duration of cell exposure to the drug with
an AGAT nadir after 2448 h (DAtri et al., 2000).
The promoting region of AGAT gene is rich in guanine and cytosine (GC) sequences that have a high
affinity for cisplatin. The presence of cisplatin reduces
the activity of the gene encoding for AGAT and then
impair one of the potential mechanisms of resistance
to temozolomide.
F. Zustovich and G. Lombardi
This is the basis for a synergistic effect between

temozolomide and cisplatin. Furthermore, platinum
compounds have their own cytotoxic activity against
malignant gliomas (Yung et al., 1991; Bertolone et al.,
1989).
A combination of cisplatin and temozolomide in
chemo-nave patients with glioblastoma was investigated in a first-line phase II study. Cisplatin was
administered at 75 mg/m2 every 28 days followed by
oral temozolomide at a dose of 130 mg/m2 bolus followed by nine doses of 70 mg/m2 every 12 h escalated
to 1000 mg/m2 in 5 days if no toxicity occurred. The
ORR was 20.4% with a PFS-6 of 34% (Brandes et al.,
2004).
Thalidomide has a well known anti-angiogenic
activity in vivo (DAmato et al., 1994). Malignant
gliomas are rich in blood vessels and vascularization
is proportional to the grade of malignancy (Takahashi
et al., 1992). Many authors evaluated the feasibility
and effectiveness of thalidomide in phase III studies. Fine et al. (2000) administered daily doses of
thalidomide from 800 to 1200 mg in 39 patients with
relapsed high grade malignant gliomas, obtaining two
partial responses (PR), two minor responses (MR),
and 12 instances of stable disease (SD). Among 18
patients with relapsed high grade malignant gliomas
treated with daily 100 mg of thalidomide, Short et al.
(2001) obtained 1 PR and 2 SD. Marx et al. (2001)
treated 38 patients with relapsed glioblastoma with
a daily escalating dose of thalidomide from 100 to
500 mg, obtaining 2 PR and 16 SD. Morabito et al.
(2004) administered daily 400 mg of thalidomide to
18 patients with relapsed glioblastoma obtaining 1 MR
response and 8 SD.
A synergism between thalidomide and cisplatin was
documented in vitro. Thalidomide seems to enhance
the vascular permeability and to improve the tumour
vascularization enhancing intratumoral cisplatin concentration (Murphy et al., 2007). Unfortunately, so
far, there are no published clinical data evaluating
the combination of thalidomide and cisplatin other
than a phase I study published by Zustovich et al.
(2007).
In conclusion, the combination of temozolomide,
cisplatin and thalidomide is based on two different in
vitro demonstrated synergies, one between temozolomide and cisplatin and the other between cisplatin and
thalidomide.
26 Glioblastoma Patients: Chemotherapy with Cisplatin, Temozolomide and Thalidomide
Clinical Experience
The clinical experience using temozolomide, cisplatin,
and thalidomide in patients with high grade malignant gliomas is based on a published dose-finding
phase I study (Zustovich et al., 2007) and the following attempt to perform a phase II trial. Unfortunately,
this study concluded prematurely due to the high
incidence of vascular thrombo-embolic events (VTE).
Nevertheless, the data analysis led to interesting conclusions that overcome the mere dose-finding investigation and revealed some interesting concerns regarding myelotoxicity and vascular venous complications,
as described in the following paragraphs.
Patients with relapsed malignant primitive brain
tumors (not eligible for further local treatment) were
enrolled. Patients that had received prior chemotherapy
were admitted, including those treated with temozolomide if not combined with other alkylating agents.
Treatment was started with cisplatin at 75 mg/m2 day
1 in combination with daily temozolomide 100 mg/m2
for 5 days every 3 weeks. Progressively increasing
dose levels were established to verify, firstly, the tolerance to the 3-week schedule of cisplatin and temozolomide with the usual cisplatin dosage of 75 mg/m2
and temozolomide at the same dose-intensity of the
standard 4-week schedule; thus, 150 mg/m2 , for 5
days, instead of 200 mg/m2 . Thalidomide was then
added in 50 mg step escalating doses starting from
100 mg total dose per day. From April 2002 to March
2005, we enrolled 17 consecutive eligible patients. Of
Fig. 26.1 Left inferior limb

arteriography of the patients
with artery occlusion probably
induced by thalidomide
administration
257
the 14 patients with glioblastoma, 11 had previously

undergone cytoreductive surgery and 13 had received
radiotherapy.
All 17 patients were evaluable for toxicity, which
was generally mild. The most frequent grade 12 toxicities were anemia, nausea, and vomiting. Four patients
had grade 12 cutaneous/allergic toxicity, one of them
also complained of dyspnoea. Grade 34 toxicity was
febrile neutropenia in one patient, asymptomatic neutropenia in 6 patients, grade 4 thrombocytopenia in 1
patient, deep venous thrombosis (DVT) in 3 patients
and pulmonary embolism (PE) in one patient, with an
overall VTE incidence of ~23.5% and of 36.4% in
the sub-group receiving thalidomide. In these patients,
thalidomide was discontinued and not readministered,
even after the onset of an anti-coagulant therapy.
Sixteen patients were evaluable for response and
all for time to progression (TTP) and overall survival
(OS). Among the 14 patients with glioblastoma, 1
patient had a PR, 8 patients had SD, and 5 patients had
progressive disease. Median TTP was 4 months (range
0.914.9 months), PFS-6 was 28% and median OS was
7.8+ months (range 2.624.2+ months).
As stated before, we planned to continue with a
phase II study but an early case of artery thrombotic occlusion occurred (see Fig. 26.1). As a consequence, and in consideration of the high rate VTE
in the dose-finding trial, we decided to continue with
the sole administration of temozolomide and cisplatin, without thalidomide (see the paragraph Future
perspectives).
258
Mielotoxicity
The dose-finding escalation of thalidomide in combination with cisplatin and temozolomide (75 mg/m2
day 1 and 150 mg/m2 days 15, respectively, every
3 weeks) was concluded at the level of 200 mg daily
total dose. Dose limiting toxicity (DLT) was hemathological with grade 4 thrombocytopenia in one patient
and grade 4 febrile neutropenia in another patient.
Moreover, 6 out of the 7 patients who developed
grade 34 neutropenia belonged to the group treated
with thalidomide. Thus, we can infer that thalidomide
impairs the bone-marrow function in proportion to the
dose administered.
In a series of 44 patients with multiple myeoloma
treated with thalidomide there were 10 cases of grade
34 neutropenia and 5 patients had a bone-marrow
hypoplasia without any increase of myeloma cell count
(Hattori et al., 2004).
The dynamic contrast enhanced magnetic resonance
(dMRI) in patients with multiple myeloma evidenced
as thalidomide is able to reduce the microcirculation
in the bone marrow (Scherer et al., 2002). This reduction is greater when thalidomide is administered in
combination with chemotherapy (Wasser et al., 2004).
A reduction of bone marrow micro vessels density in
patients with multiple myeloma tretaed with thalidomide is also been demonstrated (Kumar et al., 2004).
Vascular Complications
At first, we did not consider vascular venous grade
34 toxicity as a dose limiting toxicity in consideration of the high incidence of VTE in patients with
high-grade glioma. In a series of 68 patients with highgrade gliomas, the rate of VTE was 19% and authors
reported as risk factors chemotherapy administration
(p = 0.027) and the presence of paresis (p = 0.031)
(Dhami et al., 1993). The overall rate of VTE in our
series of patients treated with the combination of temozolomide, cisplatin and thalidomide was 36.4% and we
supposed that thalidomide might have been a VTE risk
factor in high grade glioma patients.
There are previous reports regarding thalidomideinduced vascular venous grade 34 toxicity in cancer patients. Data from FDAs MedWatch showed an
incidence of VTE of 4.6% in patients treated with
thalidomide alone, of 15.0% in patients receiving concomitant dexamethasone and of 30.9% in patients
receiving concomitant chemotherapy (Bennet et al.,
2001). In multiple myeloma patients there was a higher
incidence of VTE when thalidomide was associated
to chemotherapy regimen containing dexamethasone,
vincristine, doxorubicin, cyclophosphamide, etoposide
and cisplatin (28% vs. 4%; p = 0.002) (Zangari
et al., 2001). A low incidence of VTE occurred when
thalidomide was administered in combination with
a non-doxorubicin containing chemotherapy regimen
(dexamethasone, cyclo-phosphamide, and etoposide)
(Moehler et al., 2002). A comparison between 2
chemotherapy regimens differing only for the presence
or not of doxorubicin revealed as doxorubicin was a
VTE risk factor (16.0% vs 2.5%; p = 0.02) (Zangari
et al., 2002).
A phase II study of temozolomide and thalidomide in patients with brain metastases of melanoma
was prematurely interrupted after the enrolment of 16
patients for a high incidence of VTE. Authors reported
3 cases of PE and 1 case of DVT with an overall VTE
rate of 25% (Krown et al., 2006). That was not far
from 19% reported in high-grade glioma patients series
(Dhami et al., 1993). Unfortunately, besides ours, no
published data regarding the combination of cisplatin
and thalidomide are available. We suppose that the
concomitant administration of cisplatin and thalidomide, like doxorubicin in multiple myeloma patients,
might be a VTE risk factor in patients with malignant
high-grade gliomas.
Future Perspectives
The new generation of drugs classified as Target
therapy have been tested in malignant gliomas. The
inhibitors of the EGFR pathways gave poor result but
antiangiogenetic drugs, such as bevacizumab, seemed
to be very promising. Bevacizumab is a monoclonal
antibody targeted against VEGF. In a phase II study,
35 patients with progressive refractory glioblastoma
received bevacizumab in combination with irinotecan.
Overall response rate was 57%, PFS-6 was 46% and
the 6-month OS was 77% (Goli et al., 2007). In a
recent phase II study bevacizumab alone gave an ORR
of 35% in heavily pre-treated patients with a PFS-6 of
26 Glioblastoma Patients: Chemotherapy with Cisplatin, Temozolomide and Thalidomide
29% while the addition of irinotecan did not seem to

improve the outcome (Kreisl et al., 2008).
I believe that the combination of temozolomide,
cisplatin, and bevacizumab could be safe and even
more effective. Bevacizumab has been combined to
temozolomide in high-grade glioma patients and to
cisplatin in non-small cell lung cancer patients without the evidence of unexpected or increased vascular
and haematological toxicity. Our still unpublished data
treating temozolomide refractory glioblastoma patients
with the combination of temozolomide and cisplatin
are promising with a PR rate of 36% with 2 long lasting PFS of 16.9 and 13.2+ months after 14 enrolled
patients.
A phase III study evaluating the combination of
temozolomide, cisplatin, and bevacizumab in glioblastoma patients has been designed, but until sufficient
clinical data are gathered, it will remain an appropriate
suggestion.
References
Bennet CL, Schumok GT, Kwaan HC, Raisch DW (2001) High
incidence of Thalidomide-associated deep vein thrombosis
and pulmonary emboli when chemotherapy is also administered. Blood 98:863a (Abstract)
Bertolone SJ, Baum ES, Krivit W, Hammond GD (1989) A
phase II study of cisplatin therapy in recurrent childhood
brain tumors. A report from the childrens cancer study group.
J Neurooncol 1:511
Brandes AA, Basso U, Reni M, Vastola F, Tosoni A, Cavallo G,
Scopece L, Ferreri AJ, Panucci MG, Monfardini S, Ermani M
(2004) First-line chemotherapy with cisplatin plus fractionated temozolomide in recurrent glioblastoma multiforme:
a phase II study of the Gruppo Italiano Cooperativo di
Neuro-Oncologia. J Clin Oncol 9:15981604
Brandes AA, Pasetto LM, Vastola F, Monfardini S (2000)
Temozolomide in patients with high grade gliomas.
Oncology 3:181186
DAmato RJ, Laughnan MS, Flynn E, Folkman J (1994)
Thalidomide is an inhibitor of angiogenesis. Proc Natl Acad
Sci USA 9:40824085
DAtri S, Graziani G, Lacal PM, Nistico V, Gilberti S, Faraoni I,
Watson AJ, Bonmassar E, Margison GP (2000) Attenuation
of O(6)-methylguanine-DNA methyltransferase activity and
mRNA levels by cisplatin and temozolomide in jurkat cells.
J Pharmacol Exp Ther 2:664671
Dhami MS, Bona RD, Calogero JA, Helman RM (1993) Venous
thromboembolism and high grade gliomas. Thromb Haemost
3:393396
DIncalci M, Citti L, Taverna P, Catapano CV (1988) Importance
of the DNA repair enzyme O6 -alkyl guanine alkyltransferase (AT) in cancer chemotherapy. Cancer Treat Rev 4:
279292
259
Fine HA, Figg WD, Jaeckle K, Wen PY, Kyritsis AP, Loeffler
JS, Levin VA, Black PM, Kaplan R, Pluda JM, Yung WK
(2000) Phase II trial of the antiangiogenic agent thalidomide
in patients with recurrent high-grade gliomas. J Clin Oncol
4:708715
Goli KJ, Desjardins JE, Herndon JE, Rich JN, Reardon DA,
Quinn JA, Sathornsumetee S, Bota DA, Friedman HS,
Vredenburgh JJ (2007) Phase II trial of bevacizumab and
irinotecan in the treatment of malignant gliomas. Proc ASCO
2007:2003 J Clin Oncol 25(18s) (abstract)
Hattori Y, Kakimoto T, Okamoto S, Sato N, Ikeda Y (2004)
Thalidomide-induced severe neutropenia during treatment of
multiple myeloma. Int J Hematol 3:283288
Bromberg JE, Cairncross JG, Janzer RC, Stupp R (2005)
MGMT gene silencing and benefit from temozolomide in
glioblastoma. N Engl J Med 10:9971003
Albert SA, Fine HA (2008) Phase II trial of single-agent
15(10):36173623 (Published on line ahead of print Dec 29
2008)
Krown SE, Niedzwiecki D, Hwu WJ, Hodgson L, Houghton
AN, Haluska FG et al (2006) Phase II study of temozolomide and thalidomide in patients with metastatic melanoma
in the brain. Cancer 107(8):18831890
Kumar S, Witzig TE, Dispenzieri A, Lacy MQ, Wellik LE,
Fonseca R, Lust JA, Gertz MA, Kyle RA, Greipp PR,
Rejkumar SV (2004) Effect of thalidomide therapy on
bone marrow angiogenesis in multiple myeloma. Leukemia
3:624627
Marx GM, Pavlakis N, McCowatt S, Boyle FM, Levi JA,
Bell DR, Cook R, Biggs M, Little N, Wheeler HR (2001)
Phase II study of thalidomide in the treatment of recurrent
glioblastoma multiforme. J Neurooncol 1:3138
Moehler TM, Dchlenzka J, Kaspar B, Neben K, Egerer G, Ho
AD, Goldschmidt H (2002) Low incidence of deep venous
thrombosis in poor prognosis multiple myeloma patients
treated with thalidomide and CED chemotherapy. Blood
100:5136a (abstract)
Morabito A, Fanelli M, Carillio G, Gattuso D, Sarmiento R,
Gasparini G (2004) Thalidomide prolongs disease stabilization after conventional therapy in patients with recurrent
glioblastoma. Oncol Rep 1:9395
Murphy S, Davey RA, Gu XQ, Haywood MC, McCann LA,
Mather LE, Boyle FM (2007) Enhancement of cisplatin
efficacy by thalidomide in a 9L rat gliosarcoma model.
Scherer A, Strupp C, Wittsack HJ, Engelbrecht V, Willers R,
Germing U, Gattermann N, Haas R, Modder U (2002)
Dynamic contrast-enhanced MRI for evaluating bone marrow microcirculation in malignant haematological diseases
before and after thalidomide therapy. Radiologie 3:222230
Short SC, Traish D, Dowe A, Hines F, Gore M, Brada M (2001)
Thalidomide as an anti-angiogenic agent in relapsed gliomas.
J Neurooncol 1:4145
Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B,
260
U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeir
A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff RO
(2005) European organisation for research and treatment of
cancer brain tumor and radiotherapy groups; National Cancer
Institute of Canada clinical trials group. 2005. Radiotherapy
plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 10:987996
Takahashi JA, Fukumoto M, Igarashi K, Oda Y, Kikuchi
H, Hatanaka M (1992) Correlation of basic fibroblast
growth factor expression levels with the degree of malignancy and vascularity in human gliomas. J Neurosurg 5:
792798
Wasser K, Moehler T, Neben K, Nosas S, Heiss J, Goldschmidt
H, Hillengass J, Duber C, Kauczor HU, Delorme S (2004)
Dynamic MRI of the bone marrow for monitoring multiple
myeloma during treatment with thalidomide as monotherapy or in combination with CED chemotherapy. Rofo 9:
12851295

Yung WK, Mechtler L, Gleason MJ (1991) Intravenous carboplatin for recurrent malignant glioma: a phase II study. J Clin
Oncol 5:860864
Zangari M, Anaissie E, Barlogie B, Badros A, Desikan R, Gopal
AV, Morris C, Toor A, Siegel E, Fink L, Tricot G (2001)
Increased risk of deep-vein thrombosis in patients with
multiple myeloma receiving thalidomide and chemotherapy.
Blood 5:16141615
Zangari M, Siegel E, Barlogie B, Anaissie E, Saghafifar
F, Fassas A, Morris C, Fink L, Tricot G (2002)
Thrombogenic activity of doxorubicin in myeloma patients
receiving thalidomide: implications for therapy. Blood 4:
11681171
Zustovich F, Cartei G, Ceravolo R, Zovato S, Della Puppa A,
Pastorelli D, Mattiazzi M, Bertorelle R, Gardiman MP (2007)
A phase I study of cisplatin, temozolomide and thalidomide in patients with malignant brain tumors. Anticancer Res
27(2):10191024
Chapter 27
Glioblastoma: Role of Galectin-1 in Chemoresistance

Florence Lefranc and Robert Kiss
Abstract Malignant gliomas, especially glioblastomas, are associated with a poor prognosis. This prognosis can be at least partly explained by the fact that
glioma cells diffusely infiltrate the brain parenchyma
and exhibit decreased levels of apoptosis and are thus
resistant to cytotoxic drugs. Galectin-1, the expression
of which is stimulated by hypoxia, is a potent modulator of glioblastoma cell migration and a pro-angiogenic
molecule. Galectin-1 participates in the resistance of
cancer cells, including glioma cells, to chemotherapy
and to radiotherapy and is involved in the activation of the Ras oncogenic pathway. Our recent data
reveal that temozolomide, the standard treatment for
glioma patients, increases Galectin-1 expression in
various glioblastoma models both in vitro and in
vivo. Consequently, reducing Galectin-1 expression in
these models increases the anti-tumor effects of various chemotherapeutic agents, in particular temozolomide. Reducing Galectin-1 expression in glioblastoma
cells does not induce apoptotic or autophagic features,
but rather modulates p53 transcriptional activity and
decreases p53-targeted gene expression. The decrease
in Galectin-1 expression also impairs the expression
levels of several genes implicated in chemoresistance: ORP150, HERP, GRP78/Bip, TRA1, BNIP3L,
GADD45B and CYR61. The involvement of Galectin1 in different steps of glioma malignant progression,
such as migration, angiogenesis or chemoresistance,
makes it a potential target for the development of new
drugs to combat these malignant tumors.
F. Lefranc ()
Laboratoire de Toxicologie, Facult de Pharmacie, Universit
Libre de Bruxelles (ULB), 1050 Brussels, Belgium
e-mail: fllefran@ulb.ac.be
Keywords GBM Apoptosis Galectin-1 Cell

death Chemoresistance Ras
Introduction
Malignant gliomas, especially glioblastomas (GBM),
are characterized by the diffuse invasion of distant
brain tissue by a myriad of single migrating cells with
reduced levels of apoptosis (type I programmed cell
death [PCD]) and consequent resistance to the cytotoxic insults of pro-apoptotic drugs (Lefranc et al.,
2005). In contrast, GBM cells are less resistant to
autophagy-related cell death (type II PCD) than to
apoptosis (Lefranc et al., 2005, 2006). Current recommendations are that patients with GBM should undergo
maximum surgical resection followed by concurrent radiation and chemotherapy with temozolomide
(Lefranc et al., 2006; Stupp et al., 2009). Galectin-1
(Gal1), a lectin with specificity for -galactosides
(Liu and Rabinovich, 2005; Camby et al., 2006;
(LeMercier et al., 2009), markedly influences glioma
cell migration both in vitro and in vivo (Camby et al.,
2001, 2002, 2005). High-grade glioma patients whose
gliomas markedly express Gal1 have a significantly
shorter survival period than individuals whose gliomas
express less Gal1 (Camby et al., 2002). Decreasing
Gal1-expression in human orthotopic GBM xenografts
significantly increases the survival of GBM tumorbearing mice (Camby et al., 2002; LeMercier
et al., 2008b). Gal1-expression is increased under
hypoxic conditions; hypoxia also confers cellular
resistance to conventional chemotherapy and accelerates malignant progression (Le et al., 2005). Gal1 is
negatively regulated by p53 (Puchades et al., 2007),

261
262
and Gal1 reciprocally has been shown to modify p53

biological functions (Camby et al., 2006). The reciprocal control exerted by Gal1 or p53 on the other
could be implicated in GBM chemoresistance, as could
also be the case in the partnership between Ras and
Gal1. Ras is implicated in gliomagenesis and modulates glioma aggressiveness (Goldberg and Kloog,
2006). Ras signaling and oncogenesis depend on the
dynamic interplay of Ras with distinctive plasma membrane micro-domains and various intracellular compartments. Such interactions are dictated by individual
elements in the carboxy-terminal domain of the Ras
proteins, one of which is recognized by Gal1, galectin3 and cGMP phosphodiesterase delta (Ashery et al.,
2006). Gal1 thereby promotes H-Ras signaling to Raf
at the expense of phosphoinositide 3-kinase (PI3K)
and Ral guanine nucleotide exchange factor (RalGEF)
(Ashery et al., 2006). Gal1 could therefore be involved
in these Ras-related pathways implicated in GBM
resistance to apoptosis (Lefranc et al., 2005; LeMercier
et al., 2009).
F. Lefranc and R. Kiss
when on the current standard treatment (surgical resection to the extent feasible, followed by adjuvant radiotherapy plus temozolomide chemotherapy, given concomitantly with and after radiotherapy) (Lefranc et al.,
2005, 2006; Stupp et al., 2009).
Malignant gliomas are associated with such dismal prognoses because glioma cells can actively
migrate through the narrow extracellular spaces in the
brain, often traveling relatively long distances, making
them elusive targets for effective surgical management
(Lefranc et al., 2005, 2006). Additionally, after surgical resection and adjuvant treatment of malignant
gliomas, the residual cancer cells peripheral to the
excised lesion give rise to a recurrent tumor that in
more than 90% of cases develops immediately adjacent
to the resection margin (Lefranc et al., 2005).
Clinical and experimental data have also demonstrated that invasive malignant glioma cells show a
decrease in proliferation rate and a relative resistance
to apoptosis as compared with the highly dense cellular
center of the tumor. This may contribute to their resistance to conventional pro-apoptotic chemotherapy and
radiotherapy (Lefranc et al., 2005, 2006).
Gliomas: An Overview
Galectins: An Overview
Gliomas account for more than 50% of all primary
brain tumors and are by far the most common primary brain tumor in adults (Lefranc et al., 2005,
2006). Gliomas include tumors that are composed predominantly of astrocytes (astrocytomas), oligodendrocytes (oligodendrogliomas), ependymal cells (ependymomas) or a mixture of various glial cells (e.g.,
oligoastrocytomas) (Louis et al., 2007). The World
Health Organization grading system classifies gliomas
as grade I to IV based on the degree of malignancy,
as determined by histopathological criteria. Grade I
gliomas are generally well-circumscribed and behave
in a benign fashion, whereas grade II through IV
gliomas are malignant and diffusely infiltrate the brain
(Louis et al., 2007). Among gliomas, astrocytomas
are the most common and are comprised of pilocytic
astrocytomas (grade I), diffuse astrocytomas (grade
II), anaplastic astrocytomas (grade III) and GBM
(grade IV) (Louis et al., 2007). Glioblastomas are
characterized by a very dismal prognosis (Lefranc
et al., 2005, 2006; Stupp et al., 2009). GBM patients
have a median survival expectancy of only 14 months
Galectins are a structurally-related family of animal lectins defined by two properties: (i) an affinity
for -galactoside sugars; and (ii) sequence homology (Barondes et al., 1994; Liu and Rabinovich,
2005; Camby et al., 2006; Le Mercier et al.,
2009, 2010). This consensus sequence corresponds to
the carbohydrate-recognition domain (CRD), which
is a beta sandwich of about 135 amino acids
long and is responsible for -galactoside binding (Barondes et al., 1994; Liu and Rabinovich,
2005; Camby et al., 2006; Le Mercier et al.,
2009). To date, 15 galectins have been characterized; they are numbered according to the chronology of their discovery (galectin-1 to galectin-15)
(Barondes et al., 1994; Liu and Rabinovich, 2005;
Camby et al., 2006; Le Mercier et al., 2009, 2010).
The galectins known so far have either one or two
CRDs within a single polypeptide chain, and neither
CRD is associated with other types of well-defined
protein domains. The mono-CRD galectins can be
biologically active as monomers (galectin-5, -7, -10)
27 Glioblastoma: Role of Galectin-1 in Chemoresistance
or as homodimers (galectin-1, -2, -11, -13, -14, -15);

the bi-CRD galectins (galectin-4, -6, -8, -9, 12) are
active as monomers and might also associate into
oligomers (Leffler et al., 2004). Galectin-3, a monoCRD galectin, is unique in that it contains a short
proline, glycine and tyrosine rich N-terminal domain
fused onto the CRD that therefore allows the formation
of oligomers (Leffler et al., 2004).
Galectins can segregate into multiple cell compartments. Although these proteins lack the signal
sequence that would be required for secretion through
the classical secretory pathway, some galectins show
extracellular localization, suggesting that they are
secreted through a non-classical pathway (Camby
et al., 2006). Galectins are present both inside and
outside cells. They function extracellularly by interacting with cell surface and extracellular matrix (ECM)
glycoproteins and intracellularly by interacting, in a
carbohydrate-independent manner, with cytoplasmic
and nuclear proteins (Elola et al., 2007). They play a
role in a wide range of processes, including cell adhesion, regulation of cell growth, apoptosis, embryonic
development and immune processes-like inflammation
(Perillo et al., 1998; Moiseeva et al., 1999; Liu and
Rabinovich, 2005; Camby et al., 2006).
A large amount of experimental evidence has
been reported to support the important roles of
galectins in cancer biology (Liu and Rabinovich, 2005;
Rabinovich, 2005; Lahm et al., 2001; Stillman et al.,
2005), including tumor angiogenesis (Moiseeva et al.,
1999; Thijssen et al., 2006; Le Mercier et al., 2008b),
tumor immune escape (Liu and Rabinovich, 2005) and
cancer cell migration (Camby et al., 2001, 2002, 2005;
Elola et al., 2007; Jung et al., 2008). In the following
section, we focus our attention on the biological roles
exerted by Gal1 in gliomas.
Galectin-1 in Glioma Biology

The role of Gal1 in glioma biology was first suggested
by Yamaoka et al. (2000) and Gunnersen et al. (2000).
These groups analyzed the mRNA expression of Gal1
by northern blot in glioma specimens and glioma cell
lines. Increased expression of Gal1 mRNA was shown
to correlate with increased malignancy in human astrocytic tumors ranging from low-grade astrocytomas to
malignant gliomas (Yamaoka et al., 2000). However,
263
no statistical analysis was made (Yamaoka et al.,

2000). Two studies from our own group using clinical samples have shown that Gal1 is expressed in
all glioma types and that the level of Gal1 expression makes a distinction between non-invasive (WHO
grade I) and invasive (WHO grade II-IV) gliomas
of the astrocytic tumor (Rorive et al., 2001; Camby
et al., 2001). Specifically, we quantitatively determined
(by computer-assisted microscopy) the immunohistochemical expression of Gal1 in 220 gliomas, including
151 astrocytic, 38 oligodendroglial and 31 ependymal
tumors (Rorive et al., 2001). Our data revealed the
expression of Gal1 in all human glioma types with no
striking variation in levels among astrocytic, oligodendroglial and ependymal tumors; the level of galectin-1
expression within astrocytic tumors, however, significantly correlated with tumor grade (Rorive et al.,
2001). Furthermore, expression levels of Gal1 in highgrade astrocytic tumors from patients with short-term
survival periods were significantly higher than those in
tumors from patients with long-term survivals (Rorive
et al., 2001).
The expression of Gal1 was shown to be higher in
the invasive parts of xenografted GBM than in the less
invasive parts, suggesting their involvement in tumor
astrocyte invasion of the brain parenchyma (Camby
et al., 2001). Our group has focused on the role of
Gal1 in glioma cell migration. We xenografted three
human glioblastoma cell lines (H4, U87 and U373)
into the brains of nude mice in order to characterize
the in vivo galectin-1 expression pattern in relation to
the tumor invasion of the normal brain parenchyma.
Immunohistochemical analysis of galectin-1 expression in human U87 and U373 glioblastoma xenografts
revealed a higher level of galectin-1 expression in invasive areas as compared to the non-invasive areas of the
xenografts (Camby et al., 2001; Rorive et al., 2001).
Moreover, nude mice intracranially grafted with U87
or U373 cells that were constitutively expressing low
levels of galectin-1 (by stable transfection with an
expression vector containing the antisense galectin-1
mRNA) had longer survival periods than those grafted
with U87 or U373 cells unchanged in expression levels
of galectin-1 (Camby et al., 2002). Complementary in
vitro studies have shown that adding purified galectin-1
to U87 human GBM cells enhanced tumor cell motility in a lactose-inhibited manner (Camby et al., 2005).
This effect appeared to be related to an increase
264
in polymerized filamentous actin and the expression

of the small guanosine triphosphatase RhoA (Camby
et al., 2002).
Finally, using cDNA microarray analysis and confirmation at protein levels, we observed that the U87
GBM cells that were galectin-1 deficient by means
of an antisense galectin-1-stable transfection displayed
increased protein levels for p21waf/cip1, cullin-2, p53,
ADAM-15 and MAP-2 (Camby et al., 2005). Major
differences in ADAM-15 expression and the actin
stress fiber organization were also observed (Camby
et al., 2005). The ADAM family of membraneanchorage glycoproteins encompasses a catalytically
active MMP domain and a disintegrin domain and
may thus be involved both in the proteolytic cleavage
of cell-surface proteins and in integrin-mediated cell
adhesion (including alpha9beta1 integrin/ADAM-15
interactions) via the RGD-dependent and -independent
binding.
Collectively, these data indicate that galectin-1
enhances the migratory capabilities of tumor astrocytes
and, therefore, their biological aggressiveness. These
features have been recently confirmed by Strik et al.
(2007) and by Jung et al. (2008).
Progression-associated genetic alterations are common to different glioma types and target growthpromoting and cell-cycle control pathways resulting in
focal hypoxia, necrosis and angiogenesis (Louis et al.,
2007). GBM is distinguished pathologically from
lower grade tumors by necrosis and microvascular

hyperplasia (Louis et al., 2007). Necrotic foci are
typically surrounded by pseudopalisading cells a
configuration that is relatively unique to malignant
gliomas and has long been recognized as an ominous
prognostic feature (Louis et al., 2007). Recent investigations have demonstrated that pseudopalisades are
severely hypoxic, overexpress hypoxia-inducible factor 1 (HIF-1), and secrete proangiogenic factors such
as vascular endothelial growth factor (VEGF) and IL-8
(Louis et al., 2007). HIF-1 is a potent activator of
angiogenesis and invasion by upregulating target genes
critical for these functions; activation of the HIF-1
pathway is a common feature of gliomas and may
explain the intense vascular hyperplasia often seen
in GBM.
Galectin-1 is also a hypoxia-regulated protein (Le
et al., 2005) that has been shown recently to display major roles in angiogenesis (Thijssen et al.,
2006) in both gliomas (Le Mercier et al., 2008b)
and melanomas (Mathieu et al., 2007) (Fig. 27.1).
Galectin-1 involvement in tumor angiogenesis was
first suggested after the discovery that both vascular smooth muscle and endothelial cells express the
protein (Moiseeva et al., 1999). Clausse et al. (1999)
have previously shown that galectin-1 was upregulated in capillaries associated with carcinoma cells.
In addition, they found that galectin-1 could mediate
interactions between tumors and endothelial cells in
Fig. 27.1 In stress conditions (black squares) such as

hypoxia, radiotherapy and/or chemotherapy, increase of
galectin-1 expression in GBM cells may activate migration, radio/chemoresistance, immune escape and angiogenesis.
Intracellular galectin-1 could be targeted by siRNA strategies.

Alternatively, extracellular galectin-1 could be the target of small
inhibitors
vitro, suggesting a potential role for galectin-1 in

modulating angiogenesis. Finally, Thijssen et al.
(2006) have shown that treatment with either galectin1-specific antisense oligodeoxynucleotides or with
polyclonal anti-galectin-1 antibodies resulted in inhibition of endothelial cell proliferation and migration,
demonstrating an essential role for galectin-1 during
angiogenesis. The role of galectin-1 in tumor angiogenesis is further highlighted in galectin-1-null mice,
in which tumor growth is markedly impaired because
of insufficient tumor angiogenesis (Thijssen et al.,
2006).
We ourselves have also put forward evidence for
the role of galectin-1 in the process of angiogenesis using human glioma cells. To determine how
galectin-1 exerts its pro-angiogenic effects, we investigated galectin-1 signaling in the human Hs683
glioma cell line. We observed that galectin-1 signals through the endoplasmic reticulum transmembrane kinase/ribonuclease inositol-requiring 1alpha
(IRE1alpha) that regulates the expression of oxygenregulated protein 150 (ORP150), which in turn controls VEGF maturation. Galectin-1 also modulates
the expression of six other hypoxia-related genes
(CTGF, ATF3, PPP1R15A, HSPA5, TRA1 and CYR61)
that are implicated in angiogenesis. Moreover, we
have recently reported that downregulating galectin-1
expression in Hs683 human glioma cells through
targeted small interfering RNAs provokes a marked
decrease in the expression of the brain expressed Xlinked gene (BEX2), a feature which confers increased
survival in Hs683 orthotopic xenograft-bearing mice.
This decrease in BEX2 expression impairs vasculogenic mimicry channel formation in vitro and angiogenesis in vivo, and modulates glioma cell adhesion, motility and invasive features (Le Mercier
et al., 2009).
Galectin-1 in Glioblastoma
Chemo/Radioresistance
Resistance of human tumors to anticancer drugs is
most often ascribed to gene mutations, gene amplification or epigenetic changes that influence the uptake,
metabolism or export of drugs from single cells
(Trdan et al., 2007). Another important, yet littleappreciated, cause of anticancer drug resistance is the
limited ability of drugs to penetrate tumor tissue and
265
to reach all of the tumor cells at a potentially lethal

concentration (Trdan et al., 2007). Although now
recognized as a major contributor to malignant cancer progression and to treatment failure, the precise
role of hypoxia signaling in cancer and in prognosis still needs to be further defined (Koumenis, 2006).
As emphasized earlier, intra-tumoral hypoxia causes
genetic changes in malignant gliomas that produce
a microenvironment that selects for cells of a more
aggressive phenotype. Hypoxia can initiate cell demise
by apoptosis/necrosis but also prevent cell death by
provoking adaptive responses that, in turn, facilitate
cell proliferation or angiogenesis, thus contributing
to tumor malignant progression (Zhou et al., 2006).
Zhou et al. (2006) emphasize that considering that
activation of HIF-1 provokes pro-survival as well as
pro-death decisions under hypoxia, it will be crucial
to understand decision making processes in regulating
cell death, adaptation and chemoresistance. Hypoxia is
also known to modulate the unfolded protein response,
a coordinated program that promotes cell survival
under conditions of endoplasmic reticulum (ER) stress
(Koumenis, 2006), and which is known to contribute
to tumor malignant progression and drug resistance of
solid tumors.
As mentioned earlier, hypoxia is known to activate
galectin-1 expression (Le et al., 2005), and galectin-1
was found to be negatively regulated by transfection
with TP53 in glioma cells (Puchades et al., 2007).
We recently reported that temozolomide, the standard treatment for glioma patients (Lefranc et al.,
2006; Stupp et al., 2009), increases galectin-1 expression in Hs683 glioma cells both in vitro and in vivo. In
addition, galectin-1 expression was shown to be upregulated by ionizing irradiation of glioma cell lines in
vitro (Strick et al., 2007). Reducing galectin-1 expression in these cells by siRNA increases the antitumor
effects of various chemotherapeutic agents, in particular temozolomide both in vitro and in vivo in an
orthotopic xenograft mouse model (Le Mercier et al.,
2008a). We also observed that decreasing galectin-1
expression by means of an anti-galectin-1 siRNA
in the mouse B16F10 metastatic melanoma model
(a syngenic model whereby B16F10 melanoma cells
are injected in the tail vein) also increases the therapeutic benefits contributed by temozolomide in vivo
(Mathieu et al., 2007). This decrease of galectin1 expression in the B16F10 mouse melanoma cells
does not modify their sensitivity to apoptosis nor
266
autophagy. However, it does induce heat shock protein 70-mediated lysosomal membrane permeabilization (LMP), a process associated with cathepsin B
release into the cytosol, which in turn is believed to
sensitize the cells to the proautophagic effects of temozolomide when grafted in vivo (Mathieu et al., 2007).
In Hs683 glioma cells a decrease in galectin-1 expression does not induce apoptotic or autophagic features
and does not induce LMP, but is found to modulate
p53 transcriptional activity and decrease p53-targeted
gene expression including DDIT3/GADD153/CHOP,
DUSP5, ATF3 and GADD45A (Le Mercier et al.,
2008a). In addition, the decrease in galectin-1 expression impairs the ER stress response, which is believed
to play a role in drug resistance, and also impairs the
expression levels of seven other genes implicated in
chemoresistance: ORP150; HERP; GRP78/Bip; TRA1;
BNIP3; GADD45B; and CYR61, some of which are
also known to be modified by hypoxia (Le Mercier
et al., 2008a).
Conclusions
Galectins are known to play an important role in
malignant cancer progression and especially malignant
gliomas (Camby et al., 2006; Le Mercier et al., 2009;
Le Mercier et al., 2010). Galectin-1 is involved in
many different steps of glioma biology, such as migration, angiogenesis and resistance to chemotherapy and
radiotherapy. We have already shown that decreasing galectin-1 expression in human GBM orthotopic
xenografts in mouse brains by siRNA administration
enhances the therapeutic benefits of temozolomide
(Le Mercier et al., 2008b). Thus, galectins, especially
galectin-1, could be important targets for the development of new anticancer drugs not only for gliomas
but for other types of cancer as well (Ingrassia et al.,
2006). The novel aspects of Gal1-related function in
the ERS response highlighted in the present study
and pertinent to the chemoresistance of glioma cells
may be amenable to therapeutic manipulation. This
manipulation might be achieved either through in vivo
delivery of anti-Gal1 siRNA as demonstrated in one
of our preliminary studies (Le Mercier et al., 2008b)
or through compounds that suppress Gal1 biological
activity (Ingrassia et al., 2006; Camby et al., 2008).
The in vivo delivery of anti-Gal1 siRNA directly into
the brain could be achieved in GBM patients by means

of Ommaya reservoirs, thus minimizing/avoiding systemic exposure and potential hepatotoxicity.
Acknowledgments F. Lefranc is a clinical research fellow and
R. Kiss a director of research with the Belgian National FNRS
(Belgium). The present chapter was supported by grants awarded
by the Fonds de la Recherche Scientifique Mdicale (FRSM,
Belgium) and by the Fonds Yvonne Bol (Brussels, Belgium).
References
Ashery U, Yizhar O, Rotblat B, Elad-Sfadia G, Barkan B, Haklai
R, Kloog Y (2006) Spatiotemporal organization of ras signaling: rasosomes and the galectin switch. Cell Mol Neurobiol
26:471495
Barondes SH, Castronovo V, Cooper DN, Cummings RD,
Drickamer K, Feizi T, Gitt MA, Hirabayashi J, Hughes
C, Kasai K, Leffler H, Liu FT, Lotan R, Mercurio AM,
Monsigny M, Pillai S, Poirer F, Raz A, Rigby PWJ, Rini
JM, Wang JL (1994) Galectins: a family of animal betagalactoside-binding lectins. Cell 76:597598
Camby I, Belot N, Lefranc F, Sadeghi N, de Launoit Y, Kaltner
H, Musette S, Darro F, Danguy A, Salmon I, Gabius HJ,
Kiss R (2002) Galectin-1 modulates human glioblastoma cell
migration into the brain through modifications to the actin
cytoskeleton and levels of expression of small GTPases. J
Neuropathol Exp Neurol 61:585596
Camby I, Belot N, Rorive S, Lefranc F, Maurage CA, Lahm
H, Kaltner H, Hadari Y, Ruchoux MM, Brotchi J, Zick Y,
Salmon I, Gabius HJ, Kiss R (2001) Galectins are differentially expressed in supratentorial pilocytic astrocytomas,
astrocytomas, anaplastic astrocytomas and glioblastomas,
and significantly modulate tumor astrocyte migration. Brain
Pathol 11:1226
Camby I, Decaestecker C, Lefranc F, Kaltner H, Gabius HJ, Kiss
R (2005) Galectin-1 knocking down in human U87 glioblastoma cells alters their gene expression pattern. Biochem
Biophys Res Commun 335:2735
Camby I, Le Mercier M, Lefranc F, Kiss R (2006) Galectin1: a small protein with major functions. Glycobiology 16:
137R157R
Camby I, Le Mercier M, Mathieu V, Ingrassia L, Lefranc F,
Kiss R (2008) Galectin-1 as a potential therapeutic target for
cancer progression. Drug Fut 33:12
Clausse N, van den Brle F, Waltregny D, Garnier F, Castronovo
V (1999) Galectin-1 expression in prostate tumor-associated
capillary endothelial cells is increased by prostate carcinoma cells and modulates heterotypic cell-cell adhesion.
Angiogenesis 3:317325
Elola MT, Wolfenstein-Todel C, Troncoso MF, Vasta GR,
Rabinovich GA (2007) Galectins: matricellular glycanbinding proteins linking cell adhesion, migration, and survival. Cell Mol Life Sci 64:16791700
Goldberg L, Kloog Y (2006) A Ras inhibitor tilts the balance between Rac and Rho and blocks phosphatidylinositol
3-kinase-dependent glioblastoma cell migration. Cancer Res
66:1170911717

Gunnersen JM, Spirkoska V, Smith PE, Danks RA, Tan SS
(2000) Growth and migration markers of rat C6 glioma
cells identified by serial analysis of gene expression. Glia
32:146154
Ingrassia L, Camby I, Lefranc F, Mathieu V, Nshimyumukiza P,
Darro F, Kiss R (2006) Anti-galectin compounds as potential
anti-cancer drugs. Curr Med Chem 13:35133527
Jung TY, Jung S, Ryu HH, Jeong YI, Jin YH, Jin SG, Kim IY,
Kang SS, Kim HS (2008) Role of galectin-1 in migration
and invasion of human glioblastoma multiforme cell lines. J
Koumenis C (2006) ER stress, hypoxia tolerance and tumor
progression. Curr Mol Med 6:5569
Lahm H, Andre S, Hoeflich A, Fischer JR, Sordat B, Kaltner
H, Wolf E, Gabius HJ (2001) Comprehensive galectin fingerprinting in a panel of 61 human tumor cell lines by
RT-PCR and its implications for diagnostic and therapeutic
procedures. J Cancer Res Clin Oncol 127:375386
Le QT, Shi G, Cao H, Nelson DW, Wang Y, Chen EY, Zhao S,
Kong C, Richardson D, OByrne KJ, Giaccia AJ, Koong AC
(2005) Galectin-1: a link between tumor hypoxia and tumor
immune privilege. J Clin Oncol 23:89328941
Leffler H, Carlsson S, Hedlund M, Qian Y, Poirier F (2004)
Introduction to galectins. Glycoconj J 19:433440
Lefranc F, Brotchi J, Kiss R (2005) Possible future issues in the
treatment of glioblastomas: special emphasis on cell migration and the resistance of migrating glioblastoma cells to
apoptosis. J Clin Oncol 23:24112422
Lefranc F, Sadeghi N, Camby I, Metens T, Dewitte O, Kiss R
(2006) Present and potential future issues in glioblastoma
treatment. Expert Rev Anticancer Ther 6:719732
Le Mercier M, Fortin S, Mathieu V, Kiss R, Lefranc F (2010)
Galectins and gliomas. Brain Pathol 20:1727
Le Mercier M, Fortin S, Mathieu V, Roland I, Spiegl-Kreinecker
S, Haibe-Kains B (2009) Galectin-1 pro-angiogenic and promigratory effects in the Hs683 oligodendroglioma model are
partly mediated through the control of BEX2 expression.
Neoplasia 11:485496
Le Mercier M, Lefranc F, Mijatovic T, Debeir O, Haibe-Kains
B, Bontempi G, Decaestecker C, Kiss R, Mathieu V (2008a)
Evidence of galectin-1 involvement in glioma chemoresistance. Toxicol Appl Pharmacol 229:172183
Le Mercier M, Mathieu V, Haibe-Kains B, Bontempi G,
Mijatovic T, Decaestecker C, Kiss R, Lefranc F (2008b)
Knocking down galectin 1 in human hs683 glioblastoma
cells impairs both angiogenesis and endoplasmic reticulum
stress responses. J Neuropathol Exp Neurol 67:456469
Liu FT, Rabinovich GA (2005) Galectins as modulators of
tumour progression. Nat Rev Cancer 5:2941
Louis DN, Ohgaki H, Wiestler OD, Cavenee WK (2007) WHO
classification of tumours of the central nervous system.
International Agency for Research on Cancer (IARC), Lyon
Mathieu V, Le Mercier M, De Neve N, Sauvage S, Gras
T, Roland I, Lefranc F, Kiss R (2007) Galectin-1 knock-
267
down increases sensitivity to temozolomide in a B16F10
mouse metastatic melanoma model. J Invest Dermatol 127:
23992410
Moiseeva EP, Spring EL, Baron JH, de Bono DP (1999) Galectin
1 modulates attachment, spreading and migration of cultured
vascular smooth muscle cells via interactions with cellular
receptors and components of extracellular matrix. J Vasc Res
36:4758
Perillo NL, Marcus ME, Baum LG (1998) Galectins: versatile
modulators of cell adhesion, cell proliferation, and cell death.
J Mol Med 76:402412
Puchades M, Nilsson CL, Emmett MR, Aldape KD, Ji Y, Lang
FF, Liu TJ, Conrad CA (2007) Proteomic investigation of
glioblastoma cell lines treated with wild-type p53 and cytotoxic chemotherapy demonstrates an association between
galectin-1 and p53 expression. J Proteome Res 6:869875
Rabinovich GA (2005) Galectin-1 as a potential cancer target.
Br J Cancer 92:11881192
Rorive S, Belot N, Decaestecker C, Lefranc F, Gordower L,
Micik S, Maurage CA, Kaltner H, Ruchoux MM, Danguy
A, Gabius HJ, Salmon I, Kiss R, Camby I (2001) Galectin-1
is highly expressed in human gliomas with relevance for
modulation of invasion of tumor astrocytes into the brain
parenchyma. Glia 33:241255
Stillman BN, Mischel PS, Baum LG (2005) New roles for
galectins in brain tumorsfrom prognostic markers to therapeutic targets. Brain Pathol 15:124132
Strik HM, Schmidt K, Lingor P, Tonges L, Kugler W, Nitsche
M, Rabinovich GA, Bhr M (2007) Galectin-1 expression in
human glioma cells: modulation by ionizing radiation and
effects on tumor cell proliferation and migration. Oncol Rep
18:483488
Stupp R, Mayer M, Kann R, Weder W, Zouhair A, Betticher DC,
Roth AD, Stahel RA, Majno SB, Peters S, Jost L, Furrer
M, Thierstein S, Schmid RA, Hsu-Schmitz SF, Mirimanoff
RO, Ris HB, Pless M (2009) Neoadjuvant chemotherapy and
radiotherapy followed by surgery in selected patients with
stage IIIB non-small-cell lung cancer: a multicentre phase II
trial. Lancet Oncol 10:785793
Thijssen VL, Postel R, Brandwijk RJ, Dings RP, Nesmelova I,
Satijn S, Verhofstad N, Nakabeppu Y, Baum LG, Bakkers
J, Mayo KH, Poirier F, Griffioen AW (2006) Galectin-1 is
essential in tumor angiogenesis and is a target for antiangiogenesis therapy. Proc Natl Acad Sci USA 103:1597515980
Trdan O, Galmarini CM, Patel K, Tannock IF (2007) Drug
resistance and the solid tumor microenvironment. J Natl
Cancer Inst 99:14411454
Yamaoka K, Mishima K, Nagashima Y, Asai A, Sanai Y, Kirino
T (2000) Expression of galectin-1 mRNA correlates with
the malignant potential of human gliomas and expression of
antisense galectin-1 inhibits the growth of 9 glioma cells. J
Neurosci Res 59:722730
Zhou J, Schmid T, Schnitzer S, Brune B (2006) Tumor hypoxia
and cancer progression. Cancer Lett 237:1021
Chapter 28
Glioma-Initiating Cells: Interferon Treatment

Atsushi Natsume, Masasuke Ohno, Kanako Yuki, Kazuya Motomura,
and Toshihiko Wakabayashi
Abstract Type I interferons (IFNs) are cytokines

that exhibit immunomodulatory, cell differentiative,
antiangiogenic, and antiproliferative effects against
various neoplasms, particularly glial tumors, by classically activating the Janus Kinase (JAK)/Signal
Transducers and Activators of Transcription (STAT)
pathways. In contrast to other chemotherapeutic
agents, responses to IFN are slow and gradual, often
requiring years to develop. However, such responses
can be durable. Emerging evidence suggests that the
activity of IFNs is directed principally at small populations of cancer stem cells. Glioma stem cells
are reported to be resistant to a wide variety of
chemotherapeutic agents and possess the remarkable
ability to recover from cytotoxic therapy. Interestingly,
IFN- elicits significant antiproliferative effects on
glioma stem cells although such effects are not
elicited by temozolomide. This cytokine induces terminal differentiation of glioma stem cells to the
oligodendroglial cell lineage and downregulates
an anti-apoptotic microRNA, miR-21, which is
overexpressed in glioma stem cells. IFN-mediated
STAT3 phosphorylation may play a crucial role in
these events. Previous studies have indicated that
chemokines such as bone morphogenetic protein
(BMP) and leukemia inhibitory factor/ciliary neurotrophic factor (LIF/CTNF) induce STAT3-mediated
differentiation of glioma stem cells. The results of
A. Natsume ()
Department of Neurosurgery, Nagoya University School of
Medicine, 65 Tsurumai-cho, Showa-ku,
Nagoya 466-8550, Japan
e-mail: anatsume@med.nagoya-u.ac.jp
these studies provide a new strategy for therapeutic approaches that can induce glioma stem cells to
undergo terminal differentiation.
Keywords INFs Cytokine STAT pathways Glioma
stem cells microRNA Phosphorylation
Introduction
Type I interferons (IFNs), including IFN- and
IFN-, are cytokines that exhibit immunomodulatory, cell differentiative, antiangiogenic, and antiproliferative effects against various neoplasms, particularly glial tumors, by classically activating the Janus
Kinase (JAK)/Signal Transducers and Activators of
Transcription (STAT) pathways (Borden et al., 2007).
One of the type I IFNs, IFN- has multifaceted functions related to antitumor activity, such as cytostatic
effects, participation in the differentiation of cytotoxic
T lymphocytes and potentiation of their antitumor
immunological responses, and ability to act as a drug
sensitizer to enhance toxicity against various malignant neoplasms. Combination therapy with IFN- and
nitrosourea has been particularly useful in the treatment of malignant gliomas. We previously reported
that IFN- markedly enhances chemosensitivity to
temozolomide, an alkylating agent (Natsume et al.,
2005). This revealed that a major mechanism by which
IFN- enhances chemosensitivity is via the downregulation of O6-methylguanine-DNA methyl-transferase
(MGMT) transcription through augmentation with
TP53. This effect was also confirmed in an experimental animal model (Natsume et al., 2008). In this

269
270
chapter, we show that IFN- elicits a remarkable antiproliferative effect on glioma stem cells (GSCs) and
induces terminal differentiation of GSCs to the oligodendroglial cell lineage by the activation of STAT3.
Further, we focus on microRNA regulation by type I
IFNs in GSCs.
Cancer Stem Cells and Type I IFNs

Cancers harbor small cell populations that have the
potential to sustain growth and initiate tumorigenesis. These cells, which are known as cancer stem cells
(CSCs) or cancer initiating cells, have been identified in leukemia, multiple myeloma, breast cancer, and
glioma. In solid tumors, CSCs share many properties with normal stem cells, including self-renewal and
multi-potency, and furthermore are known to initiate
the growth of tumors.
A long-recognized paradox of response and survival in cancer therapies is that although most cancer
patients respond to therapy, few are completely cured.
The objective clinical response to treatment may not
predict substantial improvements in overall survival.
Huff et al. (2006) coined the term dandelion phenomenon to describe this paradox in cancer. The
pattern of the clinical responses displayed by patients
with chronic myeloid leukemia (CML) toward imatinib and IFN are quite different. The rapid responses
induced by imatinib are probably a consequence of
its impressive activity against differentiated CML cells
that make up the bulk of the leukemia. However these
early effective responses often do not last. The evidence could be explained by the development of CML
stem cell resistance to imatinib. Conversely, responses
to IFN are slow and gradual, often requiring years to
develop, but can be durable; this is consistent with
the data suggesting that IFNs activity is directed principally at small populations of CML stem cells. The
pattern of imatinib activity may be analogous to cutting a dandelion off at ground level and leaving the
unseen root behind, while IFN mimics specific attack
of the root of the dandelion. This responsesurvival
paradox also applies to other malignancies. In ovarian
cancer, Moserle et al. (2008) demonstrated that IFN
exerts marked antiproliferative effects on side population cells with higher proliferation rates. However,
the mechanism of action underlying this effect remains
undefined.
A. Natsume et al.
Establishment of Glioma Stem Cells

Glioblastoma multiforme (GBM) is the most lethal
primary glioma. Patients diagnosed with this type of
glioma typically have a median survival of less than
12 months, despite various therapies, including surgical resection, radiotherapy, and chemotherapy. For
many years, gliomas such as GBM have been considered to represent the bulk of heterogeneous tumors that
consist of differentiated cells and undifferentiated cells
with the capability of self-renewal and partial differentiation (Lee et al., 2006). Therefore, failure to eradicate
glioma cells may be responsible for the small fraction
of undifferentiated tumor cells that re-grow.
Nestin is an intermediate filament cytoskeletal protein found in neuroepithelial stem cells and progenitor
cells (Reynolds and Weiss, 1992). It has been shown
that glial-derived neoplasms also express nestin, and
the level of expression is higher in high-grade gliomas
compared to low-grade gliomas (Dahlstrand et al.,
1992). Using the neurosphere assay, Ignatova et al.
(2002) were able to isolate nestin-expressing cells
capable of forming clones that exhibit intraclonal heterogeneity in the expression of neural lineage-specific
proteins. Subsequently, when Singh et al. (2003) successfully cultured glioma cells in serum-free medium
containing stem cell growth factors bFGF and EGF, it
was observed that a small percentage of all glioma cells
ranging from 0.3 to 25.1%, depending on the aggressiveness of the tumor, were capable of self-replication
and formed non-adherent neurospheres and maintained
tumor culture over time via multiple passages.
These self-renewing and tumor culture-maintaining
cells not only stained positive for the undifferentiated
neural stem cell marker nestin, they also stained positive for CD133, a hematopoietic stem cell marker
also present on normal human neural stem cells.
However, these cells were found to lack the expression of tubulin III and glial fibrillary acidic protein (GFAP), which can be used as markers for
differentiated neuronal lineage. In stark contrast to
this small group of CD133+ cells, the majority of
glioma cells were found to be CD133- and incapable of forming self-sustaining neurospheres. As a
general observation, the fraction of CD133+ GSCs
from aggressive tumors exhibited an increased rate
of self renewal relative to less aggressive tumors.
Under culture conditions that promote differentiation,
GSCs lose the expression of primitive markers like
28 Glioma-Initiating Cells: Interferon Treatment
CD133 and nestin and instead express differentiated

markers for the cell of origin. When serially transplanted into the brain of a NOD-SCID (non-obese
diabetic, severe combined immunodeficient) mouse,
these GSCs were able to produce exact phenocopies
of the original tumor with all the histopathological
features and cell surface markers of the original tumor.
Immunohistological staining of these GBM xenografts
that showed differential staining for CD133 and GFAP
underlines the fact that GSCs can differentiate into
mature progeny.
Even though the presence of GSC can clearly
account for the inherent heterogeneous nature of
gliomas, cellular and genetic analysis of GSCs showed
that these cells were genetically transformed with
enhanced self-renewal properties and possessed an
abnormal karyotype, which is not only limited to
CD133+ cells, but also present in both CD133+ and
CD133- cells. This suggests that all cancer cells were
clonally derived. Even though consensus has not yet
been reached on the exact mechanism of gliomagenesis, our new found understanding of the presence
of functional hierarchy in glioma indicates that it will
be important to investigate the slow growing mutated
GSCs in order to gain an understanding of the mechanisms of treatment failure. GSCs are reported to be
resistant to a wide variety of chemotherapeutic agents
and possess the remarkable ability to recover from
cytotoxic therapy (Eramo et al., 2006). Kang et al.
(2007) reported that a small population of multipotent
CD 133+ cancer cells can survive and proliferate upon
exposure of GBM cells to a lethal dose of carmustine (BCNU). When transplanted into a SCID mouse
brain, the original tumor was reproduced. A significantly higher level of CD133+ cells was reported to
be present in previously treated GBM when compared
with newly diagnosed GBM (Liu et al., 2006). The
gene profile of CD133+ cells exhibits a high level of
expression of anti-apoptotic genes and chemotherapy
resistance genes such as BCRP1, and MGMT (Liu
et al., 2006), rendering these cells resistant to many
commonly used chemotherapeutic agents, including
temozolomide, carboplatin, paclitaxel (Taxol), and
etoposide (VP16). In the same tumor, examples
of genes such as multi drug resistance-associated
proteins 1 and 3 were found to be markedly
elevated in GSCs when compared to non-GSCs
(Salmaggi et al., 2006). This emphasizes their role in
chemoresistance.
271
GSCs not only play a crucial role in chemoresistance, but are vital with respect to the failure of
radiation therapy since tumors surviving radiotherapy
are found to be enriched in CSCs. In a study conducted
by Bao et al. (2006), irradiation of an in vivo glioma
xenograft led to a 35-fold increase in the CD133+
cell population relative to untreated xenografts. This
suggests that irradiation leads to the enrichment of
CD133+ cells in the tumor and subsequent formation of more aggressive tumors with decreased latency
after serial transplantation. Given the pattern of treatment failure observed with current standard therapy, an
alternative strategy involving selectively targeting this
functionally distinct chemo- and radiation-resistant
small group of GSCs instead of the bulk of the tumor
might provide better success in treating this deadly
disease.
Terminal Differentiation of GSC by IFN

STAT3 activation is crucial for stem cell function,
differentiation, and survival. Recent studies demonstrated that bone morphogenetic protein (BMP) and
leukemia inhibitory factor/ciliary neurotrophic factor
(LIF/CTNF) induced STAT3-mediated differentiation
of neural and GSCs (Lee et al., 2008). These studies
indicate the potential for therapeutic approaches that
can induce GSCs to undergo terminal differentiation.
Our previous study showed that STAT3 expression is strongly upregulated in GSCs, and that IFN-
phosphorylates a tyrosine residue of STAT3. IFN-
treatment elicits a remarkable antiproliferative effect
and enhances the expression of oligodendroglial markers, including oligodendrocyte-specific protein (OSP),
galactosylceraminidase (GalC), and myelin basic protein (MBP), on GSCs. However, the expression of
these markers was inhibited by a peptide that specifically inhibits STAT3 (Fig. 28.1). Furthermore, IFN treatment did not induce the expression of OSP,
GalC, and MBP in human neural stem cells. In contrast, while the expression of other neural lineage
markers (i.e., GFAP and neuron-specific class III tubulin) was unaffected, the expression of nestin and
CD133 was reduced in the GSCs after IFN- treatment. IFN--treated GSCs were unable to form tumors
in NOD/SCID mice (Yuki et al., 2009). Therefore,
IFN may represent a potential therapeutic agent for
inducing terminal differentiation of GSCs.
272
A. Natsume et al.
Fig. 28.1 IFN- phosphorylates a tyrosine residue of

STAT3. IFN- treatment elicits a remarkable antiproliferative effect. IFN- induces glioma stem cells to undergo
terminal differentiation to mature oligodendrocytes that
express oligodendrocyte-specific protein (OSP), galactosylceraminidase (GalC), and myelin basic protein (MBP), and inhibits
gliomagenesis
Regulation of MicroRNAs by IFN in GSCs
its translation (Esquela-Kerscher and Slack, 2006).

Emerging evidence suggests that miRNAs are involved
in crucial biological processes, including development, differentiation, apoptosis, and proliferation of
mammalian cells. In humans, miRNAs have been proposed to contribute to oncogenesis because they possess multifaceted functions either as tumor suppressors
or as oncogenes (Schickel et al., 2008).
RNA interference induces a multitude of responses
in addition to the knockdown of a gene. This
is best understood in the context of an antiviral immune response. In particular, double-stranded
RNA, a nucleic acid associated with viral replication, is involved in numerous interactions contributing to the induction, activation, and regulation of antiviral mechanisms. One particularly
intriguing function of double-stranded RNA is to
stimulate important protective responses such as
the activation of Dicer-related antiviral pathways,
induction of type I IFN (IFN-/), and stimulation
of double-stranded RNA-activated protein kinase
and oligoadenylate synthase (Karpala et al., 2005).
MicroRNAs (miRNAs) are small noncoding RNAs

consisting of 2022 nucleotides that participate in
the posttranslational regulation of gene expression
by means of RNA interference (RNAi). The miRNA
genes are transcribed by RNA polymerase II in
the nucleus to form large pri-miRNA transcripts.
These pri-miRNA transcripts are processed by the
Drosha enzyme to release the pre-miRNA precursor product, which is less than 70 nucleotides in
length. After the pre-miRNA is transported into the
cytoplasm, another enzyme known as Dicer processes the intermediate to generate a mature 22nucleotide miRNA. This mature miRNA is integrated into the RNA-induced silencing complex and
forms double-stranded RNA with complementary
mRNAs. Depending on the degree of homology
between the miRNA and the mRNA, the RNA-induced
silencing complex could inhibit mRNA function
by either promoting its cleavage or by inhibiting
IFN-/ regulates the levels of crucial mediators of

the antiviral response, such as protein kinase R,
the 2 5 oligoadenylate synthase/RNase L system,
adenosine deaminase ADR1, and Mx GTPase (Katze
et al., 2002). Thus, RNA interference might be
involved in the IFN-mediated antiviral response. It
was recently reported that the levels of liver-specific
miRNA, miR-122, and several other miRNAs are regulated by IFN- in human hepatoma cells, and that
IFN- rapidly modulates the expression of miRNAs
that target the hepatitis C virus genomic RNA, thereby
inhibiting viral replication (Pedersen et al., 2007). In
addition to its ability to interfere with viral replication,
IFN- is also known for its antiproliferative effects in a
variety of neoplasms such as hepatocellular carcinoma,
chronic myeloblastic leukemia, melanoma, renal cancer, and glioma (Borden et al., 2007). Therefore, the
possibility is recognized that IFN- might induce or
downregulate cellular miRNAs in human neoplasms
and thereby use the RNA interference system in its
action against tumor progression.
We tested whether IFN- can alter the expression of
cellular miRNAs in human glioma cells by using the
data obtained from a genome-wide microarray (Ohno
et al., 2009). On the basis of the initial screening, we
found that the expression levels of several miRNAs
were increased (miR-187 and miR-194) or attenuated (miR-100, miR-21, and let-7 family miRNAs) in
response to IFN- treatment. Of the miRNAs regulated
by IFN-, we focused on miR-21 because it is one of
the best known miRNAs associated with tumorigenesis and progression in gliomas. miR-21 also modulates
tumorigenesis through the regulation of genes such as
bcl-2, PTEN, tropomyosin-1, and PDCD4 (Chen et al.,
2008; Meng et al., 2007; Zhu et al., 2007). Previously,
miR-21 was suggested to be aberrantly expressed and
was recognized as one of the major anti-apoptotic factors in malignant gliomas (Chan et al., 2005). We
demonstrated that the overexpression of miR-21 occurs
in a surgical specimen of glioblastoma by performing
an in situ hybridization (Fig. 28.2ac). We also compared the miR-21 expression levels in glioma cell lines,
GSCs, and normal brain tissue. The miR-21 was found
to be overexpressed in glioma cells relative to normal
brain cells. Notably, the expression level of miR-21
was found to be greater in GSCs than in the glioma
cell lines (Fig. 28.2d). This finding may indicate that
miR-21 plays a crucial role in the initiation and progression of glioma.
273
Investigations using quantitative real timepolymerase chain reaction (qRT-PCR) revealed

that IFN- downregulates miR-21 in cultured glioma
cell lines and GSCs, and that the systemic delivery of
IFN- reduces the level of miR-21 in intracranial GSC
xenografts in mice. In the time course experiments, the
IFN- treatment showed a relatively fast response in
reducing miR-21 levels, suggesting that the negative
regulation of miR-21 might be mediated directly by
IFN-, for example, through the phosphorylation of
JAK/STAT. Our finding that IFN- also suppresses the
expression of pri-miR-21 and pre-miR-21 suggests
that it regulates miR-21 transcription. The putative
regulatory region of the miR-21 gene is located within
an intron of the overlapping transmembrane protein
49 gene (TMEM49). This regulatory region contains
2 consensus STAT3-binding sites located approximately 800 bp upstream from the transcription start
site. We used a luciferase reporter gene system with
a fused full-length pri-miR-21 promoter, including
the STAT3 binding site, to demonstrate that IFN-
inactivates STAT3-mediated miR-21 promoter activity
and that the transcription activity could be recovered
by using a STAT3 inhibitor. These data suggest that
miR-21 expression is negatively regulated by STAT3
activation induced by IFN-.
In conclusion, the downregulation of miR-21 in
response to IFN- treatment contributes to the antitumor effects of this cytokine in GSCs. Furthermore,
miR-21 expression is negatively regulated by STAT3
activation (Fig. 28.3). These results highlight the
importance of understanding the transcriptional regulation of miRNAs, which would be involved in oncogenesis of gliomas.
Discussion
Here, we show that STAT3 activation is crucial
for oligodendrogenesis and miR-21 downregulation
in GSCs. However, the role of STAT3 activation
is debatable because its overactivation has been
reported to be oncogenic in other neoplasms (Frank,
2007). Loffler et al. (2007) demonstrated that IL-6dependent STAT3 activates the transcription of miR-21
in multiple myeloma cells. While IL-6 induces proliferation of myeloma cells, IFN- reduces the growth
of glioma cells or induces apoptosis in these cells.
274
Fig. 28.2 (a) miR-21 overexpression in glioma-initiating cells

(GICs). miR-21specific probe was hybridized in situ with
glioblastoma tissue obtained from a surgical specimen. The
miR-21specific probe clearly stains the glioblastoma tissue
but doesnot stain the normal cortextissue. (b) Tumor cells
express significant amounts of miR-21, as observed under high
A. Natsume et al.
magnification, (c) whereas nontumorous tissue does not express

miR-21. (d) qRT-PCR shows that miR-21 is overexpressed to a
great extent in glioma cells than in normal brain cells. Notably,
the amount of miR-21 is greater in GICs than in the glioma cell
lines. Columns, mean; bars, SD (normal brain expressed as 1)
275
Fig. 28.3 IFN- inactivates STAT3-mediated miR-21 promoter

activity. The miRNA genes are transcribed by RNA polymerase
II in the nucleus to form large pri-miRNA transcripts. These
pri-miRNA transcripts are processed by the Drosha enzyme to
release the pre-miRNA precursor product, which is less than 70
nucleotides in length. After the pre-miRNA is transported into
the cytoplasm, another enzyme known as Dicer processes the

intermediate to generate a mature 22-nucleotide miRNA. This
mature miRNA is integrated into the RNA-induced silencing
complex and forms double-stranded RNA with complementary
mRNAs of genes such as bcl-2, PTEN, tropomyosin-1, and
PDCD4
A possible explanation for this seemingly paradoxical

role of STAT3 activation is that the STAT pathway is
context-dependent and that various intracellular and/or
environmental cues play a pivotal role in determining the outcome of the activation of the pathway. This
discrepancy may arise from differences in cytokine
stimulus and cell type (Loffler et al., 2007).
An unresolved question pertains to the identity
of the regulators of oligodendrogenesis of GSCs.
Oligodendroglial lineage arises from undifferentiated
progenitor cells, which are predominately found in
the subventricular zone. These progenitor cells mature
into myelinating oligodendrocytes as a result of interactions between BMP, sonic hedgehog, and proteins
of the Notch signaling pathways (Nicolay et al.,
2007). IFN may activate the BMP/Smad pathway and
Notch signaling. There is a strong evidence suggesting
that thyroid hormones (THs) directly regulate the
differentiation and maturation of oligodendroglial
cells. Triiodothyronine (T3) is an active hormone that
regulates gene expression by binding to specific intracellular TH receptors. Our group previously showed
that 95% of glioma cells express a member of the thyroid/steroid hormone receptor superfamily peroxisome proliferative-activated receptor gamma (PPAR),
which activates the transcription of target genes
after forming heterodimers with retinoid X receptors
(RXR). PPARs are known to be responsible for deciding the fate of specific neural stem cells. It has been
reported that PPARs regulate the proliferation, migration, and differentiation of neural stem cells via STAT3
activation. Although further confirmative studies are
warranted in this regard, we were able to show that
treatment of GSCs with IFN- causes the release of
T3 and that this is inhibited by the STAT3 inhibitor.
Treatment with exogenous T3 results in the formation
of MBP-positive mature oligodendrocytes displaying a
multibranched morphology. This suggests that GSCs
may possess neuroendocrinal properties and differentiate into mature oligodendrocytes.
276
References
Borden EC, Sen GC, Uze G, Silverman RH, Ransohoff RM,
Foster GR, Stark GR (2007) Interferons at age 50: past, current and future impact on biomedicine. Nat Rev Drug Discov
6:975990
Chan JA, Krichevsky AM, Kosik KS (2005) MicroRNA-21 is an
antiapoptotic factor in human glioblastoma cells. Cancer Res
65:60296033
Chen Y, Liu W, Chao T, Zhang Y, Yan X, Gong Y, Qiang
B, Yuan J, Sun M, Peng X (2008) MicroRNA-21 downregulates the expression of tumor suppressor PDCD4 in
human glioblastoma cell T98G. Cancer Lett 272:197205
Dahlstrand J, Collins VP, Lendahl U (1992) Expression of
the class VI intermediate filament nestin in human central
nervous system tumors. Cancer Res 52:53345341
Eramo A, Ricci-Vitiani L, Zeuner A, Pallini R, Lotti F, Sette G,
Pilozzi E, Larocca LM, Peschle C, De Maria R (2006)
Esquela-Kerscher A, Slack FJ (2006) Oncomirs microRNAs
with a role in cancer. Nat Rev Cancer 6:259269
Frank DA (2007) STAT3 as a central mediator of neoplastic
cellular transformation. Cancer Lett 251:199210
Huff CA, Matsui W, Smith BD, Jones RJ (2006) The paradox
of response and survival in cancer therapeutics. Blood 107:
431434
Ignatova TN, Kukekov VG, Laywell ED, Suslov ON, Vrionis
FD, Steindler DA (2002) Human cortical glial tumors contain neural stem-like cells expressing astroglial and neuronal
markers in vitro. Glia 39:193206
Kang MK, Kang SK (2007) Tumorigenesis of chemotherapeutic
drug-resistant cancer stem-like cells in brain glioma. Stem
Cells Dev 16:837847
Karpala AJ, Doran TJ, Bean AG (2005) Immune responses
to dsRNA: implications for gene silencing technologies.
Immunol Cell Biol 83:211216
Katze MG, He Y, Gale M Jr (2002) Viruses and interferon: a
fight for supremacy. Nat Rev Immunol 2:675687
Lee J, Kotliarova S, Kotliarov Y, Li A, Su Q, Donin NM,
Pastorino S, Purow BW, Christopher N, Zhang W, Park JK,
Fine HA (2006) Tumor stem cells derived from glioblastomas cultured in bFGF and EGF more closely mirror the
phenotype and genotype of primary tumors than do serumcultured cell lines. Cancer Cell 9:391403
Lee J, Son MJ, Woolard K, Donin NM, Li A, Cheng CH,
Kotliarova S, Kotliarov Y, Walling J, Ahn S, Kim M,
Totonchy M, Cusack T, Ene C, Ma H, Su Q, Zenklusen
JC, Zhang W, Maric D, Fine HA (2008) Epigeneticmediated dysfunction of the bone morphogenetic protein
pathway inhibits differentiation of glioblastoma-initiating
cells. Cancer Cell 13:6980
A. Natsume et al.
Loffler D, Brocke-Heidrich K, Pfeifer G, Stocsits C,
Hackermuller J, Kretzschmar AK, Burger R, Gramatzki M,
Blumert C, Bauer K, Cvijic H, Ullmann AK, Stadler PF,
Horn F (2007) Interleukin-6 dependent survival of multiple
myeloma cells involves the Stat3-mediated induction of
microRNA-21 through a highly conserved enhancer. Blood
110:13301333
Meng F, Henson R, Wehbe-Janek H, Ghoshal K, Jacob ST,
Patel T (2007) MicroRNA-21 regulates expression of the
PTEN tumor suppressor gene in human hepatocellular cancer. Gastroenterology 133:647658
Moserle L, Indraccolo S, Ghisi M, Frasson C, Fortunato E,
Canevari S, Miotti S, Tosello V, Zamarchi R, Corradin A,
Minuzzo S, Rossi E, Basso G, Amadori A (2008) The side
population of ovarian cancer cells is a primary target of
IFN-alpha antitumor effects. Cancer Res 68:56585668
Natsume A, Ishii D, Wakabayashi T, Tsuno T, Hatano H, Mizuno
M, Yoshida J (2005) IFN-beta down-regulates the expression
of DNA repair gene MGMT and sensitizes resistant glioma
cells to temozolomide. Cancer Res 65:75737579
Natsume A, Wakabayashi T, Ishii D, Maruta H, Fujii M, Shimato
S, Ito M, Yoshida J (2008) A combination of IFN-beta and
temozolomide in human glioma xenograft models: implication of p53-mediated MGMT downregulation. Cancer
Chemother Pharmacol 61:653659
Nicolay DJ, Doucette JR, Nazarali AJ (2007) Transcriptional
control of oligodendrogenesis. Glia 55:12871299
Ohno M, Natsume A, Kondo Y, Iwamizu H, Motomura K,
Toda H, Ito M, Kato T, Wakabayashi T (2009) The modulation of microRNAs by type I IFN through the activation of
signal transducers and activators of transcription 3 in human
glioma. Mol Cancer Res 7:20222030
Pedersen IM, Cheng G, Wieland S, Volinia S, Croce CM,
Chisari FV, David M (2007) Interferon modulation of cellular microRNAs as an antiviral mechanism. Nature 449:
919922
Reynolds BA, Weiss S (1992) Generation of neurons and astrocytes from isolated cells of the adult mammalian central
nervous system. Science 255:17071710
Salmaggi A, Boiardi A, Gelati M, Russo A, Calatozzolo C,
Ciusani E, Sciacca FL, Ottolina A, Parati EA, La Porta C,
Alessandri G, Marras C, Croci D, De Rossi M (2006)
Glioblastoma-derived tumorospheres identify a population of
tumor stem-like cells with angiogenic potential and enhanced
multidrug resistance phenotype. Glia 54:850860
Schickel R, Boyerinas B, Park SM, Peter ME (2008)
MicroRNAs: key players in the immune system, differentiation, tumorigenesis and cell death. Oncogene 27:
59595974
Yuki K, Natsume A, Yokoyama H, Kondo Y, Ohno M, Kato
T, Chansakul P, Ito M, Kim SU, Wakabayashi T (2009)
Induction of oligodendrogenesis in glioblastoma-initiating
cells by IFN-mediated activation of STAT3 signaling. Cancer
Lett 284:7179
Zhu S, Si ML, Wu H, Mo YY (2007) MicroRNA-21 targets the
tumor suppressor gene tropomyosin 1 (TPM1). J Biol Chem
282:1432814336
Chapter 29
Glioblastoma: Anti-tumor Action of Natural and Synthetic

Cannabinoids
Aleksandra Ellert-Miklaszewska, Iwona Ciechomska, and Bozena Kaminska
Abstract The past few decades have seen renewed

interest in medicinal cannabis or rather cannabinoids
active compounds derived from the Cannabis plant,
as well as their endogenous counterparts and a still
growing set of synthetic derivatives. One of the
most extensively studied and promising applications
of cannabinoids is their potential use as anti-cancer
agents in malignant tumors, such as glioblastomas.
The anti-tumor action of cannabinoids is mediated
via the CB1 and CB2 cannabinoid receptors. Growing
array of data suggest that significant alterations of a
balance in the cannabinoid system between the levels of endogenous ligands and their receptors occur
along with the malignant transformation in various
types of cancer. Increased CB2 receptor expression
has been observed in glioblastoma cells, invading
microglia/macrophages and endothelial cells of the
tumor blood vessels as compared to non-tumor brain
samples. Thus, among various approaches to avoid
CB1-receptor-mediated psychodysleptic side effects of
some cannabinoids, special attention is paid to substances, which selectively stimulate the CB2 receptors, putatively overexpressed in target tumor cells.
Induction of cell death by cannabinoid treatment relies
on the generation of a pro-apoptotic sphingolipid
ceramide and disruption of signaling pathways crucial
for regulation of cellular proliferation, differentiation
or apoptosis. Increased ceramide levels lead also to
ER-stress and autophagy in drug treated glioblastoma
B. Kaminska ()
Laboratory of Transcription Regulation, Department of Cell
Biology, Nencki Institute of Experimental Biology, 02-093
Warsaw, Poland
e-mail: bozenakk@nencki.gov.pl
cells. Cannabinoids have displayed a great potency in

reducing glioma tumor growth in experimental animal
models without producing the generalized toxic effects
unavoidable with most conventional chemotherapeutic drugs. Apparently their effectiveness in vivo has
been attributed to several mechanisms of action.
Cannabinoids have recently emerged as compounds
that beyond inhibition of tumor cell proliferation and
survival impair tumor angiogenesis, invasiveness and
even gliomagenesis. The good safety profile observed
in a pilot clinical trial, together with remarkable antitumor effects reported in preclinical studies may set
the basis for further research aimed at better evaluation
of the potential anti-cancer activity of cannabinoids.
A unique mechanism of cannabinoid action among
standard oncology remedies justifies further research
on their anti-tumoral properties either alone or in
combined therapies.
Keywords Glioblastoma Cannabinoids Anti-cancer
agents Receptors Ligands Autophagy
Introduction
Preparations from Cannabis sativa, the hemp plant,
have been used for centuries for both medicinal and
recreational purposes (Howlett et al., 2002; Mackie,
2006). Isolation of the active components of the plant,
called cannabinoids, in 1960s, as much as subsequent cloning of cannabinoid receptors, discovery of
their endogenous ligands and development of synthetic cannabinoids contributed to an intensive burst
of cannabinoid research. Along with our expanding
comprehension of mechanisms of cannabinoids action,

277
278
targeting cannabinoid signaling for therapeutic purposes has inevitably emerged as an interesting area of
scientific and clinical investigations.
One of the most extensively studied applications
of cannabinoids is their potential use as anti-cancer
agents. Anti-proliferative effects of cannabinoids have
been reported in various cultured cancer cells, including neural, breast, prostate, skin, thyroid cancer cells
and leukemia cells. Several studies demonstrated
anti-tumor activity of cannabinoids in animal models (Guzman, 2003). Since the first publication by
Sanchez and co-workers in 1998 providing evidence
that cannabinoids can be an effective tool against
glioma cells, growing interest of several groups including ours has been focused on therapeutic potential of
cannabinoids in glial tumors. Current scientific data
relevant to the use of cannabinoids in treatments of
glioblastomas are reviewed here including both in vitro
and in vivo results, proposed mechanisms of action,
reports from limited human studies and prospects for
patients therapy in clinics.
Cannabinoid System in Health and Brain

Tumors
Cannabinoids are a group of structurally heterogenous
but pharmacologically related compounds classified
into three subtypes. Plant-derived cannabinoids (phytocannabinoids) are uniquely found in the cannabis
plant. ()-trans-9 -tetrahydrocannabinol (9 -THC) is
the most potent and abundant out of approximately
70 identified phytocannabinoids. Endogenous cannabinoids (also known as endocannabinoids) such anandamide and 2-arachidonoylglycerol are produced by
mammals as rapidly inactivated lipid mediators, which
levels are strictly controlled by a transporter system
and hydrolyzing enzymes. Various modifications of the
chemical structure of natural cannabinoids based on
structure-activity relationship studies have led to generation of a still growing set of synthetic cannabinoids
(Fig. 29.1).
Cannabinoids elicit a wide range of central and
peripheral effects, which are mediated mostly through
cannabinoid receptors (Howlett et al., 2002). There
are two types of specific Gi/o -protein-coupled receptors cloned so far, called CB1 and CB2, although an
existence of additional cannabinoid-binding receptors
A. Ellert-Miklaszewska et al.
has been suggested (Howlett et al., 2002; Stella, 2004).

CB1 and CB2 differ in their predicted amino acid
sequence, tissue distribution, physiological role and
signaling mechanisms. CB1 is abundantly expressed in
the central nervous system and peripheral nerve terminals, where it mediates inhibition of neurotransmitter
release, but it is also present in some extra-neural
sites, such as testis, uterus, vascular endothelium, eye,
spleen, and tonsils (Howlett et al., 2002).
By contrast, the CB2 receptor is predominantly
expressed in cells and organs of the immune system (Howlett et al., 2002). The role of peripheral
CB2 receptor activation under baseline physiologic
conditions is still not well defined. It is suggested
to be involved in B-cell differentiation and modulation of immune response. Increased levels of CB2 are
reported in tissues during development, inflammation,
injury and cancer, revealing a critical role for the CB2
receptor in regulating these processes (Howlett et al.,
2002). The CB2 receptor was believed to be absent
from healthy brain, however, its expression has been
detected recently in microglia brain macrophages
(Stella, 2004; Gong et al., 2006), as well as in a small
subpopulation of neurons (Van Sickle et al., 2005;
Gong et al., 2006). The function of CB2 receptor in
the brain is completely unknown. Nevertheless, experiments performed in animals show that CB2-selective
agonists appear to be free from widespread psychoactive effects attributed to activation of the CB1 receptor
(Guzman, 2003; Valenzano et al., 2005).
Growing array of data suggest that significant alterations of a balance in the cannabinoid system between
the levels of endogenous ligands and their receptors
occur along with the malignant transformation in various types of cancer. While non-transformed astrocytes
were shown to express only the CB1 cannabinoid
receptor, both types of functional cannabinoid receptors have been found in several established human
glioblastoma cell lines, as well as in primary cultures
derived from the most malignant brain tumor, glioblastoma multiforme (GBM) (Galve-Roperh et al., 2000;
Sanchez et al., 2001; Howlett et al., 2002) (and our
unpublished data). Immunohistochemical analysis of
human low and high grade glioma surgical specimens
revealed increased CB2 receptor expression in tumor
cells, invading microglia/macrophages and endothelial cells of the tumor blood vessels as compared to
non-tumor brain samples (Sanchez et al., 2001; EllertMiklaszewska et al., 2007; Schley et al., 2009). We
29 Glioblastoma: Anti-tumor Action of Natural Synthetic Cannabinoids
279
Fig. 29.1 Chemical structures of plant-derived 9 -THC

((-)-trans-9 -tetrahydrocannabinol), endocannabinoid anandamide (N-arachidonoylethanolamine) and synthetic cannabinoids JWH133, selective for non-psychoactive CB2 receptor
((6aR,10aR)-3-(1,1-Dimethylbutyl)-6a,7,10,10a-tetrahydro-6,6,
9-trimethyl-6H-dibenzo[b,d]pyran), and WIN55,212-2, a CB1/

CB2 receptor agonist ((R)-(+)-[2,3-Dihydro-5-methyl-3-(4morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxazin-6-yl]-1naphthalenylmethanone)
detected the presence of CB2 receptors in all analyzed

biopsies of astrocytomas and glioblastomas. The proportion of malignant tumors expressing high levels of
CB2 (10 out of 16, 62.5%) was over 2-fold higher
than that seen in the tumors of lower grade (7 out
of 29, 24%) (Ellert-Miklaszewska et al., 2007). Thus,
the extent of CB2 expression correlated with tumor
malignancy. Moreover, as observed by Sanchez et al.
(2001) CB2 receptor immunoreactivity markedly prevailed over detected CB1 receptor levels in grade IV
astrocytomas (Sanchez et al., 2001). The levels of CB1
receptor expression in tumor and tumor-associated
endothelial cells were not significantly different from
the control tissue and showed no dependence on tumor
grade (Sanchez et al., 2001; Schley et al., 2009).
As brain tumors constitute the second most common
malignancy in children and the prevalence of histological types of brain tumors vary significantly between
the adult and pediatric populations, we compared the
expression of the CB2 receptor in paraffin-embedded
sections from primary brain tumors of adult and
pediatric patients (Ellert-Miklaszewska et al., 2007).
Most glioblastomas obtained from children expressed
very high levels of CB2 receptors. Interestingly,
some benign pediatric astrocytic tumors, such as

subependymal giant cell astrocytoma (SEGA), which
may occasionally cause mortality owing to progressive
growth in some patients, also displayed high CB2imunoreactivity. The high levels of CB2 expression
would predestine astrocytic tumors to be vulnerable to
cannabinoid treatment.
Significant CB2 immunostaining was also found
in limited cases of other histological types of
brain tumors, primarily in high grade ependymomas
and meningiomas (Ellert-Miklaszewska et al., 2007).
However, we observed that the CB2 staining intensity was much higher in astrocytomas than in oligodendrogliomas, ependymomas or meningiomas of the
same grade. It is worth to mention that all examined cases of embryonal tumors (medulloblastoma and
S-PNET), the most frequently diagnosed malignant
brain tumors in childhood, showed no or trace CB2
immunoreactivity. Our results suggest that the level of
CB2 receptor expression correlates primarily with the
histopathological origin of the brain tumor cells and
their differentiation state, reflected in the tumor grade.
A link between CB2 expression and malignancy
grade of the tumor has been reported also in prostate,
280
breast and pancreatic cancer, and the level of CB2

expression in transformed cells was higher than in the
respective normal tissue (Sarfaraz et al., 2005; Caffarel
et al., 2006; Carracedo et al., 2006a). In examined
tumors, CB2 receptors were usually located in the
areas of intense tissue proliferation and invading cells.
The enhancement of cannabinoid receptor expression
in malignant versus healthy tissues, might suggest a
possible role of the endocannabinoid system in the
tonic suppression of cell divisions and cancer growth.
This hypothesis is partly supported by the finding of
increased levels of anandamide and decreased levels
of one of endocannabinoid metabolizing enzymes in
human glioblastoma compared to human non-tumor
brain tissue (Petersen et al., 2005).
Anti-tumoral Action of Cannabinoids

in Glioma Cells and In Vivo Tumor
Models
The pioneer study to show that cannabinoids had antitumor effect and prolonged the life of mice bearing
Lewis lung adenocarcinoma was reported by Munson
et al. in 1975 (Munson et al., 1975). Further research
in this area was not performed until late 1990s, when
an interest in the role of cannabinoids in cancer therapy was renewed. Since then, plant-derived, synthetic
and endogenous cannabinoids have been successfully
used to block proliferation and invasive potential of
various cancer types (glioma, lymphoma, lung, thyroid, skin, pancreas, breast and prostate carcinoma) in
both in vitro and in vivo models (Guzman, 2003). The
research on therapeutic potential of cannabinoids in
cancer treatment is the most advanced in gliomas.
Programmed cell death of glioma cells after
cannabinoid treatment was first described by Manuel
Guzman and his co-workers (Sanchez et al., 1998).
They showed that 9 -THC is able to inhibit growth
of rat C6 glioma cells in vitro and induce cell death
with features typical for apoptosis (Sanchez et al.,
1998). We reported an apoptotic death triggered by a
mixed CB1/CB2 synthetic agonist WIN 55,212-2 in rat
glioma cells (Ellert-Miklaszewska et al., 2005). Recent
studies (Gomez del Pulgar et al., 2002; Salazar et al.,
2009) and our own observations show effectiveness of
9 -THC, WIN55,212-2 and a CB2-selective synthetic
cannabinoid to induce apoptosis of cultured human
glioblastoma cells and tumor-derived primary cultures.
The anti-tumor action of cannabinoids, mediated

via the CB1 or CB2 cannabinoid receptors, has been
also proved in vivo revealing a significant regression of malignant gliomas in cannabinoid-treated animals (Galve-Roperh et al., 2000; Sanchez et al.,
2001; Massi et al., 2004; Duntsch et al., 2006).
Local administration of 9 -THC or WIN55,212-2
reduced the size of tumors generated by intracranial
inoculation of C6 glioma cells in rats, leading to
complete eradication of gliomas and increased survival in one third of the treated rats (Galve-Roperh
et al., 2000). There has been no recurrence in any of
the surviving rats as monitored periodically by MRI
scan. Studies performed in mouse xenograft models
with intratumoral and intraperitoneal drug administration demonstrated that non-psychoactive phytocannabinoid cannabidiol (Massi et al., 2004), a CB2selective agonist JWH133 (Sanchez et al., 2001)
or a novel cannabinoid KM-233 (Duntsch et al.,
2006) blocked the proliferation of human astrocytoma cells implanted subcutaneously in the flank of
immune-deficient mice. Our preliminary data suggest that systemic cannabinoid administration can
effectively hamper intracranial tumor growth in rats
(unpublished).
Mechanism of Cannabinoids
Pro-apoptotic Action Inhibition
of Pro-survival Pathways
Several events and signal transduction pathways triggered mostly by stimulation of the CB1 and CB2
receptors have already been described to participate
in the cannabinoid-induced apoptosis, a programmed
cell death process (Guzman et al., 2001; Guzman,
2003). They include inhibition of PKA, superoxide
generation, and strong increase in intracellular calcium concentration (Howlett et al., 2002). However,
the best characterized mechanism of cannabinoidinduced cell death involves sustained accumulation of
pro-apoptotic sphingolipid ceramide, which modulates
signaling pathways crucial in the control of tumour
cell growth and survival (Galve-Roperh et al., 2000;
Sanchez et al., 2001).
Cannabinoid receptor activation triggers two peaks
of ceramide generation in glioma cells (Galve-Roperh
et al., 2000; Sanchez et al., 2001; Gomez del Pulgar
et al., 2002). Treatment with 9 -THC or another

CB1/CB2 receptor agonist produces a rapid release of
ceramide via enzymatic hydrolysis of sphingomyelin
from the cell membrane. The second ceramide peak is
generated within hours or days after receptor activation
and depends on increase of ceramide synthesis de novo
via induction of serine palimitoyltransferase, a regulatory enzyme of sphingolipid biosynthesis (Gomez
del Pulgar et al., 2002). Selective CB2 receptor agonists, such as JWH133, are supposed to stimulate
only the ceramide synthesis process, which is sufficient for turning on the cell death program (Sanchez
et al., 2001; Gomez del Pulgar et al., 2002). Thus,
enhanced production of ceramide de novo is considered as an important event in cannabinoids-induced
apoptosis.
Galve-Roperh and co-workers (2000) postulated
that the increased ceramide levels reported upon
cannabinoid challenge led to prolonged activation of
Raf-1/MEK/Erk signaling cascade and thus mediated
glioma cell cycle arrest and cell death. The same
authors showed also that pharmacological inhibition
of ceramide synthesis de novo prevented the inhibition of protein kinase B (also known as Akt) triggered
by cannabinoids (Gomez del Pulgar et al., 2002). Our
studies revealed that rather down-regulation of Erk
activity, together with inhibition of PI3K/Akt pathway, contributed to C6 glioma cell death induced
by WIN55,212-2 (Ellert-Miklaszewska et al., 2005).
The serine/threonine protein kinase Akt, activated
downstream of phosphoinositide 3-kinase (PI3K), as
well as the Ras-activated Raf1/MEK/Erk pathway
are widely recognized as key mediators of growth
factor-promoted cell survival in gliomas (Kapoor and
ORourke, 2003). Both survival pathways converge
on a small pro-apoptotic member of a Bcl-2 family of proteins, Bad. Phosphorylation of Bad by Akt
and Erk retains the protein in the cytosol, where
it is recognized by certain regulatory proteins and
sequestered (Zha et al., 1996). Otherwise, Bad translocates to mitochondria, and formation of heterodimers
between nonphosphorylated Bad and anti-apoptotic
proteins, such as Bcl-XL or Bcl-2, may result in a
loss of integrity of the outer mitochondrial membrane (Zha et al., 1996). The release of apoptogenic
proteins, including cytochrome c, triggers the executive phase of programmed cell death. We proposed a
mechanism, in which the decrease of mitogenic/prosurvival signaling evoked by the synthetic cannabinoid
281
WIN55,212-2 promoted the pro-apoptotic function

of Bad (Fig. 29.2). Accordingly, we demonstrated
changes in Bad phosphorylation level followed by collapse of the mitochondrial membrane potential in C6
glioma cells treated with WIN55,212-2. These events
preceded activation of caspase 9 by factors released
from disrupted mitochondria, subsequent processing
of effector caspases, cysteine proteases, which play
essential roles in apoptosis, and finally oligonucleosomal DNA fragmentation.
Our further studies, as well as some published
data suggest that human glioma cells treated with
cannabinoids enter the suicide cell death using
the same mitochondria-dependent (intrinsic) pathway
(Carracedo et al., 2006b). This mechanism contributes
to the induction of apoptosis by cannabinoids also
in other types of tumor cells (Velasco et al., 2007).
However, an involvement of an alternative, death
receptordependent (extrinsic) pathway in the apoptotic process triggered by these compounds cannot be
ruled out.
The Role of ER Stress and Autophagy

in Cannabinoid-Induced Cell Death
Different experimental approaches shown that the
pro-apoptotic and tumor growth-inhibiting activity of
cannabinoids relies on the accumulation of de novosynthesized ceramide, an event that occurs in the ER
(endoplasmic reticulum) and eventually leads to execution of cell death via mitochondrial intrinsic pathway.
A recent study by Carracedo et al. 2006b) suggested a
new link between the two events, alternative to inhibition of pro-survival pathways. They showed that 9 THC treatment of glioma cells leads to up-regulation
of the stress-regulated protein p8 and ER stress-related
downstream targets: ATF-4 (activating transcription
factor 4), CHOP (the C/EBP-homologous protein) and
TRB3 (tribbles homologue 3). Selective knockdown
of ATF-4 and TRB3 blocked cannabinoid-induced
apoptosis in glioma cells. Inhibition of ceramide
synthesis de novo prevented 9 -THC-induced p8,
ATF-4, CHOP and TRB3 up-regulation as well as
ER dilation, indicating that ceramide accumulation
is an early event in the cannabinoid-triggered ER
stress and apoptosis in glioma cells (Carracedo et al.,
2006b). Furthermore, ER-stressevoked stimulation of
282
Fig. 29.2 Mechanisms of multimode anti-tumoral action of

cannabinoids. Natural and synthetic cannabinoids induce tumor
cell death, block angiogenesis and invasiveness crucial for tumor
progression. Increased ceramide synthesis de novo via induction of serine palimitoyltransferase (SPT) cell death. Inhibition
of pro-survival pathways (Akt and Erk signaling) stimulates
translocation of Bad to the outer mitochondrial membrane and its
pro-apoptotic function. Interaction between Bad and Bcl-2 triggers a decrease of the mitochondrial membrane potential ()
and release of pro-apoptotic factors (such as cytochrome c)
to the cytosol, where apoptosis is executed by caspase cascade. Alternatively, induction of apoptosis by cannabinoids
can be mediated by ER (endoplasmic reticulum)-stress and
autophagy. Up-regulation of the stress-regulated protein p8 and
ER-stress-related downstream targets: ATF-4 (activating transcription factor 4), CHOP (the C/EBP-homologous protein) and
TRB3 (tribbles homologue 3) leads to inhibition of Akt, an
upstream activator of mTOR. Decreased activity of Akt/mTOR
pathway contributes to initiation of autophagy, that precedes
apoptosis of glioma cell
the p8/TRB3 pathway preceded the inhibition of the

Akt/mTOR (mammalial target of rapamycin) axis,
considered a key step in the early triggering of
autophagy (Salazar et al., 2009).
Irradiation and chemotherapeutic drugs kill cancer cells typically through induction of apoptosis. However most cancers (including gliomas) are
resistant to therapies that induce apoptosis (type
I programmed cell death). Macroautophagy, hereafter called autophagy, is an evolutionarily conserved
catabolic process, where a cell self-digests its cytoplasmic contents. Autophagy is accompanied by the
progressive development of vesicle structures from
autophagosomes to autolysosomes. Autophagy is activated in response to multiple stresses during cancer
progression, such as nutrient starvation, the unfolded
protein response (ER stress) and hypoxia.
Indeed, a detailed analysis of human astrocytoma

cell lines and a primary culture of human glioma cells
indicated that 9 -THC treatment led to formation of
autophagosomes in tumor cells (Salazar et al., 2009).
The authors also observed that ER stress occurred
earlier than autophagy and it preceded apoptosis in
cannabinoid-induced human and mouse cancer cell
death. Administration of 9 -THC to mice bearing
tumors derived from human astrocytoma cells produced increased TRB3 expression, inhibition of mTOR
signaling pathway, occurrence of autophagy markers and caspase-3 activation. These findings indicate
that cannabinoid promotes the autophagy-mediated
cell death through stimulation of ER stress in human
glioma cells.
Classically autophagy has been proposed to promote cell survival but paradoxically autophagy has
been reported to contribute to cell death (type II

programmed cell death). Recently it was also demonstrated that autophagy, not apoptosis, is induced in
glioblastoma cell by radiation and several chemotherapeutic agents (Aoki et al., 2008). The most striking
evidence for pro-autophagic chemotherapy to overcome apoptosis resistance in cancer cells comes from
the use of temozolomide (Aoki et al., 2008), a cytotoxic drug, which has demonstrated therapeutic benefits in glioblastoma patients. However, the role of
autophagy in causing cell death, rather than occurring along with cell death, is still not clear. As
reported by Salazar et al. (2009) pharmacological
or genetic inhibition of autophagy at different levels prevented 9 -THC-induced cell death, suggesting
that autophagy contributes to cannabinoid anti-tumoral
action.
Other Targets of Cannabinoid Action

Cannabinoids have displayed a great potency in
reducing glioma tumor growth in experimental animal models. Apparently their effectiveness in vivo
has been attributed to several mechanisms of action.
Cannabinoids have recently emerged as compounds
that beyond inhibition of tumor cell proliferation and
survival impair tumor angiogenesis, invasiveness and
even gliomagenesis (Blazquez et al., 2003; Aguado
et al., 2007).
Increased demand for oxygen and nutrients supply
to proliferating cancer cells makes angiogenesis a critical factor for the progression of solid tumors and a
popular target for oncologic therapies. Local administration of the CB2 selective cannabinoid JWH133 in a
mouse flank inoculation model of glioma turned the
vascular hyperplasia characteristic of actively growing tumors to a pattern of blood vessels characterized
by small, differentiated and impermeable capillaries,
thus proving antiangiogenic potential of the cannabinoid (Blazquez et al., 2003). This was associated
with a reduced expression of pro-angiogenic factors,
such as vascular endothelial growth factor (VEGF)
and angiopoietin-2 in cannabinoid treated tumors and
could be partly related to a direct influence of the
cannabinoid on endothelial cells migration and survival as well as on VEGF signaling in tumor cells. This
and subsequent studies of the same authors (Blazquez
283
et al., 2003, 2008) showed that 9 -THC or JWH133

administration to glioma-bearing mice decreased the
activity and expression of matrix metalloproteinase-2
(MMP-2). MMP-2 is a proteolytic enzyme that allows
tissue breakdown and remodeling during angiogenesis
and metastasis and its up-regulation is associated with
high progression and poor prognosis of gliomas. 9 THC inhibited MMP-2 expression and cell invasion in
cultured glioma cells (Blazquez et al., 2008). Blocking
the release of pro-angiogenic factors or the extracellular matrix degrading enzyme by in vivo administered
cannabinoids correlated with decreased tumor volumes
(Blazquez et al., 2003, 2008).
Although the identification of the cell of origin of
gliomas is still a matter of debate, recent findings
support the existence of brain cancer stem cells generated by transformation of the normal neural stem
cell. Aguado and coworkers (2007) showed that activation of CB1 and CB2 receptors with HU-210 and
JWH133 promoted differentiation of glioma stem-like
cells derived from GBM biopsies and from glioma cell
lines. Moreover, cannabinoid challenge decreased neurosphere formation and cell proliferation in secondary
xenografts, which correlated with the decreased efficiency of cannabinoid-treated glioma stem-like cells to
initiate glioma formation in vivo (Aguado et al., 2007).
Multimode action of cannabinoids, including inhibition of gliomagenesis may have important implications
for development of cannabinoid-based therapeutic
strategies.
Perspectives for the Clinical Use of

Cannabinoids in Glioma Patients
The past few decades have seen renewed interest in
medicinal cannabis. Capsules of THC and its synthetic
analogue nabilone are approved in several countries
as a prescription-only medicine for nausea and vomiting caused by cytotoxic chemotherapy, unresponsive
to other anti-emetics. Other potential palliative effects
of cannabinoids in oncology supported by phase III
clinical trials include apetite stimulation and pain
inhibition (Guzman, 2003). However, discussions on
controlled medical use of cannabis and cannabinoids
are now beyond its application for symptomatic relief.
There is quite a large number of emerging evidence
that cannabinoids can in fact modify the progression
284
of certain diseases. The efficacy of intra-tumorally

or systemically administered cannabinoids against
high-grade glioma in rat and mouse models (GalveRoperh et al., 2000; Sanchez et al., 2001; Massi
et al., 2004; Duntsch et al., 2006), raised promises of
using cannabinoid-based therapies against malignant
gliomas.
Safety Profile of Therapeutic

Cannabinoids
Standard chemotherapeutics are a double edged sword,
they eliminate cancer cells but affect severely healthy
cells in the body. Based on evidence from in vitro and
in vivo studies cannabinoids appear to have a favorable safety profile and do not produce the generalized
toxic effects as most conventional chemotherapeutic
drugs. Several mechanisms could be responsible for
cannabinoid targeting only the cancer cells. In contrast to pro-apoptotic action of CB1/CB2-activating
cannabinoids, such as 9 -THC and WIN55,212-2, on
transformed cells, treatment of primary cultured astrocytes with these compounds did not trigger ceramide
accumulation or induction of ER stress-related genes.
Furthermore, cannabinoids promote survival of glial
cells and neurons in different models of injury, suggesting that the anti-proliferative effect of cannabinoids is selective for brain tumor cells, viability
of normal brain cells remains unaffected or even
favored by cannabinoid challenge (Guzman, 2003).
In our studies astrocytes and glioma cells were vulnerable to WIN55,212-2 which was in line with
data on human primary mixed glial cultures and
glioblastoma cells treated with the synthetic cannabinoid (McAllister et al., 2005). Administration of
JWH133 although effective toward tumor cells, did
not affect survival or morphology of normal astrocytes
(unpublished).
Apparently negligible expression of CB2 receptor
in the brain and its abundance in high grade gliomas
seems to confer a relative safety of CB2-selective agonists for targeted glioma therapy. Concomitantly, a
novel cannabinoid chemotherapeutic exhibiting over
10-times higher affinity for the CB2 vs. CB1 receptor produced a minimal toxicity to healthy cultured
brain slices at micromolar concentrations required to
eradicate U87 glioma cells (Duntsch et al., 2006).
As reported by the same authors, clinical monitoring of mice with U87-derived subcutaneous tumors
proved that treatment with the cannabinoid was effective in reducing tumor volume and produced no general
toxic effects under intratumoral or systemic administration. More importantly, some animal studies show
also a good safety profile of CB1-activating agents.
In the studies by Galve-Roperh et al. (2000) 9 THC and WIN55,212-2 were delivered intratumorally
via an osmotic pump for 7 days leading to partial remission or complete eradication of the tumors
generated intracranially by inoculation of C6 glioma
cells The same effective doses of cannabinoids were
evaluated in tumor-free animals for potential adverse
effects in the brain. MRI analysis showed no signs
of damage related to necrosis, edema, infection or
trauma. Cannabinoid administration to tumor-bearing
and control rats induced no substantial modification
in behavioral parameters, in food and water intake or
in body weight. No abnormalities in biochemical and
hematological parameters nor markers of tissue damage have been revealed during 7-day delivery period
and for at least 2 months after cannabinoid treatment
(Galve-Roperh et al., 2000).
The strategy to use CB2 selective compounds in
glioma treatment has at least one more advantage
over the non-selective cannabinoids. The medical use
of cannabinoids is limited mainly by their undesirable side-effects attributed to marijuana abuse. Due
to the well known psychotropic activity of 9 -THC
and related compounds mimicking its action on the
CB1 receptors, potential application of these agents to
patients raises a number of clinical and ethical considerations. Among various approaches to avoid CB1receptor-mediated psychodysleptic side effects, special
attention is paid to substances, which selectively stimulate the CB2 receptors, putatively overexpressed in
target tumor cells (Guzman et al., 2001; Duntsch et al.,
2006).
Results from a Pilot Clinical Study

On the basis of successful preclinical findings Guzman
and his collaborators conducted a pilot phase I clinical trial, in which nine patients with actively growing
recurrent glioblastoma multiforme were administered
THC intratumorally via an infusion catheter placed
into the resection cavity (Guzman et al., 2006). The

enrolled patients had previously failed standard therapy (surgery and radiotherapy) and had clear evidence
of tumor progression. The primary endpoint of the
study was to determine the safety of intracranial 9 THC administration in a dose escalation regime. The
initial dose of 9 -THC delivered to the patients was
2040 g at day 1, increasing progressively for 25
days up to 80180 g/day. Over the median treatment
duration of 15 days patients underwent continuous
physical, neurological, biochemical and hematological
examinations as well as frequent magnetic resonance
and computed tomography scans of the brain. The
cannabinoid treatment did not result in any marked
alterations of clinical and laboratory tests. There was
no evidence of hemorrhage, oedema or injury.
The good safety profile observed in this pilot clinical trial, together with remarkable anti-tumor effects
reported in preclinical studies may set the basis for further research aimed at better evaluation of the potential
anti-cancer activity of cannabinoids.
Risks, Constrains and Benefits

The controlled clinical use of natural or synthetic
cannabinoids should not be perceived equally as the
use of smoked cannabis, which is a declared addictive drug. Noteworthy, central and peripheral effects
of cannabinoids pronounced in marihuana smokers are
not apparent in a controlled clinical setting (Guzman,
2003). Moreover, tolerance to these unwanted effects
of cannabinoids develops rapidly in humans and laboratory animals. The possibility that tolerance also
develops after therapeutic application as well as psychotic and addictive potential of cannabinoids have not
been thoroughly examined.
Although the general consensus indicates that
cannabinoids have anti-tumor effects there are a few
studies that have shown contradictory results or duality
of cannabinoids effects on tumor cell fate depending of the dose range used (Guzman, 2003). Some
authors questioned whether cannabinoid receptors do
constitute relevant molecular targets to treat glioblastoma in humans, as the high concentrations of the
currently available cannabinoids required to trigger
apoptosis in glioma cells in culture are not pharmacologically relevant (Held-Feindt et al., 2006). This
285
emphasizes the need for comprehensive dose-response

studies in the future correlated with analysis of the
cannabinoid receptor expression. In general, existing
data encourage to survey of new synthetic compounds
in order to replace 9 -THC or related cannabinoids
with more potent and selective mimetics. Another
possibility would be to prolong the effectiveness of
endocannabinoids by development of the inhibitors of
their breakdown.
Due to genetic and epigenetic alterations malignant glioblastomas are highly resistant to radiation and
chemotherapy. Mainstream therapeutic strategies for
the management of all primary brain tumors are still
mostly palliative, known to leave survivors with devastating neurological deficits and frequently with a high
risk of the disease recurrence. The potency of synthetic
cannabinoids to induce apoptosis in glioblastoma cells
has been tested by us and others on several cell
lines and primary cell cultures derived from biopsies
of human tumors, which to some extent may reflect
the heterogeneity of glioma molecular characteristics (Galve-Roperh et al., 2000; Sanchez et al., 2001;
Guzman, 2003; Massi et al., 2004; Duntsch et al.,
2006; Ellert-Miklaszewska et al., 2005; McAllister
et al., 2005). Thus, cannabinoids were able to override alterations of growth regulatory and apoptotic
pathways, caused by common mutations reported in
primary and secondary glioblastomas. Cannabinoid
apoptotic action relies on the generation of ceramide
and disruption of signaling pathways crucial for regulation of cellular proliferation, differentiation or apoptosis (Galve-Roperh et al., 2000; Gomez del Pulgar
et al., 2002; Ellert-Miklaszewska et al., 2005; Salazar
et al., 2009). Their unique mechanism of action among
standard oncology remedies justifies further research
on their anti-tumoral properties either alone or in
combined therapies.
References
Aguado T, Carracedo A, Julien B Velasco G, Milman G,
Mechoulam R, Alvarez L, Guzman M, Galve-Roperh I
(2007) Cannabinoids induce glioma stem-like cell differentiation and inhibit gliomagenesis. J Biol Chem 282:68546862
Aoki H, Kondo Y, Aldape K, Yamamoto A, Iwado E, Yokoyama
T, Hollingsworth EF, Kobayashi R, Hess K, Shinojima N,
Shingu T, Tamada Y, Zhang L, Conrad C, Bogler O, Mills
G, Sawaya R, Kondo S (2008) Monitoring autophagy in
286
glioblastoma with antibody against isoform B of human
microtubule-associated protein 1 light chain 3. Autophagy
4:467475
Blazquez C, Casanova ML, Planas A, Del Pulgar TG, Villanueva
C, Fernandez-Acenero MJ, Aragones J, Huffman JW,
Jorcano J L, Guzman M (2003) Inhibition of tumor angiogenesis by cannabinoids. FASEB J 2:2
Blazquez C, Salazar M, Carracedo A, Lorente M, Egia
A, Gonzalez-Feria L, Haro A, Velasco G, Guzman M
(2008) Cannabinoids inhibit glioma cell invasion by downregulating matrix metalloproteinase-2 expression. Cancer
Res 68:19451952
Caffarel MM, Sarrio D, Palacios J, Guzman M, Sanchez C
(2006) Delta9-tetrahydrocannabinol inhibits cell cycle progression in human breast cancer cells through Cdc2 regulation. Cancer Res 66:66156621
Carracedo A, Gironella M, Lorente M, Garcia S, Guzman M,
Velasco G, Iovanna JL (2006a) Cannabinoids induce apoptosis of pancreatic tumor cells via endoplasmic reticulum
stress-related genes. Cancer Res 66:67486755
Carracedo A, Lorente M, Egia A, Blazquez C, Garcia S, Giroux
V, Malicet C, Villuendas R, Gironella M, Gonzalez-Feria L,
Piris MA, Iovanna J L, Guzman M, Velasco G (2006b) The
stress-regulated protein p8 mediates cannabinoid-induced
apoptosis of tumor cells. Cancer Cell 9:301312
Duntsch C, Divi MK, Jones T, Zhou Q, Krishnamurthy M,
Boehm P, Wood G, Sills A, Ii BM (2006) Safety and efficacy
of a novel cannabinoid chemotherapeutic, KM-233, for the
treatment of high-grade glioma. J Neurooncol 77(2):143152
Ellert-Miklaszewska A, Grajkowska W, Gabrusiewicz K,
Kaminska B, Konarska L (2007) Distinctive pattern of
cannabinoid receptor type II (CB2) expression in adult and
pediatric brain tumors. Brain Res 1137:161169
Ellert-Miklaszewska A, Kaminska B, Konarska L (2005)
Cannabinoids down-regulate PI3K/Akt and Erk signalling
pathways and activate proapoptotic function of bad protein.
Cell Signal 17:2537
Galve-Roperh I, Sanchez C, Cortes ML, del Pulgar TG,
Izquierdo M, Guzman M (2000) Anti-tumoral action of
cannabinoids: involvement of sustained ceramide accumulation and extracellular signal-regulated kinase activation. Nat
Med 6:313319
Gomez del Pulgar T, Velasco G, Sanchez C, Haro A, Guzman
M (2002) De novo-synthesized ceramide is involved in
cannabinoid-induced apoptosis. Biochem J 363:183188
Gong JP, Onaivi ES, Ishiguro H, Liu QR, Tagliaferro PA, Brusco
A, Uhl GR (2006) Cannabinoid CB2 receptors: immunohistochemical localization in rat brain. Brain Res 1071:1023
Guzman M (2003) Cannabinoids: potential anti-cancer agents.
Nat Rev Cancer 3:745755
Guzman M, Duarte MJ, Blazquez C, Ravina J, Rosa MC, GalveRoperh I, Sanchez C, Velasco G, Gonzalez-Feria L (2006)
A pilot clinical study of Delta9-tetrahydrocannabinol in
patients with recurrent glioblastoma multiforme. Br J Cancer
95:197203
Guzman M, Sanchez C, Galve-Roperh I (2001) Control of the
cell survival/death decision by cannabinoids. J Mol Med
78:613625
Held-Feindt J, Dorner L, Sahan G, Mehdorn HM, Mentlein R
(2006) Cannabinoid receptors in human astroglial tumors.
J Neurochem 98:886893
Howlett AC, Barth F, Bonner TI, Cabral G, Casellas P, Devane
WA, Felder CC, Herkenham M, Mackie K, Martin BR,
Mechoulam R, Pertwee RG (2002) International union of
pharmacology. XXVII. Classification of cannabinoid receptors. Pharmacol Rev 54:161202
Kapoor GS, ORourke DM (2003) Receptor tyrosine kinase
signaling in gliomagenesis: pathobiology and therapeutic
approaches. Cancer Biol Ther 2:330342
Mackie K (2006) Cannabinoid receptors as therapeutic targets.
Annu Rev Pharmacol Toxicol 46:101122
Massi P, Vaccani A, Ceruti S, Colombo A, Abbracchio MP,
Parolaro D (2004) Anti-tumor effects of cannabidiol, a
nonpsychoactive cannabinoid, on human glioma cell lines.
J Pharmacol Exp Ther 308:838845
McAllister SD, Chan C, Taft RJ, Luu T, Abood ME, Moore
DH, Aldape K, Yount G (2005) Cannabinoids selectively
inhibit proliferation and induce death of cultured human
glioblastoma multiforme cells. J Neurooncol 74:3140
Munson AE, Harris LS, Friedman MA, Dewey WL, Carchman
RA (1975) Antineoplastic activity of cannabinoids. J Natl
Cancer Inst 55:597602
Petersen G, Moesgaard B, Schmid PC, Schmid HH, Broholm
H, Kosteljanetz M, Hansen HS (2005) Endocannabinoid
metabolism in human glioblastomas and meningiomas compared to human non-tumour brain tissue. J Neurochem
93:299309
Salazar M, Carracedo A, Salanueva IJ, Hernandez-Tiedra S,
Lorente M, Egia A, Vazquez P, Blazquez C, Torres S, Garcia
S, Nowak J, Fimia GM, Piacentini M, Cecconi F, Pandolfi PP,
Gonzalez-Feria L, Iovanna JL, Guzman M, Boya P Velasco G
(2009) Cannabinoid action induces autophagy-mediated cell
death through stimulation of ER stress in human glioma cells.
J Clin Invest 119:13591372
Sanchez C, de Ceballos ML, del Pulgar TG, Rueda D, Corbacho
C, Velasco G, Galve-Roperh I, Huffman JW, Ramon y Cajal
S, Guzman M (2001) Inhibition of glioma growth in vivo
by selective activation of the CB(2) cannabinoid receptor.
Sanchez C, Galve-Roperh I, Canova C, Brachet P, Guzman M
(1998) Delta9-tetrahydrocannabinol induces apoptosis in C6
glioma cells. FEBS Lett 436:610
Sarfaraz S, Afaq F, Adhami VM, Mukhtar H (2005) Cannabinoid
receptor as a novel target for the treatment of prostate cancer.
Schley M, Stander S, Kerner J, Vajkoczy P, Schupfer G, Dusch
M, Schmelz M, Konrad C (2009) Predominant CB2 receptor
expression in endothelial cells of glioblastoma in humans.
Brain Res Bull 79:333337
Stella N (2004) Cannabinoid signaling in glial cells. Glia
48:267277
Valenzano KJ, Tafesse L, Lee G, Harrison JE, Boulet JM,
Gottshall SL, Mark L, Pearson MS, Miller W, Shan S, Rabadi
L, Rotshteyn Y, Chaffer SM, Turchin PI, Elsemore DA,
Toth M, Koetzner L, Whiteside GT (2005). Pharmacological
and pharmacokinetic characterization of the cannabinoid
receptor 2 agonist, GW405833, utilizing rodent models
of acute and chronic pain, anxiety, ataxia and catalepsy.
Neuropharmacology 48:658672
Van Sickle MD, Duncan M, Kingsley PJ, Mouihate A, Urbani
P, Mackie K, Stella N, Makriyannis A, Piomelli D, Davison
JS, Marnett LJ, Di Marzo V, Pittman QJ, Patel KD, Sharkey

KA (2005) Identification and functional characterization of
brainstem cannabinoid CB2 receptors. Science 310:329332
Velasco G, Carracedo A, Blazquez C, Lorente M, Aguado T,
Haro A, Sanchez C, Galve-Roperh I, Guzman M (2007)
Cannabinoids and gliomas. Mol Neurobiol 36:6067
287
Zha J, Harada H, Yang E, Jockel J, Korsmeyer SJ (1996) Serine

phosphorylation of death agonist BAD in response to survival factor results in binding to 14-3-3 not BCL-X(L). Cell
87:619628
Chapter 30
Patients with Recurrent High-Grade Glioma: Therapy

with Combination of Bevacizumab and Irinotecan
Benedikte Hasselbalch, Ulrik Lassen, and Hans Skovgaard Poulsen
Abstract Despite surgery, radiotherapy, and

chemotherapy, most patients with recurrent high-grade
glioma have a poor prognosis. High-grade gliomas
are known to be vascular tumors that produce the
vascular endothelial growth factor (VEGF), one of the
mediators necessary for angiogenesis which facilitates
for tumor growth. Treatment of recurrent high-grade
gliomas with the VEGF neutralizing antibody, bevacizumab in combination with the topoisomerase
inhibitor irinotecan has demonstrated promising
response rates and enhanced progression free survival
when compared to historical controls. However, it
still needs to be confirmed, that bevacizumab and
irinotecan improves overall survival. Moreover, as not
all patients benefit from this treatment, there is an
ongoing search for one or more predictive biomarkers.
Keywords HGG Bevacizumab Irinotecan VEGF
Topoisomerase I Hypoxia
Introduction
Malignant primary brain tumors represents 12% of all
newly diagnosed tumors and account for about 2% of
all cancer-related deaths. Overall, malignant gliomas
account for 78% of malignant brain tumors. Primary
brain tumors (PBT) are of neuroepithelial origin and
according to WHO classification there are three main
H.S. Poulsen ()

Department of Radiation Biology, The Finsen Center, Sec
6321, Copenhagen University Hospital, DK-2100 Copenhagen,
Denmark
e-mail: hans.skovgaard.poulsen@rh.regionh.dk
types which usually can be distinguished by their histological features; oligodendrogliomas, mixed oligoastrocytomas and astrocytomas (or gliomas) (Louis et al.,
2007). Through analyzing the most malignant region
of the tumors, PBT are graded as low-grade tumors
(WHO grades I and II), or as high-grade tumors
(WHO grades III and IV) dependent on four main
features: nuclear atypia, mitoses, microvascular proliferation, and necrosis. By the degree of increasing
anaplasia, the types of astrocytomas usually include
pilocytic astrocytoma (grade I), diffuse astrocytoma
(grade II), anaplastic astrocytoma (grade III) and
the most malignant form, glioblastoma multiforme
(grade IV/GBM). In addition, grade III malignant
tumors include anaplastic oligodendroglioma, anaplastic oligoastrocytoma and mixed glioma.
Areas of vascular proliferation and/or necrosis are
mandatory for GBM whereas occasional proliferation
of tumor vessels can occur in grade III astrocytomas
(Louis et al., 2007). The pronounced vascularization
arises because of increased angiogenesis. The dense
vascularity does not prevent the HGG tumor from
being hypoxic, partly because of the dysfunctional
nature of the tumor vessels.
Standard treatment for HGG is debulking surgery
if possible. For GBM, this is followed by concomitant
R
temozolomide (Temodal
); an oral alkylating agent,
plus radiotherapy and adjuvant temozolomide (Stupp
et al., 2005). The introduction of temozolomide has
improved the survival of GBM significantly, increasing
the 2-year survival from 10 to 27% and 5-year survival
from 1.9 to 9.8%, compared to previous treatment regimens (Stupp et al., 2009). However, nearly all patients
with GBM will eventually relapse. The prognosis for
recurrent GBM is even worse with a median survival of

289
290
39 months when using traditional chemotherapeutic

agents.
The survival benefit of temozolomide as part of a
primary setting for grade III astrocytomas still needs
to be determined. Accordingly, the primary treatment
of grade III astrocytoma varies between the continents, e.g. most patients in north America receives
the same treatment regimen as GBM, whereas most
places in Europe use surgery and radiotherapy as primary treatment, giving temozolomide at recurrence,
-if possible, after tumor-reductive surgery. The median
survival is longer for this group of patients (24 years),
although most typically recur and may progress to
GBM. The primary treatment of anaplastic oligodendroglioma includes debulking surgery followed by
radiotherapy, whereas the timing and survival benefit
from the addition of chemotherapy still remains to be
established. Two large randomized control trials failed
to show that the addition of chemotherapy with procarbazine, lomustin and vincristin (PCV) either adjuvant
or neoadjuvant to radiotherapy significantly prolonged
disease free survival (Cairncross et al., 2006; van den
Bent et al., 2006). However, as most patients randomly
assigned to RT alone received chemotherapy at progression, interpretation of the survival data from these
studies are difficult.
HGG are still considered incurable and accordingly there is a pivotal need for improved treatment
strategies for this malignancy. The aim of therapy
for patients with relapsed HGG is to offer a better
quality of life in terms of improvement in neurological function, neuro-cognitive function and steroid
requirements, ideally coupled with a prolongation of
survival.
Despite numerous studies using temozolomide,
PCV, carboplatin, lomustin (CCNU), carmustine
(BCNU), or imatinib with or without the combination
of hydroxyurea, no regimen has emerged as a standard
for relapsed HGG after radiation and temozolomide.
Recently, the use of bevacizumab and irinotecan has
shown promising results in recurrent HGG, improving progression free survival (PFS) and quality of
life compared with historical results (Friedman et al.,
2009; Wong et al., 1999). However, it still needs to
be established if a regimen including bevacizumab for
recurrent HGG, improve overall survival (OS) compared to other treatment regimens used at recurrence.
Furthermore, there is an ongoing search for one or
more biomarkers that can predict which patients will
benefit from treatment with bevacizumab.
B. Hasselbalch et al.
Hypoxia
Hypoxia plays a prominent role in tumor growth,
invasion, angiogenesis, resistance to chemo- and
radiotherapy and decreased patient survival in various
cancer types, including HGG. The characteristic
necrotic regions of GBM are assumed to be regions of
hypoxia, although this involvement is not conclusively
determined.
When available blood flow cannot fulfill the requirements for maintaining oxygen homeostasis, the partial
oxygen pressure of these tumor areas become low,
i.e. hypoxic, or close to zero, anoxic. The diffusion
limit for oxygen is approximately 100 m and oxygen transport over further distances requires red blood
cells. Tumor hypoxia evolves as a consequence of
insufficient oxygen delivery and is a feature of most
solid tumors. Tumor growth results in increased distance to existing blood vessels, which in combination
with insufficient neo-vascularization, contributes to a
tumor microenvironment with low oxygen tension.
Moreover, tumor vessels are leaky, leading to tumor
edema and increased intratumoral pressure, which further increases hypoxia. Cancer cells undergo numerous
changes that enable them to adapt to and survive
hypoxia, contributing to a more aggressive behavior
of the tumor. The hypoxia inducible factors (HIF),
HIF-1 and HIF-2, are critical for this adaptive
response.
The characteristic necrotic regions of GBM are
surrounded by a cluster of cells known as pseudopalisading that are suspected to be regions of hypoxia,
although this has not been conclusively proven. These
necrotic areas do not seem to be related to tumor
size, as they have been found in both small and large
tumors. Furthermore, it has been demonstrated in animal glioma models that tumors <1 mm in diameter are
more hypoxic and poorly perfused with sparse vasculature as compared to larger tumors (14 mm in
diameter) (Jensen, 2009). This suggests that necrosis might not be simply due to inadequate vascular
supply, but instead a result of intrinsic molecular or
genetic changes within the tumor. Hypoxia also seems
to induce HGG cell migration and invasion (Jensen,
2009). However, the extent of hypoxia in HGG has still
to be elucidated.
HIF-1 overexpression and angiogenesis have been
shown to correlate in brain tumors (Jensen, 2009),
and there is a significant association between HIF-1
overexpression and tumor grade (Semenza, 2003).
30 Patients with Recurrent High-Grade Glioma
Cellular Response to Hypoxia

The HIF-1 transcription factor mediates adaptive
responses to changes in tissue oxygenation by regulating numerous genes involved in e.g. angiogenesis,
cell proliferation, vascular reactivity and remodeling,
and survival. HIF-1 is a member of the basic helixloop-helix-PAS (bHLH-PAS) family, which includes
the hypoxia regulated HIF-1, 2, 3 and the constitutively expressed HIF-1 (also known as ARNT)
(Semenza, 2001).
HIF-2 and, especially HIF-1 have been most
intensively studied, and the two factors display some
differences, which will be mentioned below when
relevant (otherwise described as HIF-). HIF-1 is
expressed in an apparently ubiquitous fashion, whereas
HIF-2 expression is restricted to particular cell
types, including brain and vascular endothelial cells.
HIF-1 and HIF-2 are both implicated in tumor
growth and are frequently coexpressed in human
tumors.
The HIF- proteins form a transcriptional active
heterodimer with HIF-1 during hypoxia, which initiates transcription by binding to hypoxia response
elements (HREs) in promoters or enhancers of target
genes (Semenza, 2001). The HIF- expression is determined by the rate of protein synthesis regulated via
O2 -independent mechanisms whereas protein degradation and activity is regulated via O2 -dependent mechanisms. HIF-1 synthesis is regulated by activation of
receptor tyrosine kinases (RTKs), non-receptor tyrosine kinases og G-protein-coupled receptors through
the phosphatidylinositol 3-kinase (PI3K) and the
Ras/mitogen activated protein kinase (MAPK) pathways (Semenza, 2003). Furthermore, HIF-1 is known
to induce the expression of the epidermal growth factor
receptor (EGFR) ligand, transforming growth factor
(TGF-) thus providing an autocrine loop regulating the hypoxic response. During normoxia, the
HIF- proteins are degraded through the proteasomal
machinery. HIF- is also regulated at the level of transcriptional activity, as hypoxia (12% O2 ) results in
stabilization of the HIF- proteins allowing HIF- to
form a complex together with HIF-1 which initiate transcription among others of vascular endothelial
growth factor (VEGF) (Semenza, 2003). Exclusive target genes for HIF-2 has yet to be identified, however,
it has been shown that HIF-2 can regulate cancer
stem cell function and/or differentiation through the
291
octamer-binding transcription factor (Oct-4) which in

turn contributes to HIF-2 activity.
Angiogenesis
High micro vessel density (MVD) is a hallmark for
GBM and pronounced tumor vascularity is significantly correlated with poor survival (Jain et al., 2007).
Angiogenesis is development of new vessels from preexisting ones by sprouting or by intussusception from
their vessels of origin. Many molecules are implicated
in the positive regulation of angiogenesis, e.g. acidic
fibroblast growth factor (aFGF), basic FGF (bFGF),
epidermal growth factor (EGF), transforming growth
factor (TGF-), TGF-, angiogenin, interleukin 8
(IL-8), angiopoitins and VEGF (Fig. 30.1) (Bergers
and Benjamin, 2003). In tumor development, angiogenesis is essential for tumor growth and being one
of the most vasculized tumors, angiogenesis seems
fundamental for HGG.
Tumor vasculature is characterized as immature and
malformed with abnormal branching resulting in a
chaotic structure. Moreover, the leaky nature of tumor
vessels induce edema, which further promotes the
hypoxic tumor milieu and induction of pro-angiogenic
factors like VEGF thus creating a positive paracrine
loop, maintaining angiogenesis and conditions necessary for sustained tumor growth. Besides activating
tissue endothelial cells, angiogenic factors also activates circulating endothelial cells (CECs) and endothelial progenitor cells (EPCs) from the bone marrow, that
enter the circulation and generates new blood vessels in
tumor tissue.
VEGF/VEGFR Signaling Pathway

in Pathological Conditions
VEGF is the major endothelial mitogen in central
nervous system neoplasms. Of endogenous angiogenic factors identified, the VEGF family and the
angiopoietins are the endothelium-specific angiogenic
factors. The VEGF family consists of at least five ligands (VEGF-A, -B, -C, -D, and placenta-like growth
factor (PlGF)) and three tyrosine kinase receptors
(VEGFR-1, -2, -3). VEGF-A (also known as VEGF) is
a 3445 kDa dimeric glycosylated protein. Alternative
292
Fig. 30.1 Tumor induced VEGF release gives rise to angiogenesis and increased vessel permeability. Modified from Tabernero
(2007)
exon splicing of the VEGF gene results in at least five

isoforms of VEGF (VEGF121 , VEGF145 , VEGF165 ,
VEGF189 and VEGF206 ), with VEGF165 being the predominant form in general and the most common form
found in HGG (Ferrara et al., 2003). VEGF is a survival factor for endothelial cells of newly formed but
not established vessels within the tumor. Moreover,
VEGF is a major permeability factor partly responsible
for the loss of blood-brain barrier (BBB) during tumor
growth.
The elevated expression of VEGF in human cancer
is likely induced by numerous mechanisms, of which
hypoxia via HIF-1 plays a key role as described
above. Tumor cells are the main source of VEGF in
HGG, whereas VEGF receptors are predominantly
expressed by endothelial cells. Furthermore, several
major growth factors, including EGF, TGF- and
TGF-, insulin-like growth factor-1 (IGF-1), bFGF,
IL-8 and PDGF up regulate VEGF mRNA expression
in a paracrine or autocrine manner, in cooperation
with hypoxia (Fig. 30.1). In addition to VEGF, glioma
cells are known to produce a variety of pro-angiogenic
factors, including bFGF, PDGF, IL-8, HIF-1
and hepatocyte growth factor (HGF) (Fig. 30.1).
Although tumor cells represent the major source of
VEGF, tumor-associated stroma is also an important

site of VEGF production. The expression of VEGF is
especially prominent in tumor cells around necrotic
areas in GBM, and increased concentrations of VEGF
have been found to correlate with malignancy grade
(Steiner et al., 2004) and in one study (Sathornsumetee
et al., 2008) with radiological response to bevacizumab
in HGG.
VEGF binds to two related RTKs, VEGFR-1
(Flt-1) and VEGFR-2 (KDR or Flk-1) which are both
expressed on endothelial cell surface and are found
to have increased expression in GBM, when compared to normal brain (Steiner et al., 2004). VEGFR3 is not a receptor for VEGF, but instead binds
VEGF-C and VEGF-D and its expression is largely
restricted to lymphatic endothelial cells. Despite being
the first VEGF receptor to be identified, the precise role of VEGFR-1 in angiogenesis is to be fully
elucidated.
VEGFR-2 is expressed both on endothelial cells
and tumor cells. VEGFR-2 undergoes dimerization
and ligand-dependent tyrosine phosphorylation inducing phosphorylation of among others; protein kinase
C- (PKC-), phosphatidylinositol 3-kinase (PI3K),
Ras and the Src kinase familiy. VEGFR-2 is the major
receptor involved in angiogenesis and VEGF activation leads to endothelial cell survival, proliferation,
endothelial cell migration, and vascular permability.
VEGF also interacts with the co-receptors neurophilin1 and neurophilin-2.
Anti-angiogenic Drugs
R
Bevacizumab (Avastin
) is a humanized immunoglobulin (Ig) G1 that binds to and inhibits the biologic activity of all active isoforms of the human VEGF
ligand (VEGF-A) (Fig. 30.2). It is administrated intravenously and the terminal half-life of bevacizumab
in humans is 1721 days. Bevacizumab was the first
inhibitor of angiogenesis to be approved by FDA,
based on the survival benefit observed in a randomized
phase III trial, for patients with metastatic colorectal
cancer in combination with conventional chemotherapy. Bevacizumab has also been FDA approved as first
line treatment of advanced non-small-cell lung cancer in combination with standard chemotherapy. In
addition it has been approved for metastatic HER2
negative breast cancer in combination with paclitaxel, and for metastatic renal cancer in combination
with interferon alpha. Bevacizumab is well tolerated
and lacks the major toxicities typically associated
Fig. 30.2 Examples of

different targeted therapies
investigated in HGG. The
different types of therapeutic
compounds are for
simplification assembled in
the yellow boxes according to
their respective targets. Tumor
angiogenesis can be inhibited
by monoclonal antibodies
against the pro-angiogenic
factor VEGF or by TKI
targeting VEGFR. See text for
further details
293
with cytotoxic chemotherapies. Common toxicities

related to anti-angiogenic drugs are hypertension, proteinuria, fatigue, tromboembolic events, hemorrhage
and wound-healing complications. Hypertension is
normally manageable with ordinary anti-hypertensive
treatment, whereas treatment delay can be necessary at
impaired wound heeling.
In addition to bevacizumab, another anti-VEGF
therapy in clinical trials is aflibercept (also known
as VEGF Trap), a soluble decoy VEGFR that binds
VEGF, VEGF-B and placenta-like growth factor
(PlGF) (Fig. 30.2).
Another strategy for anti-angiogenic treatment is
inhibition of the VEGFR. Several inhibitors of VEGFR
are either underway for approval in clinical trials or already approved for cancer therapy. One
such drug is cediranib, a multi-targeted TKI which
blocks VEGFR-1, VEGFR-2, and VEGFR-3 signaling and shows a response rate of 56% as singleagent therapy in recurrent GBM (Batchelor et al.,
R
R
2007). Sorafenib (Nexavar
) and sunitinib (Sutent
)
are FDA approved small molecule TKIs targeting
multiple receptor tyrosine kinases, including VEGFR
and PDGFR (Norden et al., 2009). An overview of
VEGF/VEGFR inhibitors are presented in Fig. 30.2.
Finally, thalidomide and the analog lenalidomide
are also under investigation as anti-angiogenic therapy
294
for HGG either as monotherapy or in combination

with e.g. temozolomide or BCNU. However, the preliminary results so far are not superior to the abovementioned treatment.
Irinotecan
R
R
Irinotecan (Camptosar
, Campto
) is a topoisomerase I inhibitor, used as first-line treatment in for
metastatic colorectal cancer and which has demonstrated high activity against solid tumors of the gastrointestinal tract. Irinotecan is administrated intravenously and is able to cross the BBB but demonstrates
only limited effect against HGG when used as singleagent therapy, with response rates between 0 and 15%
(Vredenburgh et al., 2008).
Topoisomerase I play a crucial role in the replication of DNA. In the chromosome, the DNA helix
is supercoiled and tightly packed into chromatin.
Topoisomerase I transiently remove the negative supercoils, facilitating transcription and replication of the
parent DNA. Topoisomerase I and II activities are significantly enhanced in malignant gliomas following
DNA damage.
Irinotecan undergo carboxylesterase-mediated
breakdown to the active metabolite, SN-38 that is 10
1,000 times more potent than irinotecan. Glioma cells
can convert irinotecan to SN-38 directly (Vredenburgh
et al., 2008). SN-38 is eliminated primarily by the lever
to the inactive metabolite SN-38 glucuronide. The
terminal half-lives for both irinotecan and SN-38 are
7.914.2 h and 13.013.8 h respectively. In addition to
biotransformation of to SN-38, irinotecan is oxidated
by cytochrome P450 enzymes (especially CY3A4/5)
forming a number of relative inactive metabolites.
The use of anticonvulsant in CNS tumor patients
are frequent and several antiepileptic medications
have been shown to induce cytochrome P450 enzyme
activity and thereby also increase systemic clearance
of irinotecan, reducing systemic exposure to irinotecan
and SN-38 considerably. Accordingly, a higher dose of
irinotecan is necessary for patients receiving enzymeinducing antiepileptic drugs (EIAEDs) (Vredenburgh
et al., 2008).
The main toxicities observed with irinotecan are
myelosuppression that can affect all the bloodforming elements of the bone marrow. Leukopenia,
neutropenia, anemia, and thrombocytopenia have all

been reported following irinotecan administration with
neutropenia being the most clinically significant consequence of myelosuppression. Moreover, nausea, vomiting, and diarrhea are common adverse events following treatment with irinotecan. Especially diarrhea can
be severe and a dose-limiting toxicity of irinotecan.
Bevacizumab in HGG
It is generally accepted that tumor growth is dependent
of angiogenesis in most solid tumors, and being one
of the most vasculized tumors, angiogenesis appears to
be of importance in malignant glioma growth. VEGF
secreted by the glioma cells acts by paracrine mechanisms upon endothelial cells resulting in endothelial
cell proliferation, survival and migration. VEGF might
also have an impact on VEGFR expressing glioma
cells, although the effect has not been clarified.
The efficacy of bevacizumab in recurrent HGG, was
first described by Stark-Vance (2005), which combined bevacizumab and irinotecan. Subsequently, several studies have shown the efficacy of bevacizumab in
recurrent HGG (Norden et al., 2009) and in May 2009,
the American food and drug administration (FDA)
approved the use of bevacizumab as a single agent
for patients with recurrent GBM based on two phase
II studies showing durable objective response rates
(Friedman et al., 2009; Kreisl et al., 2008). The promising results obtained with bevacizumab are -however,
shown to be temporary, as recurrence is inevitable
and despite prolonged PFS, overall survival remains
largely unchanged when compared to the historic controls often used as reference for phase II trials using
bevacizumab (Wong et al., 1999).
Jain et al. found that VEGF inhibition transiently normalizes the disorganized and dysfunctional tumor vasculature in some experimental models, potentially improving delivery of oxygen and
cytotoxic drugs to tumor cells (Jain, 2005). Given
the transient nature of this phenomenon, it remains
unknown whether the proposed vasculare normalization model has relevance in the long-term therapeutic
effects of bevacizumab-chemotherapy. Nevertheless,
as described below, this effect is of importance
when evaluating treatment response radiographically
on computed tomography (CT) or magnetic resonance
imaging (MRI).
The clinical and radiological benefit of bevacizumab and other anti-angiogenic therapies is indisputable. However, only 2030% of HGG patients experience this effect and there is a compelling need to
select and stratify patients most likely to benefit from
the treatment. Consequently, there is an ongoing search
for one or more valid biomarkers, which could prove to
be predictive for response to treatment.
Treatment with Bevacizumab

and Irinotecan
The combination of irinotecan and bevacizumab was
shown to be feasible and demonstrating a significant survival benefit when used phase III trial of
patients with metastatic colorectal cancer comparing
bevacizumab plus irinotecan/5-fluoruracil/leucovorin
versus irinotecan/5-fluoruracil/leucovorin alone.
In one of the first reported studies, 29 patients
with recurrent malignant glioma were treated with
bevacizumab and irinotecan, of which three patients
had complete responses (CR), 16 partial responses
(PR) and seven stable diseases (SD) according to
the MacDonald criteria (Macdonald et al., 1990;
Vredenburgh et al., 2007). These promising results
have subsequently been confirmed by several other
groups (Poulsen et al., 2009; Vredenburgh et al.,
2007).
The proposed dose of bevacizumab for patients with
relapsed HGG is 10 mg/kg every other week (q2w)
or 15 mg/kg every third week (q3w), with 10 mg/kg
q2w being the most common regimen used. No direct
comparison of the efficacy of bevacizumab administrated q2w or q3w has been performed in patients with
relapsed HGG. Irinotecan is administrated in combination with bevacizumab using 340 mg/m2 for patients
receiving EIAEDs and 125 mg/m2 for patients not
receiving EIAEDs.
To date, the importance of irinotecan in the regime
of BI still needs to be elucidated. The FDA approval of
bevacizumab as monotherapy in recurrent GBM was
based on two phase II studies performed by Kreisl
et al. (2008) and Friedman et al. (2009). The first
study demonstrated that the effect of bevacizumab
as monotherapy in recurrent GBM was feasible and
responses were comparable with previous reports
using BI in recurrent GBM (Kreisl et al., 2008). In
295
the second study, patients with recurrent GBM were

randomized in a noncomparative phase II trial to
bevacizumab alone or in combination with irinotecan
(Friedman et al., 2009). The primary objective of this
trial was evaluation of safety and efficacy, and there
was no intension of comparing the outcome of the two
treatment groups, although it was observed that data
did not indicate a treatment benefit of the addition
of irinotecan. Subsequently, bevacizumab was FDA
approved as monotherapy in recurrent GBM. However,
a randomized phase III study (randomizing between
bevacizumab versus bevacizumab and irinotecan) is
still necessary to determine the effect of irinotecan.
Even better would be a randomized trial between
bevacizumab and one or more of the previously used
regimens like PVC, BCNU or CCNU since a phase
III trial showing a survival benefit of bevacizumab in
recurrent GBM still remains to be performed. For this
reason, European Medicines Agency (EMEA) recently
did not approve bevacizumab for this indication in
2009 (http://www.emea.europa.eu), in contrast to the
FDA approval.
However, due to the promising effect of bevacizumab at recurrent disease, there are expectations
that bevacizumab would be beneficial at the primary
setting. Thus there are two ongoing phase III trials,
randomizing between the standard regime (temozolomide and radiotherapy (Stupp et al., 2005)) or standard
regime plus bevacizumab as first line treatment for
GBM (http://www.clinicaltrials.gov).
The Clinical Effect of Bevacizumab

Bevacizumab, with or without concomitant irinotecan,
has indisputably shown activity in recurrent HGG in
terms of clinical response. Corticosteroid use, neurocognetive function and performance status are likely
to affect quality of life and accordingly it is reasonable
to use these as surrogate markers. An improvement
in or maintenance of steroid dose, performance status
or clinical status is almost invariable in patients with
clinical response (CR + PR) and frequently found in
patients with SD whereas it is uncommon in patients
with PD (Friedman et al., 2009; Poulsen et al., 2009).
This indicates a benefit in terms of the patients health
related quality of life although this has not been profoundly investigated so far.
296
Whether bevacizumab is truly superior to other

available therapies is unknown due to the lack of
directly comparative trials. As the patient characteristics are not identical in the historical controls compared
to patients in uncontrolled phase II studies with bevacizumab, it is not possible to determine if bevacizumab
would be more efficient than previous treatments in
terms OS.
The Difficulty in Evaluating Clinical

Response When Using Anti-angiogenic
Therapy
The effects of bevacizumab on tumor vasculature
have given rise to challenges in response evaluation. Disruption of the BBB by the tumor results
in increased accumulation of fluid and plasma proteins in the surrounding brain. Because of the lack of
lymphatic vasculature in the brain, and the fact that
it is located in a confined space, the fluid leakage
leads to increased interstitial pressure within the tumor
and accumulation of fluid outside the tumor, resulting
in vasogenic brain edema. Corticosteroids have been
used for decades as temporary control of vasogenic
brain edema, with moderate efficacy but also numerous side effects. The vascular normalization induced
by anti-angiogenic agents like bevacizumab has shown
to alleviate brain edema (Vredenburgh et al., 2007).
This steroid effect might also improve drug delivery. However, the steroid effect from anti-angiogenic
therapy gives rise to additional challenges when evaluating tumor load and response by CT or MRI scans.
The MacDonald criteria is still used for evaluation
of response to treatment (Macdonald et al., 1990).
These criteria are based on the WHO criteria using the
contrast-enhanced largest cross-sectional area of tumor
on CT or MRI scans in combination with corticosteroid
use and changes in neurological function. However,
enhancement is nonspecific and primarily reflects a
disrupted BBB. Besides bevacizumab, enhancement
can be influenced by changes in corticosteroid dose
and radiologic technique. Thus, when treating HGG
with anti-angiogenic therapy like bevacizumab, the
response to treatment observed may result at least
partially from the bevacizumab induced normalization of abnormally permeable blood vessels and not
from anti-tumor activity. Furthermore, anti-angiogenic
treatment might control the contrast enhancing tumor

more effectively than non-enhancing tumor, causing
problems in interpretation of CT or MRI scans (LucioEterovic et al., 2009). This is also reflected in the
overall survival data from patients treated with antiangiogenic therapy, which fail to show prolonged OS
in recurrent HGG despite a promising response rate
and PFS (Norden et al., 2008a). Accordingly, other
response measurements are needed in evaluation of
tumor response in HGG taking into account both
enhancing and non-enhancing tumor, the latter being
best visualized on T2 weighted and fluid-attenuated
inversion recovery (FLAIR) MRI sequences. In order
to improve endpoints in clinical trials and response
criteria, an international working party [Response
Assessment in Neuro-Oncology[(RANO)] has been
established and recommendations are still to come.
Biomarkers as Surrogate Markers

for Clinical Response
As a consequence of the development and use of
targeted therapies, the identification of biomarkers correlating with disease activity and the efficacy of treatment is rapidly evolving. However, so far none have
been shown to predict a group of patients must likely
to benefit from bevacizumab treatment, nor predict
treatment response.
Because of heterogeneity and insufficiency of tumor
vasculature in HGG tumors, hypoxia within the tumor
can be chronic or acute (fluctuating) although the
importance of this is not known with respect to survival and/or response to anti-angiogenic treatment in
HGG. This could also influence the potential information obtained from tissue related hypoxic biomarkers,
as their precise role in HGG when treated with antiangiogenic compounds, needs to be revealed before
they can be used as predictive biomarkers. When
using an anti-angiogenic agent for the treatment of a
highly vascularized tumor as GBM, it is intriguing to
assume that MVD would be of importance for response
and hence could be used as a predictive biomarker.
However, as emphasized in the review by Hlatky et al.
(2002) MVD is not equivalent to the degree of tumor
angiogenic activity and measurement of MVD is not
predictive of tumor response under anti-angiogenic
treatment. Consequently the level of MVD in the tumor
should not be used to decide which patients would

benefit from anti-angiogenic treatment.
Immunohistochemical methods have been the most
used approach in the search for predictive and/or
prognostic biomarkers. VEGF and carbon anhydrase
9 (CA9) have been found to be associated with
radiologic response and survival outcome in 27
HGG patients treated with BI in one previous study
(Sathornsumetee et al., 2008). However, these observations have not been confirmed in a larger study.
This could be caused by heterogeneity of HGG tumors
where the small tumor sample investigated might not
be representative for the gross mass. Moreover, small
differences in techniques and tissue handling could
influence antigen preservation and thereby staining
(Hasselbalch et al., 2010).
In the search of biomarkers that predict response
and resistance to anti-angiogenic therapy, additional
modalities are under investigation. This include systemic, circulating, tissue and imaging biomarkers
(reviewed in Jain et al. (2009)). However, as described
above, the use of anti-angiogenic treatment has given
rise to difficulty in establishing adequate criteria of
response. Accordingly, this issue must be solved before
any information regarding response can be used in
combination with potential predictive biomarkers.
Moreover, the biomarkers investigated might not
be representative for the underlying biological mechanisms inducing response (or no response) to antiangiogenic therapy in HGG. This area still remains
to be thoroughly investigated. In addition, it must
be emphasized that most new drugs including bevacizumab are often tested in recurrent disease, from
which tissue is not available. Recurrent tumor may be
different from the primary tumor in terms of genetic
expression and relevance of specific targets. Thus correlative studies based in tumor tissue and expression of
tissue biomarkers may not capture meaningful associations of response and/or survival data.
Resistance to VEGF Pathway Inhibitors

Is Inevitable
VEGF pathway inhibitors induce only temporary
tumor stasis or shrinkage but fail to produce enduring
clinical responses in HGG. Despite the frequent benefit of bevacizumab and other anti-angiogenic drugs
297
used for treatment of HGG, tumor progression is

inevitable.
The traditional concept of drug resistance involves
mutational alterations of the target gene or alteration
in the drug uptake and/or efflux resulting in treatment
failure. This appears to be different when using angiogenic inhibitors, as the tumor functionally evade the
therapeutic blockade of angiogenesis even though the
specific therapeutic target remains inhibited. Instead,
alternative pathways are activated resulting in angiogenesis and sustained tumor growth.
In order to improve the effect of anti-angiogenic
treatment and achieve improved OS, we need to understand this mechanism of resistance. In a recent study
by Lucio-Eterovic et al. (2009) it was demonstrated
in vitro that bevacizumab was able to sequester the
majority of secreted VEGF in glioblastomas (LucioEterovic et al., 2009). In addition, it was observed
that bevacizumab induced upregulation of several proangiogenic molecules in vivo (bFGF) and in vitro
(i.e. bFGF, IL-1, angiogenin and TGF), which
supports the idea that one of the reasons for lack
of sustained effect from anti-angiogenic treatment is
caused by up-regulation of additional pro-angiogenic
molecules. Another concern is that inhibition of angiogenesis leads to an infiltrative tumor growth pattern with co-option of existing cerebral blood vessels
thereby achieving vascular sufficiency (Lucio-Eterovic
et al., 2009). Although not pathologically confirmed,
this observation is supported in several clinical studies, which suggest an increased invasive growth on
MRI scans from bevacizumab treated HGG patients
(Norden et al., 2008b). In addition, it has been demonstrated in an in vitro model, that addition of bevacizumab induced an increased migration/invasion of
glioma cells in a concentration-dependent manner
(Lucio-Eterovic et al., 2009). As glioma cells releases
VEGF, this is suggesting that autocrine VEGF signaling blockade plays an important role in glioma cell
invasion. This could also be of importance in the resistance pattern observed upon anti-angiogenic therapy,
thus reduced levels of VEGF gives rice to increased
invasion. In addition, these findings also indicate
that VEGF can influence glioma cells directly which
are known to express both VEGFR-1 and VEGFR-2
and that the effects are not restricted to the influence on endothelial cells. Moreover, Lucio-Eterovic
et al. (2009) demonstrated both in vitro and in vivo,
that bevacizumab treatment induced up-regulation
298
of invasion-related proteins (e.g. matrix metalloproteinase (MMP) -2, -9 and -12 and tissue inhibitor
of metalloproteinases 1 (TIMP1) which further
supports the idea that GBM cells can escape from antiangiogenic treatment by up-regulating molecules that
allow them to invade into surrounding brain areas.
The mechanism of resistance to anti-angiogenic
therapy has been profoundly reviewed by Bergers and
Hanahan (2008) who suggest four adaptive mechanisms that induce resistance to anti-angiogenic therapy. The first two overrule the necessity of VEGF
by (1) activation and/or upregulation of alternative
pro-angiogenic pathways as mentioned above or (2)
recruitment of bone marrow-derived pro-angiogenic
cells. Next, (3) the increased pericyte coverage of the
tumor vasculature, which is known to occur, is serving to support its integrity, attenuating the necessity
for VEGF-mediated survival signaling. Finally, and
discussed above (4) they also state that inhibition of
angiogenesis leads to an infiltrative tumor growth. This
could originate from the activation and increased invasion of tumor cells into normal tissue, by co-option of
normal blood vessels thereby achieving vascular sufficiency, and could explain the frequently observed
decrease in neurological status, despite the relative
stability of contrast-enhancing tumor on MRI scans.
Furthermore, tumor recurrence could also originate from brain cancer stem cell (bCSC) that are
not known to be influenced by anti-angiogenic treatment. The self-renewing, multipotent and tumorigenic
capacities of bCSC are yet another option for inducing tumor recurrence. In addition, bCSC are able to
migrate throughout the brain parenchyma, which along
with the above mentioned infiltrative growth induced
by anti-angiogenic treatment might explain the diffuse
recurrence pattern observed by MRI scan. Thus, the
absence of response to anti-angiogenic therapy could
be due to intrinsic (pre-existing) resistance or reflect
a rapid adaptation to the above-mentioned evasive
resistance mechanisms.
Conclusion
Several studies have demonstrated notable anti-tumor
activity of bevacizumab in combination with irinotecan in recurrent HGG with promising response rates
and prolongation of PFS when compared to historical
controls. However, overall survival has not improved

significantly, when compared to previous regimens.
In addition, enhanced quality of life, measured by
improvement in neurocognetive status, performance
status and decrease in corticosteroid dose, are almost
always observed in patients with clinical response
(CR + PR). The regimen is well tolerated, with hypertension, tromboembolic events, diarrhea and neutropenia being to must common side effects. Thus, the
effect of irinotecan in this regimen still needs to be
determined. The MacDonald criteria used for evaluation might be misleading since normalization of
tumor vessels and BBB reduce the contrast-enhancing
tumor volume. This could lead to an apparent response,
which is not due to decrease in number of clonogenic
cells. Moreover, as only one third of the patients experience a benefit from the treatment, the ongoing search
for predictive biomarkers is important.
References
Batchelor TT, Sorensen AG, di Tomaso E, Zhang WT, Duda
DG, Cohen KS, Kozak KR, Cahill DP, Chen PJ, Zhu M,
Ancukiewicz M, Mrugala MM, Plotkin S, Drappatz J, Louis
DN, Ivy P, Scadden DT, Benner T, Loeffler JS, Wen PY, Jain
RK (2007) AZD2171, a pan-VEGF receptor tyrosine kinase
inhibitor, normalizes tumor vasculature and alleviates edema
in glioblastoma patients. Cancer Cell 11:8395
Bergers G, Benjamin LE (2003) Tumorigenesis and the angiogenic switch. Nat Rev Cancer 3:401410
Bergers G, Hanahan D (2008) Modes of resistance to antiangiogenic therapy. Nat Rev Cancer 8:592603
Broholm H, Christensen IJ, Grunnet K, Horsman MR, Poulsen
HS, Stockhausen MT, Lassen U (2010) Prospective evaluation of angiogenic, hypoxic and EGFR-related biomarkers
in recurrent glioblastoma multiforme treated with cetuximab,
bevacizumab and irinotecan. APMIS 118:58594
Cairncross G, Berkey B, Shaw E, Jenkins R, Scheithauer B,
Brachman D, Buckner J, Fink K, Souhami L, Laperierre N,
Mehta M, Curran W (2006) Phase III trial of chemotherapy plus radiotherapy compared with radiotherapy alone for
pure and mixed anaplastic oligodendroglioma: Intergroup
Radiation Therapy Oncology Group Trial 9402. J Clin Oncol
24:27072714
Ferrara N, Gerber HP, LeCouter J (2003) The biology of VEGF
and its receptors. Nat Med 9:669676
Friedman HS, Prados MD, Wen PY, Mikkelsen T, Schiff D,
Abrey LE, Yung WK, Paleologos N, Nicholas MK, Jensen
R, Vredenburgh J, Huang J, Zheng M, Cloughesy T (2009)
Bevacizumab alone and in combination with irinotecan in
recurrent glioblastoma. J Clin Oncol 28:47334740
Hlatky L, Hahnfeldt P, Folkman J (2002) Clinical application of
antiangiogenic therapy: microvessel density, what it does and
doesnt tell us. J Natl Cancer Inst 94:883893

Jain RK (2005) Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy. Science 307:5862
Jain RK, Duda DG, Willett CG, Sahani DV, Zhu AX, Loeffler JS,
Batchelor TT, Sorensen AG (2009) Biomarkers of response
and resistance to antiangiogenic therapy. Nat Rev Clin Oncol
6:327338
Jain RK, di Tomaso E, Duda DG, Loeffler JS, Sorensen AG,
Batchelor TT (2007) Angiogenesis in brain tumours. Nat Rev
Neurosci 8:610622
Jensen RL (2009) Brain tumor hypoxia: tumorigenesis, angiogenesis, imaging, pseudoprogression, and as a therapeutic
target. J Neurooncol 92:317335
5:740745
Lucio-Eterovic AK, Piao Y, de Groot JF (2009) Mediators
of glioblastoma resistance and invasion during antivascular endothelial growth factor therapy. Clin Cancer Res 15:
45894599
Macdonald DR, Cascino TL, Schold SC Jr, Cairncross JG
(1990) Response criteria for phase II studies of supratentorial
malignant glioma. J Clin Oncol 8:12771280
Norden AD, Drappatz J, Muzikansky A, David K, Gerard M,
McNamara MB, Phan P, Ross A, Kesari S, Wen PY (2008a)
An exploratory survival analysis of anti-angiogenic therapy
for recurrent malignant glioma. J Neurooncol 2:149155
Norden AD, Drappatz J, Wen PY (2009) Antiangiogenic therapies for high-grade glioma. Nat Rev Neurol 5:610620
Norden AD, Young GS, Setayesh K, Muzikansky A, Klufas R,
Ross GL, Ciampa AS, Ebbeling LG, Levy B, Drappatz J,
Kesari S, Wen PY (2008b) Bevacizumab for recurrent malignant gliomas: efficacy, toxicity, and patterns of recurrence.
Neurology 70:779787
Poulsen HS, Grunnet K, Sorensen M, Olsen P, Hasselbalch B,
Nelausen K, Kosteljanetz M, Lassen U (2009) Bevacizumab
plus irinotecan in the treatment patients with progressive
recurrent malignant brain tumours. Acta Oncol 48:5258
Sathornsumetee S, Cao Y, Marcello JE, Herndon JE,
McLendon RE, Desjardins A, Friedman HS, Dewhirst
MW, Vredenburgh JJ, Rich JN (2008) Tumor angiogenic and
hypoxic profiles predict radiographic response and survival
in malignant astrocytoma patients treated with bevacizumab
and irinotecan. J Clin Oncol 26:271278
Semenza GL (2001) Hypoxia-inducible factor 1: oxygen homeostasis and disease pathophysiology. Trends Mol Med 7:
345350
299
Semenza GL (2003) Targeting HIF-1 for cancer therapy. Nat Rev
Cancer 3:721732
Stark-Vance V (2005) Bevacizumab and CPT-11 in the treatment of relapsed malignant glioma. Neuro Oncol 7:279402
(Abstract 342)
Steiner HH, Karcher S, Mueller MM, Nalbantis E, Kunze S,
Herold-Mende C (2004) Autocrine pathways of the vascular
endothelial growth factor (VEGF) in glioblastoma multiforme: clinical relevance of radiation-induced increase of
VEGF levels. J Neurooncol 66:129138
(2009) Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in
glioblastoma in a randomised phase III study: 5-year analysis
of the EORTC-NCIC trial. Lancet Oncol 10:459466
adjuvant temozolomide for glioblastoma. N Engl J Med 352:
987996
Tabernero J (2007) The role of VEGF and EGFR inhibition: implications for combining anti-VEGF and anti-EGFR
agents. Mol Cancer Res 5:203220
Vredenburgh JJ, Desjardins A, Herndon JE, Dowell JM,
Reardon DA, Quinn JA, Rich JN, Sathornsumetee S,
Gururangan S, Wagner M, Bigner DD, Friedman AH,
Friedman HS (2007) Phase II trial of bevacizumab and
irinotecan in recurrent malignant glioma. Clin Cancer Res
13:12531259
Vredenburgh JJ, Desjardins A, Reardon DA, Friedman HS
(2008) Experience with irinotecan for the treatment of malignant glioma. Neuro Oncol 1:8091
Wong ET, Hess KR, Gleason MJ, Jaeckle KA, Kyritsis AP,
Prados MD, Levin VA, Yung WK (1999) Outcomes
and prognostic factors in recurrent glioma patients
enrolled onto phase II clinical trials. J Clin Oncol 17:
25722578
van den Bent MJ, Carpentier AF, Brandes AA, Sanson M,
Taphoorn MJ, Bernsen HJ, Frenay M, Tijssen CC, Grisold
W, Sipos L, Haaxma-Reiche H, Kros JM, van Kouwenhoven
MC, Vecht CJ, Allgeier A, Lacombe D, Gorlia T (2006)
Adjuvant procarbazine, lomustine, and vincristine improves
progression-free survival but not overall survival in newly
diagnosed anaplastic oligodendrogliomas and oligoastrocytomas: a randomized European Organisation for Research
and Treatment of Cancer phase III trial. J Clin Oncol
24:27152722
Chapter 31
Monitoring Gliomas In Vivo Using Diffusion-Weighted

MRI During Gene Therapy-Induced Apoptosis
Timo Liimatainen, Olli Grhn, and Kimmo Lehtimki
Abstract Diffusion as detected by magnetic

resonance imaging is typically self diffusion of water
in biological samples or in vivo. Diffusion MRI has
been widely studied and applied in clinical settings
including detection of therapy responses in cancers.
Malignant brain tumors especially gliomas are significant cause of cancer related deaths even though
they are a reasonably rare type of cancer. Different
kinds of therapies comprising of surgery, radiation-,
chemotherapy or combined therapies have been
applied routinely for treatment of these malignancies.
Despite the developed therapy methods, the prognosis
for the patients with malignant gliomas remains
poor. Recently, gene therapy methods have also been
developed for the treatment of the malignant glioma.
When gene therapy becomes more widely available,
the value of the imaging methods to detect treatment
response will increase. Changes in the diffusion
of water and metabolites, as measured with MRI
and MRS, have been shown to serve as marker for
positive treatment response in experimental settings.
The main but simplified results from these studies
in cell suspensions and in vivo are: water diffusion
increases as a result of excessive cell loss and increase
in extracellular water, and the diffusion of intracellular
metabolites (total choline) decreases most likely due
to cell shrinkage. The clinical studies are limited
to chemotherapy and radiation therapy responses,
which show similar responses as gene therapy in
experimental models. As a conclusion, amongst
T. Liimatainen ()
Department of Biotechnology and Molecular Medicine,
University of Eastern Finland, Kuopio, Finland
e-mail: timo.liimatainen@uef.fi
the magnetic resonance markers water diffusion is

one of the most extensively studied and it has been
found to be sensitive to treatment response in several
tumor treatment protocols, including gene therapy
approaches.
Keywords Diffusion MRI Gene therapy
Apoptosis Water diffusion Cell death ADC
Introduction
The diffusion of molecules as measured by magnetic
resonance imaging (MRI) or spectroscopy is most
often defined as self diffusion i.e. the displacement
of molecules by random thermal, Brownian motion
without a concentration gradient. Diffusion of water
molecules is most often assessed since water is the
most abundant molecule in the human body and it
contains magnetic resonance (MR) detectable spinhalf (1 H) atoms. For the free non-restricted diffusion,
Stokes-Einsteins well known relation claims that the
diffusion constant D of molecules is proportional to
temperature, inversely proportional to the viscosity of
a solution and inversely proportional to the hydrodynamic radius of a molecule. Stokes-Einsteins relation
holds well in solutions, however diffusion in tissue is
far more complex due to permeable or semi permeable cell membranes, cell organelles, active transport
of molecules, flow, charge, variation in protein content,
viscosity, and so on. The method for measuring diffusion using MR was introduced by Stejskall and Tanner
in the early sixties (Stejskal and Tanner, 1965). They
suggested a pulsed field gradient method for diffusion detection, which is still widely in use. Since those

301
302
T. Liimatainen et al.
times MR techniques and devices have been developed

to provide a number of routine clinical applications.
One of the most lethal cancers are malignant
brain tumors, which include high-grade gliomas
(Germano and Binello, 2009). Surgery, radiation therapy, chemotherapy or combined therapies have been
applied routinely for the treatment of these malignancies. Despite the ongoing effort to develop therapy
methods, the prognosis for patients with malignant
gliomas remains poor (Germano and Binello, 2009).
Gene therapy is a rapidly developing method to treat
malignant tumors when location or other reasons prevent the use of surgical therapy and/or the close proximity of tissues sensitive to ionizing radiation prevents
the use of radiation therapy. Also, the inefficiency or
side effects of chemotherapy for certain types of malignancies may hamper its use. No matter which therapy
method is applied, the aim of the anticancer therapy
is to remove malignant tissue or kill malignant cells.
The response to therapy is likely to vary between
individuals and therefore a non-invasive method for
visualizing and optimizing the treatment protocols is
required. Medical imaging is typically used to follow up the treatment response during such therapies.
Tumor volume, derived from anatomical MRI or CT
images has often been used as a surrogate marker
for treatment response. Tumor volume measurements
show whether or not a treatment was effective in
stabilizing or reversing tumor growth. However, far
before volume response, several cellular level events
take place. This cascade of physicochemical processes
includes numerous components indicating the physicochemical state of the tumor tissue and providing a basis
for imaging markers for therapy response. Among the
markers, diffusion MRI is one of the most extensively studied and it has been found to be sensitive to
treatment response in several tumor treatment protocols, including gene therapy approaches (Kettunen and
Grhn 2005).
insult such as toxicity, hypoxia, ischemia or physical

injury. The insult leads to failure in cell energy
metabolism in the early phase, resulting in swelling
of mithochondria and ruptured nuclei, organelles,
and plasma membranes. Widespread necrosis leads
to the release of lysosome contents into the extracellular space causing inflammation and damage of
neighboring cells.
In contrast, apoptosis is an active and well controlled cell death mechanism that requires adenosine
triphosphate to maintain cell integrity and cell removal
in a controlled manner (Savill and Fadok, 2000 and
references therein). Apoptosis can be divided into
two main stages. First, apoptotic bodies are formed
through a cascade of molecular events and second,
the degradation and phagocytosis of the formed apoptotic bodies takes place. Apoptosis is present in several
processes that modify and remodel tissue architecture
(Kerr et al., 1972). Several pathological changes occur
during apoptotic cell death, such as cell shrinkage,
DNA fragmentation, nuclear condensation, and the
formation of membrane-encapsulated apoptotic bodies
(Kerr et al., 1972), in which the content of the demised
cells are recycled by macrophages and/or neighboring cells to serve the needs of the other cells (Savill
and Fadok, 2000). Interplay between external stimuli such as hormones, growth factors, cytokines and
cell internal sensing mechanisms are needed for a cell
to undergo apoptosis (Hale et al., 1996). Due to its
role in life and death, apoptosis can be thought of as
the default cell death pathway (Kinloch et al., 1999).
For instance, in many cancers apoptosis is turned
off, leading to uncontrolled cell proliferation (Kinloch
et al., 1999). In terms of non-surgical anti-cancer therapy strategies, in many cases, apoptosis is a desired
way to cells die, since harmful intracellular metabolites
and electrolyte fluxes, which could elicit inflammatory reactions and necrosis, remain within the apoptotic bodies. This means that apoptosis is relatively
non-toxic for neighboring cells.
Cell Death and Gene Therapy
Gene Therapy
Cell Death Mechanisms

Cell death may take place through two distinguishable
mechanisms, namely apoptosis or necrosis. Typically
necrosis occurs when the cell is exposed to an acute
Gene therapy is a promising new way to treat patients

with cancer and some other serious diseases. The goal
of the gene therapy is an insertion or modification
of genes in the individuals cell for the purpose of
treatment (Germano and Binello, 2009). The definition
31 Monitoring Gliomas In Vivo Using Diffusion-Weighted MRI
of the gene therapy can be phrased as the transfer of

nucleic acids to the somatic cells of a patient with a
therapeutic intent (Wirth et al., 2009). Gene therapy
can be used alone or together with surgical removal
of the cancerous tissue, which is the most common
target for gene therapy. Gene delivery is performed
using viral vectors, stem cells, and synthetic vectors
like nanoparticles or liposomes. The most often used
approach exploits so-called suicide genes, which cause
wide-spread apoptosis in the malignant tissue leading
to structural changes in tissue that can be detected by
diffusion based MRI. The first gene therapy clinical
protocol was approved by the FDA in 1989. In 2009
there were more than 1400 ongoing gene therapy trials
worldwide. Most of them were cancer gene therapies,
but only few for the treatment of brain cancers. Five
of them were gene therapies targeted to brain cancer in
phase III (Wirth et al., 2009).
Magnetic Resonance Methods for

Detection of Diffusion
The first pulsed gradient diffusion measurement using
magnetic resonance (MR) consisted of a spin echo
pulse sequence with diffusion field gradients inserted
both sides of a 180 degree refocusing pulse (Stejskall
and Tanner, 1965). The diffusion weighting factor b of
the pulse sequence depends on the gyromagnetic ratio
of the nucleus ( ) and user-controllable field gradient parameters; amplitude of the gradients (G), their
duration and time interval () between them

b= G
2
2 2
.

3
In this equation, the diffusion time is tdiff =

(/3). By varying b value in the pulse sequence, the
observed signal intensity in the diffusion experiment
can be altered according to equation
S
= ebD ,
S0
where D is the diffusion coefficient of the molecule.
Usually the compound under investigation is a water
molecule, since its abundance in the many tissues
is high. The idea in diffusion weighting is to attenuate the MR signal in the measurement due to
303
Brownian motion of the imaged molecules (Fig. 31.1).

In addition to this weighting, a readout sequence, for
instance a spin echo sequence, causes an additional
weighting to the MR signal. Therefore, contrast in the
diffusion weighted images is mixed. A common way to
eliminate readout weighting is to measure the MR signal (S) with several diffusion weightings followed by
plotting of natural logarithm of S as a function of the
b value and the diffusion coefficient (D) is determined
by the slope of the curve.
In many cases with the diffusion times used in MRI
the diffusion appears restricted. If the compartment
diameter is far larger than the average diffusion distance during tdiff , diffusion can be approximated as
free diffusion and average diffusion distance can be
calculated by the relation
x0 =

2Dtdiff .
In the cases where the average diffusion distance is

in order of compartment size, diffusion is restricted,
which leads to decreased D. When the measured
diffusion coefficient is affected by restriction, active
transfer, etc., which is usually the case in tissue, it
is often called apparent diffusion coefficient (ADC).
Compartment size and/or other aspects affecting the
measured D might not be homogeneous in all directions, causing difference between Ds measured in
different directions (i.e. using pulsed field gradients in
different physical orientations x, y, z). Several measurements (at least six) with diffusion gradients in
different orientations are needed to define a full diffusion tensor matrix, which includes more descriptive
information of the directional dependency of the diffusion. From the diffusion tensor, trace, fractional
anisotropy, and diffusivity along main axis can be calculated and information can be used to resolve, for
instance, fiber tracts in the brain tissue. It is worth
noting that although the most common parameter to
study molecular diffusion in the tumor tissue still
remains to be the average diffusion of water in single or all physical axes (x, y, z; the trace of the
diffusion tensor), there is body of evidence that at
least in experimental intracranial gliomas that there
are distinct patterns of cellularity (spherical, radial)
that are specific for different tumors (Zhang et al.,
2007). These characteristics may be detected using
diffusion tensor MRI but it remains to be shown
whether the possible changes in the cellular patterns
304
Fig. 31.1 The schematic interpretation of the MR diffusion measurement. Lets consider the ensemble of stationary
molecules; during the first field gradient pulse the magnetization
dephases an amount depending on molecule(s) location (here i,
ii or iii) in both stationary and moving molecules; during the
second field gradient the magnetization fully rephases in the
stationary molecule ensemble. Despite that moving molecules
experience similar dephasing, the movement of the molecules

between the field gradients changes the magnetic field and therefore angular velocity of rephasing experienced by the molecules,
causing only partial rephasing in this population. Partial rephasing leads to signal loss compared with stationary molecules
and this signal loss forms the base of diffusion contrast and
measurements in MRI
in tumors can be exploited to evaluate the treatment

responses.
Multi-compartmental Diffusion
in the Tissue
Water Diffusion in the Tissue
The slope of the natural logarithm of signal intensity log(S) remains close to linear up to 1000 s/mm2 ,
but if we apply higher b values the decay will begin
to bend. This is due to multi-exponentiality of the
water diffusion in the tissue environment. In a simplified tissue model the slow diffusing water component originates mostly from intracellular space and
the fast decaying component originates mostly from
extracellular space. The most obvious way to analyze
such data is a two compartment bi-exponential decay
model
Water in the Brain Tissue

Water can be located in several compartments with
varying physical and chemical environments with
characteristic ADCs. The volume ratio of the intravascular space is 34% in healthy brain tissue. Intra
and extracellular spaces occupy the remainder by 80
and 20% (Norris, 2001). Intracellular space comprises
mostly of cytoplasm, which is a mixture of 70% of
water, 20% of proteins, 10% of lipids and inorganic
compounds (Alberts et al., 1994).
S = AI exp( bDI ) + AE exp( bDE ),
where AI + AE = S0 , DI and DE refer to intraand extracellular diffusion coefficients, respectively.

Over the years debate has continued about the illconditioned nature of the multi-exponential decay
equation, thus several other models to analyze multicompartmental diffusion data has been suggested with
experimental verifications (reviewed by Mulkern et al.,
2009).
Diffusion in compartments is affected by tortuosity, especially extracellular space, volume fractions of
the compartments, cell membrane permeability and
exchange times between compartments, all these having an impact on the measured ADC. Therefore, water
ADC in extra- and intracellular spaces is not trivial
to determine experimentally. Using the cell cultures,
the intracellular ADCs can be determined, first demonstrated by van Zijl et al. (van Zijl et al., 1991). The
idea is to use high b values to spoil out the signal from the rapidly diffusing cell medium. Work
was extended to cover different cell types, and dataanalysis was performed taking permeable cell membranes into account. Water analogs can also be used
to determine intra- and extracellular ADCs separately.
The first approach was to use 133 Cs as a water diffusion analog in the tissue. The 133 Cs is a potassium
analog accumulating into intracellular space and shedding the light to intracellular water diffusion. The
ADC of 0.91103 mm/s2 was measured in normal rat brain tissue (Neil et al., 1996). The use of
analogs was continued by administration of 2FDG6P in the rats intravenously. 2FDG-6P metabolizes
when it is taken up by the cells and fluoride is kept
in the intracellular space allowing diffusion measurements. The same compound administered directly into
the ventricle allows the measurement of extracellular
ADC. Despite a massive amount of effort to figure out
mechanisms behind the water diffusion in the healthy
brain using MRI and other methods, the relationship between tissue microstructure, water dynamics
and ADC is not yet fully understood (Norris, 2001).
This fact does not prevent us from exploiting diffusion MRI to study several different diseases. In acute
stroke, a dramatic acute drop in ADC is detected a
phenomenon that is associated with a shift in tissue
water from extracellular space to intracellular space
due to cytotoxic edema while the degradation of
the cellular structures in a mature stroke lesion or in
many neurodegenerative diseases leads to increased
diffusion.
305
Assessing Response to Gene Therapy

Through Water Diffusion
Biological Aspects in Tumor and Gene
Therapy Detection
Several differences can be itemized between normal
tissue and tumor or neoplastic tissue. One of the most
characteristic features is excessive proliferation, which
typically leads to high cell density in the tumor tissue. Also increased metabolic activity, angiogenesis,
dysregulation of cellular homeostasis and formation
of distant colonies from the primary site are typical for fast proliferating tissue. Expression of specific cell surface receptors may occur as well. All of
these features of tumor can be detected using imaging
applications with several physical approaches, such as
magnetic resonance, x-rays, and radioactive isotopelabeled compounds (Rudin, 2007). Cell death and
changes in vasculature are early markers of successful
therapy response. Both of these alter the water environment in the tissue, indicating the potential of water
ADC as an effective surrogate marker for the therapy
outcome both for agents that disrupt vasculature and
therapies that induce apoptosis (Padhani et al., 2009).
Changes in cell density alter the ratio between extra
and intracellular spaces and therefore also the ADC
of water in the tissue. Therefore it is not surprising
that a correlation between cell density and ADC has
been found during tumor treatment. Cell density often
decreases in the tumor before any shrinkage of the
tumor can be detected and therefore diffusion measurements can provide an early marker for the treatment
response.
Early Research for the Detection of Tumor

Treatment Response by Diffusion MRI
In the very first studies of water ADC on tumor
treatment, Zhao et al. (1996) found that a dosage
of 300 mg/kg of cycliphosphamine in breast tumor
xenografts resulted in the doubling of water ADC in
4 days, suggesting water ADC as an early marker for
chemotherapy outcome. The increase in water diffusion, as measured using spectroscopic method, preceded the decrease in tumor volume. In the 9L tumor
306
model treated with a single injection of 13.3 mg/kg of

1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU or carmustine), the water ADC was found to increase by
roughly 50% more than the baseline. For confirmation
there were no significant changes in ADC in normal
tissue during the treatment or in non-treated tumors
(Chenevert et al., 1997). The recurrence of the tumor
took place 8 days after a single shot of BCNU; at the
same time as water ADC began to decline towards
its initial pretreatment value, which was confirmed
with histological analysis (Chenevert et al., 1997). In
this study, increased water ADC was explained by
increased mobility of the water due to increased extracellular space as a result of ongoing apoptosis and the
formation of apoptotic bodies.
The first report of water diffusion changes during gene therapy is by Poptani et al. (Poptani et al.,
1998). In the study, BT4C thymidine kinase transfected tumor cells were inoculated into the BDIX rat
brain and 34 weeks after inoculation, animals were
treated with ganciclovir to induce apoptosis. Gene
therapy in these tumors results in elevation in water
ADC in an early phase of treatment (Fig. 31.2) and
later the tumor volume response to gene therapy, as
defined from T2 weighted images. Similar results were
reported also in a refined cytosine deaminase (CD)/5fluorocytosine (5FC) chemosensitization gene therapy
paradigm in orthotopic 9L gliomas stably expressing
the cloned S. cerevisiae CD gene (Stegman et al.,
2000).
Association Between Water Diffusion and

Cell Density
In many experimental studies and clinical settings
water ADC has been found to increase with decreasing cell density in tumor tissue (Kettunen and Grhn
2005). The increase in water ADC is often explained
by the increased extracellular space i.e. decreased
number of the cells in the tissue (Chenevert et al.,
1997) and/or decreased size of the cells (Galons et al.,
1999). Increase in water diffusion was also confirmed
for the extracellular water component by Valonen
and colleagues (Valonen et al., 2004). In vitro data
measured in macrophages undergoing apoptosis by
Hortelano and colleagues show that the apoptotic cell
shrinkage was most likely the reason for decreased
ADC after apoptotic caspase activation (Hortelano
et al., 2001).
Water ADC Initial Decrease in Therapies
Fig. 31.2 The first published MRI diffusion study of gene therapy response is shown. The experiment introduces water T2 and
ADC changes during gene therapy in the rat BT4C tumor model
using HSV tk+ transfected cells (ac and ef) and wildtype
cells with intratumoral injection of PA317/3.OD5 cells to mimic
clinical gene therapy (d and h). Both groups were treated with
ganciclovir and both T2 (ac) and water ADC (eg) increased
during therapy in HSV tk+ transfected tumors, whereas no significant change was detected in wild type tumors. The increase in
ADC and T2 corresponded to the later detected decrease in tumor
volume. Adapted by permission from Macmillan Publishers
Ltd: Cancer Gene Therapy 5(2):101109, 1998, copyright
(1998)
Decreased water ADC values before treatment are

detected in nearly all malignant tumors (Padhani et al.,
2009). This is most likely due to higher cell density
and decreased extracellular space in the tumor tissue when compared with normal tissue. A decrease in
ADC soon after the initiation of therapy has also been
detected, which seems to be due to cellular swelling,
reduction in blood flow, or reduction in extracellular
space (Padhani et al., 2009). The extent and duration of such ADC reductions depend most likely on
the tumor type and timing of imaging with respect
to the used treatment (Padhani et al., 2009). Cellular
swelling might occur in the early phase of apoptosis in the response to treatment; for instance, water
ADC values are reduced by fibrosis, dehydration in
rectal cancer and brain glioma therapies (Padhani et al.,
2009). These findings indicate the complex interplay
between physicochemical events reflected by changes
in ADC maps.
Inhomogeneous ADC Response within the

Tumor Tissue During Therapy
It should be noted here, that typically changes in ADC
are not uniform across the tumors (Moffat et al., 2006).
The imaging postprocessing technique to image nonuniform therapy response is introduced for diffusion
by Moffat and colleagues (Moffat et al., 2006). They
found that by increasing the dosage of BCNU in 9L
tumor bearing rats, and calculating the change of water
ADC between initial and treated stages of the same
tumors pixel wise, the tumor response to therapy can
be evaluated. The percentage of increased water diffusion was demonstrated to predict histology derived cell
loss and overall survival of the animals. The method
is called functional diffusion mapping and it has been
extended into other anticancer treatments and clinical
studies (Hamstra et al., 2005 and references therein).
Metabolite Diffusion in the Tissue

Magnetic resonance spectroscopy (MRS) can be used
instead of imaging to detect metabolites in the tissue
based on the difference in chemical shift. A spectroscopic experiment can be sensitized to molecular
diffusion in a similar manner to imaging. For in vivo
use, the signal needs to be localized to originate from
the tissue of interest (the tumor). In the case of 1 H spectroscopy, metabolite levels are in the range of 103
of the water signal in the normal brain tissue therefore also effective water suppression has to be used to
enhance dynamic range of metabolite detection.
MRS signal for choline containing compounds
(Cho) at 3.2 ppm consists of signals originating
predominantly from glycerophosphocholine, phosphocholine and to a lesser extent from free choline. These
compounds are important constituents in plasma membrane build-up and/or catabolism and increased Cho
signal is typically detected in tumor as a consequence
of active cell membrane metabolism. The resonance
for choline containing compounds in the 1 H NMR
spectrum of brain tissue includes also resonances
from taurine, myo-inositol and phosphoethanolamine
(Remy et al., 1994). The diffusion of tCho may be
interesting in terms of the response to therapy, since
apoptosis alters the physical size of the cell while
maintaining the integrity of the cell membrane at the
307
same time. The single voxel spectroscopic study in

BT4C gliomas undergoing apoptosis in vivo showed
that the diffusion of tCho decreases as a response to
the therapy and in the opposite direction to water diffusion (Hakumki et al., 1998). In the same study, the
reduction of the smaller pool of water, most likely
from within intracellular space, has been detected
together with a decrease in tCho indicating the tCho
signal origin from intact cells. Unpublished diffusion
weighted spectroscopic imaging data from the same
tumor-therapy model shows that tCho ADC decreases
during the treatment and this response is the most dramatic in the center of the tumor (Fig. 31.3). Reduction
most likely reflects a reduction in the percentage of
intracellular volume and cell diameter, which were
measured using histological methods. The result may
indicate that the intracellular compartment where tCho
resides decreases in volume, or that the intracellular
hindrance/viscosity changes during the treatment. The
analysis of cell diameter and detected cell shrinkage
using histological methods confirms the first option,
although it is expected that also the latter has its own
contribution to the effect.
The signals from fatty acids at 0.9 and 1.3 ppm
originate mostly from mobile lipids located in lipid
vesicles (Hakumki et al., 1999), and contain significant contribution from cholesterol (Liimatainen et al.,
2006). The diffusion of the fatty acids is relatively
slow and detection of restricted diffusion of these compounds needs long diffusion times, which make the
experiment prone to movement artifacts in vivo. No
significant changes in ADCs of mobile lipids were
found in the BT4C glioma model during apoptotic cell
death (Hakumki et al., 1998). In the same model,
an increase in visible triglycerides has been reported
using single voxel and spectroscopic imaging techniques (Hakumki et al., 1999; Liimatainen et al.,
2006). In summary, these findings support the idea of
an increase in the number of lipid vesicles without a
change in size during therapy. However, most likely the
restriction was not reached for these resonances and
the result reflects the constant viscosity in the vesicles
during the treatment. The recent study (Liimatainen
et al., 2009) showed that in the early phase of the
HSV tk+ gene therapy, water diffusion correlates especially with polyunsaturated lipid resonances. The loss
of relationship between unsaturated lipid signals and
water diffusion in the late phase was explained by a low
number of cells in the remaining tissue (dead tissue).
308
Fig. 31.3 Diffusion weighted spectroscopic imaging data measured from rat BT4C HSV tk+ tumors during ganciclovir
induced apoptosis and histological parameters: (a) representative
diffusion weighted spectra from an intra tumoral voxel at baseline (day 0) showing well resolvable choline containing moieties
(CCM) signal at 3.2 ppm and signal attenuation due to diffusion
at different b values; (b) ADC map of CCM at day 0; (c) ADC

map of CCM at day 8; (d) ADC of CCM resolved for border, center and whole tumor area; and (e) histologically derived
intracellular volume (bars) and cell diameter (dotted lines).
Figure by Kimmo Lehtimki, Timo Liimatainen and Juhana
Hakumki
Other MR Contrast Methods for

Treatment Response Detection
parameters, such as average exchange time of bulk

water and macromolecule pools (Sierra et al., 2008).
Even though the measurements allow extraction of
fundamental parameters, the mechanisms behind these
findings are still partly unknown. Nevertheless, these
methods may track changes in water-protein interactions, cell microarchitecture, protein/lipid content and
composition as well as changes in pH levels. These
approaches are less sensitive to motion than diffusion
MRI, after all clinical exploitation has been delayed
because of potential tissue heating effects (high specific absorption rate, SAR).
Despite the wide use of diffusion for detection of

therapy response, several other MRI methods can be
applied for the same purpose. For instance, water T1
and T2 increase during apoptosis in the tumor tissue, as shown in a rodent tumor gene therapy model
(Hakumki et al., 2002; Poptani et al., 1998). However,
these changes are typically delayed compared with
diffusion changes. Also water T1 in a rotating frame
of reference (T1 ) increases significantly even in the
earlier phase of the gene therapy before diffusion
(Hakumki et al., 2002; Kettunen et al., 2007). With
certain assumptions, the several measurements using
rotating frame contrasts (T1 and T2 with varying
RF amplitude and phase modulation functions) can
be used to calculate fundamental characteristic tissue
Clinical Aspects
The clinical use of gene therapy for brain tumors is
still limited to clinical trials. However, the first gene
therapy products are waiting for the permission of
authorities to be released for commercial use. As the

efficacy of gene therapy, like other anticancer treatments, needs to be controlled in every patient, there
is evidently an urge for non-invasive imaging of treatment response. Water diffusion MRI is one potential
candidate for that purpose. While there is currently
only a few clinical gene therapy trials ongoing in brain
cancers and none of these utilize diffusion MRI to our
knowledge, the use of water ADC has been evaluated
in many other treatment protocols.
The following paragraph is based on the review article by Hamstra et al. (2007) and references therein.
Increased water ADC values have been measured
after chemotherapy in patients, indicating the possible
importance of diffusion weighted MRI as a biomarker
for therapy response in clinical settings. The increased
ADC has been shown to predict positive treatment outcomes in oncotherapy, irrespective of whether the cell
death occurs by apoptosis or necrosis. In two patients,
ADC was found to increase several weeks before
actual tumor response. It should be noted that during
the time when tumors were responding with conventional markers, the water ADC had already begun to
decrease. In another study, the tumors were treated
with convection-enhanced chemotherapy administered
intratumorally for several days. Diffusion changes
were observed as early as after 24 h from the beginning
of chemotherapy administration. Significant increase
in water diffusion was also measured after 3 days of
radiotherapy in six patients with primary or metastatic
tumors. A preliminary study in three children with
medulloblastoma supports the clinical value of diffusion MRI for the detection chemotherapy response.
The water diffusion was increased also after stereotactic radiation therapy. In a study using fractionated
radiation therapy, the initial water ADC was found
to predict treatment response; low ADC was associated with high viability, high cell density and probable response to therapy whereas high ADC indicated
necrosis and non-successful therapy outcome, which
is called functional diffusion mapping. Functional diffusion mapping has been used in a prospective trial
in patients with primary brain tumors. Patients underwent diffusion imaging 1 week before chemo and/or
radiation therapy and a second time was set approximately 3 weeks from the start of therapy when half
of the fractionated radiation therapy was delivered.
The weak correlation between water diffusion and
later radiographic response was found. Responsive,
309
stable and progressive tumor growth patients were

able to discriminate using functional diffusion mapping. The other study from the same research group
indicated that functional diffusion mapping correlated
with radiographic response, progression free and overall survival times. The patients divided into responders
and non-responders based on functional diffusion mapping showed overall survival of 18.2 and 8.2 months
(p < 0.01; log rank test).
Future Aspects
In addition to a standardized ADC measurement protocol to follow up treatment response, more advanced
methods that measure and analyze the diffusion data
are needed. For instance, routine clinical use of
functional diffusion mapping to follow up the treatment response may be feasible with little effort.
Diffusion MRI has been evaluated as a part of
study protocol to predict gene therapy outcome in
a rat glioma model with 1 H MRS of lipid compounds. The study showed that diffusion, together with
other MRI/imaging parameters, may provide an even
more complete picture of tissue response to treatment. One interesting and arising diffusion measurement strategy is to separate different motion regimes
by repeating measurement with several frequencies
of oscillating field gradients, allowing application
of short diffusion times while simultaneously keeping a reasonable gradient amplitude (Colvin et al.,
2008). The authors demonstrated their method in an
experimental glioma model and suggested its use
also for monitoring treatment response. All of these
methods may provide complementary information on
water/metabolite biophysical status during the treatment and add to the clinical value of diffusion
measurements.
In conclusion, diffusion MRI has been evaluated
widely for the detection of the gene therapy response
in preclinical settings. For imaging the gene therapy response, clinical applications are promising, even
though the current results are only from chemo and/or
radiation therapies. In general, evaluation of diffusion MRI data from clinical settings is hampered
by the variety of diffusion measurement protocols
(b values, imaging readout, etc.) and also variation
in analytical methods (absolute ADC, mean ADC,
310
normalized ADC, threshold changes in ADC and functional diffusion maps). The true value of diffusion
MRI for the assessment of gene therapy response will
be seen when gene therapy is more widely available
and tested in multi-institutional trials incorporating
diffusion MRI with uniform protocols.
References
Alberts B, Bray D, Lewis J, Ralf M, Roberts K, Watson JD
(1994) Molecular biology of the cell. Garland Publishing,
New York, NY
Chenevert TL, McKeever PE, Ross BD (1997) Monitoring early
response of experimental brain tumors to therapy using
diffusion magnetic resonance imaging. Clin Cancer Res
3:14571466
Colvin DC, Yankeelov TE, Does MD, Yue Z, Quarles C, Gore
JC (2008) New insights into tumor microstructure using
temporal diffusion spectroscopy. Cancer Res 68:59415947
Galons J-P, Altbach M-I, Paine-Murrieta GD, Taylor CW, Gillies
RJ (1999) Early increases in breast tumor xenograft water
mobility in response to paclitaxel therapy detected by noninvasive diffusion magnetic resonance imaging. Neoplasia
1:113117
Germano IM, Binello E (2009) Gene therapy as an adjuvant
treatment for malignant gliomas: from bench to bedside.
Hakumki JM, Grhn OHJ, Tyynel, Valonen P, Yl-Herttuala S,
Kauppinen RA (2002) Early gene therapy-induced apoptotic
response in BT4C gliomas by magnetic resonance relaxation contrast T1 in the rotating frame. Cancer Gene Ther 9:
338345
Hakumki JM, Poptani H, Puumalainen AM, Loimas S, Paljrvi
LA, Yl-Herttuala S, Kauppinen RA (1998) Quantitative 1 H
nuclear magnetic resonance diffusion spectroscopy of BT4C
rat glioma during thymidine kinase-mediated gene therapy
in vivo: identification of apoptotic response. Cancer Res
58:37913799
Hakumki JM, Poptani H, Sandmair AM, Yl-Herttuala S,
Kauppinen RA (1999) 1 H MRS detects polyunsaturated fatty
acid accumulation during gene therapy of glioma: implications for the in vivo detection of apoptosis. Nat Med
5:13231327
Hale AJ, Smith CA, Sutherland LC, Stoneman VE, Longthorne
VL, Culhane AC, Williams GT (1996) Apoptosis: molecular
regulation of cell death. Eur J Biochem 236:126
Hamstra DA, Chenevert TL, Moffat BA, Johnson TD, Meyer
CR, Mukherji SK, Quint DJ, Gebarski SS, Fan X, Tsien
CI, Lawrence TS, Junck L, Rehemtulla A, Ross BD
(2005) Evaluation of the functional diffusion map as an
early biomarker of time-to-progression and overall survival in high-grade glioma. Proc Natl Acad Sci USA 102:
1675916764
Hamstra DA, Rehemtulla A, Ross BD (2007) Diffusion magnetic resonance imaging: a biomarker for treatment response
in oncology. J Clin Oncol 25:41044109
Hortelano S, Garcia-Martin ML, Cerdan S, Castrillo A, Alvarez
AM, Bosca L (2001) Intracellular water motion decreases in
apoptotic macrophages after caspase activation. Cell Death
Diff 8:10221028
Kerr JFR, Wyllie AH, Currie AR (1972) Apoptosis: a basic biological phenomenon with wide-rangingimplications in tissue
kinetics. Br J Cancer 26:239257
Kettunen MI, Grhn OHJ (2005) Tumour gene therapy monitoring using magnetic resonance imaging and spectroscopy.
Curr Gene Ther 5:685696
Kettunen MI, Sierra A, Nrvinen MJ, Valonen PK, YlHerttuala S, Kauppinen RA, Grhn OHJ (2007) Low spinlock field T1 relaxation in the rotating frame as a sensitive
MR imaging marker for gene therapy treatment response in
rat glioma. Radiology 243:796803
Kinloch RA, Treherne JM, Furness LM, Hajimohamadreza I
(1999) The pharmacology of apoptosis. Trends Pharmacol
Sci 20:3542
Liimatainen T, Hakumki JM, Kauppinen RA Ala-Korpela M
(2009) Monitoring of gliomas in vivo by diffusion MRI and
1 H MRS during gene therapy-induced apoptosis: interrelationships between water diffusion and mobile lipids. NMR
Biomed 22:272279
Liimatainen T, Lehtimki K, Ala-Korpela M, Hakumki JM
(2006) Identification of mobile cholesterol compounds in
experimental gliomas by 1 H MRS in vivo: effects of
ganciclovir-induced apoptosis on lipids. FEBS Lett 580:
47464750
Moffat BA, Chenevert TL, Meyer CR, McKeever PE, Hall
DE, Hoff BA, Johnson TD, Rehemtulla A, Ross BD (2006)
The functional diffusion map: an imaging biomarker for
the early prediction of cancer treatment outcome. Neoplasia
8:259267
Mulkern RW, Haker SJ, Maier SE (2009) On high b diffusion
imaging in the human brain: ruminations and experimental
insights. Magn Reson Med 27:11511162
Neil JJ, Duong TQ, Ackerman JJH (1996) Evaluation of intracellular diffusion in normal and globally-ischemic rat brain
via 133 Cs NMR. Magn Reson Med 35:329335
Norris DG (2001) The effects of microscopic tissue parameters
on the diffusion weighted magnetic resonance experiment.
NMR Biomed 14:7793
Padhani AR, Liu G, Mu-Koh D, Chenevert TL, Thoeny HC,
Takahara T, Dzik-Jurasz A, Ross BD, Van Cauteren M,
Collins D, Hammoud DA, Rustin GJS, Taouli B, Choyke
PL (2009) Diffusion-weighted magnetic resonance imaging
as a cancer biomarker: Consensus and recommendations.
Neoplasia 11:102125
Poptani H, Puumalainen AM, Grhn OHJ, Loimas S,
Kainulainen R, Yl-Herttuala S, Kauppinen RA (1998)
Monitoring thymidine kinase and ganciclovir-induced
changes in rat malignant glioma in vivo by nuclear magnetic
resonance imaging. Cancer Gene Ther 5:101109
Remy C, Arus C, Ziegler A, Lai ES, Moreno A, Le Fur Y,
Decorps M (1994) In vivo, ex vivo, and in vitro one- and twodimensional nuclear magnetic resonance spectroscopy of an
intracerebral glioma in rat brain: assignment of resonances.
J Neurochem 62:166179
Rudin M (2007) Imaging readouts as biomarkers or surrogate
parameters for the assessment of therapeutic interventions.

Savill J, Fadok V (2000) Corpse clearance defines the meaning
of cell death. Nature 407:784788
Sierra A, Michaeli S, Niskanen JP, Valonen PK, Grhn HI,
Yl-Herttuala S, Garwood M, Grhn OHJ (2008) Water
spin dynamics during apoptotic cell death in glioma gene
therapy probed by T1 and T2 . Magn Reson Med 59:
13111319
Stegman LD, Rehemtulla A, Hamstra DA, Rice DJ, Jonas SJ,
Stout KL, Chenevert TL, Ross BD (2000) Diffusion MRI
detects early events in the response of a glioma model to the
yeast cytosine deaminase gene therapy strategy. Gene Ther
7:10051010
Stejskal EO, Tanner JE (1965) Spin diffusion measurements:
Spin echoes in the presence of a time-dependent field gradient. J Chem Phys 42:288292
Valonen PK, Lehtimki KK, Visnen TH, Kettunen MI,
GrhnOHJ, Yl-Herttuala S, Kauppinen RA (2004) Water
diffusion in a rat glioma during ganciclovir-thymidine kinase
311
gene therapy-induced programmed cell death in vivo: correlation with cell density. J Magn Reson Imaging 19:389396
Van Zijl PC, Moonen CT, Faustino P, Pekar J, Kaplan O, Cohen
JS (1991) Complete separation of intracellular and extracellular information in NMR spectra of perfused cells by
diffusion-weighted spectroscopy. Proc Natl Acad Sci USA
15:32283232
Wirth T, Samaranayake H, Pikkarainen J, Mtt A-M, YlHerttuala S (2009) Clinical trials for gliblastoma multiforme
using adenoviral vectors. Curr Opin Mol Ther 11:485492
Zhang J, van Zijl PCM, Laterra J, Salhotra A, Lal B, Mori S,
Zhou J (2007) Unique patterns of diffusion directionality in
rat brain tumors revealed by high-resolution diffusion tensor
MRI. Magn Reson Med 58:454462
Zhao M, Pipe JG, Bonnett J, Evelhoch JL (1996) Early detection of treatment response by diffusion-weighted 1 H-NMR
spectroscopy in a murine tumour in vivo. Br J Cancer
73:6164
Chapter 32
High-Grade Gliomas: Dendritic Cell Therapy

Hilko Ardon, Steven De Vleeschouwer, Frank Van Calenbergh,
and Stefaan W. Van Gool
Abstract The relationship between the development

of a tumor and the immune system is depicted by
the fundamental model of cancer immunoediting:
a tumor will only grow if the immune system is
not capable of eliminating immune-resistant tumor
cells (tumor immune escape). Both the adaptive
and innate immunity play a key-role in this process. Multiple tumor-associated antigens have been
described in gliomas and the expression of these
antigens is very heterogeneous with important interand intra-individual differences. This is in contrast
with other tumors, such as malignant melanoma,
where tumor-associated antigens are almost universally expressed in all tumor cells. Several studies have
shown that antigens can drain from the central nervous
system to the cervical lymph nodes, where antigenspecific T cells can be activated by antigen-presenting
cells. These activated T cells can subsequently migrate
into the brain parenchyma to the sites of antigen
challenge. This points to possible, naturally occurring immune responses eventually also against tumors,
and indicates that the central nervous system is not
an immune-silent environment. Therefore, it seems
that the immune system can provide us with some
interesting tools to treat gliomas. Several types of
immunotherapy have been used and the most promising strategy for the induction of durable immune
responses seems to be active, specific immunotherapy,
or the so-called tumor vaccination. Further research
S. De Vleeschouwer ()
Department of Neurosurgery, Laboratory of Experimental
Immunology, Catholic University Leuven, B-3000 Leuven,
Belgium
e-mail: steven.devleeschouwer@uzleuven.be
has to be done, however, to establish the value of this

form of immunotherapy in the treatment of gliomas.
Keywords Immunity Cancer immunoediting
Dendritic cells Antigens Blood-brain barrier
Lymphatic vessels
Introduction
The concept that the immune system can recognize and
eliminate primary developing tumors in the absence
of external therapeutic intervention has existed for
nearly 100 years. Although tumor-genetics and cancer
cell biology have claimed the greatest interest in cancer research the last decades, tumor-immunology has
recently gained renewed attention. Initially the question of specificity was raised, i.e. the capacity of the
immune system to distinguish normal cells from tumor
cells. In recent years though, many tumor-associated
antigens (TAA) have been described and these antigens can be recognized by innate and adaptive components of the immune system (Dunn et al., 2007).
Furthermore, innate immune cells can recognize specific molecular structures in tumor cells. The specificity of the immune system has therefore been widely
accepted. However, new questions have arisen, most
importantly on how a tumor can escape the immune
system (tumor immune escape), and whether and
how the immune system can be (re-)activated against
the tumor.
The first immunotherapeutical approach against
cancer was adapted by Wiliam Coley at the end
of the nineteenth century (Copier et al., 2009). He

313
314
used bacterial toxins (Coleys toxins) to activate

the immune system against tumors. The current
immunotherapeutical approaches are mainly aimed at
inducing a tumor-specific immune response. Several
types of cancer (including glioma, melanoma, prostate
carcinoma, renal cell carcinoma and non-small cell
lung cancer) have been treated with immunotherapy
and the largest experience has been gained in malignant melanoma (Engell-Noerregaard et al., 2009). So
far, overall results seem promising and intriguing,
although efficacy of immunotherapy cannot be measured by classical response criteria.
Until recently, the central nervous system (CNS)
has always been looked upon as an immune-privileged
organ in which the immune system could not play
a role of importance due to the blood-brain barrier (BBB) and the absence of lymphatic vessels
and immune cells. However, recent data do point to
immune responses in the CNS and it would therefore
be better to define the CNS as an immune-distinct
environment. In this chapter we will discuss cancer
immunoediting in a broader sense, with the main focus
on glioma immunity and immunotherapy.
Cancer Immunoediting
The growth of tumor cells is on the one hand regulated
and prevented by intrinsic nonimmune surveillance
systems, such as DNA-repair and intracellular control mechanisms (e.g. p53 pathway). On the other
hand, both innate and adaptive immunological control
mechanisms exist which form the so-called immunosurveillance (Dunn et al., 2002, 2004a, b; Zitvogel
et al., 2006). The interactions between a tumor and
the immune system can be divided into three different phases of cancer immunoediting (Fig. 32.1). This
process is responsible for both eliminating tumors and
sculpting the immunogenic phenotypes of tumors that
eventually form in immune competent hosts (Dunn
et al., 2002, 2004a, b; Zitvogel et al., 2006). In a first
phase of elimination, equivalent to immune surveillance, nascent transformed cells are eradicated by
the immune system. The innate immune system is
activated by inflammatory cytokines released by the
tumor, macrophages and stromal cells surrounding
tumor cells. This results in the activation of natural
H. Ardon et al.
killer (NK) cells, natural killer T (NKT) cells and T cells. These cells produce other pro-inflammatory
cytokines (such as interleukin (IL)-12 and interferon
(IFN)-) that enhance the immune response. Tumor
cell death by NK cells results in release of TAA that
can lead to activation of the adaptive immune system. Uptake of TAA by dendritic cells (DC) will lead
to tumor-specific antigen-presentation resulting in a
clonal expansion of tumor-specific CD8+ cytotoxic T
cells. In a second phase, there is a balance between the
elimination of immune-sensitive tumor cells and the
outgrowth of newly-formed immune-resistant tumor
cells: equilibrium. Due to continuous, genetic adaptations of the tumor cells (mutation-drift), tumor cell
clones with a non-immunogenic phenotype will arise
(immune resistance). The immune selection pressure will favor the growth of these tumor cell variants
resulting in ongoing tumor formation. Eventually, the
balance will tip in favor of the immune-resistant tumor
cells and then the final phase will set in: immune
escape. Most of the tumor cells will now evade the
immune system as they are insensitive to immunological detection and subsequent elimination. Hence, this
will result in clinically detectable malignant lesions.
Three forms of tumor immune escape can be recognized: 1. loss of TAA-recognition by the immune system; 2. diminished susceptibility of tumor cells to cell
death; 3. induction of immune dysfunction (Malmberg,
2004). Tumor cells can release tumor-derived soluble
factors that can suppress the immune response (e.g.
transforming growth factor (TGF-), prostaglandin
E2 (PGE2 ) and IL-10). The extracellular matrix can
bind TAA in competition with DC leading to decreased
antigen-presentation by DC and subsequently less activated T cells. Also, regulatory T cells (Treg) can be
induced by the tumor resulting in a suppression of the
immune system. Moreover, the functional phenotype
of intratumoral macrophages can be switched to M2type macrophages which can promote tumor growth
and angiogenesis. Finally, myeloid-derived suppressor cells (MDSC) can further disturb the immune
response.
Based on these strong interactions between tumor
cells and the immune system, one could speculate
about a remarkable hypothesis; tumors that are very
potent in suppressing the immune system might mainly
rely on this immune suppression for their immune
escape. The need for editing the tumor cells to
evade the immune system might therefore be less in
32 High-Grade Gliomas: Dendritic Cell Therapy
315
Fig. 32.1 Cancer immunoediting and relevance to tumor vaccination against malignant glioma (a).General principle of cancer
immunoediting: 1. In the first phase of elimination, the immune
system is capable of eliciting an effective immune response
(illustrated as a black box) against an immunogenic tumor (illustrated as a white ellipse), and tumor growth is prevented. 2. In the
second phase, there is an equilibrium between tumor growth and
elimination of tumor cells by the immune system (illustrated by a
gray ellipse and gray box). Due to immunoediting and sculpting
of the tumor cells, some tumor cells become non-immunogenic
and evade the immune response. Tumor-induced immune suppression counteracts the immune response as well. 3. In the
third and final phase, the tumor is fully immunoedited and
has become non-immunogenic (illustrated as a black ellipse).
This leads, in combination with the ongoing tumor-induced
immune suppression, to tumor immune escape and outgrowth
of the tumor. The immune system and response have become
ineffective (illustrated by a white box). (b). Tumor vaccination
(general): At the time of immune escape, the immunoedited
tumor is diagnosed and surgery is performed leading to a state
of minimal residual disease (1). The immunoedited, nonimmunogenic tumor cells are used to generate a tumor vaccine
(2), but tumor vaccination does not lead to re-activation of the
immune system (3), since stimulation of the immune system
is done with non-immunogenic cells. Thus, there is no effective immune response to eliminate the tumor (4), resulting in
renewed tumor immune escape (5). (c) Glioma immunoediting:
From the start of gliomagenesis tumor-induced immune suppression is very potent, leading to an ineffective immune response.
Consequently, the tumor can grow and evade the immune system, without the need for editing/sculpting of the tumor cells.
(d).Tumor vaccination (glioma): At the time of diagnosis, the
tumor consists of a non-immunoedited, still immunogenic
phenotype. Surgery is performed leading to a state of minimal
residual disease (1). The non-immunoedited, immunogenic
tumor cells are used to generate a tumor vaccine (2), which is
able to activate the immune system (3), in combination with inhibition of the tumor-induced immune suppression (4). Activation
of the immune system leads to an effective immune response and
the tumor is eliminated (4)
these immune suppressive tumors. Since the tumor

cells will not be attacked by the suppressed immune
system, the immune selection pressure will be less
and fewer genetic adaptations of the tumor cells will
occur. This will result in less tumor cell clones with
a non-immunogenic phenotype and more cell clones
with an immunogenic phenotype. Following this line

of thought, an active immunotherapeutical approach to
such a tumor at time of minimal residual disease (e.g.
after total resection of the tumor), in combination with
a strategy to block the tumor-mediated immune suppression, might be of more benefit. This could be in
316
contrast to a tumor with an immunoedited phenotype, since the former tumor might still be immune
sensitive (Fig. 32.1). In other words, tumors that do
not lead to spontaneous immune reactions because
of strong immune suppressive characteristics, might
be best suited for immunotherapy at time of minimal
residual disease.
Central Nervous System and Immunity

In contrast to antigen-specific immune responses in
peripheral tissues, the details are less clear for immune
responses to antigens in the CNS. The biology of
immune responses within the CNS would appear to be
distinct due to the lack of secondary lymphoid tissues
within the brain parenchyma, as well as to the possible
obstacles presented by the BBB. In general, TAA have
to be acquired and presented by antigen-presenting
cells (APC), leading to the activation of tumor-specific
lymphocytes that will have to infiltrate the brain tumor.
Research in this area is ongoing and much is still to be
learned.
H. Ardon et al.
microglia, perivascular macrophages, choroid plexus

epithelial cells, neurons, and DC have been put
forward as potential APC. Recent publications point
especially to DC as most important APC in the CNS
(McMahon et al., 2006), although their presence in
the CNS is limited. DC capture TAA from dying
tumor cells in the CNS and migrate to draining lymph
nodes in the cervical region. There, antigens are presented to different subtypes of T cells in a major
histocompatibility complex (MHC)-I and MHC-II
context. Antigen-presentation in a MHC-I molecule
(cross-presentation, a typical feature of DC) leads to
activation of CD8+ T cells, presentation in a MHC-II
molecule to activation of CD4+ T cells. Antitumoral
CD4+ T cells support a strong and prolonged CD8+ T
cell reaction, promote the maturation of DC and stimulate B cells to produce antitumoral antibodies (De
Vleeschouwer et al., 2006; Heath and Carbone, 2001;
Maldonado-Lopez and Moser, 2001). In several animal studies it has been shown that antigen-specific T
cells are primed by immigrant, antigen-loaded DC in
the cervical lymph nodes, after which the activated T
cells traffic to sites of antigen challenge in the brain
(Dunn et al., 2007). However, further research in brain
tumor models is necessary to validate this principle.
Cervical Lymphatic Drainage of the Brain

Blood-Brain Barrier
Although lymphatic vessels cannot be identified in the
brain parenchyma, several animal studies have shown
that antigens can be drained to the cervical lymph
nodes (Bradbury et al., 1981; Cserr and Knopf, 1992).
Possibly, the antigens run through the subarachnoid
space along cranial nerves and cross the cribiform
plate to the nasal mucosa where they access lymphatic drainage basins. The potential of antigens to
drain to cervical lymph nodes points to possible interactions between the CNS and peripheral lymphoid
tissue, highlighting a theoretical starting point for the
initiation of brain tumor-specific immune responses.
Antigen-Presenting Cells in the Central

Nervous System
Several cell types in the CNS including vascular
endothelial cells, smooth muscle cells, astrocytes,
Activated, antigen-specific T cells have to traffic from

the cervical lymph nodes into the brain parenchyma
to infiltrate the brain tumor. Three pathways by which
immune cells can access the CNS have been proposed:
1. from blood to the cerebrospinal fluid (CSF) via
the choroid plexus; 2. from blood to the subarachnoid
space; 3. from blood to brain parenchyma (Ransohoff
et al., 2003). The most important obstacle is formed
by the BBB that prevents passive transit of molecules
between the CNS parenchyma and the systemic circulation. The barrier is formed by tight junctions of
the capillary endothelial cells, creating a tight seal
which limits the paracellular transport of molecules.
Moreover, more than 90% of the capillary endothelium
is covered by astrocyte foot processes that form the
glia limitans. The capillaries of the circumventricular organs of the brain represent areas in which there is
no BBB, and these are thus sites that more freely permit transport of molecules across vascular endothelial
cells. Additionally, the blood-CSF barrier restricts

free movement of cells and molecules between the CSF
and the capillaries of the choroid plexus (Dunn et al.,
2007).
Within the glioma microenvironment, the vasculature is substantially altered and structures such as the
BBB are significantly compromised. Therefore, the
BBB in gliomas is more porous than the intact BBB
in healthy subjects. This increased permeability might
lead to increased immune cell infiltration. However,
efficient lymphocyte trafficking into the tumor may
be significantly reduced due the fact that tumor tissue
is severely dysregulated and may contain poor target
substrates to attract T cells.
Lymphocyte-Trafficking to Glioma
Although it has been shown that T cells can traffic to the CNS in autoimmune diseases, it is less
clear how lymphocytes traffic to developing CNS neoplasms. One possible model is the multi-step model
of migration, characterized by the rolling, sticking
and migrating paradigm (Balkwill, 2004). T cells
homing to the brain parenchyma slow down in a first
step by gradually tethering to capillary endothelium in
a rolling fashion mediated by interactions between
several selectins and their ligands. Subsequently, lymphocytes adhere to the vascular endothelium in the
sticking step mediated by integrins that become
activated by chemokines. Lymphocytes ultimately
transmigrate to the brain parenchyma in the diapedesis (migrating) step, which may occur via transor para-cellular endothelial transport.
Several immune cells have been shown to infiltrate gliomas, including tumor-infiltrating lymphocytes
(TIL), MDSC and M2-macrophages. Among the TIL,
CD4+ T cells, CD8+ CTL and immune suppressive
Treg have been documented (Abou-Ghazal et al., 2008;
Dunn et al., 2007; Heimberger et al., 2008).
Gliomas and Immunity

Several TAA have been identified in human gliomas:
Aim-2, Trp-1, Trp-2, Sart-1, Sart-2, Sart-3, Her2/neu,
GP100, Mage-1, Galt-1, Galt-3, Sox2, Sox6, Survivin,
317
EGFRvIII, Gage-1, WT1 (Dunn et al., 2007). These

antigens can be recognized by T cells and form the
basis for potential spontaneous immune responses
against CNS tumors. Two subgroups of gliomaspecific antigens have been described: 1. antigens
that are solely expressed by tumor tissue; 2. antigens that are expressed both by normal and tumor
tissue. Only the antigens from the first subgroup
are of immunotherapeutical interest, since antigens
that are expressed by normal tissue could lead to
possible autoimmune phenomena. The ideal antigens
for immunotherapy are antigens that are immunogenic and tumor-specific, i.e. not expressed by normal tissue. Moreover, these antigens should ideally
be present on all tumor cells or at least at the
tumor cells that are responsible for recurrence of the
tumor.
There are no universal glioma-specific antigens
that are expressed by all gliomas, and inter-individual
differences are the rule. Therefore, it seems important to give every patient an individual and gliomaspecific treatment using TAA that are expressed by
the tumor of the patient. Besides, even within the
tumor of one patient expression of antigens is very
heterogeneous and there are no known TAA that are
universally expressed by all tumor cells. Moreover,
due to the plasticity of the tumor cells with genetic
alterations and loss of expression of antigens, the
immunotherapeutical use of one glioma-specific antigen would lead to a selection of tumor cells that do
not express that antigen (any longer) and that are
not targeted by the immunotherapy (antigen-loss variants) (De Vleeschouwer et al., 2005; Gilboa et al.,
1998).
Furthermore, it is important to know which tumor
cells are responsible for recurrence of the tumor. In
other words, do all tumor cells have the same tumorigenic potential or does a subgroup of glioma stem cells
exist? Several studies have shown that neurosphereforming, CD133+ glioma cells are chemo- and radioresistant and might be seen as so-called brain tumor
initiating cells (Bao et al., 2006; Eramo et al., 2006;
Hemmati et al., 2003; Liu et al., 2006; Singh et al.,
2003, 2004). If this truly is the case, then immunotherapeutical approaches should focus on the TAA of these
tumor initiating cells.
Finally, high-grade gliomas (HGG) can actively
suppress the immune system through different
immune escape mechanisms: production of immune
318
suppressive cytokines, negative regulatory pathways

in lymphocytes, lack of glioma-specific antigen
recognition by immune cells (loss of MHC-I expression on tumor cells), and induction of Treg (Grauer
et al., 2009; Yamanaka, 2009). This glioma-induced
immune suppression is not only evident within the
tumor, but systemically as well (Dix et al., 1999;
Elliott et al., 1987; Fecci et al., 2006; Roszman et al.,
1991).
Immunotherapy
Restorative, Passive and Adoptive
Immunotherapy
Immunotherapy for cancer covers a broad field of
diverse approaches, all using different aspects of the
immune system repertoire. The common goal is to
destroy remaining cancer cells and prevent tumor
(re)growth (De Vleeschouwer et al., 2005).
Restorative immunotherapy consists of systemic
or local (intratumoral) administration of cytokines
(including IL-2, tumor necrosis factor (TNF-) and
IFN-) to enhance non-specific immunity. As for other
tumors, this type of treatment showed considerable,
treatment-limiting, systemic toxicity and neurotoxicity for gliomas. No conclusive evidence of efficacy was shown and treatments had to be stopped
(De Vleeschouwer et al., 2005; Grauer et al., 2009;
McDermott, 2009).
In case of passive immunotherapy targetspecific
antitumoral
monoclonal
antibodies
(mAb) are used with or without a conjugated
ligand. Antibody-dependent cellular cytotoxicity,
complement-dependent cytotoxicity and induction/blockade of intracellular signals have been
described as working mechanisms of tumor cell
killing. Several mAb have been shown to be active,
often in combination with chemotherapy: rituximab
(non-Hodgkin lymphoma and chronic lymphocytic
leukemia), trastuzumab (metastasized breast carcinoma), cetuximab (colorectal carcinoma, head/neck
carcinoma and non-small cell lung carcinoma) and
bevacizumab (colorectal carcinoma, breast carcinoma
and lung carcinoma) (Borghaei et al., 2009). Although
some studies have shown moderate results with this
H. Ardon et al.
type of immunotherapy for glioma, especially in

patients with minimal disease burden, it is unclear
whether the mAb can infiltrate sufficiently into the
tumor to induce an optimal antitumoral effect (De
Vleeschouwer et al., 2005; Grauer et al., 2009).
Recently, several studies have focused on the treatment of gliomas with bevacizumab, which is an
angiogenesis inhibitor targeting VEGF. Although
results show that symptoms are tempered, it seems
that furtive invasion of the disease might continue,
unrecognized or underestimated by standard imaging
modalities. Moreover, while treatment of other tumor
types is improved by combining chemotherapy with
anti-angiogenic drugs, inhibiting angiogenesis in
glioblastoma multiforme (GBM) may antagonize the
efficacy of chemotherapeutic drugs by restoring the
BBB. Several studies suggest an antitumoral effect
based on improved response rates and prolonged
progression-free survival (PFS). However, these data
are derived from non-randomized trials with PFS as
primary endpoint and the true antitumoral effect is not
yet clear. Available data on overall survival (OS) are
less robust (phase II) and sometimes even conflicting.
Furthermore, serious side-effects of angiogenesis
inhibitors (e.g. venous and arterial thrombo-embolism,
arterial hypertension and hemorrhage) have been
demonstrated in phase II studies (Norden et al., 2009;
Verhoeff et al., 2009).
Adoptive immunotherapy is the infusion of ex vivo
(antigen-specific or non-specific) stimulated lymphocytes or NK cells. This can be done locally in the
brain or systemically. An important advantage of this
strategy is that autoimmune reactions can be avoided.
Possible drawbacks, however, are the risk of contamination with Treg and the short survival of ex vivo
stimulated cells. Studies have shown that adoptive
transfer of autologous T cells held little or no value
for leukemia, but was moderately effective for malignant melanoma. On the other hand, adoptive transfer of
allogeneic NK cells could play a role in selected cases
of acute myeloid leukemia (Borghaei et al., 2009). For
gliomas local (intracerebral) administration of different effector immune cells (including NK cells, CTL
and activated killer cells) has been widely studied. This strategy was well-tolerated but did not yield
clinical responses. Additional local administration of
IL-2 as immuno-adjuvant resulted in significant toxicity (De Vleeschouwer et al., 2005; Grauer et al.,
2009).
Active Specific Immunotherapy

Finally, active specific immunotherapy, or the so-called
tumor vaccination, comprises the in vivo induction
of tumor-specific CTL (and/or humoral responses)
and this approach has gained considerable interest
the last decade (De Vleeschouwer et al., 2006).
Four types of tumor vaccination can be distinguished
(Table 32.1). 1. Autologous or allogeneic tumor cell
vaccines are based on the assumption that the TAA
that are expressed by the tumor can induce a cytotoxic T cell response. The vaccine consists of nonproliferating tumor cells in combination with an
immuno-adjuvant to enhance the immune response
(Bradbury and Shepherd, 2008; Grange et al., 2009;
Van Poppel et al., 2009). 2. In case of genetically
modified tumor cell vaccines, autologous or allogeneic tumor cell vaccines have been genetically engineered; genes encoding immunostimulatory cytokines
or positive co-stimulatory molecules have been integrated into the tumor cells by transfection. This strategy is based on evidence that local cytokine secretion
can activate T cells and NK cells as well as induce
inflammatory responses against tumors (Bradbury and
Shepherd, 2008; Terando et al., 2007; Van Poppel
et al., 2009). 3. Peptide-based vaccines make use of
one or more immunogenic TAA that are universally
expressed by the tumor cells (Bradbury and Shepherd,
2008; Terando et al., 2007; Van Poppel et al., 2009). 4.
DC-based tumor vaccines use DC as APC to activate
CD4+ and CD8+ T cells. Moreover, DC can activate NK and NKT cells as well (Dhodapkar et al.,
2004). DC are generated ex vivo and subsequently
loaded with a source of TAA, that will be presented
by the DC. Several cellular products can be used to
load the DC: purified defined peptides, undefined acideluted peptides from autologous tumor, tumor cell
lysate, viral vector containing genes, apoptotic bodies,
tumor homogenate, mRNA, and necrotic tumor cells
(Bradbury and Shepherd, 2008; Grange et al., 2009;
Mu et al., 2003; Sorg et al., 2003; Terando et al., 2007;
Thurner et al., 1999; Tuyaerts et al., 2002; Van Poppel
et al., 2009). From a theoretical point of view, the use
of total tumor-derived material (lysate/homogenate) as
a source of tumor antigens may be preferable because
of the emergence of antigen-loss variants if the immunization is only targeted at one TAA, which might not
be expressed by a subpopulation of tumor cells that
319
will be positively selected (De Vleeschouwer et al.,

2005; Gilboa et al., 1998).
For glioma immunotherapy, the best responses
have been reported with tumor vaccination as adjuvant postoperative therapy (Choi et al., 2009; Grauer
et al., 2009; Van Gool et al., 2009). Autologous
tumor cell vaccines in combination with cytokinesecreting fibroblasts and genetically modified autologous tumor cell vaccines have been used to treat
gliomas, but especially the results of peptide- and DCbased immunotherapeutical approaches seem promising. Peptide-based vaccines use epidermal growth factor receptor class III variant (EGFRvIII) to induce an
immune response. EGFRvIII is expressed by many,
but not all malignant gliomas and can induce specific cellular and humoral immunity. Preclinical and
clinical studies have shown that the use of EGFRvIII
vaccines is safe and feasible. Results of an ongoing
phase III study are eagerly anticipated (Choi et al.,
2009). Furthermore, much preclinical and clinical
research has been done on the role of DC-based tumor
vaccination in the treatment of malignant gliomas.
So far, seventeen clinical phase I/II studies and case
reports have been published in the literature with
emphasis on feasibility and toxicity of the treatment approach (Table 32.2). No autoimmune phenomena were observed and the treatment was welltolerated. As for the EGFRvIII-based vaccines, results
are promising.
Experimental Glioma Models

for Dendritic Cell Therapy
Over the last decades, numerous reports on heterotopic and orthotopic rat and mouse GBM models
have been published, and it is important to know that
there are some fundamental differences between those
models. For example, models exploiting spontaneous
tumor formation in genetically engineered mice differ from engrafted tumor models that are established
by implantation of primary tumor cells or tumor cell
lines. Spontaneous tumor models mimic gliomagenesis much better than engrafted models, but the main
disadvantages are the poor reproducibility, low tumor
penetrance, prolonged and unpredictable latency for
tumor formation, and the need for advanced in vivo
320
H. Ardon et al.
Table 32.1 Active specific immunotherapy

Vaccine type
Advantages
Disadvantages
Tumor
typea
Autologous tumor
cell vaccine
Patient specific
Not widely applicable
MM
Potential to generate
immunity to any
TAA
Time-consuming and
technically
challenging
preparation
RCC
Humoral and cellular

antitumoral
immunity
Requires adequate
tumor tissue for
manufacture
NSCLC
HGG
Allogeneic tumor
cell vaccine
Widely applicable
Not patient specific
MM
Patients tumor tissue

need not be available
Moderately
complex/demanding
preparation
Response dictated by
similarity between
TAA of vaccine and
TAA of patients
tumor
Requires additional
exvivo manipulation
with issues of
altered cell viability
and/or functionality
RCC
Humoral and cellular

antitumoral
immunity
Gene modified
cellular vaccine
May enhance efficacy

of non-gene
modified approaches
NSCLC
Limited clinical response

(Grange et al., 2009; Terando
et al., 2007)
Immunological and clinical
response in selected
patientsPhase III study
(Reniale) with significant
improved PFS and OS (Van
Poppel et al., 2009)
Limited data available
ongoing studies (Bradbury
and Shepherd, 2008; Kakimi
et al., 2009)
Limited data available (De
Vleeschouwer et al., 2005,
Grauer et al., 2009)
Limited clinical response
et al., 2007)
Only used after gene
modification (Van Poppel
et al., 2009)
Only used after gene
modification (Bradbury and
Shepherd, 2008; Kakimi
et al., 2009)
MM
Limited immunological and

clinical response (Grange
et al., 2009; Terando et al.,
2007)
RCC
Possible limited response (Van

Belagenpumatucel-L. clinical
response GVAX
(granulocyte
macrophage-colony
stimulating factor secreting
tumor cells): no response
(Bradbury and Shepherd,
2008; Kakimi et al., 2009)
Limited data available (De
Vleeschouwer et al., 2005;
Grauer et al., 2009)
Extensively studied measurable T cell response rarely clinical response
et al., 2007)
Limited number of studies
phase III study (Vitespen)
showed no better PFS (Van
NSCLC
HGG
Peptide vaccine
Resultsb
Acellular technique
TAA have to be
immunogenic and
tumorigenic
MM
Widely applicable
Tumor cells not

expressing TAA
evade immune
system (tumor
immune escape)
RCC
321

Vaccine type
Advantages
Disadvantages
Tumor
typea
Use of known epitopes

simplifies immune
monitoring
HLA restriction
NSCLC
Resultsb
Mucin-1 vaccine promising

results ongoing phase III
study (Bradbury and
Shepherd, 2008; Kakimi
et al., 2009)
HGG
EGFRvIII: immunological and
clinical (better PFS and OS)
response ongoing phase III
study (De Vleeschouwer
et al., 2005; Choi et al., 2009)
DC vaccine
Utilizes most potent
Leukapheresis
MM
Clinical response; 7% (other
antigen-presenting
necessary
active specific
cells
immunotherapy 34%)
(Engell-Noerregaard et al.,
2009; Terando et al., 2007)
Can be loaded with
Time-consuming and
Phase III study with positive
TAA through a
technically
results (Editorial note Future
variety of techniques
challenging
Oncol, 2009) and phase III
preparation
study with negative results
(Schadendorf et al., 2006)
Yield can be variable
RCC
and patient
response (Van Poppel et al.,
dependent
2009)
NSCLC
Immunological response no
clinical Response (Bradbury
and Shepherd, 2008; Kakimi
et al., 2009)
HGG
response (no clear
correlation) (Ardon et al.,
2010a; Ardon et al., 2010b;
De Vleeschouwer et al.,
2008; Van Gool et al., 2009;
De Vleeschouwer et al.,
2006; De Vleeschouwer
et al., 2005; Grauer et al.,
2009)
DC. dendritic cells. EGFRvIII. epidermal growth factor receptor class III variant. HGG. high-grade ghoma. HLA. human leukocyte
antigen. MM. malignant melanoma. NSCLC. non-small cell lung carcinoma. PFS. progression-free survival. OS. overall survival.
RCC. renal cell carcinoma. TAA. tumor-associated antigen
a not exhaustive
b phase III studies and case reports. unless otherwise stated
imaging techniques. Engrafted models on the other

hand lack the stepwise genetic changes occurring during tumor progression, resulting in tumors that are well
circumscribed and less invasive, lack the human counterpart of microvascular proliferation and rarely recapitulate the tumor-of-origin phenotype. Nevertheless,
based on greater reproducibility, engrafted models are
better suited for evaluating preclinical therapies such
as DC-based immunotherapy, as long as the models are
studied in immune competent animals.
The GL261 glioma model has been widely studied and Ni et al. (2001) underscored the immunogenicity of GL261 tumor cells by treating mice with
intracranial glioma with tumor extract-pulsed cloned
DC. Cured animals showed increased delayed-type
hypersensitivity (DTH) responses to GL261 cells, with
long-term tumor protection. Aoki et al. (2001) showed
that pulsing DC with a complex of tumor extract and
cationic liposomes induces an antitumoral immune
response against intracranial glioma in which CD8+
Neurosurgical
Focus 9: e8
(2000)
Cancer Res
61:842847
(2001)
Cancer Immunol.
Immunother.
50: 337344
(2001)
Br J Cancer
89:11721179
(2003)
J Immunol
171:4927
4933
(2003)
Liau et al.
Yu et al.
Kikuchi et al.
Yamanaka et al.
Wheeler et al.
17 (phase II)
17 (phase I)
10 (phase I-II)
8 (phase I)
9 (phase I)
1 (case report)
Table 32.2 Overview of reported clinical trials

No. of patients
Author
Publication
(type of trial)
Autologous
tumor
freeze-thaw
lysate
Autologous
tumor lysate
DC fusion with
autologous
glioma cells
Tumor-specific
MHC-I
associated
peptides
Allogeneic MHC
class
I-matched
GBM peptides
Cell product
Not reported
Not reported
Intradermal
and/or
intratumoral
(Ommaya)
Intradermal
Subcutaneous
Intradermal
Administration
3 vaccinations
at 2-week
intervals with
fourth
vaccination 6
week after the
third (n=10)
3 vaccinations
at 2-week
intervals
110
vaccinations
at 3-week
intervals
37 vaccinations
at 3-week
intervals
3 vaccinations
at 2-week
intervals
3 biweekly
vaccinations
Treatment
schedule
Increase in NK cells
in peripheral
blood (n=5);
positive DTH
reaction to tumor
lysate (n=3);
increased T-cell
mediated
antitumoral
activity (n=2);
Elispot IFN
Not reported (study
on CD8+ RTE)
T-cell proliferation
response to
allogeneic
acid-eluted tumor
peptides
Systemic CTL
cytotoxicty
against tumor
(n=4) (JAM
assay)
Increase in NK cells
in peripheral
blood (n=4);
increased IFN in
supernatant
(n=6)
Immune response
Clinical response
Not reported
Mixed response
(n=1);
steroids during
vaccination
(n=5);
2 minor
responses
Minor response
(n=2)
Prolonged
survival
compared to
control group
None
322
H. Ardon et al.
Publication
Cancer Res
64:49734979
(2004)
J. Neurosurg.
(Pediatrics)
100: 492497
(2004)
Br J. Cancer
91:16561662
(2004)
J Immunother
27:452459
(2004)
Clin Cancer Res

11:55155525
(2005)
Author
Yu et al.
De Vleeschouwer
et al.
Rutkowski et al.
Kikuchi et al.
Liau et al.
12 (multicohort
dose-escalation
study phase I)
15 (phase I-II)
12 (phase I)
1 (case report)
14 (phase I-II)
No. of patients
(type of trial)
Acid-eluted
tumor
associated
peptides
(autologous)
DC fusion with
autologous
glioma cells
Autologous
tumor lysate
Autologous
tumor lysate
Autologous
tumor lysate
Cell product
Intradermal
Intradermal
Intradermal
Intradermal
Subcutaneous
Administration
3 vaccinations
Vaccination +
rhIL-12
Weeks 1, 3, and
further with
4-week
interval (6
total)
Weeks 1, 3, and
further with
4-week
interval (27
vaccinations)
3 vaccinations
at 2-week
intervals
Treatment
schedule
Immune response
Positive DTH
reaction to tumor
lysate (n=15);
increased
cytotoxic activity
(n=2); increased
intracellular IFN
in CD8+ T cells
(n=1)
CTL response
(n=6)
Positive DTH
reaction to tumor
lysate (n=6)
IFN release in
PBMC (n=6);
expansion of
CD8+ antigen
specific T cell
clones (n=4);
systemic T cell
cytotoxicity
against tumor
(n=1)
Positive DTH
reaction to tumor
lysate
Clinical response
Median TTP
19.9 mo; OS
18 to >58 mo;
median OS
35.8 mo
Long-lasting
tumor-free
survival (> 5
year after
vaccination)
Partial response
(n=1);
tumor-free
survival (n=2;
5 year after
vaccination)
Partial response
(n=4); mixed
response
(n=1)
Significant
increase in
median
survival

323
Publication
Clin Cancer Res

11:41604167
(2005)
J Transl Med
5:67 (2007)
Cancer Res
68:59555964
(2008)
J Clin Neurosci
15:114121
(2008)
Author
Yamanaka
et al.
Okada et al.
Wheeler et al.
Walker et al.
13 (phase I)
34 (phase II)
2 (A)/5 (B)
24 (phase I-II)
No. of patients
(type of trial)
Single-cell
suspension of
autologous
tumor cells
IL-4 transfected
fibroblasts +
apoptotic
glioma cells
(A)/autologous
tumor lysate
(B)
Autologous
tumor lysate
Autologous
tumor lysate
Cell product
Intradermal
Subcutaneous
Intradermal
Intradermal or
intradermal +
intratumoral
(Ommaya)
Administration
Priming phase
consisting of 6
vaccinations at
2-week
intervals;
further
vaccinations at
6-week
intervals
3 vaccinations at
2-week
intervals with
fourth
vaccination 6
week after the
third
2 vaccinations
with 2-week
interval
110
vaccinations at
3-week
intervals
Treatment
schedule
Post-vaccine
antigen-directed
IFN response in
PBMC
(qPCR-based
assay) (n=17);
DTH-test resulted
in cutaneous
GBM in 1 pat
(DTH was
subsequently
discontinued)
Increased T cell
infiltration in
post-vaccination
tumor specimens
compared to
pre-vaccination
specimens (n=3)
A: increased T cell
reactivity; B: no
response
Positive DTH
reaction to tumor
lysate (n=8);
positive IFN
Elispot (n=7)
Immune response
9-month survival
(9/13); 12-month
survival (6/13);
18 months or
longer survival
(3/13)
Significant positive
correlation
between
post-vaccine
response
magnitude on one
hand and TTS
and TTP
spanning
chemotherapy on
the other hand
Partial response
(n=1); minor
response (n=3);
significant
increase in
median survival
A: clinical and
radiological
response in both
patients; B: no
response
Clinical response
324
H. Ardon et al.
Clin Cancer Res

14:30983104
(2008)
Pediatr Blood
Cancer
54:519525
(2010a)
J Neurooncol
99:261272
(2010b)
De Vleeschouwer
et al.
Ardon et al.
Ardon et al.
8 (pilot study)
45 children
(phase I-II)
(33 HGG)
56 (phase I-II)
No. of patients
(type of trial)
Autologous
tumor lysate
Autologous
tumor lysate
Autologous
tumor lysate
Cell product
Intradermal
Intradermal
Intradermal
Administration
Cohort
comparison
Cohort
comparison
Treatment
schedule
Not reported
Positive DTH
reaction to tumor
lysate (9/21 at
time of diagnosis
and 9/17 after
two vaccinations)
Immune response
PFS 3 months; OS
9.6 months; 2-year
OS 14.8%; total
resection is
predictor for better
PFS; younger age
and total resection
are predictors for
better OS in
univariable
analysis; tendency
towards improved
PFS when faster
DC vaccination
schedule with
tumor lysate
boosting was
applied
PFS HGG 4.4 months
GBM 4.3
months; OS HGG
13.5 months
GBM 12.2 months;
2-year OS 18%
6-month PFS 75%;
PFS 18 months;
OS 24 months
Clinical response
Incorporated in
Positive DTH
RChT 4
reaction to tumor
induction
lysate (3/7);
vaccinations
positive IFN
at 1-week
Elispot (5/8);
interval; then
increase CTL
3 boost
(6/7)
vaccinations
with 1-month
interval,
followed by
3-monthly
boostvaccinations
CTL, cytotoxic T lymphocyte; DC, dendritic cell; DTH, delayed-type hypersensitivity; GBM, glioblastoma multiforme; IFN, interferon-; IL-4, interleukin-4; JAM, just another
method; MHC, major histocompatibility complex; mo, month; NK, natural killer; no., number; OS, overall survival; pat, patient; PBMC, peripheral blood mononuclear cell; PFS,
progression-free survival; qPCR, quantitative polymerase chain reaction; rhIL-12, recombinant human interleukin-12; RTE, naive recent thymic emigrant T cell; TTP, time to
progression; TTS, time to survival; w, week
Publication
Author

325
326
T cells are involved. Protective immunity against

intracranial glioma growth obtained through immunization with either lysate- or RNA-loaded DC was
reported by Insug et al. (2002), while adding recombinant IL-12 to the vaccine regimen further improved
its efficacy. Vaccination with lysate-pulsed DC combined with IFN- gene therapy resulted in a survival
benefit, as was demonstrated by Saito et al. (2004).
Similarly, the group of Okada et al. (2007) revealed
that the sequential intratumoral delivery of an IFN-
encoding adenoviral vector and DC induced longterm survival and specific CTL activity. Furthermore,
the same group showed that intratumoral administration of DC, genetically engineered to secrete IFN, enhanced the efficacy of peripheral vaccines with
cytokine gene-transduced tumor cells. Using vaccines created through electrofusion of DC and irradiated tumor cells, Kjaergaard et al. (2005) observed
complete tumor regression of established intracranial
tumors with infiltration of both CD4+ and CD8+ T
cells. Ciesielski et al. (2006) have focused on survivin, a member of the apoptosis inhibition family of
proteins, as GL261 TAA. In particular, the authors
exploited the xenogeneic differences between human
and murine survivin sequences to develop a more
immunogenic tumor vaccine. The efficacy of systemic immunotherapy with DC loaded with GL261
antigens was confirmed by Pellegatta et al. (2006a).
Additionally, these authors introduced the concept of
cancer stem cells in this model and reported that DC
targeting of such stem cells within the GL261 tumor
cell pool provides a higher level of protection against
GL261 glioma (Pellegatta et al., 2006b). Recently,
Grauer et al. (2007 and 2008) illustrated the pronounced impact of FoxP3+ Treg in the GL261 model
and suggested that Treg elimination is a prerequisite
for successful eradication of established glioma using
tumor lysate pulsed DC. Maes et al. (2009) showed
that DC loaded with tumor antigens in the form of
RNA molecules are capable of inducing a T cellmediated antitumoral immune response. Furthermore,
the impact of Treg in this model was evident and
elimination of Treg resulted in long-term surviving
mice. Strikingly, upon rechallenge of long-term surviving mice after Treg depletion and/or DC vaccination,
only those mice that were treated with DC vaccination
depicted a prolonged antitumoral protective immune
response.
H. Ardon et al.
Dendritic Cell Therapy for Human

Gliomas
Seventeen clinical phase I/II studies and case reports
have been published in the literature since 2000
(Table 32.2). The median patient number in these
reports was only 12, ranging from 1 to 56. The inclusion criteria, immunotherapeutic designs and interpretations varied significantly among and even within
the reports, being based entirely on single-center
approaches and hypotheses. Hence, no firm conclusions can be drawn regarding the optimal immunotherapeutic strategy, nor would it make sense to perform
a meta-analysis on these data. The importance of a
minimal residual disease setting for adjuvant postoperative DC vaccination in HGG patients was already
stressed in the first reports. In the series published
by De Vleeschouwer et al. (2008), the importance
of a gross total resection prior to DC vaccination
for GBM patients was confirmed in a multivariate
analysis. General awareness and agreement on this
important issue is mounting and even culminating in
the assumption that therapeutic tumor vaccinations
should be used for early disease states. Most of the
reports describe patients treated with immunotherapy
at a stage of minimal residual disease, after gross total
or near-total resection. Interestingly, patients have been
under maintenance corticosteroid treatment just prior
and/or during immunotherapy in some trials, which
one would expect to have a negative effect on the
generation of an effective immune response. In order
to avoid this confounding variable, other trials have
been restricted to patients who could obtain a total
or near-total resection and be rapidly weaned from
corticosteroids. There are multiple arguments supporting the need for these restrictions. First, because there
are major immune-suppressive mechanisms at play
within the tumor microenvironment and even systemically, only a gross total resection results in a
clinically effective recovery of normal immune system function. Second, it might be difficult to produce
good-quality DC out of monocytes isolated at the time
of corticosteroid treatment. Third, an overwhelming
peritumoral inflammatory reaction was seen in one
patient with bulky residual tumor (Rutkowski et al.,
2004). In this patient, the vaccine-induced inflammatory immune reaction occurred with a progressively
327
increasing delay after the vaccine injection. This is

in line with the hypothesis that the growing tumor
and, maybe as such, the increasingly induced immune
suppression counteract the vaccine-induced immune
response. This observation points to a tumor-specific
reaction induced by the vaccine, but maybe also to a
gradually shifting balance in favor of tumor-induced
immune suppression.
monitoring tools directed toward the affected organ,

that is, the brain, will have to take into account that the
presence of tumor-infiltrating lymphocytes and other
immune effector cells does not unequivocally correlate with a good or a bad outcome. Proof of efficacy of
immunotherapy asks for impact on OS in patients with
HGG, of whom the different prognostic risk factors are
clearly categorized.
Selection of Outcome Measures
Nature of Dendritic Cells
Immunological responses, assessed by a wide variety

of immune monitoring tools, have been demonstrated
in 50% of reported cases, ranging from 20 to 100%.
Only recently, Wheeler et al. (2008) found stronger
evidence between immunological responses and survival in vaccinated glioma patients. However, the use
of peripheral immune monitoring tools might prove
inadequate in case the peripheral immune system does
not mirror in any way local immune events in the
brain tumor microenvironment. Although immunological responses are mandatory to map the way and to
provide proof of the principle for this therapeutic strategy, proof of efficacy can only be established in
large well-designed comparative clinical trials, ideally
in controlled randomized trials. Especially in HGG,
an old paradigm states that pre-treatment prognostic variables might have more impact on outcome
than any (new) potentially active therapy or treatment
strategy. To that end, recursive partitioning analysis (RPA) as originally described and validated for
newly diagnosed malignant gliomas by the Radiation
Therapy Oncology Group (RTOG) in 1993, provides
us with an excellent model of prognostic classes
of pre-treatment and treatment-related patients variables. Using this RPA classification, the impact of
new treatments on OS in HGG can be examined in
clinically similar patient groups and compared with
large databases of conventionally treated patients. To
date, there is consensus that clinical responses, as
defined by the Macdonald criteria (Macdonald et al.,
1990), do not apply for biological treatment modalities
such as tumor vaccination. The paradigm of a therapeutic vaccine leading to (detectable) immunological
responses and hence to (detectable) clinical responses
with increased OS is at least incomplete. Even immune
All clinical reports have used monocyte-derived DC

for their trials. The methodology for DC preparation
is now fairly well established, and gives a sufficient
yield of mature DC (DCm) for injections into patients.
Most of the older trials have used immature DC,
while others have used maturation stimuli like TNF, penicillin-killed Streptococcus pyogenes (OK-432),
monocyte-conditioned medium, IFN- and TNF- in
combination with IL-4-secreting fibroblasts (Van Gool
et al., 2009). Only one phase I study focused on
the dose of DC and could not find any dose-limiting
toxicity (Liau et al., 2005).
At the time of the design of most published clinical trials, it was not known if and which lymph nodes
would be the optimal destination of injected DC. Later
on, however, data became available that priming of T
cells by DC within the cervical lymph nodes induced
an integrin homing pattern toward intracerebral locations. Other preclinical in vivo brain tumor models
clearly showed an enrichment of tumor-specific Treg
in the blood and cervical lymph nodes. As pointed out
by Tuyaerts et al. (2007), the question whether DC
should be injected intradermally, intravenously or in
the lymph nodes is not yet resolved. This becomes
particularly important as only a small amount of the
intradermally injected DC ultimately reaches the T cell
area of lymph nodes. However, taking into account the
documented failure rate of even experienced radiologists injecting DC in lymph nodes under ultrasound
guidance, and the desire to spread DC vaccination
technology to more centers in order to perform largescale clinical trials, most researchers in the field favor
the intradermal route for DC injection.
A very interesting issue is the potential role that
local injection of DC directly into the tumor region
328
might play in immunotherapeutical approaches. In

their study, Yamanaka et al. (2005) showed some benefit from intratumoral plus intradermal injections of
DC as compared with only intradermal injections,
although one should note that this observation was
not derived from a prospective randomized approach.
Nevertheless, a local change from an immune suppressive into a more immune stimulatory microenvironment might be induced by intratumoral administration
of DC.
Monitoring
Immune responses following vaccination have been
monitored in most trials. These analyses have included
positive DTH skin reaction, T cell reactivity and NK
cell enrichment in peripheral blood, as well as measuring T cell infiltration in tumoral tissue taken after
vaccination. The reported immune monitoring remains
very global in these clinical trials, mainly due to the
lack of specific antigens to be targeted. Although data
are still preliminary, advanced magnetic resonance
imaging (MRI) eventually combined with positron
emission tomography (PET) may soon provide better
tools to monitor the effects of immunotherapy of HGG.
Results of Clinical Studies

Because all clinical trials reported thus far comprise
case reports, phase I, phase I/II or phase II trials, the questions answered in these trials relate to
feasibility, toxicity and early efficacy, both immunologically and/or oncologically (Table 32.2). In all
trials, it seemed to be feasible to administer DCbased vaccines. Moreover, all reports conclude that
the toxicity is minimal except for the case reported
by Rutkowski et al. (2004) where an overwhelming
inflammatory reaction around a residual tumor was
observed. Of major importance, none of the reports
describes autoimmune phenomena induced by DC vaccination.
Besides feasibility, toxicity and clinical outcome,
all published studies on DC-based immunotherapy
have aimed to assess immune responses. Several
immune assays have been developed for this purpose,
H. Ardon et al.
but so far, results have been difficult to interpret.

Moreover, immunological responses are not detected
in all patients. A correlation between immunologically
measurable effects and clinical outcome was reported
in several papers. In the study by Wheeler et al. (2008),
a significant progressive (logarithmic) association was
found between time to survival after inclusion and
post-vaccine IFN- enhancement. This allowed them
to propose that single-vaccine response metrics can
quantitatively reflect inhibition of tumor progression
by T cells in GBM patients.
Only a limited number of research groups are
actively performing clinical trials world-wide. All
these trials are small, and include heterogeneous
groups of patients with regard to histology and/or
time of vaccination. Some studies included patients
with all subtypes of HGG, while others focused solely
on patients with GBM. Some studied patients with
relapsed HGG, while others included patients with a
first diagnosis of HGG, either exclusively or combined
with patients treated at recurrence. Only one group
reported on children with relapsed HGG. Moreover,
as pointed out before, the vaccination protocol varied enormously among research groups. Due to this
heterogeneity in study design and execution, sound
comparative analyses on clinical outcome are difficult. Nevertheless, in most trials, some clinical effect is
observed and some important clues for future clinical
studies are provided, especially from trials where there
are comparisons with historical controls or between
internal subgroups. For example, Yu et al. (2001)
reported that seven patients with newly diagnosed
GBM who received three biweekly intradermal vaccinations with peptide-pulsed DC had median survival
times of 455 days as compared to 257 for 42 nonvaccinated but otherwise comparably treated patients
in the same institute. Although compared with historical control patients and although no statistics are
provided, the prolongation of the median survival is
an interesting observation in this phase I clinical trial.
The same group reported on vaccinations at the time
of relapse in eight GBM patients, and made a similar comparison to 26 historical control patients (Yu
et al., 2004). Both patient groups underwent second
craniotomy for relapsed GBM. Whereas the median
survival was 30 weeks in the control group, subsequent vaccination after new resection resulted in a
significant change of median survival to 133 weeks,
with some patients surviving for more than 250 weeks.
In a recent paper, the same group demonstrated that

vaccine responders exhibited more favorable clinical
outcomes relative to non-responders. Moreover, the
vaccine-induced responses elicited therapeutic benefits primarily by sensitizing tumors to subsequent
chemotherapy (Wheeler et al., 2008). In the phase I
study reported by Liau et al. (2005), clinical outcome
in the patients with stable tumors or no residual disease
at time of DC vaccination compared favorably, even
when compared with historical/concurrent data for the
best prognostic subgroup of GBM patients treated during the same time period at their institute. In contrast,
Okada et al. (2007) did not show any benefit in PFS in
patients treated with DC loaded with autologous tumor
lysate in combination with IL-4 producing fibroblasts.
In the study on recurrent HGG patients reported by
Yamanaka et al. (2003), radiological responses were
observed in the group of five patients who received
intratumoral plus intradermal vaccinations, while no
response was detected in the five patients who were
only injected intradermally.
De Vleeschouwer et al. (2008) report on a large
group of patients with relapsed GBM treated with DC
loaded with tumor lysate in a stepwise process using
a cohort-comparison study concept (HGG-IMMUNO2003), comprising of four cohorts (A-D). Following
this strategy, the authors have been able to compare
each cohort within the trial with the other cohorts.
Thus far, three major issues in the complexity of
DC immunotherapy have been addressed: the vaccination schedule, the boosting and the maturation
cocktail. In this large group of patients, a complete
resection improved the median PFS and OS as compared to an incomplete resection. Interestingly, a stepwise improvement in PFS curves could be seen by
shortening the interval between the induction DC vaccines, with further improvements noted by adapting
the boosting intervals and technology, and finally by
changing the maturation cocktail. Reported 2-year survival was 16.7% in cohort C and 27.8% in cohort D.
In the report by Ardon et al. (2010b), DC-based
immunotherapy was integrated into the primary treatment regimen for eight pilot patients with newly
diagnosed GBM. After radiochemotherapy, 4 weekly
DC-based vaccines were injected, and further boosting
was done with lysate vaccines during the maintenance temozolomide phase. Median PFS in these eight
patients was 17.8 months and median OS was 24.3
months. PFS at 6 months was seen in six of the eight
329
patients (75%). The report on the results of a large

phase I/II trial HGG-2006 with 77 patients included,
is in preparation.
Optimization of Immunotherapy
Although several studies have reported promising results, and activation of the immune system
against gliomas seems possible, immunotherapeutical approaches are counteracted by rapid tumor
progression and glioma-induced immune suppression. There is a clear need for improvement of the
immunotherapeutical strategies and several options
exist to enhance the immune response. It has been
shown in murine models that co-stimulation of T cells
can be enhanced by using agonistic antibodies to costimulatory molecules, resulting in a better activation
of T cells. The use of antagonistic antibodies to coinhibitory molecules (such as cytotoxic lymphocyteassociated antigen-4 (CTLA-4) and programmeddeath-1 receptor) can also lead to a stronger T cell
activation (Grauer et al., 2009; McDermott, 2009).
It is also important to target Treg that suppress the
antitumoral immune response. As already mentioned,
it has been shown in mouse models that elimination of
Treg, either in combination with anti-CTLA-4 antibodies or not, results in increased immunity against glioma
and a stronger response to DC-based immunotherapy
(Grauer et al., 2008; Maes et al., 2009). In humans,
studies have been carried out using daclizumab
(anti-CD25 antibody), ONTAK (denileukin diftitox)
and CD25-specific immunotoxins (LMB-2) (Grauer
et al., 2009; Rech and Vonderheide, 2009). Clinical
responses, however, have been limited. Low-dose
metronomic temozolomide diminished Treg in a rodent
glioma model and in melanoma patients, and therefore seems a promising tool to use in glioma patients.
Unfortunately, this tool cannot be combined with DCbased immunotherapy since metronomic temozolomide administration would counteract the induction of
an immune response (Banissi et al., 2009; Su et al.,
2004). On the other hand, low-dose metronomic oral
cyclophosphamide has been propagated as an effective Treg-depleting regimen that does not counteract
immunotherapy (Ghiringhelli et al., 2004, 2007).
The local immune suppression by the glioma itself
can be targeted via suppressive cytokines, such as
330
TGF-2, that are produced by the tumor. TGF- is

capable of inhibiting T cell and B cell activation and
proliferation, it suppresses the activity of NK cells,
reduces the production of cytokines like IL-2, IL-6,
IL-10, IFN-, and suppresses the expression of MHCII molecules on glioma cells. In preclinical studies
it has been shown that the expression of TGF-2 by
GBM cells can be diminished by exposing them to
antisense oligonucleotides (AON) against mRNA of
TGF-2 The most advanced AON for the therapy
of high-grade gliomas is a phosphorothioate-modified
AON, AP 12009 (trabedersen), which targets mRNA
encoding TGF-2. AP 12009 is administered intratumorally using convection-enhanced delivery. A series
of Phase I and II clinical trials have evaluated the toxicity profile and optimal dose of the substance, and a
randomized, controlled Phase III study is ongoing in
patients with recurrent or refractory anaplastic astrocytoma after standard radio- and chemotherapy (Hau
et al., 2007, 2009; Schneider et al., 2008).
Finally, production of DC can be optimized; for
example, maturation of DC can be improved by
using Toll-like receptor agonists (De Vleeschouwer
et al., 2008; Grauer et al., 2009; Van Gool et al.,
2009).
Conclusion
In recent years insights into tumor immunology have
been refined and it is now clear that both the adaptive and innate immune system play an important role.
Moreover, it has been shown that immunological processes take place within the CNS and the brain should
be looked upon as an immune-distinct environment.
Many glioma-specific antigens have been described
and interactions between gliomas and the immune
system are being elucidated: gliomas will grow and
evade the antitumoral immune responses if the balance between tumor cell proliferation and elimination tips in favor of immune-resistant cells (immune
escape).
Immunotherapy for patients with HGG is a novel
therapeutic approach that opens new opportunities
for enhanced survival without major toxicity. This
is of particular importance, taking into account that
HGG cause a relatively high community burden, not
only with many years of life lost due to cancer,
H. Ardon et al.
but also because of the morbidity from the tumor

and subsequent treatments. In the assessment of
immunotherapeutical results for patients with HGG,
the improvement of PFS and, particularly, the significant increase of OS with satisfactory quality of
life are, of course, by far the most important variables, much more than immunologic surrogate markers
for response. The studies in preclinical animal models for human gliomas, especially the murine models,
are interesting in order to obtain more insight into
the basic biology underlying disease evolution and
immunotherapeutic mechanisms. Still, timely translation of new concepts into clinical practice should be
the primary objective. With several groups entering
into clinical practice in this field, new data are now
being generated. However, it is of paramount importance to study this promising approach based on data
from well designed trials with appropriate end points.
Thus far, the observations with regard to both immunological responses and clinical responses are promising
and beneficial effects are reproducible. Of particular
interest is the fact that no induction of autoimmunity
or other major toxicities have been observed to date.
However, larger and more homogeneous patient groups
with more refined inclusion criteria should be studied for more rapid progress in this field. Moreover,
confirmatory studies with an appropriate randomization versus a control patient group, preferably even
stratified for prognostic markers, including molecular
tumor signature, should be implemented in the near
future. For this, collaborative efforts between wellorganized and experienced vaccination centers should
be established.
Acknowledgements This translational research program

has been supported by the Olivia Hendrickx Research Fund
(http://www.olivia.be). Support was also obtained from
Electrabel Netmanagement Vlaanderen, CAF Belgium, Baxter,
the Herman Memorial Research Fund (http://www.hmrf.be),
the James E. Kearney Memorial Fund and gifts from private
families and service clubs. Additionally, grants were obtained
from Stichting tegen Kanker, IWT (TBM project), the
Stem Cell Institute Leuven, the Emmanuel van der Schueren
Fund, the International Union against Cancer, the Klinisch
Onderzoeksfonds UZ Leuven, and the Fund for Scientific
Research Flanders (FWO-V). We are very grateful for the
technical assistance from KatjaVandenbrande, Goedele Stegen,
Vallentina Schaiko, Elke Nackers and Anas Van Hoylandt. We
thank the neurooncology team in the hospital for fruitful patient
discussion, and the staff of the Laboratory of Experimental
Immunology for basic scientific discussions.
References
Abou-Ghazal M, Yang DS, Qiao W, Reina-Ortiz C, Wei J,
Kong LY, Fuller GN, Hiraoka N, Priebe W, Sawaya R,
Heimberger AB (2008) The incidence, correlation with
tumor-infiltrating inflammation, and prognosis of phosphorylated STAT3 expression in human gliomas. Clin Cancer Res
14:82288235
Aoki H, Mizuno M, Natsume A, Tsugawa T, Tsujimura K,
Takahashi T, Yoshida J (2001) Dendritic cells pulsed with
tumor extract-cationic liposome complex increase the induction of cytotoxic T lymphocytes in mouse brain tumor.
Cancer Immunol Immunother 50:463468
Ardon H, De Vleeschouwer S, Van Calenbergh F, Claes L,
Kramm CM, Rutkowski S, Wolff JE, Van Gool SW (2010a)
Adjuvant dendritic cell-based tumour vaccination for children with malignant brain tumours. Pediatr. Blood Cancer
54:519525
Ardon H, Van Gool S, Lopes IS, Maes W, Sciot R, Wilms G,
Demaerel P, Bijttebier P, Claes L, Goffin J, Van Calenbergh F,
De Vleeschouwer S (2010b) Integration of autologous dendritic cell-based immunotherapy in the primary treatment for
patients with newly diagnosed glioblastoma multiforme: a
pilot study. J Neurooncol. [Epub ahead of print]
Balkwill F (2004) The significance of cancer cell expression
of the chemokine receptor CXCR4. Semin Cancer Biol 14:
171179
Banissi C, Ghiringhelli F, Chen L, Carpentier AF (2009) Treg
depletion with a low-dose metronomic temozolomide regimen in a rat glioma model. Cancer Immunol Immunother
58:16271634
Borghaei H, Smith MR, Campbell KS (2009) Immunotherapy of
cancer. Eur J Pharmacol 625:4154
Bradbury MW, Cserr HF, Westrop RJ (1981) Drainage of cerebral interstitial fluid into deep cervical lymph of the rabbit.
Am J Physiol 240:F329F336
Bradbury PA, Shepherd FA (2008) Immunotherapy for lung
cancer. J Thorac Oncol 3:S164S170
Choi BD, Archer GE, Mitchell DA, Heimberger AB, McLendon
RE, Bigner DD, Sampson JH (2009) EGFRvIII-targeted
vaccination therapy of malignant glioma. Brain Pathol 19:
713723
Ciesielski MJ, Apfel L, Barone TA, Castro CA, Weiss TC,
Fenstermaker RA (2006) Antitumor effects of a xenogeneic survivin bone marrow derived dendritic cell vaccine against murine GL261 gliomas. Cancer Immunol
Immunother 55:14911503
Copier J, Dalgleish AG, Britten CM, Finke LH, Gaudernack
G, Gnjatic S, Kallen K, Kiessling R, Schuessler-Lenz M,
Singh H, Talmadge J, Zwierzina H, Hakansson L (2009)
Improving the efficacy of cancer immunotherapy. Eur J
Cancer 45:14241431
Cserr HF, Knopf PM (1992) Cervical lymphatics,
the blood-brain barrier and the immunoreactivity of the brain: a new view. Immunol Today 13:
507512
331
De Vleeschouwer S, Fieuws S, Rutkowski S, Van Calenbergh
F, Van Loon J, Goffin J, Sciot R, Wilms G, Demaerel
P, Warmuth-Metz M, Soerensen N, Wolff JE, Wagner S,
Kaempgen E, Van Gool SW (2008) Postoperative adjuvant
dendritic cell-based immunotherapy in patients with relapsed
glioblastoma multiforme. Clin Cancer Res 14:30983104
De Vleeschouwer S, Rapp M, Sorg RV, Steiger HJ, Stummer
W, Van Gool S, Sabel M (2006) Dendritic cell vaccination
in patients with malignant gliomas: current status and future
directions. Neurosurgery 59:988999
De Vleeschouwer S, Van Calenbergh F, Demaerel P, Flamen
P, Rutkowski S, Kaempgen E, Wolff JE, Plets C, Sciot R,
Van Gool SW (2004) Transient local response and persistent tumor control of recurrent malignant glioma treated
with combination therapy including dendritic cell therapy. J
Neurosurg Pediatrics 100:492497
De Vleeschouwer S, Van Gool SW, Van Calenbergh F (2005)
Immunotherapy for malignant gliomas: emphasis on strategies of active specific immunotherapy using autologous
dendritic cells. Childs Nerv Syst 21:718
Dhodapkar KM, Cirignano B, Chamian F, Zagzag D, Miller
DC, Finlay JL, Steinman RM (2004) Invariant natural killer
T cells are preserved in patients with glioma and exhibit
antitumor lytic activity following dendritic cell-mediated
expansion. Int J Cancer 109:893899
Dix AR, Brooks WH, Roszman TL, Morford LA (1999) Immune
defects observed in patients with primary malignant brain
tumors. J Neuroimmunol 100:216232
Dunn GP, Bruce AT, Ikeda H, Old LJ, Schreiber RD (2002)
Cancer immunoediting: from immunosurveillance to tumor
escape. Nat Immunol 3:991998
Dunn GP, Dunn IF, Curry WT (2007) Focus on TILs: prognostic significance of tumor infiltrating lymphocytes in human
glioma. Cancer Immun 7:12
Dunn GP, Old LJ, Schreiber RD (2004a) The immunobiology of
cancer immunosurveillance and immunoediting. Immunity
21:137148
Dunn GP, Old LJ, Schreiber RD (2004b) The three Es of cancer
immunoediting. Annu Rev Immunol 22:329360
Editorial note Future Oncol (2009) Phase III trial indicates
potential for melanoma vaccine. Future Oncol 5:759761
Elliott LH, Brooks WH, Roszman TL (1987) Activation of
immunoregulatory lymphocytes obtained from patients with
malignant gliomas. J Neurosurg 67:231236
Engell-Noerregaard L, Hansen TH, Andersen MH, Thor Straten
P, Svane IM (2009) Review of clinical studies on dendritic
cell-based vaccination of patients with malignant melanoma:
assessment of correlation between clinical response and
vaccine parameters. Cancer Immunol Immunother 58:114
Death Differ. 13:12381241
Fecci PE, Mitchell DA, Whitesides JF, Xie W, Friedman AH,
Archer GE, Herndon JE, Bigner DD, Dranoff G, Sampson JH
(2006) Increased regulatory T-cell fraction amidst a diminished CD4 compartment explains cellular immune defects in
patients with malignant glioma. Cancer Res 66:32943302
Ghiringhelli F, Larmonier N, Schmitt E, Parcellier A, Cathelin
D, Garrido C, Chauffert B, Solary E, Bonnotte B, Martin
F (2004) CD4+CD25+ regulatory T cells suppress tumor
332
immunity but are sensitive to cyclophosphamide which
allows immunotherapy of established tumors to be curative.
Eur J Immunol 34:336344
Ghiringhelli F, Menard C, Puig PE, Ladoire S, Roux S, Martin
F, Solary E, Le Cesne A, Zitvogel L, Chauffert B (2007)
Metronomic cyclophosphamide regimen selectively depletes
CD4+CD25+ regulatory T cells and restores T and NK effector functions in end stage cancer patients. Cancer Immunol
Immunother 56:641648
Gilboa E, Nair SK, Lyerly HK (1998) Immunotherapy of
cancer with dendritic-cell-based vaccines. Cancer Immunol
Immunother 46:8287
Grange JM, Krone B, Stanford JL (2009) Immunotherapy for
malignant melanomatracing Ariadnes thread through the
labyrinth. Eur J Cancer 45:22662273
Grauer OM, Nierkens S, Bennink E, Toonen LW, Boon L,
Wesseling P, Sutmuller RP, Adema GJ (2007) CD4+FoxP3+
regulatory T cells gradually accumulate in gliomas during
tumor growth and efficiently suppress antiglioma immune
responses in vivo. Int J Cancer 121:95105
Grauer OM, Sutmuller RP, van Maren W, Jacobs JF, Bennink
E, Toonen LW, Nierkens S, Adema GJ (2008) Elimination
of regulatory T cells is essential for an effective vaccination
with tumor lysate-pulsed dendritic cells in a murine glioma
model. Int J Cancer 122:17941802
Grauer OM, Wesseling P, Adema GJ (2009) Immunotherapy of
diffuse gliomas: biological background, current status and
future developments. Brain Pathol 19:674693
Hau P, Jachimczak P, Bogdahn U (2009) Treatment of malignant
gliomas with TGF-beta2 antisense oligonucleotides. Expert
Rev Anticancer Ther 9:16631674
Hau P, Jachimczak P, Schlingensiepen R, Schulmeyer F, Jauch
T, Steinbrecher A, Brawanski A, Proescholdt M, Schlaier J,
Buchroithner J, Pichler J, Wurm G, Mehdorn M, Strege R,
Schuierer G, Villarrubia V, Fellner F, Jansen O, Straube T,
Nohria V, Goldbrunner M, Kunst M, Schmaus S, Stauder
G, Bogdahn U, Schlingensiepen KH (2007) Inhibition of
TGF-beta2 with AP 12009 in recurrent malignant gliomas:
from preclinical to phase I/II studies. Oligonucleotides 17:
201212
Heath WR, Carbone FR (2001) Cross-presentation, dendritic
cells, tolerance and immunity. Annu Rev Immunol 19:4764
Heimberger AB, Abou-Ghazal M, Reina-Ortiz C, Yang DS, Sun
W, Qiao W, Hiraoka N, Fuller GN (2008) Incidence and
prognostic impact of FoxP3+ regulatory T cells in human
gliomas. Clin Cancer Res 14:51665172
Hemmati HD, Nakano I, Lazareff JA, Masterman-Smith M,
Geschwind DH, Bronner-Fraser M, Kornblum HI (2003)
Cancerous stem cells can arise from pediatric brain tumors.
Proc Natl Acad Sci USA 100:1517815183
Insug O, Ku G, Ertl HC, Blaszczyk-Thurin M (2002) A dendritic cell vaccine induces protective immunity to intracranial
growth of glioma. Anticancer Res 22:613621
Kakimi K, Nakajima J, Wada H (2009) Active specific
immunotherapy and cell-transfer therapy for the treatment of
non-small cell lung cancer. Lung Cancer 65:18
Kikuchi T, Akasaki Y, Abe T, Fukuda T, Saotome H, Ryan
JL, Kufe DW, Ohno T (2004) Vaccination of glioma
patients with fusions of dendritic and glioma cells and
recombinant human interleukin 12. J Immunother 27:
452459
H. Ardon et al.
Kikuchi T, Akasaki Y, Irie M, Homma S, Abe T, Ohno T (2001)
Results of a phase I clinical trial of vaccination of glioma
patients with fusions of dendritic and glioma cells. Cancer
Immunol Immunother 50:337344
Kjaergaard J, Wang LX, Kuriyama H, Shu S, Plautz GE (2005)
Active immunotherapy for advanced intracranial murine
tumors by using dendritic cell-tumor cell fusion vaccines. J
Liau LM, Black KL, Martin NA, Sykes SN, Bronstein JM,
Jouben-Steele L, Mischel PS, Belldegrun A, Cloughesy TF
(2000) Treatment of a Glioblastoma patient by vaccination with autologous dendritic cells pulsed with allogeneic
Major Histocompatibility Complex Class I-matched tumor
peptides: case report. Neurosurgical Focus 9:e8
Liau LM, Prins RM, Kiertscher SM, Odesa SK, Kremen TJ,
Giovannone AJ, Lin JW, Chute DJ, Mischel PS, Cloughesy
TF, Roth MD (2005) Dendritic cell vaccination in glioblastoma patients induces systemic and intracranial T-cell
responses modulated by the local central nervous system
tumor microenvironment. Clin Cancer Res 11:55155525
Macdonald DR, Cascino TL, Schold SC Jr, Cairncross JG
(1990) Response criteria for phase II studies of supratentorial
malignant glioma. J Clin Oncol 8:12771280
Maes W, Rosas GG, Verbinnen B, Boon L, De Vleeschouwer
S, Ceuppens JL, Van Gool SW (2009) DC vaccination with
anti-CD25 treatment leads to long-term immunity against
experimental glioma. Neuro Oncol 11:529542
Maldonado-Lopez R, Moser M (2001) Dendritic cell subsets
and the regulation of Th1/Th2 responses. Semin Immunol
13:275282
Malmberg KJ (2004) Effective immunotherapy against cancer:
a question of overcoming immune suppression and immune
escape? Cancer Immunol Immunother 53:879892
McDermott DF (2009) Immunotherapy of metastatic renal cell
carcinoma. Cancer 115:22982305
McMahon EJ, Bailey SL, Miller SD (2006) CNS dendritic cells:
critical participants in CNS inflammation? . Neurochem Int
49:195203
Mu LJ, Gaudernack G, Saeboe-Larssen S, Hammerstad H,
Tierens A, Kvalheim G (2003) A protocol for generation
of clinical grade mRNA-transfected monocyte-derived dendritic cells for cancer vaccines. Scand J Immunol 58:578586
Ni HT, Spellman SR, Jean WC, Hall WA, Low WC (2001)
Immunization with dendritic cells pulsed with tumor extract
increases survival of mice bearing intracranial gliomas. J
Neurooncol 51:19
Norden AD, Drappatz J, Wen PY (2009) Antiangiogenic therapies for high-grade glioma. Nat Rev Neurol 5:610620
Okada H, Lieberman FS, Walter KA, Lunsford LD, Kondziolka
DS, Bejjani GK, Hamilton RL, Torres-Trejo A, Kalinski P,
Cai Q, Mabold JL, Edington HD, Butterfield LH, Whiteside
TL, Potter DM, Schold SC Jr, Pollack IF (2007) Autologous
glioma cell vaccine admixed with interleukin-4 gene transfected fibroblasts in the treatment of patients with malignant
gliomas. J Transl Med 5:67
Pellegatta S, Poliani PL, Corno D, Grisoli M, Cusimano
M, Ubiali F, Baggi F, Bruzzone MG, Finocchiaro G

(2006a) Dendritic cells pulsed with glioma lysates induce
immunity against syngeneic intracranial gliomas and
increase survival of tumor-bearing mice. Neurol Res 28:
527531
Pellegatta S, Poliani PL, Corno D, Menghi F, Ghielmetti F,
Suarez-Merino B, Caldera V, Nava S, Ravanini M, Facchetti
F, Bruzzone MG, Finocchiaro G (2006b) Neurospheres
enriched in cancer stem-like cells are highly effective in
eliciting a dendritic cell-mediated immune response against
malignant gliomas. Cancer Res 66:1024710252
Ransohoff RM, Kivisakk P, Kidd G (2003) Three or more routes
for leukocyte migration into the central nervous system. Nat
Rev Immunol 3:569581
Rech AJ, Vonderheide RH (2009) Clinical use of anti-CD25
antibody daclizumab to enhance immune responses to tumor
antigen vaccination by targeting regulatory T cells. Ann NY
Acad Sci 1174:99106
Roszman T, Elliott L, Brooks W (1991) Modulation of T-cell
function by gliomas. Immunol Today 12:370374
Rutkowski S, De Vleeschouwer S, Kaempgen E, Wolff JEA,
Kuhl J, Demaerel P, Warmuth-Metz M, Flamen P, Van
Calenbergh F, Plets C, Srensen N, Opitz A, Van Gool SW
(2004) Surgery and adjuvant dendritic cell-based tumour
vaccination for patients with relapsed malignant glioma, a
feasibility study. Br J Cancer 91:16561662
Saito R, Mizuno M, Nakahara N, Tsuno T, Kumabe T,
Yoshimoto T, Yoshida J (2004) Vaccination with tumor cell
lysate-pulsed dendritic cells augments the effect of IFN-beta
gene therapy for malignant glioma in an experimental mouse
intracranial glioma. Int J Cancer 111:777782
Schadendorf D, Ugurel S, Schuler-Thurner B, Nestle FO, Enk
A, Brocker EB, Grabbe S, Rittgen W, Edler L, Sucker
A, Zimpfer-Rechner C, Berger T, Kamarashev J, Burg
G, Jonuleit H, Tuttenberg A, Becker JC, Keikavoussi P,
Kmpgen E, Schuler G (2006) Dacarbazine (DTIC) versus vaccination with autologous peptide-pulsed dendritic
cells (DC) in first-line treatment of patients with metastatic
melanoma: a randomized phase III trial of the DC study
group of the DeCOG. Ann Oncol 17:563570
Schneider T, Becker A, Ringe K, Reinhold A, Firsching R, Sabel
BA (2008) Brain tumor therapy by combined vaccination
and antisense oligonucleotide delivery with nanoparticles. J
Neuroimmunol 195:2127
Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J,
Hide T, Henkelman RM, Cusimano MD, Dirks PB (2004)
Identification of human brain tumour initiating cells. Nature
432:396401
Sorg RV, Ozcan Z, Brefort T, Fischer J, Ackermann R, Muller M,
Wernet P (2003) Clinical-scale generation of dendritic cells
in a closed system. J Immunother 26:374383
Su YB, Sohn S, Krown SE, Livingston PO, Wolchok JD, Quinn
C, Williams L, Foster T, Sepkowitz KA, Chapman PB (2004)
Selective CD4+ lymphopenia in melanoma patients treated
with temozolomide: a toxicity with therapeutic implications.
Terando AM, Faries MB, Morton DL (2007) Vaccine therapy
for melanoma: current status and future directions. Vaccine
25(Suppl 2):B4B16
333
Thurner B, Roder C, Dieckmann D, Heuer H, Kruse M, Glaser
A, Keikavoussi P, Kmpgen E, Bender A, Schuler G (1999)
Generation of large numbers of fully mature and stable
dendritic cells from leukapheresis products for clinical application. J Immunol Methods 223:115
Tuyaerts S, Aerts JL, Corthals J, Neyns B, Heirman C, Breckpot
K, Thielemans K, Bonehill A (2007) Current approaches
in dendritic cell generation and future implications for
cancer immunotherapy. Cancer Immunol Immunother 56:
15131537
Tuyaerts S, Noppe SM, Corthals J, Breckpot K, Heirman C, De
Greef C, Van Riet I, Thielemans K (2002) Generation of
large numbers of dendritic cells in a closed system using Cell
Factories. J Immunol Methods 264:135151
Van Gool S, Maes W, Ardon H, Verschuere T, Van Cauter S, De
Vleeschouwer S (2009) Dendritic cell therapy of high-grade
gliomas. Brain Pathol 19:694712
Van Poppel H, Joniau S, Van Gool SW (2009) Vaccine therapy in patients with renal cell carcinoma. Eur Urol 55:
13331342
Verhoeff JJ, van Tellingen O, Claes A, Stalpers LJ, van
Linde ME, Richel DJ, Leenders WP, van Furth WR (2009)
Concerns about anti-angiogenic treatment in patients with
glioblastoma multiforme. BMC Cancer 9:444
Walker DG, Laherty R, Tomlinson FH, Chuah T, Schmidt C
(2008) Results of a phase I dendritic cell vaccine trial for
malignant astrocytoma: potential interaction with adjuvant
chemotherapy. J Clin Neurosci 15:114121
Wheeler CJ, Black KL, Liu G, Mazer M, Zhang XX, Pepkowitz
S, Goldfinger D, Ng H, Irvin D, Yu JS (2008) Vaccination
elicits correlated immune and clinical responses in glioblastoma multiforme patients. Cancer Res 68:59555964
Wheeler CJ, Black KL, Liu G, Ying H, Yu JS, Zhang W, Lee
PK (2003) Thymic CD8(+) T cell production strongly influences tumor antigen recognition and age-dependent glioma
mortality. J Immunol 171:49274933
Yamanaka R (2009) Dendritic-cell- and peptide-based vaccination strategies for glioma. Neurosurg Rev 32:265273
Yamanaka R, Abe T, Yajima N, Tsuchiya N, Homma J,
Kobayashi T, Narita M, Takahashi M, Tanaka R (2003)
Vaccination of recurrent glioma patients with tumour lysatepulsed dendritic cells elicits immune responses: results of a
clinical phase I/II trial. Br J Cancer 89:11721179
Yamanaka R, Homma J, Yajima N, Tsuchiya N, Sano M,
Kobayashi T, Yoshida S, Abe T, Narita M, Takahashi M,
Tanaka R (2005) Clinical evaluation of dendritic cell vaccination for patients with recurrent glioma: results of a clinical
phase I/II trial. Clin Cancer Res 11:41604167
Yu JS, Liu G, Ying H, Yong WH, Black KL, Wheeler CJ (2004)
Vaccination with tumor lysate-pulsed dendritic cells elicits
antigen-specific, cytotoxic T-cells in patients with malignant
glioma. Cancer Res 64:49734979
Yu JS, Wheeler CJ, Zeltzer PM, Ying H, Finger DN, Lee
PK, Yong WH, Incardona F, Thompson RC, Riedinger MS,
Zhang W, Prins RM, Black KL (2001) Vaccination of malignant glioma patients with peptide-pulsed dendritic cells elicits systemic cytotoxicity and intracranial T-cell infiltration.
Zitvogel L, Tesniere A, Kroemer G (2006) Cancer despite
immunosurveillance: immunoselection and immunosubversion. Nat Rev Immunol 6:715727
Chapter 33
Glioblastoma Multiforme: Use of Adenoviral Vectors

Thomas Wirth, Haritha Samaranayake, and Seppo Yl-Herttuala
Abstract Glioblastoma multiforme is a tumour of

the brain with extremely poor prognosis. It constitutes
approximately 6070% of all primary brain tumours
with a mean survival of 14.6 months after diagnosis.
Current standard therapy includes surgical resection
followed by adjuvant radiotherapy and/or chemotherapy. With the introduction of gene therapy a new therapeutic area has been opened and numerous approaches
are currently been studied in pre-clinical as well as
in clinical settings for the treatment of brain tumours.
These include pro-drug activation/suicide gene therapy, anti-angiogenic gene therapy, oncolytic virotherapy, immune modulation and many other strategies,
which will be discussed in this review. The use of
suicide gene therapy has been probably one of the
most studied therapeutic approaches and will be also
discussed in this review in more detail.
Keywords Glioblastoma multiforme Adenoviral
vector Pro-drug activation Proliferation Invasion
Suicide gene therapy
Introduction
One of the most lethal malignancies are gliomas,
which constitute approximately 6070% of all primary brain tumours with an overall global incidence
T. Wirth ()
Department of Biotechnology and Molecular Medicine, A. I.
Virtanen Institute, University of Kuopio, FI-70211 Kuopio,
Finland
e-mail: Thomas.Wirth@uef.fi
of 46 new cases per 100,000 population per year.

Current standard therapy includes surgical resection
followed by adjuvant radiotherapy and/or chemotherapy. The overall prognosis of patients diagnosed with
glioblastoma multiforme (GBM) and treated with current therapies has remained poor for the last few
decades, with a mean survival of 14.6 months after
diagnosis and a 5-year survival of <3% for these
patients. The diffusely infiltrating property of GBM
makes total surgical resection often virtually impossible, leading invariable to tumour recurrence at some
point of time. The maximum radiation dose that can be
given to the brain is limited to 60 Gy, which is usually not sufficient to completely eradicate the tumour.
This has provided impetus to test novel therapy modalities to achieve better responses to improve the quality
of life and to increase the survival time of GBM
patients.
With the introduction of gene therapy a new therapeutic area has been started, and so far, several gene
therapy approaches have been studied against glioma.
By September 2008 there were a total of 91 gene
therapy clinical trials registered world wide, targeting brain cancer: 62 of them in phase I, 17 in phase
I/II, 7 in phase II, 1 in phase II/III and 4 in phase
III, of which one was a study using adenoviruses as
the gene transfer vector. In that study the suicide gene
Herpes simplex virus thymidine kinase is used to target dividing cells within the brain (i.e. tumour cells).
Since this is so far the most advanced gene therapy
study for GBM, where adenoviral vector is used, it
will be discussed in more detail. In addition to the
suicide gene therapy approach, other gene therapy
approaches where adenoviral vectors are used will be
highlighted.

335
336
Brain Tumours
The brain is an important site in oncology. Many
malignancies originating from organs such as lung,
breast, colon, skin, kidney and thyroid can metastasize to the brain resulting in secondary brain tumours.
At the same time many different cell types in the
brain can ascertain a malignant phenotype giving rise
to primary brain tumours. Primary brain tumours are
classified based on histological appearance, resemblance of the tumour cells to embryonic, foetal or
differentiated mature cells, growth pattern, radiological features and surgical appearance. In the majority
of brain tumours the cell of origin is unknown due
to unavailability of identifiable pre-malignant lesions
and cellular atypia preventing the comparison with
any normal cells. Primary brain tumours can be classified into three groups: (1) neuroepithelial tumours
which include astrocytic, oligodendroglial, ependymal, choroid plexus, neuronal and pineal parenchymal tumours, (2) non-neuroepithelial tumours such as
meningeomas, nerve sheath tumours, malignant lymphoma, pituitary adenoma and germ cell tumours, and
(3) other tumours including tumours of unknown origin. Approximately 6070% of all primary intracranial
tumours are of neuroepithelial origin. About 30% are
derived from brain coverings (meningiomas) and 7.5%
are located in cranial or spinal nerves, whereas lymphomas and germ cell tumours account for 4 and 1%
respectively. The most common primary brain tumours
in adults are gliomas, which account for 6070% of
primary tumors.
In the WHO tumour grading system gliomas are
graded into four grades of malignancy from grade I
to IV based on morphological characteristics, such
as nuclear atypia, mitotic figures, microvascular proliferation and focal pseudopalisading necrosis, with
increasing biological aggressiveness with increasing tumour grade. Tumours with malignant grade
I and II are termed low grade whereas those with
grade III and IV are termed high grade tumours.
Tumours of astrocytic origin are graded in order
of increasing anaplasia into pilocytic astrocytoma
(WHO grade I), low grade diffuse astrocytoma (WHO
grade II), anaplastic astrocytoma (AA) (WHO grade
T. Wirth et al.
III) and the most frequent and most malignant subtype glioblastoma multiforme (GBM) (WHO grade
IV), which accounts for 6070% of all gliomas
(Ohgaki and Kleihues, 2005). Oligodendrogliomas and
mixed oligodendroastrocytomas are classifies into two
grades; low grade (grade II) and high grade (grade III).
Diffuse astrocytic tumours including low grade diffuse astrocytomas, AA and GBM, oligodendrogliomas
and meningiomas in adults and pilocytic astrocytomas, ependymomas and medulloblastomas in children, are in general the most frequent primary brain
tumours.
Current Therapy
Current standard therapy of GBM includes surgical resection followed by adjuvant radiotherapy and
chemotherapy (i.e. temozolomide). In the latest study
presented by Stupp and his colleagues the overall mean
survival with this therapy was shown to be 14.6 months
after diagnosis with a 5-year survival of 9.8% (Stupp
et al., 2009). More recently, in May 2009 the FDA
approved Avastin (bevacizumab) for GBM under an
accelerated approval process. The approval was based
on the results of 2 phase II clinical trials that showed
Avastin reduced tumour size in some GBM patients.
In addition, another recently published phase II trial
of single-agent bevacizumab followed by bevacizumab
plus irinotecan showed significant therapeutic activity
in patients with recurrent glioblastoma.
In general, the challenge in the treatment of malignant gliomas lies in the nature of how they grow.
Gliomas are characterized by an extensive, diffuse
infiltration of the tumour cells into the surrounding brain parenchyma. Partly because of their growth
pattern, complete surgical resection is generally not
possible. The consequence is an inevitable tumour
recurrence at some point. In addition, some tumours
due to their anatomical location are either surgically
inaccessible or have limited likelihood of near-total
surgical resection.
The maximum radiation dose that can be given
to the brain is limited to 60 Gy, which is often
not sufficient, as brain tumours are generally resistant to radiotherapy. Studies have shown that doses
above 60 Gy would not give any additional therapeutic benefit, because of the increase in adverse effects,
such as radiation induced necrosis of normal brain.
33 Glioblastoma Multiforme: Use of Adenoviral Vectors
Treatment of gliomas with systemic chemotherapy

is challenging, since the blood brain barrier (BBB)
impairs the delivery of adequate concentrations of
most chemotherapeutic compounds. The reasons for
not achieving adequate concentrations is because the
BBB lacks intercellular fenestrations, has high electrical resistance and is not permeable to ionic compounds
rendering it relatively impermeable to most water soluble compounds.
History of Gene Therapy

Scientific understanding of the molecular basis of life
increased dramatically after Oswald T. Averys discovery in 1944 that deoxyribonucleic acid (DNA) was the
transforming principle the secret code of life. Then
Francis Crick and James Watson described the double
helix structure of DNA in 1953. The process, however, by which DNA replicates itself during cellular
reproduction, or how DNA expresses its genetic information, was still a mystery in the late 1950s. A little
less than 20 years after Oswald T. Averys discovery
Marshall Nirenberg and his colleagues in 1962 deciphered UUU (one three-unit batch of uracil, which was
a code word for identifying phenylalanine) as the
first word in the chemical dictionary of life and thereby
set a pillar for gene therapy.
Gene therapy is generally defined as the transfer
of nucleic acids to somatic cells of a patient with a
therapeutic intent. According to the EMA, gene therapy medicinal product means a biological medicinal
product that contains an active substance, which contains or consists of a recombinant nucleic acid used in
or administered to human beings with a view regulating, replacing, adding or deleting a genetic sequence,
as well as if its therapeutic, prophylactic or diagnostic
effect relates to the recombinant nucleic acid sequence
it contains, or to the products of genetic expression of
this sequence. As compared to traditional medicine,
gene therapy offers unique possibilities to treat genetic
causes of diseases, as an example can be mentioned
diseases like fatal enzyme deficiencies.
In the US, the FDA approved the first gene therapy clinical protocol, which was initiated in 1989.
In the protocol, tumour-infiltrating lymphocytes were
collected from patients with advanced melanoma, and
transduced ex vivo with a marker gene, expanded in
vitro and then re-infused into the patients. A year
337
later, the first gene therapy with a therapeutic intent

was conducted on the 4-year old Ashanti de Silva, on
September 14, 1990, at the NIH Clinical Centre. The
patient had adenosine deaminase (ADA) deficiency, a
genetic disease making her defenseless against infections. Unfortunately, the efficacy of Ashantis gene
therapy was not clearly demonstrated due to simultaneous enzyme replacement therapy with polyethylene
glycol adenine deaminase (PEG-ADA), which she had
to take as a back up. In 2009 Kohn and Candotti
compared the transplantation of hematopoietic stem
cells with gene therapy, concluding that gene therapy
has been at least as successful, if not better, than the
transplantation of hematopoietic stem cells (Kohn and
Candotti, 2009). However, before that, in 1999 a tragic
incident happened, and it was ultimately after the death
of Jesse Gelsinger a patient who was treated for a disorder called partial ornithine transcarbamylase (OTC)
deficiency by adenoviral gene therapy and in whom
death could be directly linked to the viral vector
concerns of the safety of gene therapy were raised.
Another incident also raised doubts about the safety of
gene therapy. Eleven young children suffering from the
fatal X-linked SCID-XI syndrome were treated with
gene therapy. However, after the initial success 3 out
of 11 treated patients had developed a leukaemia-like
disease as a result of the use of the murine leukaemia
virus (MLV) vector.
Today, there are more than 1400 ongoing gene
therapy clinical trials worldwide, most of which are
targeted against cancer, and currently two approved
gene therapy-based medicines are approved in China,
Gendicine and H-101 (Oncorine), both indicated for
head and neck cancer. Gendicine, approved in 2003
by the Chinese State FDA, is a p53 adenovirus (Ad)
for the treatment of head and neck squamous cell cancer in combination with radiotherapy. H-101 (which
is similar to Onyx-015) is a conditionally replicative
Ad that gained marketing approval in China in 2006.
Along with this progress, gene therapy has evolved as
a new therapeutic option for the treatment of GBM.
Ideally, a gene therapy vector would target a specific
tissue with high transduction efficiency and sustain
a stable, regulated gene expression without any side
effects or immunogenic responses. Currently, none
of the most commonly used gene delivery vectors
fulfil all these criteria. Nevertheless, the majority of
gene therapy clinical trials have been performed using
viral vectors, reflecting their superiority over non-viral
systems.
338
Adenoviral Vector
Adenoviral vectors (Ads) have been widely utilized
for gene therapy studies, both, in preclinical and clinical settings. To date, 52 serotypes of human Ads
have been found and they have been classified into
seven species (human adenovirus A to G). Ads are
non-enveloped, icosahedral viruses of 6090 nm in
diameter. They have a linear double stranded DNA
genome of 3040 kb and as gene transfer vectors
they can carry relatively large fragments of foreign
DNA. The main components of Ads are the capsid
and the core. The main component of the capsid is
hexon. The homotrimeric hexon capsomers form the
20 triangular faces of the icosahedron. In each of the
12 vertices, there is a penton capsomere composed
of penton base and fiber proteins. The latter forms
a spike-shaped protrusion with a terminal globular
domain or knob that is responsible for the attachement of the virus to its primary receptor. In addition to
hexon and penton capsomers, there are several hexonassociated proteins that have capsid-stabilizing functions. The double stranded DNA genome is located
in the virion core with DNA associated proteins and
terminal proteins that serves as a primer for DNA
replication.
Human Ads are known to cause a variety of
clinical symptoms, depending on the serotype in
question. These include upper and lower respiratory
tract infections, gastroenteritis, keratoconjunctivitis,
acute hemorrhagic cystitis, meningoencephalitis and
hepatitis.
The coxsackie- and adenovirus receptor (CAR) is
the primary receptor for human Ads. The knob domain
of the homotrimeric fiber binds to CAR, mediating
virus attachment to the cell. Subsequently, an RGDmotif located in the penton base interacts with a cell
surface integrin molecule that plays a role as a secondary or internalization receptor and triggers the
virus entry via clathrin-dependent receptor mediated
endocytosis.
In addition to the CAR receptor other receptors,
such as class I major histocompatibility complex
(MHC), heparan sulphate glycosaminoglycans and
vascular cell adhesion molecule-1 have been shown
to play a role in cell entry. Ads transduce both dividing and quiescent cells and they provide an efficient, but transient, transgene expression. They exist
T. Wirth et al.
extrachromosomally within the cell, although the DNA

migrates into the nucleus. Because of their broad host
and cell range adenoviral vectors have been extensively
used in experimental models, as well as in clinical protocols. Even though there is not much data about the
long-term aspects regarding the safety of adenoviral
vectors in humans, at least one meta-analysis showed
that adenoviruses have demonstrated adequate safety
as gene transfer vectors in humans (Wirth et al., 2006).
Most adenoviral vectors used in gene therapy are
based on serotype 2 and 5, since these serotypes are
associated only with mild or asymptomatic infections.
In the first generation of adenoviral vectors, the region
encoding for early viral protein E1 is replaced by
the transgene making the vectors replication defective. By deleting the E1 gene, a transgene capacity
of 4.74.9 kb was achieved, which was then further increased to 8.3 kb by deleting the non-essential
region E3. To circumvent the problem of the host
cellular immune response against the viral proteins
produced by other early viral genes, a second generation vector was described, where the E4 region was
deleted. This second generation of adenoviral vectors
was shown to be safer in terms of not causing as much
apoptosis and inducing to a lesser extent an immune
response in vivo. The third generation adenovirual vectors were constructed by deleting other adenovirus
genes, resulting in a so called high-capacity, gutless or
helper-dependent Ads vectors.
Gene Therapy of Brain Tumours

Brain tumours are in most cases single, localized
lesions of rapidly dividing cells in a background of
non-dividing cells. They rarely metastasize outside
CNS and recurrences occur typically in the close
vicinity of the original lesion, making them highly
amenable for gene therapy. In gene therapy, the vectors can generally be classified into viral and nonviral vectors, where the viral vectors are considered
to be currently the most effective ones. Commonly
used viral vectors for brain tumour gene therapy
include retroviruses, herpes simplex viruses, Ads and
adeno-associated viruses (AAV). Apart from these
Baculoviruses, New Castles Disease viruses (NDV),
polio viruses, Semliki Forest viruses, measles viruses,
lentiviruses, reoviruses, and vaccinia viruses have been
used. In most of these vectors the viral genome is

genetically modified to make them replication defective and non-pathogenic. However, in the oncolytic
viruses the genome is modified to give them the ability
to specifically replicate only in tumour cells, thereby
lysing these cells. Some of the new viral vectors have
the ability to replicate only once in the target cells,
without the ability to disseminate, so that the transgene
copy number and expression within the transduced
cells increases.
For gene therapy of brain tumours, several
approaches have been studied. These include pro-drug
activation/suicide gene therapy, anti-angiogenic gene
therapy, oncolytic virotherapy, immune modulation,
correction of gene defects, inhibition of tumour invasion, apoptosis induction, gene therapy to enhance
chemo- and radiotherapy, myeloprotective gene therapy, antisense and RNA interference (RNAi) based
strategies. Although most of these methods have
demonstrated promising success in in vitro and preclinical studies, only a few of these approaches have
progressed up to phase III clinical trials. The very
first clinical gene therapy trials against brain tumours
have been registered in 1992. In that trial autologous
tumour cells were modified ex vivo with retrovirus
to express interleukin-2 gene in neuroblastoma. In the
following year brain tumour patients were treated with
herpes simplex virus thymidine kinase (HSV-tk) suicide gene therapy using retrovirus vectors producing
cells (VPCs) and concomitant administration of ganciclovir (GCV). By September 2008 there were a total of
91, gene therapy clinical trials registered world wide,
targeting brain tumours: 62 of them in phase I, 17
in phase I/II, 7 in phase II, 1 in phase II/III and 4
in phase III. In 14 out of 62 trials adenoviral vectors
were used and only one of them, so far, has reached
and completed phase III. That study was a randomized
multicentre phase III trial using HSV-tk adenoviral
vectors for the treatment of GBM.
Adenoviruses as a Gene Transfer Vector

in Malignant Glioma Gene Therapy
Ads vectors are next to retroviral vectors as the most
commonly used gene transfer vectors for the delivery of therapeutic genes into GBM. They have been
used in a variety of different gene therapy strategies,
339
including suicide gene therapy, anti-angiogenic gene

therapy, oncoviral therapy, gene therapy based on
immune modulation, gene therapy targeting oncogenes
or tumour suppressor genes and gene therapy targeting tumour invasion and apoptosis. Also, gene delivery
of antisense or siRNAs targeted to growth factors, to
increase the sensitivity of glioma cells to chemotherapeutics or to inhibit migration and invasion of glioma
cells, are more recent approaches that have been
developed.
Anti-angiogenic Therapy
Angiogenesis is an integral part in tumour development. As the tumour growth continues, new
blood vessels need to be developed to meet with
the new metabolic demands, and to sustain the
growth. Angiogenesis was initially described by Judah
Folkman as a complex process, tightly regulated by
coordinated expression of a variety of stimulating and
inhibiting molecules, which involves rapid proliferation of endothelial cells, resulting in new blood vessel
formation. In tumours VEGF is secreted by tumour
cells and stromal cells.
GBM is a highly vascular tumour with profound
neovascularisation, with angiogenesis playing a key
role in its development and progression. Therefore,
gene therapy approaches inhibiting angiogenesis has
naturally become a logical and widely experimented
area also for the treatment of malignant gliomas. Both,
VEGFR-1 (Flt-1) and VGFR-2 are shown to be overexpressed in high grade gliomas, of which VEFGR-2
seems to play a significant role in promoting angiogenesis (Stratmann et al., 1997). However, the significance
of VEGFR-1 (Flt-1) is yet to be elucidated.
Apart from VEGF, other angiogenic factors such as
PDGF, hepatocyte growth/scatter factor, FGF, uPAR,
and cathepsin B are frequently over expressed in GBM.
An artificial transcription factor based on Cys2His2 zinc-finger proteins (ZFPs) targeting the VEGF
promoter has been used by Kang et al. as a strategy for modulating VEGF levels in gliomas, eliciting
a pronounced antitumour effect in a human glioblastoma xenograft model (Kang et al., 2008). He showed
also that in vivo inoculation with adenoviral vectors encoding for brain-specific angiogenesis inhibitor
1 (BAI1) resulted in antiangiogenic and antitumour
340
effects of established subcutaneous or intracerebral

tumours, and that it promoted increased survival in
mice (Kang et al., 2006). Adenoviral vector containing the cDNA of wild-type p16 and antisense RNA
of uPAR were used to transduce the U251 glioblastoma and a glioblastoma xenograft cell line, resulting
in a significant inhibition of human mammary epithelial cell capillary formation and VEGF expression
(Nalabothula et al., 2007). Also, retargeting of Ads
carrying the gene for HSV/tk to endothelial cells of
the tumour vasculature by using the fibroblast growth
factor ligand was used as an antiangiogenesis strategy
(Gupta et al., 2006). Insertion of tissue inhibitor of
metalloproteinase-3 (TIMP-3) cDNA in the oncolytic
adenovirus-Delta-24-RGD by Lamfers and colleagues
has not shown to improve the therapeutic outcome
over the parent oncolytic virus without the transgene
(Lamfers et al., 2005). Another group demonstrated
that inhibition of matrix metalloproteinase-9 by using
Ads resulted in reduced in vitro invasion and angiogenesis of human microvascular endothelial cells (Jadhav
et al., 2004). The same was shown also by two other
groups, where ribozymes targeted to the mRNA of
VEGF were used (Ciafre et al., 2004). The multifunctional growth factor scatter factor/hepatocyte growth
factor (SF/HGF) and its receptor c-met have been
implicated in the tumorigenesis, malignant progression, and chemo/radioresistance of multiple human
malignancies, including gliomas. Adenovirus expressing anti-SF/HGF and anti-c-met U1snRNA/ribozymes
substantially inhibited tumour growth and promoted
animal survival in an orthotopic glioma model, as well
as in a subcutaneous mouse model using the glioma
cell line U87MG (Kargiotis et al., 2008). Also, adenoviral vector mediated gene transfer of the tumour
necrosis factor- (TNF-) was shown to improve the
antiglioma efficacy with concomitant radiation and
temozolomide therapy (Yamini et al., 2007).
T. Wirth et al.
viral proteins. Since the function of the E1b in the

wild type virus is to prevent p53 induced apoptosis of
viral infected cells, it was thought that E1b mutated
viruses would replicate only in p53 deficient tumour
cells, eventually leading to tumour cell death. After
some promising pre-clinical studies a phase I clinical
trial was conducted, showing that it was well tolerated without serious adverse effects (Chiocca et al.,
2004). However, no definitive anti-tumour efficacy
was observed in that trial. Later, several other genetically modfied oncolytic viruses have been studied.
Adenovirus-Delta-24-RGD oncolytic adenovirus has
a 24 base pair deletion within the E1A early viral
protein gene that renders the virus replication ability
only in cells with defective retinoblastoma pathway,
which is commonly altered in malignant gliomas, independent of the host cell p53 status. This oncolytic
adenovirus selectively replicates in malignant cells
with a defective p16/Rb/E2F signalling circuitry. One
phase I clinical trial has been registered so far using
adenovirus-delta-24-RGD for glioma gene therapy.
A variant of this vector, adenovirus-delta-24-hyCD
was designed to express a humanised form of the
Saccharomyces cerevisiae cytosine deaminase gene
(hyCD). It has shown better efficacy in glioma models
than the adenovirus-Delta-24-RGD, when combined
with 5-FC. Also, due to their tumour tropic properties,
human mesenchymal stem cells were used as intravascular delivery vehicles of Delta24-RGD to human
gliomas. Other mutant adenoviral vectors, with mutations in both E1A and E1B genes, demonstrated a
potent anti-tumour effect in a glioma xenograft model.
Additionally, a number of CRADs, based on tumour
selective expression of genes required for viral replication under the control of a tumour or tissue specific
promoter, have been constructed. Also, CRADs carrying an additional transgene, such as p53 or HSV-tk,
have been developed in order to potentiate the therapeutic effect of these vectors.
Oncolytic Viruses
Immune Modulation
A very interesting therapeutic approach that has
been edexploited is the natural capacity of adenoviral vectors to induce cell lysis, often termed as
oncolytic viruses or conditionally replicative adenoviruses (CRADs). The first genetically modified
oncolytic adenovirus was ONYX-015, which carries
a mutation in the E1b region that encodes for early
Like most other malignancies, gliomas are known

to suppress the host immune system. The secretion of immunosuppressive substances by the tumour,
such as transforming growth factor- (TGF-) and
interleukin-10 (IL-10), that decreases the expression of MHC molecules and FasL are some of the
tumour factors contributing to immunosuppression and

impaired anti-tumour immune response by the host.
With the aim of enhancing cell mediated immunity,
both, adoptive and active (vaccines) immunotherapy
strategies have been attempted in GBM with varying
degrees of success.
Ads expressing fms-like tyrosine kinase-3 ligand
(Flt3L), heat shock protein 72 and the activating
immunoreceptor NKG2D have been used to induce
an immune response against the tumour and to stimulate the infiltration of immune cells into the tumour
microenvironment. Ads Flt3L have been used in combination with adenovirus HSV-tk gene therapy by
several groups demonstrating an induction of cellular and humoral immune responses, as well as an
immunological memory against glioblastoma (King
et al., 2008; Ghulam Muhammad et al., 2009). Also,
Ads gene transfer encoding for the wild type p53, the
granulocyte monocyte-colony stimulating factor (GMCSF) and the co-stimulatory signal B7 have been used,
resulting in inhibition of cell growth and the induction
of apoptosis, as well as in a significant proliferation of
autologous peripheral blood lymphocytes and specific
cytotoxicity against primary glioma cells (Pan et al.,
2009). In addition, adenovirally transduced bone marrow stromal cells transduced with interleukin-12 were
used in one study, resulting in an increase in natural killer cell infiltration in brain tissue. Significant
(Hong et al., 2009) Ads gene transfer of a secreted
trimeric form of the TNF-related apoptosis-inducing
ligand (TRAIL) has shown efficacy when injected
intra-tumourally in a xenograft model of a human
glioma cell line U-87MG, and when injected into intracranial human glioma tumours established in nude
mice (Kim et al., 2006; Jeong et al., 2009).
Targeting Tumour Proliferation

and Invasion
Tumourigenesis/carcinogenesis is a multi-step process with mutations in many genes, including protooncogenes, tumour suppressor genes and genes that
promote cell cycle progression, growth factor independence, angiogenesis, enhanced motility, decreased
cell-to-cell cohesion, decreased levels of apoptosis,
and reduced chemosensitivity. The development of
gliomas is commonly associated with mutations in
four genetic pathways: the p53/ARF/human MDM2
341
pathway, the p16/cyclin D/CDK4 pathway, the receptor tyrosine kinase (RTK)/Ras pathway, and the
PI3K/PTEN/Akt pathway. Correction of these genetic
defects by delivery of the wild type gene through a
vector seems to be a logical approach for cancer gene
therapy.
The tumour suppressor gene p53 has been one of
the most common targets in cancer gene therapy. p53
is mutated or deleted in more than 50% of human cancers, including gliomas, and has been for that reason
also a target in glioma gene therapy. p53 maintains
the genetic integrity after DNA damage and functions as a gatekeeper of cellular growth, apoptosis and
senescence. p53 normally exists in cells in very low
amounts and is relatively inactive, but the amount is
increased and protein is activated during cellular stress.
Depending on the cellular environment, activation of
the protein leads to either cell cycle arrest or apoptosis. A variety of different p53 gene correction therapies
have been attempted showing promising results in preclinical studies, such as reduced tumour proliferation
in nude mice, reduction in tumour volume in rats,
increased survival, chemo and radio-sensitisation, inhibition of angiogenesis and induction of apoptosis. As
a result, the first gene therapy product was approved
for clinical use in China in 2003 (GendicineTM ).
GendicineTM is based on an adenovirus encoding for
p53 and indicated for the treatment of head- and neck
squamous cell cancer. However, after the approval by
the Chinese SFDA, there has been discussion about
the efficacy of GendicineTM treatment, since the FDA
refused to accept an approval application for a similar product: AdvexinTM (INGN 201; Introgen; Austin,
TX, USA) for the treatment of head and neck cancer.
Another gene of interest has been the tumour suppressor gene retinoblastoma (Rb). In normal cells
the retinoblastoma (Rb) gene is in a hyperphosphorylated form, and bound to the transcription factor
E2F1, which prevents transcription of genes important
for mitosis and progression of cell cycle through the
G1/S phase restriction point. P16/Rb/cyclinD/CDK4
is the most commonly mutated pathway in gliomas,
and is associated with the progression from low-grade
to intermediate-grade gliomas. Enforced expression
of MDA-7/IL-24, by use of a recombinant adenovirus, was shown by (Hamed et al., 2010) to have a
potent cancer cellspecific apoptosis-inducing ability
and tumour growthsuppressing properties in an orthotopic rat glioma model.
342
In addition to uncontrolled cell proliferation of

tumour cells, another target and challenge in glioma
therapy is the extensive invasion of the surrounding normal brain parenchyma by the glioma cells.
Modification of the extracellular matrix by proteases
and protease inhibitors, cytoskeletal reorganization,
growth factors, cytokines and intracellular signalling
molecules are the key players involved in this process. A study showed that gene transfer of tissue
inhibitors of matrix metalloproteinases (TIMPs) concomitant with PTEN inhibited the invasion of glioma
cells, mainly by inhibiting the secretion of matrix metalloproteinase 2 (MMP-2) (Lu et al., 2004). Another
way to block invasion by using adenoviral vectors was
shown in a study by Furakawa and his colleagues,
showing that PTEN was able to suppress invasion in
an organotypic brain slice model, co-culture of glioma
spheroids and rat brain slices, where adenoviral vector
transduced cells failed to invade surrounding normal
brain tissues (Furukawa et al., 2006). Urokinase-type
plasminogen activator (uPA) and its receptor (uPAR)
have been shown to play an important role in the
invasiveness of gliomas and other infiltrative tumours.
Downregulation of uPAR, uPA or MMP-9 by antisense oligonucleotides has been shown to inhibit the
invasiveness and tumourigenicity of glioma cells in an
orthotopic animal model, as well as caused regression
of glioma in a subcutaneous glioma model using the
human glioma cell line U87MG (Lakka et al., 2003;
Gondi et al., 2003).
Suicide Gene Therapy

Suicide gene therapies can generally be divided
into two steps. In the first step, a gene of a foreign
enzyme is delivered (e.g. by the use of a viral vector) to the tumour where it is to be expressed. In
the second step, a non-active pro-drug is administered, after which it is selectively metabolized to
the active form by the foreign enzyme expressed
in the tumour (Fig. 33.1). Ideally, the gene for the
enzyme should be expressed exclusively in the tumour
cells and should reach a concentration sufficient to
activate the pro-drug for a clinical benefit. Several
suicide genes have been studied with varying results.
HSV-tk, Cytosine deaminase (CD)/5-fluorocytosine
(5-FC), cytochrome P450/cyclophosphamide (CPA),
E.coli purine nucleoside phosphorylase (PNP)/
T. Wirth et al.
6-methyl-purine-2 -deoxynucleoside, and carboxypeptidase G2 (CPG2)/methotrexate--phenylalanine are

some of the pro-drug activation systems that have
been attempted on brain tumour treatment. Because
the expression of the foreign enzyme will not occur in
all cells of the targeted tumour in vivo, a bystander
effect is required, whereby the pro-drug is cleaved to
an active drug that kills not only the tumour cells in
which it is formed, but also neighboring tumour cells
that do not express the transgene.
Of all suicide gene therapy apporaches, the HSV-tk
suicide gene therapy has been the most widely studied suicide gene therapy approach for the treatment of
brain tumours. It is currently also the only adenoviral based gene therapy strategy for the treatment of
glioblastomas having finished a phase III clinical trial.
HSV-TK/Ganciclovir Therapy
HSV-tk/GCV therapy is based on the pro-drug activating enzyme HSV-tk that converts the nucleotide
analogue GCV to its toxic metabolite. Apart from
malignant glioma, HSV-tk/GCV has been studied also
in many other cancer types, both in pre-clinical models
as well as in clinical trials.
Originally, the HSV-tk gene was cloned byMcKnight in 1980 (McKnight. 1980). Their property to kill
tumour cells, when given GCV, was soon realized and
led to the idea of using this as a therapeutic strategy
to treat solid tumours. The first proof-of-concept of
the therapeutic efficacy of HSV-tk/GCV therapy was
established ten years later by Moolten and Wells, when
they used retroviral vectors expressing the HSV-tk
gene to transduce murine sarcoma and lymphoma
cells in vivo. In the initial studies murine fibroblasts
producing HSV-tk retroviral vectors were inoculated
intratumourally, followed by GCV treatment.
Ganciclovir
Mechanism of Action
Ganciclovir is a synthetic acyclic analogue of
2 -deoxy-guanosine, chemically designated as 9[[2-hydroxy-1-(hydroxymethyl)ethoxy]methyl]guanine.
Originally, GCV is known as being the first antiviral
drug to be effective in the treatment of cytomegalovirus
343
Fig. 33.1 A schematic representation of suicide gene therapy.

In the first step, the gene for an exogenous enzyme is delivered and expressed in the target cells. Subsequently, a pro-drug
is administered that is converted to the active drug by the foreign

enzyme in the target cells. The active drug will then cause the
therapeutic effect in the malignant cells
(CMV) diseases in humans. GCV has antiviral activity

against all members of the human herpes virus family.
It is more toxic than its congener, acyclovir, but has
a 820 fold higher antiviral activity against human
CMV than acyclovir, which is generally ineffective
against diseases caused by CMV.
GCV is converted to GCV-monophosphate by the
HSV-tk, which has a 1000 fold higher affinity for
GCV than the mammalian form of the enzyme. Cell
kinases continue the phosphorylation further and convert the GCV-monophosphate into a GCV diphosphate
and finally to a toxic GCV triphosphate. Cytotoxic
GCV-triphosphate results in the inhibition of the DNA
polymerase, thus preventing DNA replication. HSVtk/GCV therapy is cell cycle dependent, where only
dividing cells will be affected. This property is considered to be an advantage in cancer therapy where
dividing tumour cells are surrounded by non-dividing,
healthy cells.
An additional mechanism of action that has been
connected to HSV-tk gene therapy is the expression
of the non-human enzyme HSV-tk and the presence of
tumour antigens that become available after the death
of the transduced cells. These antigens are able to
induce an antitumoural immune response against the
tumour cells. It is believed that both local and systemic
immune responses contribute to the overall therapeutic effect of HSV-tk/GCV by stimulating the tumour
infiltration of CD4+ and CD8+ T cells, natural killer
(NK) cells and macrophages into the tumour. Work

by different groups suggests that GCV induced cell
killing occurs due to apoptosis induced by the activation of mitochondrial damage pathways, chromosomal
aberrations, and sister chromatin exchange that leads
to target cell death due to cell cycle arrest.
The Bystander Effect

As current gene therapy vectors cannot achieve 100%
gene transfer efficacy, the bystander effect is essential for an effective anti-tumour therapy. Because of
that, a very important advantage of HSV-tk/GCV gene
therapy is its bystander effect, where neighbouring
non-transduced cells are also exposed to the toxic
GCV-triphosphate form. Interestingly, it was noticed
that also HSV-tk-negative cells were killed after exposure to GCV. Moreover, it was found that even with low
concentration of HSV-tk-positive tumour cells, tumour
growth could be prevented. Several experiments in
murine models suggest that 10% of the tumour cells
have to be transduced with HSV-tk gene in order
to achieve a significant tumour regression. However,
this bystander effect seems to be dependent on the
cell type. The exact mechanism of the bystander
effect is not completely understood, but several mechanisms have been postulated. For example, it has been
observed by that the number of gap junctions plays a
344
vital role in this effect. Additionally, it was demonstrated that the number of gap junctions seems to
correlate with the efficiency of killing neighbouring
cells. There is experimental evidence that the release
of toxic GCV-triphosphate into culture medium by
dying cells results in apoptosis of non-transduced cells.
This was shown to be dependent on GCV-triphosphate
accumulation in the surrounding cells and the concentration of GCV in the medium, suggesting that cell-to
cell contacts are not essential. Other possible mechanisms that could be responsible for the bystander effect
include stimulation of the immune system, endothelial cell transduction leading to disruption of tumour
vasculature and phagocytosis of apoptotic vesicles by
neighbouring non-transduced cells.
Clinical Efficacy
So far, the only adenoviral vector that has finished a
R
phase III clinical trial (Cerepro
, by Ark Therapeutics
Group plc) is based on the suicide gene therapy HSVtk. In that trial Ads vector encoding for HSV-tk was
injected into the walls of the tumour cavity of glioma
patients after the resection of the tumour (Fig. 33.2).
R
The clinical efficacy of Cerepro
was evaluated first
in two separate phase II clinical trials; a phase IIa trial,
and a phase IIb trial (Sandmair et al., 2000) (Immonen
et al., 2004). In the phase IIa study Sandmair compared
the efficacy of retrovirus-packaging cells (VPCs) and
R
Cerepro
. In that study efficacy was shown already
3 months after the resection and gene transfer, when
R
R
Cerepro
was used. Cerepro
showed a mean survival advantage of 15 months, which was significantly
(p < 0.001) longer than with the control group (8.3
months) or the VPC-group (7.4 months).
This was the first controlled trial to show a significant survival advantage using HSV-tk gene therapy. Although the VPC approaches have been found
safe, no efficacy has been observed in GBM patients
(Rainov, 2000). The low gene transfer efficacy with
retrovirus and lack of the treatment response in a study
by Sandmair et al. indicated that retroviral HSV-tk
gene therapy may not be efficient enough in human
clinical settings. This was also further confirmed by
two independent clinical trials who reported results
from the first randomized, open-label, parallel group
phase III trial of 248 patients, where HSV-tk produced
T. Wirth et al.
retroviral producing cells did not result in an improvement of survival. However, the randomized and controlled phase IIb trial published by Immonen, carried
out in 36 patients, seventeen patients with operable or recurrent malignant gliomas receiving HSV-tk
R
adenoviral vector (Cerepro
), implicated a survival
advantage over control patients, who did not receive
R
Cerepro
. The mean survival of the patients in
R
the Cerepro
-group (70.6 weeks) was significantly
(p < 0.0095) longer when compared to the standard
care group (39.0 weeks) or a historical control group
(P < 0.0017). This study was also historically the first
randomized, controlled study done with an adenoviral
vector using HSV-tk, where increased survival of the
patients was shown when compared to standard therapy. The results from the study were very encouraging
R
and it was concluded that Cerepro
could provide an
effective adjuvant treatment for patients with operable
primary or recurrent malignant glioma and therefore
a multicentre, standard care controlled, randomized
clinical phase III trial was commenced. However, the
results coming out of that trial were not as significant
as those from the previous IIb trial. As a result, suggestions by the European Medicines Agency were given,
after which Ark Therapeutics withdrew their market
application to enrol additional patients to the trial.
Safety of Adenoviral Vectors

The safety data of adenoviral mediated gene therapy
collected from different human trials have uniformly
been satisfactory. A report shows that the tolerability towards adenoviral vectors is acceptable and that
side effects have mostly been mild without any serious
adverse events related to gene therapy (Wirth et al.,
2006). Generally, the main concerns in cancer gene
therapy of GBM are the safety issues when using
viruses to deliver the therapeutic genes into the brain.
Especially in the case of adenoviruses the inflammatory reactions that might arise in the brain are of
concern. In other clinical HSV-tk studies with retroviruses and adenoviruses there have been evidence of
increased brain edema, epileptic seizures and haemorrhages. However, these complications have not been
related directly to viral mediated gene therapy, but
more likely to the advanced disease and the surgical
procedures.
345
Fig. 33.2 The adenoviral vector based suicide gene therapy

R
by Ark Therapeutics is injected into the wall of
Cerepro
the tumour cavity of glioma patients, after the resection of
the tumour. Mainly healthy cells will be transduced producing
the HSV-tk and converting the pro-drug Ganciclovir into

Ganciclovir monophosphate. The actual cytotoxic metabolite
Ganciclovir triphosphate is killing possible residual proliferating
tumour cells
The first study evaluating the safety and efficacy

R
of Cerepro
was also the first completed recombinant adenovirus mediated HSV-tk gene therapy trial
against MG in humans. In that study serological assessment, i.e. routine blood and urine analysis, was done
as well as tests for the detection of anti-adenoviral
antibodies were performed in order to evaluate the
safety of HSV-tk therapy. The study showed that the
therapy was well tolerated and no major alterations
in routine laboratory tests were detected. No systemic
escape of the virus was detected from urine or plasma
samples with PCR. Anti-adenovirus antibodies were
measured before and 2 weeks after the gene transfer.
Some increases (4 fold) in four patients were detected,
two of the patients having also a short-term transient
fever reaction. No severe adverse events related to
viral mediated gene therapy were detected in either
of the treatment groups. The most serious adverse
events were some elevation in the frequency of epileptic seizures (2 patients), hemiparesis and aphasia (1
patient). Nevertheless, it could not be excluded that
these adverse events were related to the tumour resection itself and the mechanical irritation of the operation
procedure.
R
The phase IIb study of Cerepro
showed also a
good safety profile with no signs of significant safety
concerns. Three patients had transient reversible elevation of liver enzymes; two out of 17 patients in the
R
Cerepro
group developed localized post-opertaive
intracerebral oedema. Adenovirus was detected by
PCR in the plasma of two patients 3 days after gene

therapy, but not thereafter. A similar good safety profile
was seen also in the phase III trial.
It has now been almost a decade since the first
successful gene therapy trial using an adenoviral vector for gene therapy of GBM. Altogether 344 patients
of whom 194 have received adenoviral vector have
been participants in the AdHSV-tk/GCV clinical trials,
no treatment related deaths have occurred and only a
few serious adverse effects related to the gene therapy
has been observed. On the basis of these findings the
therapy is very safe and well tolerated.
Conclusion
Currently, most of the gene therapy strategies used
are limited to the local administration of the gene
delivery vector, or to ex vivo gene transfer approach.
For that, GBM represent an attractive target for gene
therapy because of its restricted anatomical location
and absence of metastases outside the CNS. However,
as far as it comes to local gene therapy, the greatest shortcoming in this approach appears to be the
low transduction efficiency of the gene transfer vector and its minimal distribution from the injection site.
To some extent the low transduction efficiency can be
regarded as a methodological problem, but still, if one
wants to improve the potential scope of gene therapy,
focus needs to be directed towards vector development.
346
This seems necessary, since apart from a limited number of studies little success has been made with gene
therapy of GBM today. Nevertheless, a lot of effort and
innovation can be seen and it is probably only a question of time, when gene therapy will be reckoned as a
therapy for the treatment of GBM.
References
Chiocca EA, Abbed KM, Tatter S, Louis DN, Hochberg FH,
Barker F, Kracher J, Grossman SA, Fisher JD, Carson K,
Rosenblum M, Mikkelsen T, Olson J, Markert J, Rosenfeld
S, Nabors LB, Brem S, Phuphanich S, Freeman S, Kaplan
R, Zwiebel J (2004) A phase I open-label, dose-escalation,
multi-institutional trial of injection with an E1B-Attenuated
adenovirus, ONYX-015, into the peritumoral region of recurrent malignant gliomas, in the adjuvant setting. Mol Ther
10:958966
Ciafre SA, Niola F, Wannenes F, Farace MG (2004) An
anti-VEGF ribozyme embedded within the adenoviral VAI
sequence inhibits glioblastoma cell angiogenic potential in
vitro. J Vasc Res 41:220228
Furukawa K, Kumon Y, Harada H, Kohno S, Nagato S, Teraoka
M, Fujiwara S, Nakagawa K, Hamada K, Ohnishi T (2006)
PTEN gene transfer suppresses the invasive potential of
human malignant gliomas by regulating cell invasion-related
molecules. Int J Oncol 29:7381
Ghulam Muhammad AK, Candolfi M, King GD, Yagiz K,
Foulad D, Mineharu Y, Kroeger KM, Treuer KA, Nichols
WS, Sanderson NS, Yang J, Khayznikov M, Van Rooijen N,
Lowenstein PR, Castro MG (2009) Antiglioma immunological memory in response to conditional cytotoxic/immunestimulatory gene therapy: humoral and cellular immunity lead to tumor regression. Clin Cancer Res 15:
61136127
Gondi CS, Lakka SS, Yanamandra N, Siddique K, Dinh DH,
Olivero WC, Gujrati M, Rao JS (2003) Expression of antisense uPAR and antisense uPA from a bicistronic adenoviral
construct inhibits glioma cell invasion, tumor growth, and
angiogenesis. Oncogene 22:59675975
Gupta V, Wang W, Sosnowski BA, Hofman FM, Chen TC (2006)
Fibroblast growth factor-2-retargeted adenoviral vector for
selective transduction of primary glioblastoma multiforme
endothelial cells. Neurosurg Focus 20:E26
Hamed HA, Yacoub A, Park MA, Eulitt PJ, Dash R, Sarkar D,
Dmitriev IP, Lesniak MS, Shah K, Grant S, Curiel DT, Fisher
PB, Dent P (2010) Inhibition of multiple protective signaling
pathways and Ad.5/3 delivery enhances mda-7/IL-24 therapy
of malignant glioma. Mol Ther 18(6):11301142
Hong X, Miller C, Savant-Bhonsale S, Kalkanis SN (2009)
Antitumor treatment using interleukin- 12-secreting marrow
stromal cells in an invasive glioma model. Neurosurgery
64:11391146
Immonen A, Vapalahti M, Tyynela K, Hurskainen H,
Sandmair A, Vanninen R, Langford G, Murray N,
Yla-Herttuala S (2004) AdvHSV-tk gene therapy with
T. Wirth et al.
intravenous ganciclovir improves survival in human malignant glioma: a randomised, controlled study. Mol Ther 10:
967972
Jadhav U, Chigurupati S, Lakka SS, Mohanam S (2004)
Inhibition of matrix metalloproteinase-9 reduces in vitro
invasion and angiogenesis in human microvascular endothelial cells. Int J Oncol 25:14071414
Jeong M, Kwon YS, Park SH, Kim CY, Jeun SS, Song KW,
Ko Y, Robbins PD, Billiar TR, Kim BM, Seol DW (2009)
Possible novel therapy for malignant gliomas with secretable
trimeric TRAIL. PLoS One 4:e4545
Kang YA, Shin HC, Yoo JY, Kim JH, Kim JS, Yun CO (2008)
Novel cancer antiangiotherapy using the VEGF promotertargeted artificial zinc-finger protein and oncolytic adenovirus. Mol Ther 16:10331040
Kang X, Xiao X, Harata M, Bai Y, Nakazaki Y, Soda Y, Kurita
R, Tanaka T, Komine F, Izawa K, Kunisaki R, Setoyama
M, Nishimori H, Natsume A, Sunamura M, Lozonshi L,
Saitoh I, Tokino T, Asano S, Nakamura Y, Tani K (2006)
Antiangiogenic activity of BAI1 in vivo: implications for
gene therapy of human glioblastomas. Cancer Gene Ther
13:385392
Kargiotis O, Chetty C, Gondi CS, Tsung AJ, Dinh DH, Gujrati
M, Lakka SS, Kyritsis AP, Rao JS (2008) Adenovirusmediated transfer of siRNA against MMP-2 mRNA results in
impaired invasion and tumor-induced angiogenesis, induces
apoptosis in vitro and inhibits tumor growth in vivo in
glioblastoma. Oncogene 27:48304840
Kim CY, Jeong M, Mushiake H, Kim BM, Kim WB, Ko JP, Kim
MH, Kim M, Kim TH, Robbins PD, Billiar TR, Seol DW
(2006) Cancer gene therapy using a novel secretable trimeric
TRAIL. Gene Ther 13:330338
King GD, Muhammad AK, Curtin JF, Barcia C, Puntel M, Liu C,
Honig SB, Candolfi M, Mondkar S, Lowenstein PR, Castro
MG (2008) Flt3L and TK gene therapy eradicate multifocal glioma in a syngeneic glioblastoma model. Neuro Oncol
10:1931
Kohn DB, Candotti F (2009) Gene therapy fulfilling its promise.
N Engl J Med 360:518521
Lakka SS, Gondi CS, Yanamandra N, Dinh DH, Olivero
WC, Gujrati M, Rao JS (2003) Synergistic down-regulation
of urokinase plasminogen activator receptor and matrix
metalloproteinase-9 in SNB19 glioblastoma cells efficiently
inhibits glioma cell invasion, angiogenesis, and tumor
growth. Cancer Res 63:24542461
Lamfers ML, Gianni D, Tung CH, Idema S, Schagen FH, Carette
JE, Quax PH, Van Beusechem VW, Vandertop WP, Dirven
CM, Chiocca EA, Gerritsen WR (2005) Tissue inhibitor of
metalloproteinase-3 expression from an oncolytic adenovirus
inhibits matrix metalloproteinase activity in vivo without
affecting antitumor efficacy in malignant glioma. Cancer Res
65:93989405
Lu W, Zhou X, Hong B, Liu J, Yue Z (2004) Suppression
of invasion in human U87 glioma cells by adenovirusmediated co-transfer of TIMP-2 and PTEN gene. Cancer Lett
214:205213
McKnight SL (1980) The nucleotide sequence and transcript
map of the herpes simplex virus thymidine kinase gene.
Nucleic Acids Res 8:59495964
Nalabothula N, Lakka SS, Dinh DH, Gujrati M, Olivero WC,
Rao JS (2007) Sense p16 and antisense uPAR bicistronic

construct inhibits angiogenesis and induces glioma cell
death. Int J Oncol 30:669678
Ohgaki H, Kleihues P (2005) Population-based studies on incidence, survival rates, and genetic alterations in astrocytic
and oligodendroglial gliomas. J Neuropathol Exp Neurol
64:479489
Pan D, Wei X, Liu M, Feng S, Tian X, Feng X, Zhang X 2009.
Adenovirus mediated transfer of p53, GM-CSF and B7-1
suppresses growth and enhances immunogenicity of glioma
cells. Neurol Res 32(5):502509
Rainov NG (2000) A phase III clinical evaluation of herpes
simplex virus type 1 thymidine kinase and ganciclovir gene
therapy as an adjuvant to surgical resection and radiation
in adults with previously untreated glioblastoma multiforme.
Hum Gene Ther 11:23892401
Sandmair AM, Loimas S, Puranen P, Immonen A, Kossila M,
Puranen M, Hurskainen H, Tyynela K, Turunen M, Vanninen
R, Lehtolainen P, Paljarvi L, Johansson R, Vapalahti M, YlaHerttuala S (2000) Thymidine kinase gene therapy for human
malignant glioma, using replication-deficient retroviruses or
adenoviruses. Hum Gene Ther 11:21972205
Stratmann A, Machein MR, Plate KH (1997) Anti-angiogenic
gene therapy of malignant glioma. Acta Neurochir Suppl
68:105110
347
(2009) European Organisation for Research and Treatment
of Cancer Brain Tumour and Radiation Oncology Groups
and National Cancer Institute of Canada Clinical Trials
Group. Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival
in glioblastoma in a randomised phase III study: 5-year
analysis of the EORTC-NCIC trial. Lancet Oncol 10:
459466
Wirth T, Hedman M, Makinen K, Manninen H, Immonen A,
Vapalahti M, Yla-Herttuala S (2006) Safety profile of plasmid/liposomes and virus vectors in clinical gene therapy.
Curr Drug Saf 1:253257
Yamini B, Yu X, Pytel P, Galanopoulos N, Rawlani V,
Veerapong J, Bickenbach K, Weichselbaum RR (2007)
Adenovirally delivered tumor necrosis factor-alpha
improves the antiglioma efficacy of concomitant radiation and temozolomide therapy. Clin Cancer Res 13:
62176223
Chapter 34
Fischer/F98 Glioma Model: Methodology

David Fortin
Abstract Adequate animal glioma models are

mandatory for the pursuit of preclinical research in
neuro-oncology. Many implantation models have been
described, but none perfectly emulate human malignant gliomas. This chapter reports our experience
in standardizing, optimizing and characterizing the
Fischer/F98 glioma model on the clinical, pathological, radiological and metabolic aspects. We specifically
analyzed H&E staining and immunohistochemistry for
GFAP, vimentin, albumin, TGF-b1, TGF-b2, CD3 and
CD45. We also used MR imaging at 5, 10, 15 and
20 days to correlate with the clinical and pathological evolution. 18 F-FDG and 11 C-acetate PET studies
for metabolic characterization of the tumors were also
assessed. The standardized model described in this
study behaves in a predictable and reproducible fashion, and could be considered for future pre-clinical
studies by research investigators. It adequately mimics
the behavior of human malignant astrocytomas.
Keywords Animal model Neuro-oncology
Fischer/F98 glioma model Cell line Host
Implantation
Introduction
Malignant glial tumors remain a therapeutic challenge,
despite many advances in the field. Most of these
advances have had a minor impact on the survival of
D. Fortin ()
Division of Neurosurgery, Department of Surgery, Sherbrooke
University, Sherbrooke, QC, Canada J1H5N4
e-mail: David.fortin@usherbrooke.ca
patients. Considering the aggressive nature of malignant gliomas, as well as the limited progress realized
in patients outcome over the last decades, it is obvious that research endeavors shall continue unabated. In
vivo preclinical studies require adequate representative
models to this end. The use of animal glioma models
thus remain mandatory in our pursuit of better treatment strategies, as well as a practical tool to improve
our knowledge of the disease process.
Many models have been developed for this purpose
using different approaches. Although from a theoretical point of view, transgenic models might better
emulate human tumors and present several advantages
(Lampson, 2001; Peterson et al., 1994), in practice
most researchers still use implantation models. This is
largely attributable to their simplicity of use and relatively inexpensive cost compared to transgenic models,
which require substantial resources to develop and
handle (Lampson, 2001; Peterson et al., 1994). More
so, to produce a representative glioma model using
transgenic strategies, one would require a detailed
knowledge of the different genes and genetic pathways involved in the genesis of these tumors. Although
diverse genetic anomalies have been described in
malignant glial tumors, these anomalies are multiple,
complex and highly variables from tumor to tumor,
thus rendering the use of such an approach empiric.
These transgenic models are best used to study specific
genotypic alterations, and their consequent role in the
tumor genesis and development. To study the potency
of newly designed treatment strategies, implantation
models remain actual and useful.
Implantation models, derived from cultured cell
lines that are implanted in the brain of the target animal, are widely used (Barth, 1998; Ko et al., 1980).
However, a great number of these models tend to

349
350
neglect some essential characteristics that can limit the

way they emulate the behavior of their human counterparts. The goal of any such model should be to
reproduce as best as possible the clinical situation,
so that conclusions derived from experiments using a
given animal model is likely to be transposable to the
clinic. In essence, this is the quintessential of translational research. In order to optimize glioma implantation models, some determinants should be carefully
sought (Barth, 1998; Peterson et al., 1994). A close
attention to details in the design of the model and in
the implantation technique are paramount, and flaws
should be aggressively pursued and resolved. The first
point that needs to be addressed in the design of an
adequate glioma implantation model is the choice of
the studied cell line and the host animal.
The Choice of a Cell Line and of the Host

The selection of the appropriate cell line and host animal obviously represents the first step in the design of
an implantation model. As most brain tumors develop
in relatively immunocompetent people, and given the
fact that immune modulation is a potential therapeutic approach in the treatment of brain tumors, the
neoplastic cells should be implanted in immunocompetent hosts rather than nude animals. Thus, in order
to avoid any undesired immune reaction producing
inherent tumor rejection, an essential quality of any
implantation model is syngeneicity between the tumor
cell line and the host (Barth, 1998). Nude animals
will find application when working in xenogenic conditions, with human glioma cells extracted from fresh
tumor samples. Another important characteristic of the
selected glioma model is the presence of a relative
resistance to the available treatment strategies, as are
human malignant gliomas. A large number of rodent
malignant glial cell lines are commercially available,
and models have been described using each of them.
The C6 and 9L lines are the most widely used, although
both have obvious shortcomings when it comes to
preclinical evaluation of therapeutic strategies.
The C6 cell line is the most commonly reported
model in the literature. Interestingly, it is also probably one of the most flawed model, being an infiltrative
glioma which has no known syngeneic host and is usually implanted in the Wistar rat (Barth 1998; Peterson
et al., 1994; Parsa et al., 2000). Immune responses
D. Fortin
potentially leading to tumor rejection and complete

spontaneous tumor regression have been reported with
this model (Parsa et al., 2000; Grobben et al., 2002) as
it has been shown to generate important humoral and
cellular immune responses after tumor implantation.
Spectacular in vivo experimental results using the C6
model have not translated so far in any real therapeutic breakthrough for malignant glioma patients. Thus,
the use of this model should be discouraged, as any
conclusions obtained with it is subject to bias and thus
hardly transferable to the clinic.
The 9L cell line gives rise to sarcomatous
tumors that often behave like metastatic lesions, with
poor infiltration of surrounding brain parenchyma
(Stojiljkovic et al., 2003). As the infiltrative behavior
and the migrational potential of malignant astrocytic
cells represent one of the most disconcerting phenotypical characteristics of this cancer, infiltration of
adjacent brain parenchyma is expected and should
be present in any model attempting to reproduce a
malignant astrocytic tumor.
We chose to work with the F98 cell line for several
reasons. The F98 cell line is an anaplastic glioma with
a minor sarcomatous component that was originally
produced by a single N-ethyl-N-nitrosourea (ENU)
injection to a 20-week pregnant Fischer rat (barth,
1998). The offspring developed brain tumors that were
harvested and maintained in culture. Its syngeneic host
is well known to be the Fischer rat (Barth, 1998). A
very weak immunogenicity characterizes this line, with
reports of successful short-term implantations in cat
brains (Wechsler et al., 1989). It is extremely stable
in culture, with no morphological changes observed
in over a hundred cell passages in our laboratory.
Finally, and of the outmost importance, this cell line is
extremely resistant to treatment. Few treatment strategies have produced significant tumor regression when
tested on this cell line after intracranial implantation
in its syngeneic host, and only a few investigators
have reported survival improvement using this model
in preclinical therapeutic trials (Adam et al., 2003;
Barth et al., 2003; Fournier et al., 2003; Mathieu
et al., 2005). This aggressive behavior adequately parallels the clinical situation. The F98 cell line is an
anaplastic glioma with a minor sarcomatous component depicting low immunogenicity when implanted
in its syngeneic host. Consequently to their implantation, animals develop infiltrative tumors that are very
resistant to conventional treatment.
34 Fischer/F98 Glioma Model: Methodology
The Implantation Technique

The selected cell line should obviously be implanted
in the brain of the syngeneic host. As the rat brain
can only accommodate tumor of a very small caliber, some authors still use subcutaneous flank tumor
implantation models. Needless to say that these models
are totally incompetent to investigate glioma therapeutic strategies, as the brain microenvironment plays a
fundamental role in tumor infiltration, angiogenesis
and tumor growth via complex molecular interactions
(Saini et al., 1999). Besides, the delivery issue cannot
be addressed with flank tumor implantation.
The different technical aspects leading to tumor cell
implantation are all essential steps in the success of a
reproducible and valid tumor model. These steps are
too frequently occulted from discussion and dismissed
as unimportant, with the consequence that the lack of
standardization plague a great number of implantation
models. As a result, different implantation coordinates
are used without validation, tumor takes are highly
variable, and so are tumor volumes within a given
study group (Fournier et al., 2003; Peterson et al.,
1994; Raila et al., 1999; Thorsen et al., 2003). Reports
of free-hand implantation, as well as uncontrolled
tumor cell suspension infusion are legion in the literature. Given all these variations, it can be difficult
to draw reliable conclusions from therapeutic studies
investigating tumor growth and/or tumor volumes. It is
thus paramount to standardize all the steps involved in
the elaboration of a glioma implantation model to minimize these procedural-related factors that can bias the
model. The mandatory use of a stereotactic frame for
the procedure cannot be overemphasized (Fig. 34.1). A
freehand injection has already been reported to be inaccurate and to give rise to inconsistent tumors (Maggio,
1996), but some investigators still use this approach for
their models, as illustrated by recently published work
(Stojiljkovic et al., 2003).
In designing the implantation procedure, the goals
sought are the constant and reproducible production
of tumors growing in a predictable pattern, low morbidity and no mortality for the host, and a tumor
take rate of or close to 100%. The implantation technique should be relatively non traumatic, with minimal
brain parenchymal damage set as a goal to reduce
the associated local inflammatory reaction inherent to
the implantation process. To facilitate in vivo imaging
351
with noninvasive modalities such as magnetic resonance (MR) and positron emission tomography (PET),
the implantation should be suitably located to avoid
adjacent structures that could interfere with the tumor
model imaging (e.g. Harderian glands in the rat). Very
few of the current animal models described in the literature meet all these criteria. Many authors using a
given rat strain use different implantation coordinates
and techniques (Parsa et al., 2000). Reported tumor
take is highly variable from study to study, ranging
from 50 to 100% (Fournier et al., 2003; Raila et al.,
1999). A considerable variation in tumor volumes can
also be observed within a group of implanted animals
using the same technique. A recently published study
reported tumor volume after a constant observation
period ranging from less than 10 mm3 to more than
80 mm3 , despite standardization in the implantation
technique (Thorsen et al., 2003).
The Choice of Implantation Coordinates

The choice of implantation coordinates has an obvious
impact on the model, and influences all the clinical parameters of the animal. Obviously, the set of
coordinates can be adjusted to suit specific objectives
pursued. Usually, as the coordinates are adjusted in
the rostro-caudal plane, the more caudal (or posterior) implantation will tend to produce shorter survival
and precocious clinical manifestations. Typically, in
brain tumor survival studies, the set of coordinates
targets the area of the caudate nucleus. Although an
excellent location that allows a sizeable tumor growth
prior to clinical manifestations, two pitfalls should
be avoided when implanting in the frontal lobe: ventricular implantation and implantation at proximity of
the Harderian glands. Ventricular implantation will
lead to intraventricular dissemination, and significantly
shorten the lifespan of the test animal. The Harderian
glands are exocrine glands located in the medial aspect
of the orbit that depict a highly metabolic activity.
Uptake of the [18 F]fluorodeoxyglucose used in pet
scanning of the animals by the glands could potentially
obscure tumor uptake thus abrogating all possibility at
metabolic characterization with this tracer.
When designing the initial phase of our F98-Fischer
implantation model, we used a simple method to identify the ideal implantation coordinates based on Evans
352
D. Fortin
Fig. 34.1 Final setup for the implantation. (a) The microinfusion pump is mounted on the stereotactic frame to allow a
continuous and slow infusion rate. (b) Once again care is used in
the placement of the animals head in the frame as even a slight
sagittal mis-angulation will significantly affect the implantation
coordinates. (c) The burr hole was accomplished using a 16

gauge needle after coordinates localisation using the stereotactic frame. A 27 gauge needle is used to deliver the implantation
solution
blue staining (Fig. 34.2a). This step should ideally be

performed to verify the validity of selected implantation coordinates, when designing, modifying or starting to work with a new model. Using the bregma
as a reference landmark, anterior (03 mm) and lateral (14 mm) coordinates were tested by intracerebral
infusion of an Evans blue solution using a stereotactic
frame as the targeting device. Depth of implantation
was also assessed, varying from a 5 to 7 mm depth
to the external table of the skull. Once the Evans
blue solution had been infused, the animals were euthanized, brain specimens were harvested, cut in the
coronal plane and assessed for Evans blue localization (Fig. 34.2b). The optimal implantation coordinates
identified by this approach were as follow: 1 mm anterior and 3 mm lateral to the right of the bregma, with
implantation solution infusion at a depth of 6 mm
from the outer table of the skull. When animals are
mounted on the stereotactic frame, special care should

be used in placing the head in the frame to avoid sagittal angulation, so that the head remains parallel to the
frame stand. This will prevent implanting the tumor
cell suspension away from the planned coordinates.
Tumor Cell Implantation Suspension

Finally, the volume of implantation solution, as well
as the number of cells suspended should be standardized. Serial experiments were performed in our
laboratory to assess the cellular suspension characteristics that would lead to a valid and reproducible
model with an optimal time-frame from implantation to death. The smallest infusion volume and the
smallest number of cells were pursued, the rationale
being that this would decrease the trauma associated
353
Fig. 34.2 Localization of implantation coordinates and

histopathological characerization. (a) Brain specimen harvested
after implantation coordinates testing using Evans blue as a
marker. The entry points are noticeable. For standardization, the
stereotactic frame was used in all procedures, and great care was
used to place the head of the animal in the frame in an horizontal
fashion, minimizing any sagittal angulation, so that implantation
coordinates would be reproducible. (b) Coronal cut of the brain
showing the depth of the projected implantation coordinates
and the implantation tracts. A depth of 6 mm was used on the
left side, whereas a 4 mm depth was used on the right side. (c)
Coronal H&E slide of a representative specimen obtained and
sacrificed at 25 days post-implantation, when the animal presented a lethargic condition. Notice the large right hemispheric
tumor distorting the brain parenchyma and producing a significant midline shift. This tumor also depicts central necrosis, and
peripheral brain infiltration. (d) H&E slide (40) of a representative specimen, focusing the field of view at the periphery of
the tumor nodule. Infiltration of the adjacent brain parenchyma
by neoplastic cells is obvious, as is a pattern of perivascular
clustering. (e) GFAP immunohistochemistry (40) of a representative specimen. Widespread cytoplasmic staining for GFAP
is obvious in numerous cells. The field of view is focused in the
center of the tumor nodule. (f) Albumin immunohistochemistry
(20) of a representative specimen sacrificed at 26 days post
implantation. The field of view is centered at the periphery of
the main tumor nodule to emphasize the albumin staining in the
region of the brain around tumor depicting BBB breakdown
to the implantation procedure, and thus the consequent inflammatory reaction. Moreover, we hypothesized that by delivering smaller volumes, we would
increase the infiltrative behavior of the tumor cells by

decreasing the artifactual cavity created by the implantation. For the same reason, the use of a micro-infusion
354
pump to infuse the suspension solution was introduced. Initially, manual injection of the solution was
accomplished. In an effort to standardize infusion time,
ensure a constant rate, and decrease the implantation
traumatism, a micro-infusion pump (UltraMicroPump,
World Precision Instruments Inc.) was acquired and
was used to deliver the implantation solution. By
allowing the slowest delivery time, we intended to
decrease the risks of reflux of the solution along
the implantation tract, an artifactual problem common with glioma implantation models. The following
parameters were tested: volumes of 10 L containing
5 105 and 2 105 tumor cells; volumes of 5 L
with 1 105 or 1 104 cells, and 1 L containing 1
103 cells.
Final Implantation Parameters

Based on the survival interval obtained after implantation and on the pathological examination of the
specimens, we elected to work with a final solution
concentration of 1 104 cells diluted in a volume of
5 L, and delivered at a rate of 1 L/min.
The Standardized Fisher-F98 Model

Cell Line Preparation
The F98 cell line was obtained from ATCC, and was
grown in monolayer using a solution of Dulbeccos
modified Eagle medium (DMEM) supplemented with
10% fetal bovine serum (FBS) and a mix of penicillinstreptomycin. Cells are incubated at 37 C in a humidified environment with 5% CO2 and propagated upon
confluence, every 3 days.
The implantation solution is prepared by trypsinization of the cell culture followed by re-suspension in
a DMEM solution free of FBS. Briefly, after washing
twice with PBS, the adherent cell lines are detached
from the Petri dishes by the addition of 1 ml of
0.05% trypsin-0.02% EDTA. The trypsin is then inactivated using 9 ml of DMEM without FBS. 1 104
cells are resuspended in a final volume of 5 L of
DMEM solution free of FBS. A trypan blue exclusion test is performed to assess cell viability before
implantation. Typically, tumor cell viability decreases
sharply 120 min. after suspension in solution.
D. Fortin
Implantation Technique
Adult male Fischer rats weighting 225250 g are used
for the model. Anesthesia is induced by inhalation of
a mixture of oxygen with 5% halothane, followed by
an intra-peritoneal injection of ketamine (87 mg/kg)
and xylazine (13 mg/kg) for maintenance. Animals are
then mounted on a stereotactic frame. As mentioned
previously, special care is used in the placement of the
head in the frame to avoid sagittal angulation, so that
the head is parallel to the frame stand (Fig. 34.1). This
will prevent inadequate posterior implantation, as the
head has a tendency to be tilted backward. A midline
scalp incision is performed, followed by identification
and exposure of the bregma. Using a 16-gauge needle,
a burr hole is placed on the right frontal bone, using
these coordinates: 1 mm anterior and 3 mm lateral
to the right of the bregma, with implantation solution
infusion at a depth of 6 mm from the outer table of
the skull. A 25-L SGF syringe with a 27-gauge needle secured to the frame is used to infuse the cellular
solution over a period of 5 min. The solution is delivered at a constant rate of 1 L per minute, using a
micro-infusion pump, after which the needle is slowly
withdrawn over 1 min, to minimize backflow of the
suspension solution.
Bone wax is applied to close the burr hole and the
scalp is closed with a continuous one-layer resorbable
suture.
Clinical Evolution
Animals are allowed to recover from the procedure,
and are given food and water ad libitum afterward.
They are assessed clinically bi-daily for the apparition of signs of raised intracranial pressure (lethargy,
vomiting, and cachexia) or focal neurological signs
(hemiparesis, ataxia). Symptoms typically consist of
progressive lethargy with variable degree of hemiplegia contralateral to the implantation side. In the
later phases, ataxic irregular pattern of breathing is
observed. A significant majority of test subjects lose
weight over the course of their evolution, with a mean
of 5.2% body weight loss. The median survival is
26 days. The implantation technique described herein
depicted very low morbidity in our hands with no
peri-procedural death in 58 implantations, as described
in our original study (Mathieu et al., 2007). Since then,

more than 400 implantations have been performed in
our laboratory with less than 10 procedural-related
death events.
Histopathological Characterization
Results presented here refer to a group of 68 animals comprised in our original study (Mathieu et al.,
2007).
355
Pathologic Analysis
Slides were scanned at low-power magnification to
identify the tumors, which were then examined at
higher magnification. Tumor morphology and characteristics were assessed on H&E. The number of
mitotic figures per high-power field (HPF, 40 magnification) was noted for proliferation assessment.
Immunohistochemistry labeling was assessed qualitatively for GFAP, vimentin, albumin, and TGF-b, and
quantitatively for CD3 and CD45 (number of labeling
cells per HPF).
Immunocytochemistry
Pathological Analysis Results
Upon retrieval, brain specimens are fixed in a formalin solution for 48 h, cut in the coronal plane in
1 mm-width slices using a dedicated brain matrix
and embedded in paraffin. The blocks are sectioned at 3 m intervals and the resulting slides are
stained with hematoxylin and eosin (H&E). After
deparaffinization and dehydration, a microwave antigen retrieval process is performed. Slides are placed
in 0.1 mmol/l citrate buffer in a microwaveable
pressure cooker and boiled in a 700-W microwave
oven for 30 min. Sections are incubated with the
primary antibodies selected for study. Biotinylated
species-specific secondary antibodies are applied followed by an avidin-biotin amplification and peroxides
development.
For this histopathological characterization, monoclonal antibody labeling was obtained for GFAP
(BD bioscience, San Jose, California, dilution 1/10),
vimentin (BD bioscience, San Jose, California dilution 1/100), TGF-b1 (Santa Cruz biotechnology, Santa
Cruz, California, dilution 1/100), TGF-b2 (Santa Cruz
biotechnology, Santa Cruz, California, dilution 1/80),
albumin, CD3 (BD bioscience, San Jose, California
dilution 1/175) and CD45 (BD bioscience, San Jose,
California dilution 1/800). Many publications have
reported the overexpression of TGF-b in malignant
gliomas, with a possible role in neoplastic tumor
growth, and migration (Jachimczak et al., 1996). Since
the suppression of this protein is considered paramount
in the research efforts of our laboratory, the characterization of its expression in this model was accomplished. CD3 and CD45 were labelled to study the
presence and extent of an associated inflammatory
reaction.
The specimen examination revealed the presence of a

significant tumor in 68 of a total of 68 eligible animals,
thus producing a tumor take of 100%.
Histopathological characteristics of tumors grown
from F98 cells implanted in the brain of Fischer
rats have previously been described (Reifenberger
et al., 1989; Seitz et al., 1988). These tumors grow
by diffusely infiltrating the surrounding brain. They
demonstrate extensive areas of necrosis and vascular
proliferation. The findings reported in this work are
entirely congruent with these descriptions. The tumors
obtained with this model were large, with a central
necrotic core and numerous mitotic figures, with a
mean of 8 mitoses per HPF. The very high mitotic
count translates the aggressiveness of this lineage.
Obvious distortion in the cerebral architecture was
accompanied by important brain shift and mass effect
(Fig. 34.2c). Vascular proliferative changes could be
observed in most specimens. Perivascular neoplastic
infiltration was obvious, as was the adjacent brain
parenchymal infiltration by individual tumor cells at
the edge of the lesion, in all samples (Fig. 34.2d).
Extention of the tumor along the implantation tract
all the way to the surface of the brain, with leptomeningeal spread was rarely seen in this model,
whereas it was common place when we tested solution
volumes greater than 5 L without the micro-infusion
pump.
GFAP immunohistochemistry confirmed that a significant proportion of the neoplastic cells were of astrocytic lineage (Fig. 34.2e). In addition, dense GFAP
labeling was observed in the adjacent parenchyma surrounding the tumors in reactive astrocytes. Vimentin
356
expression was diffuse in the tumor, being present in

almost every tumor cells. As was previously postulated
by Whittle et al., this observation probably suggests
that the majority of tumor cells are neoplastic astrocytes in a dedifferentiated state, as vimentin expression
is an early event in normal glial differentiation, while
GFAP appears in later stages (Whittle et al., 1998).
Profuse albumin immunostaining was observed in the
extracellular space at the periphery of the tumors, an
illustration of the pathologically increase in the bloodbrain barrier permeability (Fig. 34.2f). This reflects the
fact that the blood-brain barrier is locally disrupted by
the growth of the tumor, as albumin normally is only
minimally present in the interstitial space of the brain,
due to its large molecular weight. Indeed, albumin
immunohistochemistry has been reported to be a reliable marker of blood-brain barrier integrity (Poduslo
et al., 2001). When positive, labeling for CD3 cells
was seen at the periphery of the tumor and around
adjacent capillaries. A mean of 2.8 staining cells per
HPF was observed. CD45 staining was extremely low,
with only an occasional staining cell. This observation confirms the weakly immunogenic properties of
the F98 cells after their implantation into the brain
of the Fischer rat, and the fact that the implantation
procedure did not generate a significant inflammatory
reaction. This is in sharp contrast to an earlier study we
led, testing the implantation of F98 cells in syngeneic
(fischer) vs allogenic (Long-Evans) setting. In allogenic setting, a significant inflammatory reaction was
detected (mean of 29.1 CD3 staining cell per HPF);
this ultimately was associated with a tumor graft rejection, and a decrease tumor take (50% in Long-Evans,
compared to 100% in Fischer) (Mathieu et al., 2005).
The risk that immunologic graft rejection will produce
biases when evaluating therapeutic strategies with this
animal model in syngeneic setting is minimal. The
importance of working in a syngeneic context cannot
be over-emphasized.
TGF-b1 depicted perivascular staining in the tumor
nodules, especially at the periphery. Occasional individual neoplastic cells depicted cytoplasmic staining.
We have also documented that F98 cells produce
TGF-b1 in active isoform by rt-PCR, ELISA and
western blotting. TGF-b2 has been reported as an
important local immunosuppressive factor produced
by malignant astrocytic cells (Maxwel et al., 1992).
However, TGF-b2 staining was insignificant in this
model.
D. Fortin
MRI Evaluation Using a Clinical MRI Unit

The original study reporting our full characterization
of the present model was accomplished prior to the
acquisition of an animal high field magnet MRI unit
(Blanchard et al., 2006). Thus, in vivo tumor growth
was originally characterized using a clinical 1.5 T
unit (Siemens 1.5T magnetom Sonata with Syngo
2002B software). To determine the optimal imaging
parameters, a series of preliminary images were performed on a non implanted test animal subject. Twelve
implanted rats were then randomly assigned to one
of four groups, who were to be imaged at 5, 10, 15
and 20 days following tumor implantation. Procedures
were performed under ketamin/xylazine anesthesia.
Animals were positioned prone in the magnet, with the
antenna (Siemens small loop flex coil antenna) placed
directly on the head. Coronal T2-weighted images and
contrast-enhanced T1-weighted images were obtained.
Following the imaging procedure, the animals were
immediately sacrificed. Brains were harvested and
processed for H&E. To validate the MRI images,
contrast-enhanced T1 images were compared with the
corresponding H&E slides (Fig. 34.3). The volume and
largest cross-sectional area on contrast imaging were
assessed using the image-processing software Sigma
Scan Pro 5 (Hallogram Publishing, Aurora, USA).
Gadolinium-enhanced
T1-weighted
images
depicted a clear brain-tumor interface in all the
study specimens. Tumor progression was obvious
and measurable over the study period. The average volume calculated for each time group was as
follows: 6.15 mm3 2.74 at 5 days, 15.9 mm3
7.0 at 10 days, 59.8 mm3 18.8 at 15 days and
77.0 mm3 29.0 at 20 days. Comparisons of the
largest radiological and pathological cross-sectional
area demonstrated a parallel increase over time,
although the MRI studies consistently overestimated
tumor volumes. The reasons for this discrepancy
are multiple: a shrinking factor produced by tissue
fixation at pathology and the enhancement on MR at
the brain around tumor area, related to the presence of
neovascularization are the major mechanisms at play
(Wree et al., 1995). Moreover, and in a distinctive
behaviour observed with the pathology specimens, the
growth pattern reached a treshold between day 15 and
day 20. Interestingly, when results were plotted on a
logarithmic scaled graphic, tumor volumes at day 5, 10
357
Fig. 34.3 Composite picture displaying tumor progression in

four animals, at 5, 10, 15 and 20 days after implantation. Top row
depict coronal gadolinium-enhanced T1 MRI at 5, 10, 15 and 20
days in four different animals, whereas bottom row match the
corresponding pathology samples. The MRI adequately translates the general morphology of the tumors and progression over
the observation period
and 15 depicted a near-perfect exponential relationship

with a doubling time of 3.05 days. Using this function,
the extrapolated initial implantation volume was
estimated at 1.85 mm3 . A threshold was observed after
15 days, thus implying that tumor growth gradually
decreases after day 15 post-implantation. This effect
is likely related to the increase interstitial pressure in
and around the tumor, and the incompetence of the
vascular supply past a certain volume to accommodate
tumor growth. T2-weighted images demonstrated
significant peritumoral edema, with blurred tumor
margins.
The use of an image analysis software allowed a
precise and quantitative characterization of tumor volumes, suggesting that it could be a useful adjunct
method for the evaluation of the impact of a therapeutic strategy by allowing volumetric comparisons of the
tumors between treatment groups in living animals.
underwent this imaging modality 14 days after the

implantation. Scans were carried with the animals
under general anesthesia. The animals were placed
supine on the bed of the camera and kept warm with
a heating pad. Prior to imaging, the camera was calibrated with a phantom filled with water containing
0.98 MBq (700 uCi) of 18 F-FDG in a volume of
22 ml. Dynamic imaging was initiated 1 min prior
to injection of the isotopes, and continued for 1830
s, followed by 3 (1289 s) volumetric acquisitions.
The images were reconstructed using 25 iterations
of the maximum likelihood expectation maximization
(ML-EM) algorithm implementing position-dependent
detector response (Selivanoc et al., 2000). Isotope
uptake was analyzed on the reconstructed images using
the Sherbrooke LabTEP software (Benard F, Rousseau
E, Sherbrooke LABTEP software, v. 1.25 release 3,
November 2003). A region of interest (ROI) was drawn
over the tumor and from an equivalent size ROI traced
over an area of healthy brain tissue. Results were
expressed as the tumor to healthy brain tissue uptake
ratio.
PET scanning studies demonstrated strong 18 FFDG uptake by the tumors as compared to the adjacent
brain parenchyma. Harderian glands also demonstrated
a significant uptake of the radiotracer, showing the
importance of tumor implantation at a distance from
those glands. Fortunately, harderian glands uptake did
not obscure neoplastic uptake in such a way as to hinder adequate delineation of the tumors, as we took
Metabolic Evaluation
Metabolic evaluation of the model was carried out by
18F-FDG and 11 C-acetate positron emission tomography (PET) scanning. The Sherbrooke small animal
PET scanner (based on avalanche photodiode detectors) was used to acquire the images over the whole
brain (Lecomte et al., 1994). With proper sampling
motion, this scanner achieves 2.1 mm FWHM transaxial and 14 L volumetric resolution. Two animals
358
great care in the establishment of our implantation

coordinates to implant at a significant distance from
these glands. These results are consistent with the
extremely malignant behavior of the model, as FDGPET has been demonstrated to be a reliable prognostic
tool for pathology and survival in human low and high
grade astrocytomas (Padma et al., 2003). Tumor-tobrain uptake ratio increased over time after the isotope
injection, which reflects the normal clearance of the
isotope from the brain, and selective sequestration in
tumor cells. For 11 C-acetate, no difference in uptake
between tumor and surrounding brain was observed.
Radiotracer uptake ratio was 1.12 for 11 C-acetate, and
1.71 and 1.97 for 18 F-FDG at 30 and 50 min postinjection, respectively. Therefore, 11 C-acetate was not
considered suitable for metabolic evaluation of the
F98-Fischer model.
Conclusion
Animal models are necessary for the preclinical evaluation of antitumoral therapeutic strategies. The syngeneic Fischer/F98 rat glioma model described in this
work adequately mimics the clinical and pathological behaviour of human high-grade astrocytic tumors.
Careful and systematic implantation techniques, and
the use of a stereotactic apparatus and micro-infusion
pump, contribute to produce a predictive and constant
growth pattern that allows an accurate assessment of
treatment. The use of the micro-infusion pump allows
delivery of the implantation solution at a constant flow,
thus minimizing local tissue trauma at the time of infusion. We believe that a meticulous attention to all the
steps involved in the implantation is important, as the
accuracy and reproducibility of the model is at stake.
Ultimately, results of therapeutic experiments will be
biased if the model is inadequate. Using the parameters
evoked in this chapter, a median survival of 26 days
is obtained, with a gradual, constant and predictable
growth in tumors as documented pathologically and
radiologically. This survival length is ideal for therapeutic trials as it allows a sufficient time-window to
treat the animals, and the effect of treatment can be
recognized soon enough by using prolongation of survival as a surrogate. Moreover, magnetic resonance
imaging and PET imaging studies can easily be accomplished to complement the clinical and pathological
D. Fortin
data and generate a thorough evaluation and understanding in the tumor growth of the model, whether
it is for the purpose of a therapeutic trial or for detailed
characterization.
References
Adam JF, Elleaume H, Joubert A, Biston MC, Charvet AM,
Balosso J (2003) Synchrotron radiation therapy of malignant
brain glioma loaded with an Iodinated contrast agent: first
trial on rats bearing F98 gliomas. Int J Radiat Oncol Biol
Phys 57:14131426
Barth RF (1998) Rat brain tumor models in experimental neurooncology. The 9L, C6, T9, F98, RG2(D74), RT-2 and CNS-1
Barth RF, Yang W, Coderre JA (2003) Rat brain tumor models to
assess the efficacy of boron neutron capture therapy: a critical
evaluation. J Neurooncol 62:6174
Blanchard J, Mathieu D, Patenaude Y, Fortin D (2006) MRpathological comparison in F98-Fischer glioma model using
a human gantry. Can J Neurol Sci 33:8691
Fournier E, Passirani C, Montero-Menei C, Colin N, Breton P,
Sagodira S, Menei P, Benoit JP (2003) Therapeutic effectiveness of novel 5-fluorouracil-loaded poly (methylidene
malonate 2.1.2)-based microspheres on F98 glioma-bearing
rats. Cancer 97:28222829
Grobben B, De Deyn PP, Slegers H (2002) Rat C6 glioma as
experimental model system for the study of glioblastoma
growth and invasion. Cell Tissue Res 310:257270
Jachimczak P, Hessdorfer B, Fabel-Schulte K, Wismeth
C, Brysch W, Schlingensiepen KH (1996) Transforming
growth factor-beta-mediated autocrine growth regulation
of gliomas as detected with phosphorothioate antisense
oligonucleotides. Int J Cancer 65:332337
Ko L, Koestner A, Wechsler W (1980) Morphological characterization of nitrosourea-induced glioma cell lines and clones.
Lampson LA (2001) New animal models to probe brain
tumor biology, therapy, and immunotherapy: advantages and
remaining concerns. J Neurooncol 53:275287
Lecomte R, Cadorette J, Richard P, Rodrigue S, Rouleau D
(1994) Design and engineering aspects of a high resolution
positron tomograph for small animal imaging. IEEE Trans
Nucl Sci 41:14461452
Maggio WW (1996) Rodent glioma models. Methods Neurosci
30:8196
Mathieu D, Lamarche J, Fortin D (2005) The importance of
a syngeneic glioma implantation model: comparison of the
F98 cell line in Fischer and Long-Evans rats. J Appl Res
5:1725
Mathieu D, Lecomte R, Tsanaclis AM, Larouche A, Fortin D
(2007) Standardization and detailed characterization of the
syngeneic fischer/F98 glioma model. CJNS 34:296306
Maxwell M, Galanopoulos T, Neville-Golden J, Antoniades
HN (1992) Effect of the expression of transforming growth
factor-beta 2 in primary human glioblastomas on immunosuppression and loss of immune surveillance. J Neurosurg
76:799804

Padma MV, Said S, Jacobs M, Hwang DR, Dunigan K, Satter
M, Christian B, Ruppert J, Bernstein T, Kraus G, Mantil JC
(2003) Prediction of pathology and survival by FDG PET in
Parsa AT, Chakrabarti I, Hurley PT, Chi JH, Hall JS, Kaiser
MG, Bruce JN (2000) Limitations of the C6/Wistar rat
Intracerebral glioma model: implications for evaluating
immunotherapy. Neurosurgery 47:9931000
Peterson DL, Sheridan PJ, Brown E W Jr (1994) Animal models
for brain tumors: historical perspectives and future directions. J Neurosurg 80:865876
Poduslo JF, Curran GL, Wengenack TM, Malester B, Duff K
(2001) Permeability of proteins at the blood-brain barrier in
the normal adult mouse and double transgenic mouse model
of Alzheimers disease. Neurobiol Dis 8:555567
Raila FA, Bowles AP, Perkins E, Terrell A (1999) Sequential
imaging and volumetric analysis of an intracerebral C6
glioma by means of a clinical MRI system. J Neurooncol
43:1117
Reifenberger G, Bilzer T, Seitz RJ Wechsler W (1989)
Expression of vimentin and glial fibrillary acidic protein in
ethylnitrosourea-induced rat gliomas and glioma cell lines.
Saini M, Bellinzona M, Meyer F, Cali G, Samii M (1999)
Morphometrical characterization of two glioma models in
the brain of immunocompetent and immunodeficient rats. J
Neurooncol 42:5967
359
Seitz RJ, Deckert M, Wechsler W (1988) Vascularization of
syngenic intracerebral RG2 and F98 rat transplantation
tumors. A histochemical and morphometric study by use
of ricinus communis agglutinin I. Acta Neuropathol 76:
599605
Selivanoc V, Picard Y, Cadorette J, Rodrigue S, Lecomte R
(2000) Detector response models for statistical iterative
image reconstruction in high resolution PET. IEEE Trans
Nucl Sci 47:11681175
Stojiljkovic M, Piperski V, Dacevic M, Rakic L, Ruzdijic S,
Kanazir S (2003) Characterization of 9L glioma model of
the Wistar rat. J Neurooncol 63:17
Thorsen F, Ersland L, Nordli H, Enger PO, Huszthy PC,
Lundervold A, Standnes T, Bjerkvig R, Lund-Johansen M
(2003) Imaging of experimental rat gliomas using a clinical
MR scanner. J Neurooncol 63:225231
Wechsler W, Szymas J, Bilzer T, Hossmann KA (1989)
Experimental transplantation gliomas in the adult cat brain.
1.Experimental model and neuropathology. Acta Neurochir
98:7789
Whittle IR, Macarthur DC, Malcolm GP, Li M, Washington
K, Ironside JK (1998) Can experimental models of rodent
implantation glioma be improved? A study of pure and mixed
glioma cell line tumours. J Neurooncol 36:231242
Wree A, Goller HJ, Beck T (1995) Local cerebral glucose utilization in perfusion-fixed rat brains. J Neurosci Methods
58:143149
Chapter 35
Cellular and Molecular Characterization of Anti-VEGF

and IL-6 Therapy in Experimental Glioma
Sophie Javerzat and Martin Hagedorn
Abstract Vascular endothelial growth factor

(VEGF) and interleukin 6 (IL6) play central roles in
various angiogenesis-dependent diseases, including
malignant brain tumors. VEGF-blocking agents are
increasingly used against glioblastoma, but the potential of IL6 pathway inhibition is little explored in this
pathology. Here we resume recent work performed in
our recently developed experimental glioma model
grown on the chick chorio-allantoic membrane (CAM)
showing that (1) VEGF-inhibition leads to avascular
tumors which up-regulate genes associated with poor
survival in glioblastoma patients, (2) single inhibition
of VEGF or IL6 is equally effective in regard to
angiogenesis inhibition, activates different molecular
responses in tumor cells but increases invasion and (3)
combined inhibition of IL6 and VEGF prevents tumor
development without inducing invasion. Different
treatment and transcriptomic profiling strategies are
discussed with emphasis to alternative tumor models
such as the chicken CAM experimental glioma.
Keywords VEGF Interleukin 6 Gene profiling
Chick embryo CAM Anti-angiogenic therapies
Introduction
The capability of a growing cancer to attract new
blood vessels from the host (angiogenesis) is one of
M. Hagedorn ()
INSERM U1029, Universit Bordeaux 1, F-33405, Talence,
Cedex, France
e-mail: m.hagedorn@angio.u-bordeaux1.fr
the essential features of human tumor progression

(Hanahan and Weinberg, 2000). Blocking angiogenesis has now become a validated treatment paradigm to
increase efficacy of standard chemotherapy. Vascular
endothelial growth factor (VEGF) and its receptors
KDR (kinase insert domain receptor; also called
VEGFR-1), FLT1 and FLT4 (fms-related tyrosine
kinase 1 and 4; VEGFR-2 and -3) are central modulators of blood vessel growth and key targets of current
anti-angiogenic drugs. VEGF blockage is sufficient to
delay or inhibit tumor growth as well as formation
of metastasis in various animal models (Cheng et al.,
1996; Inoue et al., 2002). Inhibition of VEGF also
leads to tumor regression of established tumors in mice
(Huang et al., 2003) and to reduction of angiogenesis
in human rectal carcinoma (Willett et al., 2004).
However, targeting only one angiogenic factor
might not be sufficient to cure. There is evidence that
impairing two or more anti-angiogenic pathways in a
combinational therapy setting is more efficient than
monotherapy (Jendreyko et al., 2005; Saidi et al., 2009;
Strieth et al., 2006). Indeed, tumors may switch on
alternative pro-angiogenic pathways such as fibroblast growth factor (FGF) and bypass the need for
VEGF signaling (Casanovas et al., 2005), for review
see (Ferrara, 2010). Severe hypoxia caused by angiogenesis inhibitors may increase malignant potential of
cancer cells leading to tumor persistence or increased
vessel co-option and invasion (Kunkel et al., 2001;
Niola et al., 2006).
It is therefore of importance to comprehend how
tumor cells react to different anti-angiogenic therapies
on a molecular level. Such information should be useful to design novel inhibitors of individual molecules,
which are involved in anti-angiogenesis resistance.
Here, we describe an experimental system, which

361
362
allows to selectively study gene regulation in glioma

cells in response to anti-angiogenic mono (VEGF or
IL6) and combo (VEGF+IL6) therapy.
The Experimental Glioma Model in the

Chick Embryo: A Tool for Gene Profiling
Studies
Most existing in vivo GBM models are based on inoculation of glioma cells into rodent brains (Dai and
Holland, 2001) or the use of transgenic mice with modified expression of tumor suppressors or oncogenes in
glial cell lineages (Holland et al., 2000; Reilly et al.,
2000). However, rodent models require stereo-tactical
equipment, specific surgical techniques, costly animal
housing and raise ethical concerns.
We have developed an alternative experimental
glioma model based on grafting of the U87 glioma cell
line on the chick chorio-allantoic membrane (CAM).
The chorionic epithelium of the CAM constitutes a
natural barrier, which generally is difficult to penetrate for tumor cells. However, a high number of tumor
cells (35 million/CAM) and gentle laceration of the
CAM surface allow reproducible tumor development
(Hagedorn et al., 2005). Furthermore, tumor cells are
placed in a plastic ring, which restricts displacement of
the cell suspension during movements of the embryo.
The implantation of tumor cells is generally done on
day 10 of embryonic development, when the CAM
has reached a sufficient size, making displacement of
the plastic ring a rare occurrence. Developing tumors
can be examined on a daily basis and are well vascularized 4 days after implantation (see Fig. 35.1a).
Experimental tumors reproducibly form in a VEGFdependent manner, and recapitulate key morphological
features of glioblastoma in patients such as hemorrhage, necrosis and angiogenesis. On a molecular level,
newly formed chick blood vessels inside the tumor
become positive for VEGFR-2 from the second day
on, as evidenced by in situ hybridization using a probe
for the avian VEGFR-2 homolog Quek2 (Hagedorn
et al., 2005). A typical experimental glioma 4 days
after implantation is shown in Fig. 35.1b (left panel).
The rapid development of experimental glioma on
the CAM and the close resemblance to important
histopathological events observed in patients suggest
that growth may be accompanied by pertinent changes
S. Javerzat and M. Hagedorn
in gene expression pattern. An advantage of the model

is a perfect visual control of the tumor material used
for gene expression studies. The tumor is excised
in vivo under a bio-microscope using sterile scissors
and immediately placed in liquid nitrogen. This eliminates contamination with adjacent, non-relevant tissue,
which is a major point if tumor material needs to be
analyzed by sensitive molecular biology methods such
as semi-quantitative real-time PCR or microarrays.
Experimental gliomas are hybrid tissues, composed
of human tumor cells and chicken host cells, including, but not limited to endothelial cells, pericytes and
immune cells. Using species-specific GeneChips, it
is therefore possible to selectively study gene regulation of one or the other compartment (host vs. tumor
cells). This approach maximizes the chance to identify genes, which are specifically regulated in tumor
cells in response to anti-angiogenic treatments. On
the other hand, if one is interested in the molecular
reaction of the stroma during tumor progression, the
use of the chicken GeneChips helps to focus on these
genes (Dumartin et al., 2010). To illustrate the discrimination level of species-specific microarrays, we
mixed human mRNA isolated from U87 cells with
increasing amounts of chicken mRNA isolated from
the CAM. These mixtures (ratios indicated as percent
in the figure legend) were then hybridized in parallel on human and chicken Affymetrix GeneChips
(Fig. 35.1c). Chicken CAM mRNA on a human chip
hybridizes to less than 5% of the probes representing
human genes. The same result was obtained when U87
mRNA was incubated with the chicken GeneChip. We
estimate the amount of chick tissue (mainly blood vessels) accounts for about 10% of the tumor mass in the
experimental glioma. We therefore also tested different ratios of human and chick mRNA on the two chips,
keeping the amount of total mRNA constant. When
chick and human mRNA mixes were hybridized on the
human chip, the percentage of significantly detected
transcripts decreases in a dose-dependent manner. The
more chick mRNA was present in the mix, the less
positive detection on the human chip occurred. This
shows that chick mRNA does not hybridize to the
human chip in a non-specific manner. As expected,
increasing quantity of human mRNA in the mixture
decreases detection if hybridized on the chicken chips.
These results demonstrate that species-specific gene
expression changes can be detected in this system. The
experimental glioma model on the CAM thus allows
35 Cellular and Molecular Characterization of Anti-VEGF and IL-6 Therapy
363
Timeline of a typical experiment

Tumor angiogenesis
Transfection/
transduction of
tumor cells
Implantation
Day
Tissue isolation
Exp. end
Observations / treatments
Phenotypes observed after 4 days
Control
siRNA VEGF
Avastin
Considerations for gene profiling in the model
Human U87 cells
CAM
mRNA isolation and mixing
% positive signals
Hybridization
on GeneChips
Fig. 35.1 (a) Typical set-up of an experimental glioma experiment on the CAM. On embryonic day 10, cells are implanted
on the CAM. The cell suspension reforms a tumor, which starts
to acquire blood vessels approximately 2448 h after implantation. Tumors can be treated from the second day on and
photographed in vivo. At day 4, tumors can be isolated for histology or gene profiling studies. Alternatively, the tumor can
be let grown until day 7, when the experiments needs to be
ended due to the near hatching of the chick embryo. (b) Typical
results after 4 days of growth. Control tumors show strong angiogenic response, already visible by slight magnification, whereas
anti-VEGF tumors are small and appear white, due to the near
absence of blood vessels. (c) Advantage of the chick glioma
model for gene profiling studies. The experimental glioma is a
hybrid tissue, containing both mRNAs regulated in the human
cells as well as angiogenic blood vessels from the chick. Chick
and human mRNAs isolated were mixed at indicated ratios and
hybridized in parallel on both chip types. Note low significant detection of chick mRNA on the human chip (arrow) and
vice versa. Hybridization occurs in a dose-dependent manner,
in a similar manner on both chips (C=chick mRNA, H=human
mRNA)
364
fast and precise analysis on the cellular and molecular

level of major steps of human GBM progression and
response to different treatment with new anti-cancer
drugs.
Interfering with Angiogenic Regulators

in the Experimental Glioma: The
Example of VEGF
To simulate a defined anti-angiogenic setting in our
model, we transfected U87 glioma cells with shortinterfering RNAs (siRNAs) targeting all isoforms of
VEGFA. Efficacy of siRNAs has been evaluated in
vitro using semi-quantitative real-time PCR and commercially available ELISAs. siRNA-mediated blocking of VEGF has initially been chosen to circumvent
problems related to drug delivery and control of doses
in the chick system. However, more recent results in
our laboratory have shown that Avastin at daily topical administration of 100 g per CAM results in
strong reduction of tumor growth and vascularization,
comparable to siRNA-mediated effects (Fig. 35.1b).
Intriguingly, small tumors still formed albeit of a
whitish appearance as evidenced by biomicroscopy
(Fig. 35.1b). Histological analysis revealed that VEGF
knock down leads to almost complete absence of
functional blood vessels inside experimental glioma.
However, although siRNA-mediated knock down of
VEGF is extremely efficient and varies few within
one experiment, some heterogeneity in the degree of
response may occur. Over the years, we observed slight
differences in the phenotypes depending on the time,
ranging from series of tumors with no blood vessels
at all to tumors with some residual vessels. These differences are most likely due to the variable genetic
background of chick embryos.
Gene Profiling of Anti-Angiogenic

Therapy: A Glimpse of Resistance
The fact that glioma cells are still able to survive on
the CAM after VEGF withdrawal and form little
avascular tumors, suggests that angiogenesisindependent tumor growth has taken place. Tumor
cells growing in such conditions are often exposed
to metabolic stress such as hypoxia, lack of serum
and other nutrients. Hypoxia will drive cancer cells to
adapt to this unfavorable environment and eventually
become more aggressive. On a molecular level,
this means over expression of genes, which render
tumor cells resistant to apoptosis and resistant to
radiotherapy (Jensen, 2009; Knisely and Rockwell,
2002). To get more insight into the genes over
expressed in these resistant tumors, we undertook
a simple gene profiling experiment where a pool of
typical avascular glioma under VEGF knockdown
(VEGF-KD) was compared to a pool of control tumors
(Saidi et al., 2008). We then investigated whether
genes induced in VEGF-KD tumors could be related
to disease aggressiveness. Several results support
this hypothesis. IL1B, CHI3L2 and PI3 (encoding
for interleukin 1 beta, chitinase 3-like 2 and elafin,
respectively) were significantly co-regulated with
CHI3L1 encoding YKL-40, a known marker for
disease progression (Pelloski et al., 2005) in a cohort
of patient glioblastoma, suggesting functional linkage
of these factors. More, CHI3L1, CHI3L2, IL1B and
PI3 were significantly over expressed in high-grade
glioma (grade IV, glioblastoma) compared to grade
II astrocytoma (Fig. 35.2a). Most important, high
PI3 mRNA levels were significantly associated with
shorter survival in our patient cohort (Fig. 35.2b).
Immunhistochemical detection of elafin protein
could be found in hypoxic areas in gliobastoma
sections from patients (Fig. 35.2c, d), but was not
present in low-grade glioma (Fig. 35.2e). The linkage
between metabolic stress and PI3 gene expression
could also be evidenced in vitro: both, hypoxia and
low-serum induced PI3 mRNA in U87 cells, up to
30-fold. Very recently, a link between elafin expression
and poor prognosis has also been validated in ovarian
cancer (Clauss et al., 2010).
This simple experiment gives a snapshot of molecular adaptation of glioma cells exposed to anti-VEGF.
Genes induced in this condition are potentially mediators of disease progression, as they are associated with
higher tumor grade and worse prognosis.
Fig. 35.2 (a) mRNA levels of selected genes in GBM (n=42,

except for PI3, n=38) and low-grade (LG) glioma (n=10, except
for CHI3L1 n=9 and PI3, n=8) were compared using the MannWhitney U-test. Medians of mRNA levels for CHI3L1, CHI3L2,
IL1B and PI3 differed significantly between GBMs and LG
glioma. (b) High levels of PI3 (encoding elafin) are significantly associated with poor survival as determined by analysis of
Kaplan and Meier survival curves. Three censored subjects are
365
indicated by symbols. (ce) Immunohistochemistry with antielafin antibody in GBMs. Elafin-positive tumor cells are mainly
seen around necrotic areas (N), no staining is observed in lowgrade gliomas (original magnification: A and B 100, CE
400). (f) Malignant glioma cells in vitro strongly expressed
PI3/elafin 24 and 48 h after serum deprivation or exposure to
hypoxia (Mean SD)
366
Pre-clinical Evaluation of New

Candidates and Combined Therapies
in the Experimental Glioma
As discussed above, RNA interference directed against
VEGF models anti-angiogenesis in the experimental
glioma and allows evaluating tumor response both at
the morphological and molecular level. If a particular gene needs to be studied more in detail, it may be
preferable to create stably transduced cells with lentiviral vectors carrying siRNA precursors. It is however
advisable to first test transiently transfected cells and
evaluate phenotypes. In our hands, both methods show
comparable knock down efficacies in regard to transcript reduction. Vectorization may be an advantage
for cost purposes when many different assays with the
same interfering RNA are required.
RNA interference in the experimental glioma model
is ideal for investigating alternative tumor-specific targets. Another group has successfully demonstrated
inhibition of tumor growth and angiogenesis using siRNAs against OPN (osteopontin) or HDAC4 (Lamour V
et al., 2010, Mottet et al., 2009).
We chose interleukin-6 as a potential candidate
to be evaluated in the model. Inhibition of the IL-6
pathway is already broadly explored in chronic inflammatory diseases such as rheumatoid arthritis. Several
lines of evidence suggest that interfering with IL-6
may be a promising rationale for anti-glioma therapy.
Mice expressing GFAP-v-src fail to develop glioma
on an IL6/ background (Weissenberger et al., 2004).
Patient glioma produce large amounts of interleukin6 in culture and in vivo (Van Meir et al., 1990), its
expression is significantly higher in patients with poor
prognosis and amplification of the IL6 locus is associated with shorter survival (Chang et al., 2005; Tchirkov
et al., 2001). At some point the IL-6 and VEGF axis
converge in glioma cells, as interleukin-6 stimulates
VEGF production in vitro and in vivo via the signal transducer and activator of transcription 3 (STAT3)
pathway (Loeffler et al., 2005).
The experimental glioma on the CAM is particularly suited to challenge IL6 dependent pathways,
as U87 cells positively regulate the transcription of
the IL6 gene when the tumor switches to the angiogenic phenotype (Hagedorn et al., 2005). Because
VEGF and IL6 cross-regulate each other while functioning independently, inhibiting one pathway at a time
or both together can give valuable information on

a possible synergistic antitumoral effect. We showed
that knock down of each of the gene reduces corresponding transcripts in the experimental tumor with
>95% efficiency and also observed cross-inhibiton
at around 50%. Combined knock down resulted in
tumors with hardly detectable amounts of VEGF
and IL6.
On a morphological level, combined inhibition of
IL6 and VEGF is significantly more efficient than
single inhibition as evidenced by reduced size, proliferation and neovascularization of the tumors. This
comparative analysis can be easily led on significant numbers of tumors (n>20 per group) providing
robust quantitative data. On the left panel of Fig. 35.3,
we show an example of vascular changes to different treatments. In anti-IL6 or anti-VEGF conditions
alone a similar reduction in tumor angiogenesis can be
observed, compared to control tumors. Combinatory
treatment completely inhibits vascularization of the
tumors.
Interestingly, the experimental glioma allows investigating invasive properties that may be enhanced
under anti-angiogenic treatment, a feature that neurooncologists are particularly sensitized to when treating
glioma patients with Avastin (de Groot et al., 2010).
In control glioma, tumor cells are rarely detected at
distance from the primary implantation site in the supporting CAM. Under metabolic/hypoxic stress induced
by IL6 or VEGF knockdown, tumor cells invade the
CAM tissue (Fig. 35.3 right panels), with a 34 fold
increase in the number of tumor cells detected in the
CAM. Different modes of invasion can be observed.
In the anti-VEGF condition, cells invade in a collective manner, whereas IL6 knock down mostly induces
single-cell spreading. In combined treated tumors,
the invasive phenotype was significantly reduced, to
a similar degree as seen in controls. These results
have been confirmed in an orthotopic mouse glioma
model.
The experimental glioma is a fast, cost-effective
alternative to screen different combination of
inhibitors, including siRNAs for their effect on
standard tumor parameters such as angiogenesis, but
also recapitulates biological phenomena observed
in mammalian brains such as induction of infiltration after treatment with single angiogenesis
inhibitors.
367
Fig. 35.3 Left panel: representative illustrations of angiogenic

and invasive phenotypes revealed by triple immunofluorescence
staining (blue: DAPI; red: anti-vimentin; green: SNA-lectin;
scale bar left panels = 50 m; right panels = 100 m). Note
that control tumors are characterized by the presence of a dense
immature vascular network. Knockdown of IL6 and/or VEGF
strongly decreases tumor vascularization and combined knockdown is even more efficient at reducing vascular density. Right
panel: invasive behavior of tumor cells in the CAM. Sections

of tumor and surrounding CAM tissue 7 days after implantation were stained with DAPI (blue) and anti-vimentin (red).
A representative field of the margin between tumor (asterisk)
and the CAM (dashed line) is shown. Both, anti-IL6 and antiVEGF cause increased invasion in the CAM stroma, whereas
IL6/VEGF combined knockdown cells barely invade the CAM
Exploring Molecular Pathways

Associated with Inhibition
tissue than murine models and cross hybridization

between chicken and human sequences is negligible
(see Fig. 35.1c). Tumors with or without treatment
(single or double knock down) harbored a specific gene
signature. As the phenotypes of anti-VEGF and antiIL6 treated tumors are very similar, with an equivalent
reduction in tumor volume and vascularization, and
As detailed previously, the CAM experimental glioma

is well suited to perform comparative gene profiling as it allows more precise isolation of tumor
368
the two induced pathways are partially overlapping,

highlighting differences at the level of the molecular
responses may be predictive for a better therapeutic
outcome. For instance there were twice as many genes
down regulated in anti-IL6 tumors compared to antiVEGF tumors. In addition and as anticipated from the
small volumes of the tumors, we found out that the
cell cycle machinery is severely affected when both
VEGF and IL6 pathways are inhibited. Around 25 positive regulators of mitosis and 10 major components
of chromatin assembly (mainly histone subunits) are
significantly down regulated in the combined treatment. Interestingly, many genes involved in invasion
were over expressed in tumors after single inhibition,
but were strongly inhibited when VEGF and IL6 were
inhibited. Amongst those factors, CXCL12/SDF1 (Tu
et al., 2009) and PDPLN (Wicki and Christofori, 2007)
merit further investigation as potential markers of poor
prognosis and inducers of invasive behavior glioma
cells.
We showed that the experimental glioma on the
CAM allows to predict the efficiency (reduction
of tumor growth) and adverse effect (induction of
invasiveness) of anti-VEGF and anti-IL6 pathways
inhibitors. Humanized antibodies specifically raised
against Interleukin-6 receptor are available and used
in the clinic for rheumatoid arthritis (Tocilizumab,
Actemra, Roche) but no trials have yet been engaged
to estimate anti-glioma effects of this new drug. We
are currently investigating effects on tumor progression and infiltration of a combination of Avastin and
Tocilizumab in mouse glioma models.
Future Directions
The experimental glioma model is a useful tool to predict treatment efficacy of new anti-angiogenic drugs
and to validate new targets. An advantage of the
model is fast development of tumors and the possibility to visually evaluate treatment efficacy. There are
a growing number of humanized antibodies, aptamers
and small molecules inhibitors of pro-angiogenic key
factors. It will therefore be of interest to test different combinations of these drugs to find a treatment
paradigm that efficiently blocks tumor vascularization and invasion at the lowest dose possible. This
combination is likely to give similar results in other
pre-clinical models and ultimately may be of benefit for glioblastoma patients. The experimental glioma
model on the CAM might considerably speed up the
search for efficient anti-glioma treatments.
Acknowledgements This work was supported La Ligue Contre
le Cancer, Comite de la Dordogne (to SJ) and lAgence
Nationale de la Recherche, ANR (Glioma Model, JC05_0060,
to MH).
References
Casanovas O, Hicklin DJ, Bergers G, Hanahan D (2005) Drug
resistance by evasion of antiangiogenic targeting of VEGF
signaling in late-stage pancreatic islet tumors. Cancer Cell
8:299309
Chang CY, Li MC, Liao SL, Huang YL, Shen CC, Pan HC
(2005) Prognostic and clinical implication of IL-6 expression
in glioblastoma multiforme. J Clin Neurosci 12:930933
Cheng SY, Huang HJ, Nagane M, Ji XD, Wang D, Shih CC,
Arap W, Huang CM, Cavenee WK (1996) Suppression of
glioblastoma angiogenicity and tumorigenicity by inhibition
of endogenous expression of vascular endothelial growth
factor. Proc Natl Acad Sci USA 93:85028507
Clauss A, Ng V, Liu J, Piao H, Russo M, Vena N, Sheng Q,
Hirsch MS, Bonome T, Matulonis U, Ligon AH, Birrer MJ,
Drapkin R (2010) Overexpression of elafin in ovarian carcinoma is driven by genomic gains and activation of the nuclear
factor kappaB pathway and is associated with poor overall
survival. Neoplasia 12:161172
Dai C, Holland EC (2001) Glioma models. Biochim Biophys
Acta 1551:M19M27
de Groot JF, Fuller G, Kumar AJ, Piao Y, Eterovic K, Ji
Y, Conrad CA (2010) Tumor invasion after treatment of
glioblastoma with bevacizumab: radiographic and pathologic
correlation in humans and mice. Neuro Oncol 12:233242
Dumartin L, Quemener C, Laklai H, Herbert J, Bicknell R,
Bousquet C, Pyronnet S, Castronovo V, Schilling MK,
Bikfalvi A, Hagedorn M (2010) Netrin-1 mediates early
events in pancreatic adenocarcinoma progression, acting
on tumor and endothelial cells. Gastroenterology 138:
15951606
Ferrara N (2010) Pathways mediating VEGF-independent tumor
angiogenesis. Cytokine Growth Factor Rev 21:2126
Hagedorn M, Javerzat S, Gilges D, Meyre A, de Lafarge B,
Eichmann A, Bikfalvi A (2005) Accessing key steps of
human tumor progression in vivo by using an avian embryo
model. Proc Natl Acad Sci USA 102:16431648
Hanahan D, Weinberg RA (2000) The hallmarks of cancer. Cell
100:5770
Holland EC, Celestino J, Dai C, Schaefer L, Sawaya RE, Fuller
GN (2000) Combined activation of Ras and Akt in neural progenitors induces glioblastoma formation in mice. Nat
Genet 25:5557
Huang J, Frischer JS, Serur A, Kadenhe A, Yokoi A, McCrudden
KW, New T, OToole K, Zabski S, Rudge JS, Holash

J, Yancopoulos GD, Yamashiro DJ, Kandel JJ (2003)
Regression of established tumors and metastases by potent
vascular endothelial growth factor blockade. Proc Natl Acad
Sci USA 100:77857790
Inoue M, Hager JH, Ferrara N, Gerber HP, Hanahan D (2002)
VEGF-A has a critical, nonredundant role in angiogenic
switching and pancreatic beta cell carcinogenesis. Cancer
Cell 1:193202
Jendreyko N, Popkov M, Rader C, Barbas CF 3rd (2005)
Phenotypic knockout of VEGF-R2 and Tie-2 with an intradiabody reduces tumor growth and angiogenesis in vivo. Proc
Natl Acad Sci USA 102:82938298
Jensen RL (2009) Brain tumor hypoxia: tumorigenesis, angiogenesis, imaging, pseudoprogression, and as a therapeutic
target. J Neurooncol 92:317335
Knisely JP, Rockwell S (2002) Importance of hypoxia in the
biology and treatment of brain tumors. Neuroimaging Clin
N Am 12:525536
Kunkel P, Ulbricht U, Bohlen P, Brockmann MA, Fillbrandt
R, Stavrou D, Westphal M, Lamszus K (2001) Inhibition
of glioma angiogenesis and growth in vivo by systemic treatment with a monoclonal antibody against vascular endothelial growth factor receptor-2. Cancer Res 61:
66246628
Lamour V, Le Mercier M, Lefranc F, Hagedorn M, Javerzat
S, Bikfalvi A, Kiss R, Castronovo V, Bellahcene A (2010)
Selective osteopontin knockdown exerts anti-tumoral activity in a human glioblastoma model. Int J Cancer 126:
17971805
Loeffler S, Fayard B, Weis J, Weissenberger J (2005)
Interleukin-6 induces transcriptional activation of vascular
endothelial growth factor (VEGF) in astrocytes in vivo and
regulates VEGF promoter activity in glioblastoma cells via
direct interaction between STAT3 and Sp1. Int J Cancer
115:202213
Mottet D, Pirotte S, Lamour V, Hagedorn M, Javerzat S, Bikfalvi
A, Bellahcene A, Verdin E, Castronovo V (2009) HDAC4
represses p21(WAF1/Cip1) expression in human cancer
cells through a Sp1-dependent, p53-independent mechanism.
Oncogene 28:243256
Niola F, Evangelisti C, Campagnolo L, Massalini S, Bue MC,
Mangiola A, Masotti A, Maira G, Farace MG, Ciafre SA
(2006) A plasmid-encoded VEGF siRNA reduces glioblastoma angiogenesis and its combination with interleukin-4
blocks tumor growth in a xenograft mouse model. Cancer
Biol Ther 5:174179
369
Pelloski CE, Mahajan A, Maor M, Chang EL, Woo S, Gilbert M,

Colman H, Yang H, Ledoux A, Blair H, Passe S, Jenkins RB,
Aldape KD (2005) YKL-40 expression is associated with
poorer response to radiation and shorter overall survival in
glioblastoma. Clin Cancer Res 11:33263334
Reilly KM, Loisel DA, Bronson RT, McLaughlin ME, Jacks T
(2000) Nf1; Trp53 mutant mice develop glioblastoma with
evidence of strain-specific effects. Nat Genet 26:109113
Saidi A, Hagedorn M, Allain N, Verpelli C, Sala C, Bello
L, Bikfalvi A, Javerzat S (2009) Combined targeting of
interleukin-6 and vascular endothelial growth factor potently
inhibits glioma growth and invasiveness. Int J Cancer
125:10541064
Saidi A, Javerzat S, Bellahcene A, De Vos J, Bello L, Castronovo
V, Deprez M, Loiseau H, Bikfalvi A, Hagedorn M (2008)
Experimental anti-angiogenesis causes upregulation of genes
associated with poor survival in glioblastoma. Int J Cancer
122:21872198
Strieth S, Eichhorn ME, Sutter A, Jonczyk A, Berghaus A,
Dellian M (2006) Antiangiogenic combination tumor therapy
blocking alpha(v)-integrins and VEGF-receptor-2 increases
therapeutic effects in vivo. Int J Cancer 119:423431
Tchirkov A, Rolhion C, Bertrand S, Dore JF, Dubost JJ, Verrelle
P (2001) IL-6 gene amplification and expression in human
glioblastomas. Br J Cancer 85:518522
Tu H, Zhou Z, Liang Q, Li Z, Li D, Qing J, Wang H, Zhang
L (2009) CXCR4 and SDF-1 production are stimulated by
hepatocyte growth factor and promote glioma cell invasion.
Onkologie 32:331336
Van Meir E, Sawamura Y, Diserens AC, Hamou MF, de Tribolet
N (1990) Human glioblastoma cells release interleukin 6 in
vivo and in vitro. Cancer Res 50:66836688
Weissenberger J, Loeffler S, Kappeler A, Kopf M, Lukes A,
Afanasieva TA, Aguzzi A, Weis J (2004) IL-6 is required
for glioma development in a mouse model. Oncogene 23:
33083316
Wicki A, Christofori G (2007) The potential role of podoplanin
in tumour invasion. Br J Cancer 96:15
Willett CG, Boucher Y, di Tomaso E, Duda DG, Munn LL,
Tong RT, Chung DC, Sahani DV, Kalva SP, Kozin SV,
Mino M, Cohen KS, Scadden DT, Hartford AC, Fischman
AJ, Clark JW, Ryan DP, Zhu AX, Blaszkowsky LS, Chen
HX, Shellito PC, Lauwers GY, Jain RK (2004) Direct
evidence that the VEGF-specific antibody bevacizumab
has antivascular effects in human rectal cancer. Nat Med
10:145147
Chapter 36
Adult Brainstem Gliomas: Diagnosis and Treatment

Florence Laigle-Donadey and Jean-Yves Delattre
Abstract Adult brainstem gliomas constitute a heterogeneous group of tumors. A better radiological
analysis of these tumors will improve their classification and help to better distinguish prognosis subgroups.
New MRI techniques may also contribute to differential diagnosis and help neurosurgeons in removing resectable brainstem tumors. However, biopsy
remains indicated in many contrast enhancing brainstem masses in adults because of the great variety of
differential diagnosis. Conventional radiotherapy is the
standard of care of brainstem glioma and chemotherapy has been disappointing to date. Given the lack
of efficacy of conventional drugs, a better understanding of the biology of this tumor is the key to develop
targeted therapy. In the future, advances in diagnostic and treatment modalities will probably result
in improvement in the pattern of care of brainstem
gliomas which remain now associated with a poor
prognosis.
Keywords Brainstem gliomas MRI techniques
Differential diagnosis Exophytic gliomas Drugs
Radiotherapy
F. Laigle-Donadey ()
Service de Neurologie Mazarin, Hpital de la Piti-Salptrire
APHP, 75651 Paris Cedex 13, France
e-mail: Florence.laigle-donadey@psl.aphp.fr
Introduction
Brainstem tumors occur in a region located between
the mesencephalon and the medulla. This definition excludes patients with tumors originating in the
thalamus and hypothalamus, or lesions originating
from the cerebellum, cerebellar peduncles, upper cervical spinal cord, as well as tumors arising from
ventricles.
Brainstem glioma is the most frequent tumor of
the region but it constitutes a heterogeneous group
of tumors with variable prognoses. A bimodal age
distribution has been noted, allowing a clear distinction between BS gliomas in children and adults
that differ in their clinical and radiological presentation, histology, biological behaviour, and clinical
course.
This chapter will focus on adult BS gliomas which
is a rare disease. Indeed, in contrast with in children,
where brainstem tumors are frequent, accounting for
approximately 1020% of all pediatric CNS tumors,
brainstem gliomas are much less frequent in adults,
accounting for 22.5% of all intracranial tumors in
adults (White, 1963).
As for children, tissue confirmation is frequently not
feasible unless an exophytic component exists, because
of the risks associated with direct histologic examination (Packer et al., 1985) and their classification relies
mainly on clinical presentation and MRI characteristics (Guillamo et al., 2001a). A striking difference
between adults and children is the better prognosis of the diffuse intrinsic form in adults (Landolfi
et al., 1998; Selvapandian et al., 1999; Guillamo et al.,
2001a).

371
372
Diffuse Intrinsic or Diffusely

Infiltrative Low Grade Brainstem
Gliomas
Epidemiology
Diffuse intrinsic brainstem gliomas represent the most
frequent type of brainstem glioma in adults, accounting
for 4550% of cases (Guillamo et al., 2001a; Salmaggi
et al., 2008). These tumors usually occur in young
patients, between the age of 20 and 30 years.
Pathology
Biopsy is rarely performed in typical intrinsic diffuse
low-grade glioma in adults. When a biopsy is performed, a benign histology (grade II glioma) is found
in up to 82% of cases in adults, in contrast with children (Guillamo et al., 2001a, b) which is a likely
explanation for the better survival of the diffuse intrinsic form in adults (Guillamo et al., 2001a; Salmaggi
et al., 2008). For this reason, Guillamo et al. suggest
that this subgroup of tumor should be named diffuse intrinsic low-grade brainstem gliomas in adults.
However, adult patients can rarely present a rapidly
growing tumor similar to diffuse intrinsic BSG of
children (Guillamo et al., 2001a).
F. Laigle-Donadey and J.-Y. Delattre
pseudomyasthenic presentation (Dirr et al., 1989), as

well as orthostatic hypotension (Hsu et al., 1984).
Radiological Presentation
MRI reveals a pattern similar to the pattern observed
in children with a diffuse enlargement of the brainstem, with high signal on T2-weighted images and
low signal on T1-weighted images, located in the pons
(70%) or in the medulla (30%). The lesion does not
enhance after contrast infusion (Freeman and Farmer,
1998; Guillamo et al., 2001b). MRS showed an elevation of the Cho/NAA ratio at diagnosis in all nine
patients from a cohort that was investigated with MRS
(Fig. 36.1) (Salmaggi et al., 2008).
Treatment
Neurosurgery
Biopsy is rarely performed in typical intrinsic diffuse low-grade glioma in adults (Guillamo et al.,
2001b). However, this remains a matter of debate.
Some authors consider that guided stereotactic biopsy
Clinical Presentation
Despite this histological difference, there are no major
differences in clinical presentation between adults
and children (Selvapandian et al., 1999), except for
a longer duration of symptoms and signs in adults,
(Landolfi et al., 1998; Selvapandian et al., 1999).
Patients typically present with a long-lasting history of
facial palsy.
Facial myokymia (Selvapandian et al., 1999), as
well as hemifacial spasm (Elgamal and Coakham,
2005) may occur. Other symptoms rarely described
include snoring associated with Ondines curse in the
case of medulla oblongata glioma (Marin-Sanabria
et al., 2006), central neurogenic hyperventilation
(Gaviani et al., 2005), tongue tremor (Saka et al.,
2006), visual auras and migraine (Lim et al., 2005),
Fig. 36.1 Axial flair MR in a 35 year-old patient presenting with diplopia. Note diffuse swelling and increased signal
intensity due to intrinsic pontine glioma
36 Adult Brainstem Gliomas: Diagnosis and Treatment
is relatively safe and believe that an accurate diagnosis should be obtained in all cases and not only
when atypical clinical or imaging features suggesting
a non-neoplastic lesion (Thomas et al., 1988).
Resection of intrinsic diffuse brainstem tumors
remains impossible despite technical advances (Tanaka
et al., 2005; Mursch et al., 2005).
A stereotaxic approach can be a rapid and safe
method for evacuation of the contents of cysts (Thomas
et al., 1988; Rajshekhar and Chandy, 1995), providing neurological benefit in most cases. A stereotactic
placement of an Ommaya reservoir after reaccumulation of the cyst can also be performed.
Patients with hydrocephalus may require ventriculostomy or ventriculoperitoneal shunting for symptomatic relief.
Radiotherapy
Focal radiotherapy is the standard treatment for
adult brainstem gliomas and can improve or stabilize
patients for years. The conventional dose of radiotherapy ranges from 54 to 60 Gy. Diffuse intrinsic
brainstem glioma seems to be more responsive to
radiotherapy in adults than in children. However there
is an important discrepancy between the clinical and
the radiological response. In one series, a majority
of patients (62%) clinically improved for a long
period (over 6 months) after radiotherapy while radiological response was noted in only 19% of cases
(Guillamo et al., 2001a). The optimal date of treatment remains unknown. Indeed, some patients may
have mild symptoms during extended periods without
any treatment, suggesting that radiotherapy may sometimes be deferred until clear evidence of symptomatic
tumor progression (Guillamo et al., 2001a).
373
radiological result in one case report of progressive diffuse BS glioma treated with bevacizumab and irinotecan (Torcuator et al., 2009).
Symptomatic Treatment
Patients with difficulties in swallowing may need gastrostomy, such as percutaneous esophagogastrostomy.
Steroids are frequently administered to brainstem
tumor patients. However, they should be used sparingly, at the lowest necessary dosage, because of their
numerous side effects.
Follow-up
The frequently indolent growth pattern of the intrinsic
form in adults contrasts with the natural history of this
disease in children and is likely related to a lower grade
of the tumors. Median survival of diffuse intrinsic lowgrade gliomas in adults is 7.3 years, as compared to
less than 1 year in children.
Two recent retrospective studies have evaluated
the characteristics from adult BSGs. In an American
cohort including 101 adult patients, the overall survival at 5 and 10 years was 58 and 41%, respectively,
with a median survival of 85 months (range 1228)
(Kesari et al., 2008). In an Italian cohort of 34 adult
patients, the median overall survival was 59 months
(Salmaggi et al., 2008). However, it is worth noting that
rarely some adult patients can present with a rapidly
growing tumor similar to children with diffuse intrinsic
brainstem glioma (Guillamo et al., 2001a).
Malignant Brainstem Gliomas

Epidemiology
Chemotherapy
Efficacy of chemotherapy in adult brainstem gliomas
remains unproven and adjuvant chemotherapy cannot
be currently recommended. Moreover, the effectiveness of chemotherapy on relapse is uncertain, although
it may benefit some patients, sometimes with a frank
clinical improvement contrasting with a lack of radiological response. The role of new therapies such as
anti-angiogenic drugs need to be defined in this case.
Torcuator et al. reported an interesting clinical and
Malignant brainstem glioma accounts for 3139% of

adults with BSG (Guillamo et al., 2001a; Salmaggi
et al., 2008) and appear in older patients, usually after
the age of 40.
Pathology
Histological examination shows high-grade (III, IV)
gliomas in all cases.
374
Clinical Presentation
The clinical history is subacute, consisting of cranial palsies and long tract signs, rapidly leading to an
altered performance status.
Careful systemic and neurologic work up usually

leads to the correct diagnosis. However, a biopsy
should be considered in many contrast enhancing
brainstem masses when this work up is not conclusive (Abernathey et al., 1989; Rajshekhar and Chandy,
1995).
Radiological Presentation
Treatment
Contrast enhancement was found in 100% of cases
upon initial MRI in Guillamos series, and often exhibited a ring-like pattern (Guillamo et al., 2001a), suggesting central necrosis. However, the radiological
picture of malignant BS glioma is non specific and
justifies histological confirmation in most cases. In
previous studies, preoperative radiological diagnoses
were found to be incorrect in 1025% of cases in
patients over 20 years of age who presented with a
contrast enhancing lesion in the brainstem (Fig. 36.2)
(Rajshekhar and Chandy, 1995; Boviatsis et al., 2003).
Differential Diagnoses
The differential diagnosis of malignant brainstem
gliomas is often difficult and includes many various
diseases (Table 36.1).
Fig. 36.2 Axial T1-weighted MR after gadolinium infusion in

an adult patient with an exophytic brainstem glioma showing
post gadolinium enhancement
Surgery
In order to avoid the complications of an empiric
and inappropriate therapy, biopsy is indicated in many
contrast enhancing brainstem masses because the histological variety is much more important in adults than
in childhood (see paragraph above) (Abernathey et al.,
1989; Kratimenos and Thomas, 1993; GoncalvesFerreira et al., 2003; Shad et al., 2005). Moreover,
a benign, non-neoplastic pathology was diagnosed in
1317% of adults with a suspected malignant brainstem mass (Rajshekhar and Chandy, 1995).
Although it is difficult to estimate, the risk of
severe complications of a stereotactic biopsy for a suspected malignant brainstem glioma is considered to
be low (probably between 1 and 5.6%) (Rajshekhar
and Chandy, 1995; Boviatsis et al., 2003). Usually, the
lower the lesion is located in the brainstem, the greater
the risks involved. Operative mortality is exceedingly
rare (Thomas et al., 1988; Kratimenos and Thomas,
1993; Boviatsis et al., 2003).
Morbidity is usually minimal and temporary,
consisting of transient neurological deterioration
(Kratimenos and Thomas, 1993; Goncalves-Ferreira
et al., 2003), or obstructive hydrocephalus. Permanent
neurological deficit was estimated to occur in about
1.4% in Rajshekhars series (Rajshekhar and Chandy,
1995) and some patients required early re-aspiration of
a haematoma (Kratimenos and Thomas, 1993).
The optimal method and routes of biopsy are controversial (Abernathey et al., 1989; Kratimenos and
Thomas, 1993; Goncalves-Ferreira et al., 2003).
CT-or MRI-guided stereotactic surgery of the
brainstem appears safe and reliable (Rajshekhar and
Chandy, 1995; Massager et al., 2000; Boviatsis et al.,
2003). PET-scan may be helpful (Massager et al.,
2000). Biopsy provides a high yield of positive histological diagnosis, estimated in the literature around
375
Table 36.1 Main differential diagnoses of malignant brainstem gliomas

Other tumors
Infections
Vascular processes
Metastasis (Friedman et al., 1989)

Lymphoma (Friedman et al., 1989;
Goncalves-Ferreira et al., 2003)
Germinoma and other intracranial
germ cell tumors (Friedman et al.,
1989)
Ganglioglioma (Goncalves-Ferreira
et al., 2003)
Ependymoma (Abernathey et al.,
1989)
Dysembryoplastic neuroepithelial
tumor (Fujimoto et al., 2000)
Chordoma (Friedman et al., 1989)
Neurocytoma (Swinson et al., 2006)
Dermoid and epidermoid cysts
(Rajshekhar and Chandy, 1995)
Cystic trochlear nerve neurinoma
(Shenoy and Raja, 2004)
Acoustic neuroma (Yousry et al.,
2004)
Tuberculomas
(Rajshekhar and Chandy,
1995)
Toxoplasmosis
(Goncalves-Ferreira
et al., 2003)
Pyogenic abscess
(Rajshekhar and Chandy,
1995)
Aspergillus abscess
(Goncalves-Ferreira
et al., 2003)
Cryptococcal abscess
(Abernathey et al., 1989)
Haematomas, vasculitis
(Goncalves-Ferreira
et al., 2003)
Arteriovenous
malformations
(Abernathey et al., 1989),
especially intracranial
dural arteriovenous
fistula (Crum and Link,
2004) and cavernous
angiomas
Hypertensive
encephalopathy (Jurcic
et al., 2004)
Infarctions (Abernathey
et al., 1989)
Radionecrosis
(Abernathey et al., 1989)
95% (Thomas et al., 1988; Abernathey et al., 1989;

Kratimenos and Thomas, 1993; Massager et al., 2000).
Inflammations and systemic

diseases
Neuro behet (Ben Taarit
et al., 2002),
Sarcoidosis
(Goncalves-Ferreira
et al., 2003),
Progressive multifocal
leukencephalopathy
(Friedman et al., 1989;
Goncalves-Ferreira et al.,
2003),
Demyelinating disease
(Abernathey et al., 1989;
Goncalves-Ferreira et al.,
2003)
overlap between adults and children. They include

focal tectal gliomas (8%), cystic pilocytic astrocytoma
and exophytic posterior glioma.
Radiotherapy
These malignant tumors are highly resistant to treatment by radiotherapy. Indeed, after radiotherapy, only
13% of patients showed clinical and radiological
improvement (Guillamo et al., 2001a). Gamma Knife
radiosurgery has been recently used in some cases of
brainstem gliomas, sometimes with satisfying functional outcome (Fuchs et al., 2002), although it seems
to be ineffective in other cases.
Chemotherapy
Chemotherapy has been occasionally used in this subgroup of tumors but data are too scant to draw any
recommendation on this issue.
Follow-up
The overall prognosis remains very poor despite radiotherapy and chemotherapy, with a median survival
about 11.2 months (Guillamo et al., 2001a) and 25
months (Salmaggi et al., 2008).
The other types of BS tumors identified in the adult
do not seem to differ from pediatric types, suggesting
Focal Tectal Brainstem Gliomas

Epidemiology
Focal tectal gliomas constitute a small subgroup in
adults (8% of cases).
Pathology
Histopathology, in the few cases where it could be
obtained, most often reveals a grade II oligoastrocytoma (Guillamo et al., 2001a). However, some cases of
pilocytic astrocytomas have been described.
A few cases of glioblastoma have also been reported
in this location with an aggressive course.
Clinical and Radiological Presentation

Focal tectal glioma usually presents with obstructive hydrocephalus, as in children. They may rarely
present with intracranial hemorrhage, which should be
removed.
376
Treatment and Follow-up

These lesions are associated with long survival
(sometimes exceeding 10 years) following ventriculoperitoneal shunting and usually focal radiotherapy
(Chamberlain, 1999). Since simple observation after
shunting may be an appropriate option in children, the
decision to treat adult patients with radiotherapy is also
questionable to some authors (Guillamo et al., 2001a),
who propose a similar approach in adults.
progress has been made. In this group, advances in

treatment will require a better knowledge of the biology of these tumors, and innovative approaches, possibly based on the analysis of the molecular profile of the
tumor. Experimental animal models of BSG and preclinical human studies to investigate intratumoral drug
treatment as well as promising new agents, such as signal transduction inhibitors, are actively being studied
in children (Fouladi et al., 2007; Pollack et al., 2007)
and could secondly benefit to adult patients. Finally,
the criteria to evaluate tumor response during therapy
of BSG remain unclear and require novel approaches.
Posterior Exophytic Gliomas

Exophytic contrast-enhancing glioma, which is well
known in children (up to 10% of cases) and is associated with a good prognosis, is extremely rare in
adults, maybe because most exophytic gliomas are
pilocytic astrocytomas, a rare tumor type in adults
(Guillamo et al., 2001a). A case report of exophytic pontine glioma simulating acoustic neurinoma
has been reported in an adult patient (Swaroop and
Whittle, 1997). Surgical resection is reserved for such
patients with specific tumoral locations as dorsal exophytic tumors protruding into the fourth ventricle.
Improvement in neurosurgical techniques (particularly
the use of intraoperative ultrasound, intraoperative
neurophysiological mapping and computer reconstruction techniques) has facilitated partial resection of
tumors previously considered as inoperable (Tanaka
et al., 2005), or even gross total removal in some cases,
without affecting the functional prognosis.
Brainstem Gliomas Associated with

Data regarding brainstem glioma associated with NF1
in adult patients are too scarce to draw conclusions, but
these tumors may be more aggressive in adults than in
children (Guillamo et al., 2001a).
Conclusion
In conclusion, despite improvements in the understanding and classification of BSGs, the primary challenge remains diffuse BSG of adults where little
References
Abernathey CD, Camacho A, Kelly PJ (1989) Stereotaxic suboccipital transcerebellar biopsy of pontine mass lesions.
Ben Taarit C, Turki S, Ben Maiz H (2002) Pseudotumoral
neurobehcet: a case report. J Mal Vasc 27:9395
Boviatsis EJ, Kouyialis AT, Stranjalis G, Korfias S, Sakas
DE (2003) CT-guided stereotactic biopsies of brain stem
lesions: personal experience and literature review. Neurol Sci
24:97102
Chamberlain MC (1999) Adult brainstem gliomas. Neurology
53:437438
Crum BA, Link M (2004) Intracranial dural arteriovenous
fistula mimicking brainstem neoplasm. Neurology 62:
23302331
Dirr LY, Donofrio PD, Patton JF, Troost BT (1989) A falsepositive edrophonium test in a patient with a brainstem
glioma. Neurology 39:865867
Elgamal EA, Coakham HB (2005) Hemifacial spasm caused
by pontine glioma: case report and review of the literature.
Neurosurg Rev 28:330332
Fouladi M, Nicholson HS, Zhou T, Laningham F, Helton KJ,
Holmes E, Cohen K, Speights RA, Wright J, Pollack IF,
Childrens Oncology Group (2007) A phase II study of
the farnesyl transferase inhibitor, tipifarnib, in children with
recurrent or progressive high-grade glioma, medulloblastoma/primitive neuroectodermal tumor, or brainstem glioma:
a Childrens Oncology Group study. Cancer 110:25352541
Freeman CR, Farmer JP (1998) Pediatric brain stem gliomas: a
review. Int J Radiat Phys 40:265271
Friedman WA, Sceats J, Nestok B, Ballinger WE (1989)
The incidence of unexpected pathological findings in an
imageguided biopsy series: a review of 100 consecutive
cases. Neurosurgery 25:180184
Fuchs I, Kreil W, Sutter B, Papaethymiou G, Pendl G (2002)
Gamma Knife radiosurgery of brainstem gliomas. Acta
Fujimoto K, Ohnishi H, Tsujimoto M, Hoshida T, Nakazato
Y (2000) Dysembryoplastic neuroepithelial tumor of the
cerebellum and brainstem. Case report. J Neurosurg 93:
487489

Gaviani P, Gonzalez RG, Zhu JJ, Batchelor TT, Henson JW
(2005) Central neurogenic hyperventilation and lactate production in brainstem glioma. Neurology 64:166167
Goncalves-Ferreira AJ, Herculano-Carvalho M, Pimentel J
(2003) Stereotactic biopsies of focal brainstem lesions. Surg
Neurol 60:311320
Guillamo JS, Doz F, Delattre JY (2001b) Brain stem gliomas.
Curr Opin Neurol 14:711715
Guillamo JS, Monjour A, Taillandier L, Devaux B, Varlet
P, Haie-Meder C, Defer GL, Maison P, Mazeron JJ,
Cornu P, Delattre JY, Association des Neuro-Oncologues
dExpression Francaise (ANOCEF) (2001a) Brainstem
gliomas in adults: prognostic factors and classification. Brain
124:25282539
Hsu CY, Hogan EL, Wingfield W Jr, Webb JG, Perot
PL Jr, Privitera PJ, Talbert OR, Balentine JD (1984)
Orthostatic hypotension with brainstem tumors. Neurology
34:11371143
Jurcic V, Ferluga D, Jeruc J, Pogacnik T, Popovic M (2004)
Hypertensive encephalopathy mimicking brainstem tumour
in psychiatric patient. Folia Neuropathol 42:3741
Kesari S, Kim RS, Markos V, Drappatz J, Wen PY, Pruitt AA
(2008) Prognostic factors in adult brainstem gliomas: a multicenter, retrospective analysis of 101 cases. J Neurooncol
88:175183
Kratimenos GP, Thomas DG (1993) The role of image-directed
biopsy in the diagnosis and management of brainstem
lesions. Br J Neurosurg 7:155164
Landolfi JC, Thaler HT, DeAngelis LM (1998) Adult brainstem
gliomas. Neurology 51:11361139
Lim EC, Wilder-Smith EP, Chong JL, Wong MC (2005)
Seeing the light: brainstem glioma causing visual auras and
migraine. Cephalalgia 25:154156
Marin-Sanabria EA, Kobayashi N, Miyake S, Kohmura E (2006)
Snoring associated with Ondines curse in a patient with
brainstem glioma. J Clin Neurosci 13:370373
Massager N, David P, Goldman S, Pirotte B, Wikler D, Salmon
I, Nagy N, Brotchi J, Levivier M (2000) Combined magnetic resonance imaging- and positron emission tomographyguided stereotactic biopsy in brainstem mass lesions: diagnostic yield in a series of 30 patients. J Neurosurg
93:951957
Mursch K, Halatsch ME, Markakis E, Behnke-Mursch J (2005)
Intrinsic brainstem tumours in adults: results of microneurosurgical treatment of 16 consecutive patients. Br J Neurosurg
19:128136
Packer RJ, Zimmerman RA, Luerssen TG, Sutton LN, Bilaniuk
LT, Bruce DA, Schut L (1985) Brainstem gliomas of childhood: magnetic resonance imaging. Neurology 35:397401
Pollack IF, Jakacki RI, Blaney SM, Hancock ML, Kieran MW,
Phillips P, Kun LE, Friedman H, Packer R, Banerjee A,
377
Geyer JR, Goldman S, Poussaint TY, Krasin MJ, Wang Y,
Hayes M, Murgo A, Weiner S, Boyett JM (2007) Phase I
trial of imatinib in children with newly diagnosed brainstem
and recurrent malignant gliomas: a Pediatric Brain Tumor
Consortium report. Neuro Oncol 9:145160
Rajshekhar V, Chandy MJ (1995) Computerized tomographyguided stereotactic surgery for brainstem masses: a riskbenefit analysis in 71 patients. J Neurosurg 82:976981
Saka E, Ozkaynak S, Tuncer R (2006) Tongue tremor in brainstem pilocytic astrocytoma. J Clin Neurosci 13:503506
Salmaggi A, Fariselli L, Milanesi I, Lamperti E, Silvani A,
Bizzi A, Maccagnano E, Trevisan E, Laguzzi E, Rud R,
Boiardi A, Soffietti R., Associazione Italiana di Neurooncologia (2008) Natural history and management of brainstem gliomas in adults. A retrospective Italian study. J Neurol
255:171177
Selvapandian S, Rajshekhar V, Chandy MJ (1999) Brainstem
glioma: comparative study of clinico-radiological presentation, pathology and outcome in children and adults. Acta
Neurochir (Wien) 141:721727
Shad A, Green A, Bojanic S, Aziz T (2005) Awake stereotactic biopsy of brain stem lesions: technique and results. Acta
Neurochir (Wien) 147:4749
Shenoy SN, Raja A (2004) Cystic trochlear nerve neurinoma mimicking intrinsic brainstem tumour. Br J Neurosurg
18:183186
Swaroop GR, Whittle IR (1997) Exophytic pontine glioblastoma mimicking acoustic neuroma. J Neurosurg Sci 41:
409411
Swinson BM, Friedman WA, Yachnis AT (2006) Pontine atypical neurocytoma: case report. Neurosurgery 58:E990
Tanaka S, Kobayashi I, Utsuki S, Iwamoto K, Takanashi J (2005)
Biopsy of brain stem glioma using motor-evoked potential
mapping by direct peduncular stimulation and individual
adjuvant therapy. Case report. Neurol Med Chir (Tokyo)
45:4955
Thomas DG, Bradford R, Gill S, Davis CH (1988) Computerdirected stereotactic biopsy of intrinsic brain stem lesions.
Br J Neurosurg 2:235240
Torcuator R, Zuniga R, Loutfi R, Mikkelsen T (2009)
Bevacizumab and irinotecan treatment for progressive diffuse brainstem glioma: case report. J Neurooncol 93:
409412
White HH (1963) Brain stem tumors occurring in adults.
Neurology 13:292300
Yousry I, Muacevic A, Olteanu-Nerbe V, Naidich TP, Yousry TA
(2004) Exophytic pilocytic astrocytoma of the brain stem in
an adult with encasement of the caudal cranial nerve complex
(IXXII): presurgical anatomical neuroimaging using MRI.
Chapter 37
The Use of Low Molecular Weight Heparin in the Treatment

and Prevention of Thromboembolic Disease in Glioma
Patients
Bo H. Chao and H. Ian Robins
Abstract The neuro-oncologist routinely deals with

an extraordinarily high risk patient population relative
to thromboembolic disease. With the advent of easily
accessible diagnostic studies, e.g., ultrasound and/or
spiral CT scans, timely diagnosis of venous thromboembolic events (VTE) is readily accomplished.
The introduction of low molecular weight heparin
(LMWH) about 20 years ago (in contrast to unfractionated heparin and vitamin K antagonists) has provided
a class of agents with a favorable therapeutic index.
In the review to follow, the literature regarding the
use of LMWH in neuro-oncology patient populations
is summarized. Topics addressed include prophylaxis,
treatment, as well as consideration of the potential
anti-neoplastic properties of these drugs.
Keywords Low molecular
Thromboembolic disease
Prophylaxis Antagonist
weight
Glioma
heparin
VTE
Introduction
It is well recognized that cancer patients in general
(Shlebak and Smith, 1997), and in particular glioma
patients are highly predisposed to thromboembolic
phenomena (Sawaya and Highsmith, 1988; Marras
et al., 2000). Brisman and Mendell (1973) demonstrated an 8.4% incidence of pulmonary emboli in
glioma patients; this represents a three fold increase
H.I. Robins ()
Department of Medicine, Human Oncology, and Neurology,
University of Wisconsin School of Medicine and Public Health,
Madison, WI 53792, USA
e-mail: hirobins@wisc.edu
in the incidence seen in non-malignant neurosurgical patients. Similarly, the incidence of deep vein
thromboembolism (DVT) in such patients is 27.5%
compared to 17% in a control neurosurgical group
(Kayser-Gatchalian and Kayser, 1975). Ruff and
Posner (1983) noted a 31% incidence of confirmed
DVT in 264 glioma patients. Quevedo et al. (1994)
noted an incidence of 28% of thromboembolism in
a series of 64 high grade glioma patients. Sawaya
et al. (1992), using I-fibrinogen scanning, demonstrated DVTs in 60% of glioblastoma multiforme
(GBM) patients; the presence of venous thromboembolic events (VTE) did not correlate with ambulatory
status, time of surgery, length of operation, or occurrence in a paretic limb. Sawaya and Highsmith (1992)
have suggested that malignant brain tumors release
a factor responsible for this predisposition to coagulopathy. Previous work suggesting increased platelet
adhesiveness in malignant brain tumors is consistent
with this supposition (Nathanson and Savitsky, 1952;
Millac, 1967). More recent work by Iberti et al. (1994)
and Hamilton et al. (1994) supports the concept of
an increased coagulable state of brain tumor patients.
Coupled to this significant morbidity is the potential
risk for the central nervous system hemorrhage when
anti-coagulants are used as a therapeutic intervention and/or prophylaxis. In the text to follow we will
discuss the existing data for the treatment and prophylaxis of thromboembolic disease in glioma patients.
The review will focus on the use of low molecular
weight heparin (LMWH), as the published literature
and experience center on this class of anti-coagulants
relative to both safety and efficacy. Encompassed in
this discussion will be a review of the data suggesting the potential for anti-neoplastic effects with the
application of LMWH.

379
380
Low Molecular Weight Heparin as

Treatment for Thromboembolic Disease
Based on randomized clinical studies, LMWH is the
agent of choice for the initial and long term treatment
of VTE in patients with neoplastic disease (Lee, 2009;
Lyman et al., 2007). The cumulative risk of symptomatic and recurrent VTE during 6 months of anticoagulation was 17% for patients treated with a vitamin
K antagonist, such as warfarin or acenocoumarol, compared with 9% for patients treated with dalteparin, a
LMWH (Lee et al., 2003). The relative risk reduction of 52% was statistically significant (hazard ratio
= 0.48; P = 0.002). One episode of recurrent VTE is
prevented for every 13 patients treated with dalteparin.
Regarding side effects, no difference in major or minor
bleeding was detected between the groups, with 6% of
dalteparin-treated patients and 4% of control patients
having major bleeding episodes (Lee et al., 2003).
There were no confirmed cases of heparin-induced
thrombocytopenia, and patient compliance with selfinjection in these studies was good (Deitcher et al.,
2006; Meyer et al., 2002; Lee et al., 2003; Hull et al.,
2006). Comparing to unfractionated heparin, LMWH
reduces the risk of recurrent thrombosis by 32%, major
bleeding by 43%, and death by 24%, according to a
Cochrane systemic review summarizing 22 clinical trials (van Dongen et al., 2004). LMWH is the preferred
agent for treatment of cancer-associated thrombosis
because of its efficacy, safety, and convenience.
In the aforementioned studies in a general population of cancer patients with VTE, dalteparin was given
at full therapeutic dose of 200 U/kg once daily for a
month, followed by dose reduction down to 75% of the
full dose for the remainder of the 6-month period (Lee
et al., 2003). It should be noted that the concept of a
25% dose reduction after initial therapy was an arbitrary choice, and not evaluated in comparison to other
possibilities. Other studies used full therapeutic dose of
enoxaparin for 3 months (Deitcher et al., 2006; Meyer
et al., 2002). However, in patients with GBM, because
of a higher risk of bleeding, we typically treat VTE in
GBM patients with full dose LMWH until resolutions
of symptoms from their thrombosis, usually within 2
weeks of onset of VTE, followed by 50% dose reduction in the following month, and followed by another
25% dose reduction after 23 months. We recommend
dividing the total daily dose during the acute phase
B.H. Chao and H.I. Robins
of treatment into twice-a-day dosing to reduce bleeding risk. If the GBM patient develops recurrent VTE
during dose reduction of LMWH, we would check the
antifactor Xa levels 4 h after an injection and adjust the
dose accordingly. There has not been a trial to evaluate
this treatment algorithm; it has been used at our center since the introduction of LMWH with an estimated
incidence of recurrent VTE less than 5%. This use of
a lower dose of LMWH post acute treatment is consistent with the efficacy of LMWH in the prophylaxis
setting and had evolved as a conceptual extrapolation.
It has been our practice to continue LMWH indefinitely in high grade glioma patients based on the
experience in a Eastern Cooperative Oncology Group
(ECOG)/Radiation Therapy Oncology Group (RTOG)
trial, described in detail in section Anti-neoplastic
Effects of Low Molecular Weight Heparin below
(Robins et al., 2008). Patients who are on LMWH
for protracted periods of time are at risk for osteoporosis, (and often have another risk factor for bone
demineralization, i.e., steroids). We suggest monitoring such patients with bone mineral density studies,
and considering interventions such as bisphosphonates
as appropriate.
It should be noted that fondaparinux, a synthetic
selective inhibitor of activated factor X (with no activity against thrombin), is approved for the initial therapy
of VTE (Buller et al., 2004). In terms of administration,
convenience, and cost, it is comparable to LMWH.
Similar to LMWH, it is contraindicated in patients
with significant renal insufficiency. However, it has
not been studied in cancer patients. Hence its use in
cancer patients is generally limited in those who have
allergies to heparin or a history of heparin-induced
thrombocytopenia (Lee, 2009).
The role of vena cava filters in the treatment of
VTE in cancer patients remains undefined as previously reviewed (Lee, 2009). Filters do not treat the
underlying hypercoagulable state in cancer patients.
The use of filters in patients receiving anticoagulation led to a reduction in symptomatic pulmonary
embolism, but more recurrent DVTs and no difference in overall survival (Investigators, 2005). It is our
view that the use of filters should be reserved for
situations in which anticoagulant therapy is contraindicated, e.g., active bleeding, severe thrombocytopenia, or in preparation for a significant neurosurgical
intervention.
37 LMWH in the Treatment and Prevention of Thromboembolic Disease in Glioma Patients
Low Molecular Weight Heparin as

Prophylaxis for Thromboembolic
Disease
As discussed in section Introduction, glioma patients
are highly predisposed to thromboembolic phenomena
with an incidence approaching ~30%. In a phase II
study in patients with newly diagnosed glioblastoma
multiforme (GBM), patients were given dalteparin
5000 units during and post radiation, as described in
detail in section Anti-neoplastic Effects of LMWH
below (Robins et al., 2008). The primary endpoint was
survival, and the secondary end point was the incidence of VTE. This study closed prematurely with
only 42 patients accrued due to a change in standard
of care for GBM. As the study was conceived it was
estimated that over the course of the study, 30% of
the enrolled patients would develop a VTE over the
entire time course of their disease. With a planned
cohort of 72 patients, there was 81% power to detect
a decrease from 30 to 15% in VTE using a two-sided
0.05 exact test for a single proportion. In the end the
study observed no VTE in patients actively receiving dalteparin. Time on dalteparin ranged from 1.2 to
25.4 months, with a median time of 6.3 months. There
were no reports of grade 3/4 bleeding or thrombocytopenia related to dalteparin pre or post progression.
There were local site reactions to dalteparin in up to
71% of patients; two patients experienced grade 3 toxicities ascribed to radiation. Two patients developed
DVT after stopping dalteparin when they developed
progressive GBM.
A seemingly contradictory result in the context
of a phase III Canadian study was later reported at
American Society of Clinical Oncology annual meeting in 2007 (Perry et al., 2007b), and was updated later
in the year at another meeting (Perry et al., 2007a).
This trial was planned to enroll 512 patients; it closed
prematurely, secondary to lack of drug availability,
with 186 malignant glioma patients randomized to dalteparin 5000 units daily versus placebo. The primary
endpoint was the development of VTE. Although there
was a trend for LMWH to decrease VTE, it was not statistically significant (HR = 0.65, 95% CI: 0.271.54,
p = 0.33). However, the hazard ratio was impressive
at 0.65 in favor of the LMWH arm. It remains speculative whether this hazard ratio would have achieved
statistical significance if the study had met its accrual
381
objective. There was also a trend for increased bleeding in the LMWH arm 5.1% (n = 5) versus 1.2%
(n = 1). These authors in commenting on the trend
toward increased intracranial bleeding noted the
clinical significance of this is unclear since intratumoral hemorrhage is sometimes part of the natural
history of this disease (Perry et al., 2007a). Potential
noteworthy differences in these studies may relate to
patient populations. The RTOG study was restricted to
a GBM cohort (as opposed to all malignant glioma),
as well as a requirement for an unusually good performance status, i.e., 95% ECOG performance status of
0 or 1. To date a final published report regarding the
Canadian study has not become available.
At present, we agree with the authors of both of
these individual studies: The results of these studies taken collectively leave the role of anticoagulant
in thromboembolic prophylaxis as unclear for this
group of patients. Definitive conclusions regarding the
application of VTE prophylaxis strategies will require
further investigation.
There have, however, been studies of VTE prophylaxis after elective neurosurgery in patients with brain
tumors, including but not exclusive to glioma. Two
multi-center, randomized, double-blind trials studied
the use of LMWH (nadroparin 7500 units subcutaneously once daily, or enoxaparin 40 mg subcutaneously once daily, respectively) versus placebo in
conjunction with the use of compression stockings
in the prevention of VTE after elective neurosurgery
(Nurmohamed et al., 1996; Agnelli et al., 1998). The
majority of the accrued patients had brain tumors, up
to 30% of whom were diagnosed with glioma (Agnelli
et al., 1998). LMWH or placebo was given within
24 h after surgery and continued for up to 10 days
or until hospital discharge. Both studies showed that
LMWH combined with compression stockings is more
effective than compression stockings alone for the prevention of VTE after elective neurosurgery, without
inducing any significant increase of major bleeding.
The absolute risk reduction of VTE in patients receiving LMWH was 8 and 14%, whereas the absolute risk
reduction of proximal VTE or pulmonary embolism
in patients receiving LMWH was 4 and 8%, respectively (Nurmohamed et al., 1996; Agnelli et al., 1998).
However, it should be noted that LMWH in these
studies was given within 24 h after surgery.
Another randomized study conducted at the
University of Michigan, where LMWH (enoxaparin
382
30 mg subcutaneously every 12 h) was initiated before

the induction of anesthesia and was continued throughout the hospital stay, showed that enoxaparin therapy
initiated at the time of anesthesia induction significantly increased postoperative intracranial hemorrhage
(Dickinson et al., 1998). This study was terminated
early because of the increased incidence of adverse
events in the enoxaparin treatment group.
In conclusion, LMWH given within 24 h after
neurosurgery, but not prior to neurosurgery, in combination with compression stockings, is effective for
the prevention of VTE in glioma patients. However,
the role of VTE prophylaxis beyond the postoperative
period after hospital discharge is yet to be determined
and still requires further studies.
Anti-neoplastic Effects of Low Molecular

Weight Heparin
The genesis of a clinical study to test the antineoplastic proprieties of LMWH in GBM patients
occurred when investigators anecdotally noted usually
good outcome in high grade glioma patients who were
receiving LMWH for DVT. In reviewing the literature, a meta-analysis was identified that suggested a
reduced mortality in cancer patients receiving LMWH
compared with patients receiving standard heparin;
further, this benefit was not consequent to a reduced
thromboembolism rate (Siragusa, 1996). Additionally,
a potential mechanistic basis for this was identified:
(i) anti-angiogenesis (Folkman et al., 1983; Norrby
1993), (ii) effects on cellular matrix (Bitan et al.,
1995; Collen et al., 2000), and (iii) reduced cell proliferation (Castellot et al., 1986). In this regard, it is
of interest to note that based on preclinical studies,
anti-angiogenesis agents can putatively serve as radiosensitizers (Teicher et al., 1995; Wachsberger et al.,
2003). Also, a preclinical study linked up-regulated
tissue factor expression by epidermal growth factor
receptor (EGFR) and EGFRvIII and the prothrombotic events that occur in the progression of GBM
(Rong et al., 2006). As GBM is among the most
vascular of all neoplasms, abundantly expressing vascular endothelial growth factor and platelet-derived
growth factor (Dunn et al., 2004), the use of LMWH
as a therapeutic adjunct to radiotherapy was viewed as
worthy of clinical testing.
Taking aforementioned considerations into account,

the ECOG in cooperation with the RTOG initiated
a phase II study of the effect of dalteparin during and post radiation in newly diagnosed patients
with GBM (Robins et al., 2008): Dalteparin of 5000
units daily was given subcutaneously with and after
conventional radiotherapy to newly diagnosed GBM
patients. Forty-five patients were accrued. The study
was designed such that at time of progression, patients
could continue dalteparin in addition to standard regimens. Pretreatment characteristics included: Median
age 61 (range 2678); ECOG Performance status:
0 = 38%, 1 = 57%, 2 = 5%; gross total resection 45%.
There were no grade 3/4 bleeding or thrombocytopenic
events, and no VTE occurred while patients were on
dalteparin. Median time on dalteparin was 6.3 months,
median time to progression was 3.9 months; median
survival was 11.9 months. There was no significant
improvement in survival when compared to the RTOG
GBM database (with various radiation/drug doublets
including BCNU) using recursive partitioning analysis
(RPA) (Curran et al., 1993). The results were, however,
inferior to the use of temozolomide, which became
the new standard of care just at the time this study
was concluded (Stupp et al., 2004). As noted above,
investigators were provided the option of continuing
dalteparin after disease progression. The rationale for
this included its potential use as an anti-angiogenesis
agent, i.e., to slow disease progression, as well as use
for DVT prophylaxis. In this study, 22 of 38 patients
continued dalteparin post first progression. Median
survival for those who stopped dalteparin at first progression versus continued use was 3.2 and 7.8 months,
respectively. This analysis is confounded by several
factors and should be interpreted with caution. Patients
could receive other treatments at progression, as outlined in the protocol, with or without the continuation
of dalteparin. The patients who stopped dalteparin at
first progression may have been too sick to receive
other treatments. For those patients who continued
dalteparin at first progression, their survival could be
confounded by the varying lengths that they continued
dalteparin and/or by any of the other treatments they
received.
As discussed in section Low Molecular Weight
Heparin as Prophylaxis for Thromboembolic Disease
of this chapter above, historically the incidence of
VTE in GBM patients is ~30%, and in this study, no
VTE was observed in GBM patients actively receiving
37 LMWH in the Treatment and Prevention of Thromboembolic Disease in Glioma Patients
dalteparin. Because dalteparin might reduce the incidence of VTE, having no significant overlapping toxicities with most other drugs, its testing in a combined
modality approach with other medications active in the
treatment of GBM might be warranted in future trials.
Summary Comments
As outlined above, the neuro-oncologist routinely
deals with an extraordinarily high risk patient population relative to VTE. Indeed, it is recommended that
new patients be trained in the signs and symptoms of
DVT and pulmonary embolism. With the advent of
easily accessible diagnostic studies, e.g., ultrasound
and/or spiral CT scans, timely diagnosis of VTE is
less problematic than in past years. Outpatient therapy is becoming standard practice at most centers for
clinically well compensated patients.
For more than 5 decades, unfractionated heparin and
vitamin K antagonists have been used for the treatment
and prevention of VTE. The introduction of LMWH
about 20 years ago, however, has provided a class of
agents with a unique therapeutic index that is favorable in glioma patients. It is anticipated that more
convenient drugs (e.g., oral, or perhaps not requiring
monitoring) currently in development, targeting activated factor X (i.e., factor Xa) or thrombin (factor IIa),
will become available. As the optimal treatment for
VTE is prevention, it is hoped and anticipated that such
drug development of oral IIa and Xa inhibitors will
provide a new set of efficacious and safe agents for this
high risk patient population.
References
Agnelli G, Piovella F, Buoncristiani P, Severi P, Pini M,
DAngelo A, Beltrametti C, Damiani M, Andrioli GC,
Pugliese R, Iorio A, Brambilla G (1998) Enoxaparin plus
compression stockings compared with compression stockings alone in the prevention of venous thromboembolism
after elective neurosurgery. N Engl J Med 339:8085
Bitan M, Mohsen M, Levi E, Wygoda MR, Miao H, Lider
O, Svahn CM, Ekre HP, Ishai-Michaeli R, Bar-Shavit R,
Vlodavsky I, Peretz T (1995) Structural requirements for
inhibition of melanoma lung colonization by heparanase
inhibiting species of heparin. Isr J Med Sc 31:106118
383
Brisman R, Mendell J (1973) Thromboembolism and brain

tumor. J Neurosurg 38:337338
Buller HR, Davidson BL, Decousus H, Gallus A, Gent M,
Piovella F, Prins MH, Raskob G, Segers AE, Cariou R,
Leeuwenkamp O, Lensing AW (2004) Fondaparinux or
enoxaparin for the initial treatment of symptomatic deep
venous thrombosis: a randomized trial. Ann Intern Med
140:867873
Castellot JJ, Jr, Choay J, Lormeau J, Petitou M, Sache E,
Kamovsky MJ (1986) Structural determinants of the capacity of heparin to inhibit the proliferation of vascular smooth
muscle cells. II. Evidence for a pentasaccharide sequence
that contains a 3-0-sulfate group. J Cell Biol 102:19791984
Collen A, Smorenburg SM, Peters E, Lupu F, Koolwijk P, van
Noordan C, van Hinsbergh VWM (2000) Unfractionated
and low molecular weight heparin affect fibrin structure and
angiogenisis in vitro. Cancer Res 60:61966200
Curran WJ Jr, Scott CB, Horton J, Nelson JS, Weinstein AS,
Fischbach AJ, Chang CH, Rotman M, Asbell SO, Krisch
RE (1993) Recursive partitioning analysis of prognostic factors in three Radiation Therapy Oncology Group malignant
glioma trials. J Natl Cancer Inst 85:704710
Deitcher SR, Kessler CM, Merli G, Rigas JR, Lyons RM, Fareed
J (2006) Secondary prevention of venous thromboembolic
events in patients with active cancer: enoxaparin alone versus
initial enoxaparin followed by warfarin for a 180-day period.
Clin Appl Thromb Hemost 12:389396
Dickinson LD, Miller LD, Patel CP, Gupta SK (1998)
Enoxaparin increases the incidence of postoperative intracranial hemorrhage when initiated preoperatively for deep
venous thrombosis prophylaxis in patients with brain tumors.
Dunn IF, Hesse O, Black PM (2004) Growth factors in glioma
angiogenesis: FGFs, PDGF, EGF, and TGFs. J Neurooncol
50:121137
Folkman J, Langer R, Linhardt RJ, Haudenschild C, Taylor S
(1983) Angiogenesis inhibition and tumor regression caused
by heparin or a heparin fragment in the presence of cortisone.
Science 221:719725
Hamilton MG, Hull RD, Pineo GF (1994) Venous thromboembolism in neurosurgery and neurology patients: A review.
Hull RD, Pineo GF, Brant RF, Mah AF, Burke N, Dear R,
Wong T, Cook R, Solymoss S, Poon MC, Raskob G (2006)
Long-term low-molecular-weight heparin versus usual care
in proximal-vein thrombosis patients with cancer. Am J Med
119:10621072
Iberti TJ, Miller M, Abalos A, Fischer EP, Post KD, Benjamin
E, Oropello JM, Wiltshire-Clement M, Rand JH (1994)
Abnormal coagulation profile in brain tumor patients during
surgery. Neurosurgery 34:389395
Investigators, P.d.R.d.E.P.p.I.C (2005) Eight-year follow-up of
patients with permanent vena cava filters in the prevention
of pulmonary embolism: the PREPIC (Prevention du Risque
dEmbolie Pulmonaire par Interruption Cave) randomized
study. Circulation 112:416422
Kayser-Gatchalian MC, Kayser K (1975) Thrombosis and
intracranial tumors. J Neurol 209:217224
Lee AY (2009) Anticoagulation in the treatment of established
venous thromboembolism in patients with cancer. J Clin
Oncol 27:48954901
384
Lee AY, Levine MN, Baker RI, Bowden C, Kakkar AK, Prins M,
Rickles FR, Julian JA, Haley S, Kovacs MJ, Gent M (2003)
Low-molecular-weight heparin versus a coumarin for the
prevention of recurrent venous thromboembolism in patients
with cancer. N Engl J Med 349:146153
Lyman GH, Khorana AA, Falanga A, Clarke-Pearson D, Flowers
C, Jahanzeb M, Kakkar A, Kuderer NM, Levine MN,
Liebman H, Mendelson D, Raskob G, Somerfield MR,
Thodiyil P, Trent D, Francis CW (2007) American Society
of Clinical Oncology guideline: recommendations for venous
thromboembolism prophylaxis and treatment in patients with
cancer. J Clin Oncol 25:54905505
Marras LC, Geerts WH, Perry JR (2000) The risk of
venous thromboembolism is increased throughout the course
of malignant glioma: an evidence-based review. Cancer
89:640646
Meyer G, Marjanovic Z, Valcke J, Lorcerie B, Gruel Y, SolalCeligny P, Le Maignan C, Extra JM, Cottu P, Farge D (2002)
Comparison of low-molecular-weight heparin and warfarin
for the secondary prevention of venous thromboembolism in
patients with cancer: a randomized controlled study. Arch
Intern Med 162:17291735
Millac P (1967) Platelet stickiness in patients with intracranial
tumors. BMJ 4:2526
Nathanson M, Savitsky JP (1952) Platelet adhesive index studies
in multiple sclerosis and other neurological disorders. Bull
NY Acad Med 28:462468
Norrby K (1993) Heparin and angiogenesis: a low-molecularweight fraction inhibits and a highmolecular-weight fraction stimulates angiogenesis systemically. Haemostasis
23(Suppl 1):141149
Nurmohamed MT, van Riel AM, Henkens CM, Koopman MM,
Que GT, dAzemar P, Buller HR, ten Cate JW, Hoek JA,
van der Meer J, van der Heul C, Turpie AG, Haley S,
Sicurella A, Gent M (1996) Low molecular weight heparin and compression stockings in the prevention of venous
thromboembolism in neurosurgery. Thromb Haemost 75:
233238
Perry JR, Rogers L, Laperrier N, Julian J, Geerts W, Agnelli
G, Malkin M, Sawaya R, Baker R, Levine M (2007a) A
phase III randomized placebo-controlled trial of thromboprophylaxis using dalteparin low molecular weight heparin
(LMWH) in patients with newly diagonsed malignant glioma
updated results. Thromb Res 120:ABS PO40
Perry JR, Rogers L, Laperriere N, Julian J, Geerts W, Agnelli
G, Malkin M, Sawaya R, Baker R, Levine M (2007b)
PRODIGE: a Phase III randomized placebo-controlled trial
of thromboprophylaxis using dalteparin low molecular
weight heparin (LMWH) in patients with newly diagnosed
malignant glioma. J Clin Oncol 25:77s (Abstr 2011)

Quevedo JF, Buckner JC, Schmidt JL, Dinapoli RP, OFallon
JR (1994) Thromboembolism in patients with high-grade
glioma. Mayo Clin Proc 69:329332
Robins HI, ONeill A, Gilbert M, Olsen M, Sapiente R, Berkey
B, Mehta M (2008) Effect of dalteparin and radiation on
survival and thromboembolic events in glioblastoma multiforme: a phase II ECOG trial. Cancer Chemother Pharmacol
62:227233
Rong Y, Durden DL, Van Meir EG, Brat DG (2006) Differential
regulation of tissue factor expression by EGFR and
EGFRvIII in GBM. Neuro-Oncology 8:401 (ABS CB27)
Ruff RL, Posner JB (1983) Incidence and treatment of peripheral venous thrombosis in patients with glioma. Ann Neurol
13:334336
Sawaya R, Highsmith RF (1988) Brain tumors and the fibrinolytic enzyme system. In: Kornblith PL, Walker MD (eds)
Advances in neuro-oncology. Futura, Mount Kisco, NY,
pp 103157
Sawaya R, Highsmith RF (1992) Postoperative venous thromboembolism and brain tumors: Part III. Biochemical profile.
J Neuro-Oncol 14:113118
Sawaya R, Zuccarrello M, Elkalliny M, Nishiyama H (1992)
Postoperative venous thromboembolism and brain tumors:
Part I. Clinical Profile. J Neurooncol 14:119125
Shlebak AA, Smith DB (1997) Incidence of objectively diagnosed thromboembolic disease in cancer patients undergoing
cytotoxic chemotherapy and/or hormonal therapy. Cancer
Chemother Pharmacol 39:462466
Siragusa SCB, Piovella F, Hirsh J, Ginsberg S (1996) Lowmolecular-weight heparins and unfractionated heparin in the
treatment of patients with acute venous thromboembolism:
Results of a meta-analysis. J Am Med 100:269277
Stupp R, Mason WP, Van Den Bent MJ, Weller M, Fisher
B, Taphoorn M, Brandes AA, Cairncross G, Lacombe D,
Mirimanoff RO (2004) Conclusive results of a randomized
phase III trial by the EROTC Brain and RT groups and NICIC
Clinical Trails Group. Concomitant and adjuvant temozolomide and radiotherapy for newly diagnosed glioblastoma
multiforme. Proc Am Soc Clin Oncol 23:1 (Abstract 2)
Teicher BA, Holden SA, Ara G, Dupuis NP, Liu F, Yuan J, Ikebe
M, Kakeji Y (1995) Influence of an antiangiogenic treatment
on 9L gliosarcoma: oxygenation and response to cytotoxic
therapy. Int J Cancer 61:732737
van Dongen CJ, van den Belt AG, Prins MH, Lensing AW (2004)
Fixed dose subcutaneous low molecular weight heparins versus adjusted dose unfractionated heparin for venous thromboembolism. Cochrane Database Syst Rev CD001100:192
Wachsberger P, Burd R, Dicker AP (2003) Tumor response to
ionizing radiation combined with antiangiogenesis or vascular targeting agents. Clin Cancer Res 9:19571971
Part III
Prognosis
Chapter 38
Brainstem Gliomas: An Overview

Marco Antonio Lima
Abstract Brainstem gliomas are a heterogeneous

group of tumors that affect all ages, although they
are predominant in children. There are several grading systems but they are often divided in four subgroups: tectal, diffuse pontine, cervicomedullary and
dorsal exophytic. While diffuse pontine gliomas usually are high-grade tumors, the other three subtypes
are low- grade lesions (pilocytic astrocytoma or grade
II gliomas). The clinical presentation is diverse, but
involvement of cranial nerves and long tracts are frequently present. Imaging is crucial for diagnosis and
surgical planning and, in the case of diffuse pontine
gliomas, MRI has replaced biopsy as gold standard.
Treatment includes general measures (control of pain,
seizures, physical and speech therapy) that are best
addressed by a multidisciplinary team and surgical
resection or palliation depending on the extension,
grade and location. Since hydrocephalus is common,
ventricular derivation followed by observation is the
most frequent approach for tectal gliomas. Surgery
and/or radiation therapy are reserved for clinical or
radiological progression. Diffuse pontine gliomas are
inoperable and radiation therapy is the mainstay of
treatment leading to transient control of the symptoms. Most cervicomedullary and dorsally exophytic
tumors are treated with partial or complete resection
when feasible, followed by radiation therapy at time
of recurrence or progression. Prognosis is dismal for
pontine gliomas, and most patients do not survive
2 years from diagnosis. Outcome is better for tectal,
M.A. Lima ()

Department of Neurosurgery, Instituto Nacional de
Cncer-INCa, Centro, Rio de Janeiro RJ-22230-130, Brazil
e-mail: mlima@inca.gov.br
cervicomedullary and dorsally exophytic tumors and

many patients achieve a long-term survival, although
neurological sequelae are common.
Keywords Brainstem gliomas Tectal gliomas
Diffuse pontine gliomas Cervicomedullary gliomas
Dorsal exophytic gliomas CNS
Introduction
Brainstem gliomas are a group of heterogeneous
tumors that may affect either children or adults, but
are more prevalent in the first years of life. In the
pre-imaging era, they were considered a single entity
and inoperable, but newer imaging techniques allowed
the classification of subgroups with distinct behaviors.
Nowadays, the knowledge of course of the disease as
well the outcome of these different subtypes permits
a better determination of which patients would benefit
from surgery. In addition, despite the involvement of a
crucial area of the central nervous system, modern neurosurgical procedures currently permit safer resections
with minimal additional morbidity.
Epidemiology
Brainstem gliomas are more frequent in children and
represent 10% of pediatric central nervous system
tumors (Walker et al., 1999). In a recent review of 6212
pediatric patients with gliomas, 19.8% were located in

387
388
the brainstem (Qaddoumi et al., 2009). There is no gender predilection and the mean age at onset in children
is 79 years (Jallo et al., 2004). According to Smith
et al. in the 19901994 period, the age-adjusted incidence of brainstem high-grade gliomas was 0.14 per
105 person-years, while the low-grade gliomas was
0.18 per 105 person-years (Smith et al., 1998).
There are few studies of brainstem gliomas in adults
since they are a rare entity, representing less than 2%
of gliomas in this population (Guillamo et al., 2001).
Although they can occur at any age, the peak incidence
in adult patients is between the third and fourth decades
(Selvapandian et al., 1999). In a recent study of brainstem gliomas in adults, the median age at diagnosis
was 36 years (Kesari et al., 2008).
Neurofibromatosis 1 (NF1) predisposes to central
nervous system tumors, especially pilocytic astrocytomas. Although the optic pathways are the preferred
site, other common locations are the brainstem and
cerebellum (Ferner, 2007). In a retrospective study
of 104 NF1 patients, 17% had brainstem gliomas
(Guillamo et al., 2003), but lesions were more indolent
in this group.
Classification
Several grading systems for classification and staging
of brainstem gliomas have been described over the
years based on their location, focality, pattern of tumor
growth and the presence of necrosis and hydrocephalus
(Barkovich et al., 1990; Epstein, 1985). A simple classification would divide brainstem gliomas in four
major categories: diffuse intrinsic gliomas, focal tectal,
cervicomedullary and dorsal exophytic tumors (LaigleDonadey et al., 2008).
Focal tectal gliomas: these are rare tumors, representing less than 5% of brainstem gliomas in
children (Laigle-Donadey et al., 2008). However,
in a recent study with 101 adult brainstem glioma
patients revealed that 20% of the lesions were tectal (Kesari et al., 2008). Indeed, Guillamo et al.
have shown focal tectal gliomas in 8/48 adult
patients (16.7%) suggesting that this subtype might
be more frequent in adults than previously thought
(Guillamo et al., 2001). The lesions are well
M.A. Lima
circumscribed and most are low-grade gliomas

(WHO I and II).
Diffuse intrinsic gliomas: these are the most
prevalent and aggressive brainstem gliomas, being
responsible for 80% of the cases. Although they can
occur anywhere in the brainstem, they are usually
located at the pons, extending into the midbrain or
medulla over the course of the disease. Most diffuse gliomas are malignant fibrillary astrocytomas
(grade III or IV), but biopsy is not performed in all
cases, since diagnosis is frequently based solely on
clinical and radiological presentation.
Cervicomedullary gliomas: lesions in the cervicomedulary region are present in 510% with
brainstem gliomas and occur predominantely in the
pediatric population. Most tumors are low-grade
astrocytomas and are similar to intramedullary
spinal cord gliomas. According to Epstein et al.
their growth is limited rostrally by the decussating fibers and pial elements (Epstein and Farmer,
1993). The lesion is well circumscribed and may be
associated with syringomyelia caudally.
Dorsal exophytic gliomas: this subgroup represents
between 14 and 24% of the brainstem gliomas. The
epicenter is usually within the fourth ventricle and
the age at presentation is younger than in other subtypes. The biopsy or surgical resection often reveals
a low-grade glioma. Although, total resection is
not feasible in many cases, partial resection can
improve outcome.
Clinical Manifestations
Tectal lesions grow insidiously and usually produce
symptoms due to compression of cerebral aqueduct
leading to a non-communicating hydrocephalus. The
time from onset of symptoms until diagnosis is variable ranging from days to several years (Stark et al.,
2005). Headache, vomiting and ataxia are frequently
present at time of diagnosis (Daglioglu et al., 2003).
Papilledema is observed in close to 50% of the patients.
Parinaud syndrome, strabismus and long tract signs
are found in a minority of patients (Ternier et al.,
2006). Unusual presentations are tremor, irritability,
parkinsonism, seizures, vertigo and decline in school
performance. After ventricular drainage and treatment
38 Brainstem Gliomas: An Overview
of hydrocephalus, signs and symptoms can remain

stable for years.
Patients with diffuse pontine gliomas present with
rapidly progressive neurological deficits, usually in
weeks or few months, but it may be sudden, suggesting a stroke initially (Guillamo et al., 2001).
A shorter duration of symptoms is associated with a
worse outcome (Ueoka et al., 2009). The clinical triad
includes ataxia, long tract signs and multiple cranial
nerve palsies, particularly the sixth and seventh nerves.
Clinical course is more protracted in adults when compared to children (Guillamo et al., 2001). Headache is
common and can be indistinguishable from migraine
including the presence of visual aura (Lim et al., 2005).
The single involvement of the facial nerve may initially
suggest Bells palsy (Walker et al., 1999). Very rarely,
spontaneous remission can occur (Lenard et al., 1998).
Duration of symptoms in patients with cervicomedullary gliomas range from months to years and
it is shorter in fibrilary astrocytomas than in pilocytic astrocytomas and gangliogliomas (Di Maio et al.,
2009). Due to preponderance of low-grade tumors, the
clinical course is protracted and delay in diagnosis is
frequent. Clinical manifestations are related to the primary site of the lesion, medulla or spinal cord, but in
most patients, symptoms are overlapping. Involvement
of medulla leads to lower cranial nerve dysfunction.
Nasal speech, swallowing problems, head tilt, palate
deviation and facial nerve palsies dominate the clinical
picture. Nausea and vomiting are particularly common
and related to bulbar compression and not to increased
intracranial pressure. Children may present with failure to thrive. Predominant spinal cord lesions produce
weakness, muscle wasting, sphincteric disturbances
and may lead to respiratory failure.
Patients with dorsally exophytic tumors present
with headache, vomiting, failure to thrive, torticollis
and ataxia. Signs of long tract dysfunction are not common at presentation, but may appear over the course of
the disease.
Differential Diagnosis
Infectious, autoimmune, vascular, metabolic and
tumoral disorders may have a similar clinical or
radiological presentation of a brainstem glioma
389
Table 38.1 Differential diagnosis of brainstem gliomas
Infectious
Tuberculosis
Listeria monocytogenes infection
Toxoplasmosis
Pyogenic abscess
Viral rhombencephalitis
Progressive multifocal leukoencephalopathy
Autoimmune
Sarcoidosis
NeuroBehet
Multiple sclerosis
Neuromyelitis optica
Systemic eritematous lupus
Vascular
Haematomas
Vascular malformations
Vasculitis
Infarction
Posterior reversible leukoencephalopathy
Metabolic/inherited
Alexander disease
Neurofibromatosis type 1
Tumoral
Lymphoma
Metastasis
Germinoma
Acoustic neuroma
(Table 38.1) (Laigle-Donadey et al., 2008; Park et al.,

2007).
Many infectious agents can penetrate de CNS and
provoke lesions in the brainstem. Usually, the process
is acute or subacute and clinical picture evolves over
days. Conversely, brainstem tumors grow more insidiously and years may elapse until diagnosis. Fever and
systemic signs are clues for an infectious process as
well as the presence of immune system dysfunction.
Tuberculosis, toxoplasmosis and Listeria monocytogenes infection predominantly occur in immunossupressed individuals. Cerebrospinal fluid analysis can
be helpful in order to determine the specific agent.
Inflammatory lesions secondary to autoimmune
process may present with mass effect and/or contrast
enhancement, mimicking brainstem tumors. When the
neurological lesion is the first manifestation of the disease, diagnosis is particularly challenging and may be
obtained only after brain biopsy. A relapsing remitting
course, the predilection for young women, and involvement of other sites of the neuraxis suggests multiple
390
sclerosis. Neuromyelits optica affects mainly the optic

nerves and spinal cord, but is not restricted to these
sites. Anti-acquaporine 4 antibody is present in more
than half of the cases. Sole involvement of CNS in
sarcoidosis is rare and other organs such as lung and
lymph nodes are affected first. Patients with Behet
disease can present with a pseudotumoral brainstem
lesion, which can be hard to differentiate from a tumor.
History of eye lesions, arthritis and genital and oral
ulcers should be inquired and are crucial for diagnosis.
Hemorrhagic or ischemic strokes produce sudden
symptoms and signs in most patients and only rarely
can be mistaken for a brainstem glioma. However,
a progressive or relapsing course can be seen in
brainstem vasculitis and vascular malformations, casting doubt in diagnosis. In these cases, imaging (MR
angiography, CT angiography and digital subtraction
angiography) is particularly helpful.
NF1 patients may develop T2-weighted hyperintense lesions on MRI named unidentified bright objects
in the brainstem. These lesions spontaneously disappear during childhood but can be confused with a
tumoral lesion. Alexander disease is a rare degenerative leukodystrophy secondary to a mutation in the
glial fibrillary acidic protein (GFAP) gene. In juvenile and adult types, a predominant bulbar involvement
can occur. Diagnosis is made through MRI and genetic
testing.
Brainstem can be affected by other tumors including
primary CNS neoplasms and metastasis. The clinical
and radiological aspects of these lesions are reviewed
elsewhere in this book.
Diagnosis
Imaging is fundamental to determine the location and
extension of brainstem gliomas. CT imaging can show
cysts, calcification or the presence of hydrocephalus,
but its sensitivity for posterior fossa lesions is low
and definition is poor. MRI is the preferred method
since it can differentiate tumors from inflammatory and
vascular lesions and special techniques such as spectroscopy and perfusion/diffusion sequences can help to
determine tumor grade before surgery.
Tectal gliomas are iso or hypointense in
T1-weighted and hyperintense in T2-weighted
and FLAIR sequences (Fig. 38.1). Lesions are located
M.A. Lima
Fig. 38.1 A fluid attenuation inversion recovery (FLAIR) MR

image showing a hyperintense lesion in tectum. Biopsy revealed
a pilocytic astrocytoma
in the tectal plate, well circumscribed and show

contrast enhancement in less than 20% of the cases.
Extension into the tegmentum or thalamus may
occur and ventricular dilatation is frequently present
(Squires et al., 1994).
MRI is particularly important for diagnosis of
diffuse pontine gliomas in children, since it has
replaced biopsy as gold standard for treatment decisions. Schumacher et al. conducted a multicentric
study with 142 patients with brainstem disease, including 78 patients with brainstem gliomas (Schumacher
et al., 2007). When clinical, immunological, MRI and
CSF data were considered, it was possible to differentiate tumoral form non-tumoral lesions in more than
95% of the cases, obviating the need of histological confirmation. On the other hand, it is believed
that pontine gliomas are uniformly high-grade, but
in a recent study with 24 patients that underwent
biopsy for pontine gliomas, two had low-grade tumors
(grade II glioma and pilocytic astrocytoma) (Roujeau
et al., 2007). In the same study, the procedure was
considered safe with minimal morbidity. Nowadays,
391
Fig. 38.2 (a) A hypointense sagital T1-weighted and (b) a hyperintense FLAIR lesion with ill defined margins typical of a pontine
glioma
Fig. 38.4 A dorsally exophytic tumor with heterogeneous contrast enhancement is seen in T1-weighted MR sequence
Fig. 38.3 T2-weighted sagital MR sequence showing an extensive cervicomedullary tumor from the medulla to C6
most specialized centers indicate biopsy only in atypical cases. Characteristically, there is a diffuse enlargement of pons and tumor limits are ill defined due to
its infiltrative nature (Fig. 38.2). The lesion is hyperintense in T2-weighted and mostly hypointense in
T1-weighted sequences. Contrast enhancement is not
usually observed and its presence is of no prognostic significance. Engulfment of the basilar artery can
occur (Jallo et al., 2004). MR spectroscopy may be a
useful adjuvant in the diagnosis and follow-up of the
pontine gliomas. Elevated choline/creatine (Cho/Cr)
392
and Cho/N-acetylaspartate (NAA) ratios are usually

observed (Thakur et al., 2006).
Upon MRI, cervicomedullary gliomas are homogeneous T1-weighted iso or hypointense and
T2-weighted hyperintense lesions (Fig. 38.3). Di
Maio et al. reported that tumors in 8/9 patients showed
contrast enhancement (Di Maio et al., 2009). Their
growth is limited rostrally by the decussating fibers
of the corticospinal tract and tumor expands caudally
into the spinal cord.
Dorsally exophytic tumors are well-demarcated
lesions located in the floor of the fouth ventricle and
expand dorsolaterally (Fig. 38.4). Similar to other
brainstem tumor, they show T1-weighted hypointensity and T2-weighted hyperintensitiy. Most often,
gadolinium enhancement is homogeneous.
Treatment
General measures in the treatment of brainstem
gliomas include analgesia, management of bulbar
symptoms and motor impairment, relief of hydrocephalus and emotional support. Palliation is the main
component of care in patients with diffuse pontine
gliomas. These issues are best addressed by a multidisciplinary team that should include oncologists, neurologists, neurosurgeons, speech and physical therapists
in a brain tumor center.
Tectal gliomas are indolent lesions that usually
present with signs of raised intracranial pressure. The
initial approach is aimed to relieve hydrocephalus
either with ventricular shunt placement or third ventriculostomy. While shunt placement has been used
traditionally, many patients need revision due to infection or malfunction. More recently, third ventriculostomy has become the preferred method by many
neurosurgeons. It is safe and efficacious in the short
and long term. Li et al. attempted third ventriculostomy in 18 tectal glioma patients and was successful in all. At follow-up, 89% of the individuals
remained shunt-free (Li et al., 2005). The procedure
can be repeated if the ventriculostomy has closed.
Surgical resection soon after diagnosis is controversial.
While some groups favor early resection if complete
removal of the mass seems feasible, others sustain a
conservative approach and reserve surgery for clinical
or radiological (contrast enhancement, enlargement
M.A. Lima
or cystic changes) progression. Another approach is

stereotactic biopsy followed by radiation therapy as
described below. In all cases, intraoperative neurophysiological monitoring should be performed in order
to avoid excessive damage.
Radiation therapy can be used as an initial approach
soon after stereotactic biopsy in patients with high
grade lesions or, late in the course of the disease, after
clinical progression of low grade tumors. Either conformal radiation therapy or stereotactic radiosurgery
(SR) have been used depending on the size, presence
of distinct tumor margins and experience at the center. Doses are in the 912 Gy range to the tumor
margins when SR is employed. In order to avoid radiation induced damage, peripheral doses higher than
14 Gy should not be used (Kihlstrom et al., 1994).
Chemotherapy has a limited role in the management
of tectal gliomas, being used mostly after clinical
progression.
Surgery has no role for diffuse pontine gliomas,
except in cases of CSF diversion for hydrocephalus.
The diffuse and infiltrative nature of these tumors does
not produce any well-defined margins. As described
above, it is consensus that biopsy is not necessary
for most patients, and it should be performed only in
atypical cases.
The mainstay of treatment for diffuse pontine
gliomas is focal conventional radiation therapy. Doses
from 50 to 60 Gy divided in daily fraction of 1.52 Gy
over 6 weeks are used (Laigle-Donadey et al., 2008).
This approach leads to clinical improvement in over
75% of the patients initially but does not improve
survival (Hargrave et al., 2006). The median onset
of disease progression after radiation therapy is < 6
months. The radiological response occurs only in half
of the cases. In the last two decades, several studies
have investigated alternative radiation protocols such
as hyperfractionation and dose escalation. In a large
phase III trial, 130 patients were randomly assigned to
receive local field radiation at doses of 1.8 Gy daily
to a total dose of 54 Gy or a twice a day regimen
of 1.17 Gy per fraction to a total dose of 70.2 Gy
over 6 weeks (Mandell et al., 1999). The results were
disappointing: in the conventional radiation group the
median time to death was 8.5 months and in the hyperfractionation group was 8 months. In another study, 66
children underwent treatment with 1 Gy twice daily to
a total dose of 78 Gy (Packer et al., 1994). This regimen resulted in a prolonged dependency of steroids
and 9/66 patients showed radiological evidence of radiation induced necrosis. There was no improvement in
survival when compared to the standard regimen.
Various single or multiple drug regimens have
been tested in several trials for management of diffuse pontine gliomas either alone or in combination
with radiation therapy. Regimens that used chemotherapy prior to radiation therapy usually failed due
to early disease progression. Temozolomide, a drug
that has extensively used in supratentorial high grade
gliomas improving survival, had no impact in diffuse pontine gliomas either with concurrent radiation therapy or at clinical progression (De Sio et al.,
2006). Bevacizumab, an antiangiogenic agent that
has been used in recurrent high grade supratentorial gliomas also had no impact in the outcome
(Torcuator et al., 2009), but may be useful for radiation induced necrosis in pontine glioma patients
(Liu et al., 2009). Radiosensitizing drugs also had
disappointing results. Overall, chemotherapy has not
brought any substantial impact in the prognosis of
these patients. Newer approaches such as local delivery of chemotherapeutic agents in order to bypass the
blood brain barrier, immunotherapy and gene therapy are currently being evaluated. Due to the poor
response to the conventional therapies, trials with
experimental treatments should be offered to these
patients.
Surgical resection is recommended initially for cervicomedullary gliomas, but morbidity of a radical excision may be important. In most series, biopsy or subtotal excision of the tumor can be performed safely. Di
Maio et al. achieved subtotal resection in 6/7 patients
(Di Maio et al., 2009), while 9/11 patients underwent
subtotal resection in Poussaint series (Young Poussaint
et al., 1999). Neurological deficits can be transient or
permanent, especially bulbar cranial nerve dysfunction
and respiratory function. In cases of extensive spinal
cord manipulation, intubation for 4872 h after surgery
is advised (Jallo et al., 2004). Early tracheostomy and
feeding gastrostomy may be necessary in the presence
of swallowing impairment. Cervical kyphosis is a late
complication in patients with cervical laminectomies
at the time of surgery due to spinal cord extension of
the tumor.
Since most tumors in the cervicomedullary region
are low grade and grow slowly, radiation therapy is
deferred until recurrence or clinical progression with
the aim of prevents its deleterious effects in the CNS
393
(Laigle-Donadey et al., 2008). In patients with high

grade gliomas, radiation therapy can be performed
soon after biopsy or resection of the lesion.
Chemotherapy role is yet to be defined in cervicomedullary gliomas. It is usually reserved for recurrence, but a recent study showed that chemotherapy soon after subtotal resection may improve
tumor/brainstem interface, opening the possibility for
a more extensive resection (Di Maio et al., 2009).
Surgical resection is indicated for dorsally exophytic tumors. Advances in neurosurgical and imaging
techniques allow extensive resection, although total
removal of the lesion is not possible in most cases
(Jallo et al., 2004). The goal is the removal of the
extrinsic component following by debulking of the
intrinsic component in order to prevent neurological
morbidity (ataxia, long tract dysfunction and nystagmus). Due to the benign nature of these tumors, radiation therapy and chemotherapy are used in cases of
recurrence or progression.
Outcome
The outcome of brainstem gliomas is related to the
location and histological grade of the tumors. Tectal
lesions portray a good prognosis and tumors may
remain stable for years. Stark et al. followed 12 children for a median time of 9.5 years (Stark et al.,
2005). Radiological progression was observed in three
patients, but no clinical deterioration occurred. All
children had a good neurological performance at the
end of follow-up. In Poussaint series, 20/32 children were stable after a mean follow up of 5 years
(Poussaint et al., 1998). For the 12 remaining patients
who showed clinical or radiological progression, eight
showed decreased tumor size and three showed stable
residual lesion after surgery and/or radiation therapy.
The prognosis of diffuse pontine gliomas remains
dismal despite many experimental protocols of radiation and chemotherapy. The mean survival is less than
1 year in most series and the 2-year survival is only
10% (Broniscer and Gajjar, 2004). Age at onset of disease is prognostic since survival is longer in adults then
children (Guillamo et al., 2001; Kesari et al., 2008).
Tumor control can be achieved in most patients
with cervicomedullary gliomas. In a study with 39
patients, the 5-year progression-free and total survivals
394
were 60 and 89% respectively. In another study, 6/11

patients were stable after surgery at a mean follow-up
of 5.2 years (Young Poussaint et al., 1999).
Dorsally exophytic tumors have a good outcome
after surgical resection. Among 18 patients who underwent surgery, only four showed evidence of radiological progression after a median follow-up of 113
months (Pollack et al., 1993).
References
Barkovich AJ, Krischer J, Kun LE, Packer R, Zimmerman RA,
Freeman CR, Wara WM, Albright L, Allen JC, Hoffman HJ
(1990) Brain stem gliomas: a classification system based on
magnetic resonance imaging. Pediatr Neurosurg 16(2):7383
Broniscer A, Gajjar A (2004) Supratentorial high-grade astrocytoma and diffuse brainstem glioma: two challenges for the
pediatric oncologist. Oncologist 9(2):197206
Daglioglu E, Cataltepe O, Akalan N (2003) Tectal gliomas in
children: the implications for natural history and management strategy. Pediatr Neurosurg 38(5):223231
De Sio L, Milano GM, Castellano A, Jenkner A, Fidani P,
Dominici C, Donfrancesco A (2006) Temozolomide in resistant or relapsed pediatric solid tumors. Pediatr Blood Cancer
47(1):3036
Di Maio S, Gul SM, Cochrane DD, Hendson G, Sargent MA,
Steinbok P (2009) Clinical, radiologic and pathologic features and outcome following surgery for cervicomedullary
gliomas in children. Childs Nerv Syst 25(11):14011410
Epstein F (1985) A staging system for brain stem gliomas.
Cancer 56(7 Suppl):18041806
Epstein FJ, Farmer JP (1993) Brain-stem glioma growth patterns. J Neurosurg 78(3):408412
Ferner RE (2007) Neurofibromatosis 1 and neurofibromatosis
2: a twenty first century perspective. Lancet Neurol 6(4):
340351
Guillamo JS, Creange A, Kalifa C, Grill J, Rodriguez D,
Doz F, Barbarot S, Zerah M, Sanson M, Bastuji-Garin S,
Wolkenstein P (2003) Prognostic factors of CNS tumours
in Neurofibromatosis 1 (NF1): a retrospective study of 104
patients. Brain 126(Pt 1):152160
Guillamo JS, Monjour A, Taillandier L, Devaux B, Varlet P,
Haie-Meder C, Defer GL, Maison P, Mazeron JJ, Cornu P,
Delattre JY (2001) Brainstem gliomas in adults: prognostic
factors and classification. Brain 124(Pt 12):25282539
Hargrave D, Bartels U, Bouffet E (2006) Diffuse brainstem
glioma in children: critical review of clinical trials. Lancet
Oncol 7(3):241248
Jallo GI, Biser-Rohrbaugh A, Freed D (2004) Brainstem
gliomas. Childs Nerv Syst 20(3):143153
Kesari S, Kim RS, Markos V, Drappatz J, Wen PY, Pruitt AA
(2008) Prognostic factors in adult brainstem gliomas: a multicenter, retrospective analysis of 101 cases. J Neurooncol
88(2):175183
Kihlstrom L, Lindquist C, Lindquist M, Karlsson B (1994)
Stereotactic radiosurgery for tectal low-grade gliomas. Acta
M.A. Lima
Laigle-Donadey F, Doz F, Delattre JY (2008) Brainstem gliomas
in children and adults. Curr Opin Oncol 20(6):662667
Lenard HG, Engelbrecht V, Janssen G, Wechsler W, Tautz C
(1998) Complete remission of a diffuse pontine glioma.
Neuropediatrics 29(6):328330
Li KW, Roonprapunt C, Lawson HC, Abbott IR, Wisoff J,
Epstein F, Jallo GI (2005) Endoscopic third ventriculostomy
for hydrocephalus associated with tectal gliomas. Neurosurg
Focus 18(6A):E2
Lim EC, Wilder-Smith EP, Chong JL, Wong MC (2005)
Seeing the light: brainstem glioma causing visual auras and
migraine. Cephalalgia 25(2):154156
Liu AK, Macy ME, Foreman NK (2009) Bevacizumab as therapy for radiation necrosis in four children with pontine
gliomas. Int J Radiat Oncol Biol Phys 75(4):11481154
Mandell LR, Kadota R, Freeman C, Douglass EC, Fontanesi
J, Cohen ME, Kovnar E, Burger P, Sanford RA, Kepner J,
Friedman H, Kun LE (1999) There is no role for hyperfractionated radiotherapy in the management of children with
newly diagnosed diffuse intrinsic brainstem tumors: results
of a Pediatric Oncology Group phase III trial comparing
conventional vs. hyperfractionated radiotherapy. Int J Radiat
Oncol Biol Phys 43(5):959964
Packer RJ, Boyett JM, Zimmerman RA, Albright AL, Kaplan
AM, Rorke LB, Selch MT, Cherlow JM, Finlay JL, Wara
WM (1994) Outcome of children with brain stem gliomas
after treatment with 7800 cGy of hyperfractionated radiotherapy. A Childrens Cancer Group Phase I/II Trial. Cancer
74(6):18271834
Park KY, Ahn JY, Cho JH, Choi YC, Lee KS (2007)
Neuromyelitis optica with brainstem lesion mistaken for brainstem glioma. Case report. J Neurosurg
107(3 Suppl):251254
Pollack IF, Hoffman HJ, Humphreys RP, Becker L (1993)
The long-term outcome after surgical treatment of dorsally
exophytic brain-stem gliomas. J Neurosurg 78(6):859863
Poussaint TY, Kowal JR, Barnes PD, Zurakowski D, Anthony
DC, Goumnerova L, Tarbell NJ (1998) Tectal tumors of
childhood: clinical and imaging follow-up. AJNR Am J
Neuroradiol 19(5):977983
Qaddoumi I, Sultan I, Gajjar A (2009) Outcome and prognostic
features in pediatric gliomas: a review of 6212 cases from the
surveillance, epidemiology, and end results database. Cancer
115(24):57615770
Roujeau T, Machado G, Garnett MR, Miquel C, Puget S,
Geoerger B, Grill J, Boddaert N, Di Rocco F, Zerah M,
Sainte-Rose C (2007) Stereotactic biopsy of diffuse pontine
lesions in children. J Neurosurg 107(1 Suppl):14
Schumacher M, Schulte-Monting J, Stoeter P, Warmuth-Metz
M, Solymosi L (2007) Magnetic resonance imaging compared with biopsy in the diagnosis of brainstem diseases of
childhood: a multicenter review. J Neurosurg 106(2 Suppl):
111119
Selvapandian S, Rajshekhar V, Chandy MJ (1999) Brainstem
glioma: comparative study of clinico-radiological presentation, pathology and outcome in children and adults. Acta
Neurochir (Wien) 141(7):721726; discussion 726727
Smith MA, Freidlin B, Ries LA, Simon R (1998) Trends in
reported incidence of primary malignant brain tumors in
children in the United States. J Natl Cancer Inst 90(17):
12691277

Squires LA, Allen JC, Abbott R, Epstein FJ (1994) Focal tectal tumors: management and prognosis. Neurology 44(5):
953956
Stark AM, Fritsch MJ, Claviez A, Dorner L, Mehdorn HM
(2005) Management of tectal glioma in childhood. Pediatr
Neurol 33(1):3338
Ternier J, Wray A, Puget S, Bodaert N, Zerah M, SainteRose C (2006) Tectal plate lesions in children. J Neurosurg
104(6 Suppl):369376
Thakur SB, Karimi S, Dunkel IJ, Koutcher JA, Huang W (2006)
Longitudinal MR spectroscopic imaging of pediatric diffuse
pontine tumors to assess tumor aggression and progression.
AJNR Am J Neuroradiol 27(4):806809
395
Torcuator R, Zuniga R, Loutfi R, Mikkelsen T (2009)
Bevacizumab and irinotecan treatment for progressive diffuse brainstem glioma: case report. J Neurooncol 93(3):
409412
Ueoka DI, Nogueira J, Campos JC, Maranhao Filho P, Ferman S,
Lima MA (2009) Brainstem gliomasretrospective analysis
of 86 patients. J Neurol Sci 281(12):2023
Walker DA, Punt JA, Sokal M (1999) Clinical management of
brain stem glioma. Arch Dis Child 80(6):558564
Young Poussaint T, Yousuf N, Barnes PD, Anthony
DC, Zurakowski D, Scott RM, Tarbell NJ (1999)
Cervicomedullary astrocytomas of childhood: clinical
and imaging follow-up. Pediatr Radiol 29(9):662668
Chapter 39
Tumor-Associated Epilepsy in Patients with Glioma

Anja Smits and Anette Storstein
Abstract Epileptic seizures are common symptoms

in patients with glioma, occurring at disease presentation as well as during later disease. The frequency of
tumor-related seizures is strongly related to the growth
rate of the tumor. In low-grade gliomas, seizures as
initial symptom leading to the brain tumor diagnosis occur in 7090% of all patients. About 50% of
these patients continue to have seizures before operation in spite of optimal antiepileptic drug treatment.
In high-grade gliomas, seizures at disease presentation are less frequent and occur in only 3050% of
all patients but may be more difficult to control. Apart
from the tumor growth rate, the specific location of the
tumor in the brain and its proximity to the cortex affect
seizure risk and seizure control in glioma. For many
patients with tumor-related seizures, optimal seizure
control is not achieved by antiepileptic drugs only but
requires additional therapy such as surgical resection
and/or radio- and chemotherapy. In this review, we
discuss different aspects of the clinical presentation
of the disease in patients with tumor-related seizures.
We also discuss the possible therapeutic strategies
for patients with medically refractory seizures, illustrating the importance of a qualified specialized
team of neuro-oncologists, neurologists and neurosurgeons for the optimal management of this patient
group.
A. Smits ()
Department of Neuroscience, Neurology, University Hospital
Uppsala, S-751 85 Uppsala, Sweden
e-mail: Anja.smits@neuro.uu.se
Keywords Epilepsy Glioma Antiepileptic drugs

Epileptogenesis Differential diagnoses Ischemic
infarction
Introduction
Epileptic seizures are common in patients with
gliomas, affecting the majority of patients with lowgrade tumors and a substantial proportion of patients
with high-grade glioma. Seizures occur as initial symptoms at disease onset and are frequent symptoms
of chronic tumor disease as well. With tumor treatments for patients with glioma becoming more effective resulting in longer patient survival, the burden
of chronic epilepsy will add substantial morbidity to
this patient group. Several studies have shown that
persistent seizures in patients with brain tumors have
sustained negative effects on daily functioning and
quality of life (Klein et al., 2003). Optimal seizure
control is therefore an important part of the clinical
management in patients with brain tumors. Cognitive
dysfunction, another main symptom in patients with
glioma, is often closely related to both refractory
seizures and antiepileptic therapy. The clinical management of cognitive dysfunction in patients with brain
tumors is discussed elsewhere.
The incidence of seizures in patients with glioma
is highly variable and is affected by several factors.
The location of the tumor in the brain, the proximity
to adjacent cortex and the growth rate of the tumor
are important parameters determining seizure risk. The
majority of patients with slowly growing low-grade
gliomas presents with epileptic seizures at disease
onset or within the first year from tumor diagnosis,

397
398
often as the only symptom. In patients with high-grade

gliomas, epileptic seizures occur less frequently both
at disease onset as well as during later course of disease. In addition to symptomatic seizures, these tumors
cause neurological deficits as a result of the destructive
growth in the brain.
Apart from the evident variability in seizure risk
at disease onset, the seizure control that can be
obtained in patients with gliomas is highly variable. Tumor-related seizures in high-grade gliomas
are less frequent, but the epileptic symptoms of these
patients may be difficult to control. Many patients with
slowly growing gliomas suffer from chronic epilepsy
during the entire course of disease, whereas others
become seizure-free after their first initial seizure when
antiepileptic treatment is started. Add-on treatment
with new antiepileptic drugs may increase the proportion of seizure-free patients in the glioma population,
but further studies need to assess the efficiency of
these drugs in glioma-related seizures as well as the
side effects induced by polypharmacy. Other treatment
options should be considered when antiepileptic drug
treatment fails. Surgical resection including resection
of the epileptic foci is known to have beneficial effects
on seizure control, especially in low-grade gliomas.
Recent reports have shown positive effects with seizure
reduction obtained by radiotherapy and chemotherapy as well in this patient group. In general, however,
current therapy of tumor-related seizures is far from
perfect (van Breemen et al., 2007). A deeper understanding of the underlying mechanisms of brain-tumor
epilepsy is a prerequisite for improving seizure control
and optimizing the clinical management of gliomarelated seizures.
Epidemiology
In general, slowly growing tumors are significantly
more epileptogenic than fast growing tumors. Epileptic
seizures as initial symptoms leading to the brain tumor
diagnosis occur in 7090% of low-grade gliomas and
in 3050% of high-grade gliomas. Some specific histological tumor subtypes characterized by a typically
indolent course of disease, such as dysembryoblastic
neuroepithelial tumors (DNET) and gangliogliomas,
are associated with a seizure frequency of almost
100%. In contrast, the incidence of seizures as first
A. Smits and A. Storstein
symptoms in metastatic brain disease is as low as

2030% (van Breemen et al., 2007).
Around 45% of all adult-onset seizures are caused
by brain tumors (Olafsson et al., 2005). However,
seizures may go unnoticed for long time periods. The
difficulty in defining initial seizures in adults as first
symptoms at the time point of disease onset is illustrated in a study by King et al. (1998) who investigated
300 consecutive adults and children presenting with an
unexplained seizure. A clinical diagnosis of generalized or focal epilepsy could be made in 141 (47%)
of these patients, and imaging of the brain by CT or
MRI identified 38 epileptogenic lesions (17 of these
were tumors). It was demonstrated that 17% of the
patients who presented with their first tonic-clonic
seizure reported similar events in the past and 28%
had earlier experienced focal epileptic symptoms, such
as temporal lobe auras due to non-convulsive seizures.
This study confirms the common experience by many
doctors that patients may not be aware of or not appreciate the importance of minor epileptic symptoms, and
fail to mention these in an initial interview. As a consequence, the brain tumor diagnosis may be delayed
for months or, in case of slowly growing tumors, up to
several years.
Epileptogenesis in Gliomas
The mechanisms behind the development of tumorrelated seizures are likely to involve tumor-related as
well as individual patient-related factors but are incompletely understood. The seizure risk and the individual
response to antiepileptic drug treatment are variable
within subgroups of patients with similar tumor localization and histology. Thus, apart from tumor location
and tumor growth rate, other factors that are related to
the genetic susceptibility of the individual patient and
the intrinsic properties of the host brain may contribute
to the development of glioma-related seizures.
A putative mechanism of acquiring intrinsic epileptic properties is by altered expression of signaling
molecules involved in neuronal circuitries. According
to the modern theory of neuronal networks, synchronization of neurons in complex networks is the prerequisite for normal neurological functioning. Due to
modifications in such existing circuitries by for example tumor growth, active processes of axonal sprouting,
39 Tumor-Associated Epilepsy in Patients with Glioma
synaptogenesis and neurogenesis will take place in the

peritumoral region (Brogna et al., 2008). Such excessive and random synchronization of neurons may persist during critical periods of time, typically in slowly
growing tumors, and cause epilepsy. This theory provides new insights also in the underlying mechanisms
of cognitive disturbances, the other main symptom of
disease in low-grade gliomas. Thus, disarrangement
of the same neuronal cerebral networks may lead to
cognitive disturbances, as a result of destruction, and
epileptic seizures, as a result of pathological excess of
functional network structures (Brogna et al., 2008).
Pathophysiology
The pathophysiology of seizure development in
patients with glioma is thought to occur through different mechanisms in high-grade and low-grade tumors.
In fast growing high-grade gliomas, focal ischemia
and deafferentiation of cortical areas due to mass
effect are likely to be important causative factors
(Fig. 39.1), whereas gliosis and chronic inflammatory changes in peritumoral regions of slowly growing
gliomas are considered to predispose for epileptic
seizures (Fig. 39.2). Increased levels of Fe3+ ions
deposits in intra- or peritumoral areas, due to small
bleedings from pathological blood vessels, typically
occur in high-grade gliomas and may further contribute to the development of tumor-associated seizures
(Fig. 39.1) (Beaumont and Whittle, 2000; Schaller
and Regg, 2003). The almost 100% association of
epileptic seizures in patients with specific histological
types of neuroglial tumors, such as DNET and gangliogliomas, has raised the question whether these tumors
contain intrinsic epileptogenic properties. It has been
speculated that the specific neurochemical profiles of
these lesions predispose for generating epileptic activity (Beaumont and Whittle, 2000).
Tumor Characteristics
The inverse relationship between tumor growth rate
and seizure risk is true also for patients with tumors
of similar histological malignancy grade. Thus, within
the group of grade II gliomas, the slowly growing
oligodendrogliomas are more epileptogenic than the
399
diffuse astrocytomas. Other important factors underlying the development of epilepsy are the localization
of the tumor in the brain and the proximity with
the cortical gray matter (Schaller and Ruegg, 2003).
Tumors with location in the vicinity of the primary
motor cortex and with limbic and perilimbic cortical
localization are highly epileptogenic, whereas occipital tumors are less likely to manifest with seizures.
Low-grade gliomas are more frequently than de novo
glioblastomas situated in eloquent cortico-subcortical
regions, such as the supplementary motor area (SMA)
and the insular and peri-insular area. These locations
are highly epileptogenic areas and strongly associated with medically refractory seizures (Duffau and
Capelle, 2004).
Peritumoral Brain
In tumor-associated epilepsy, the tumor somehow acts
as a generator to produce an epileptic focus in the
brain. In many cases, however, the epileptic focus,
i.e. the location corresponding with the start of the
seizure, is not contiguous with the tumor. The interplay between the tumor histology and the surrounding tissue is of importance for seizure development,
especially in chronic epilepsy of patients with slowly
growing gliomas. Alterations in the ultrastructure of
the peritumoral brain regions have been demonstrated
in patients with tumor-associated seizures, as well as
changes in metabolic activity and in pH (Beaumont
and Whittle, 2000). The presence of necrotic tissue and
the hypoxia of fast growing high-grade gliomas due to
pathological vascularization may induce changes in the
adjacent regions with the potential to increase neuronal
excitability (Fig. 39.1).
Some patients with chronic tumor-related seizures
may develop secondary epileptic foci that do not correspond to the tumor or peritumoral area, although
this phenomenon is still far more controversial in
human epilepsy compared to animal models (Cibula
and Gilmore, 1997). It is clear though that epileptic
activity in regions distant to the tumor is more frequently found in patients with a long seizure history
and with temporal lobe tumors. Consistent with the
existence of separate epileptic foci, these patients tend
to have two different types of focal seizures (Cibula
and Gilmore, 1997). Unless complete resection of
400
a
c
b
d
Fig. 39.1 Micrographs of histopathological sections of a

glioblastoma (kindly provided by Dr Aronica, Department of
Neuropathology, University of Amsterdam, the Netherlands).
(a) Hematoxylin-eosin staining shows characteristic tumor
morphology with necrotic areas; (b) Immunostaining with
anti-GFAP-antibody shows invasion of the tumor in the cortex;

(c) Hematoxylin-eosin staining at higher magnification shows
polymorphic tumor cells and (d) proliferating blood vessels;
(e) Immunostaining with anti-Ki6 antibody shows a large
amount of proliferating tumor cells
Fig. 39.2 Micrographs of histopathological sections of an

astrocytoma grade II (kindly provided by Dr Aronica,
Department of Neuropathology, University of Amsterdam,
the Netherlands). (a) Hematoxylin-eosin staining shows a
well-differentiated tumor at low, respectively; (b) high
magnification; (c) and the infiltration of the tumor in the cortex;
f
(d) Immunostaining with anti-GFAP antibody shows immunoreactive astrocytic tumor cells; (e) whereas the tumor cells
lack immunoreactivity for neuronal nuclear protein; (f)
Immunostaining with anti-Ki6 antibody shows few (<4%) proliferating tumor cells
the epileptogenic zones is performed in addition to

tumor resection, identified by preoperative foci mapping, these patients will continue to have postoperative
seizures.
Genetic Factors
It is likely that genetic factors play a causative role in
the large variability of epileptic symptoms in patients
with glioma. No candidate genes have been identified
yet that can be associated with susceptibility for tumorassociated epilepsy in glioma (Berntsson et al., 2009).
A number of monogenetic diseases are known with
an increased risk for developing epilepsy. Of particular interest for glioma-related seizures are mutations
in the tumor-suppressor gene LGl1. This gene, located
at the 10q24 region, codes for the leucine-rich glioma
inactivated protein 1 (LGI1) or epitempin. LGI1 plays
a role in glioma progression, probably by regulation
of cell migration and invasion, and mutations of this
gene cause the syndrome autosomal dominant partial
epilepsy with auditory features, a rare form of lateral temporal lobe epilepsy (Kalachikov et al., 2002).
The exact mechanism of how LGII contributes to the
epileptogenesis of glioma is not known.
Clinical Manifestations
Clinical manifestations of tumor-related seizures
include all types of partial seizures with or without secondary generalization. Sometimes the duration of the
focal seizure component is extremely short and these
clinical seizure manifestations may erroneously be
classified as primary generalized tonic-clonic seizures.
Secondary generalization is more common in early
seizures, i.e. seizures that occur at disease presentation and before antiepileptic treatment is started.
Once antiepileptic drug therapy has been initiated, the
seizures tend to become focal only (Hildebrand et al.,
2005).
Seizure Types
Complex focal seizures are more common in patients
with slowly growing gliomas and chronic epilepsy,
401
whereas simple focal seizures occur more frequently

in high-grade gliomas.
There is often a good correlation between tumor
location and seizure type, although it may be extremely
difficult to differentiate on clinical grounds only,
for example between temporal lobe and frontal lobe
seizures (Manford et al., 1996).
In general, manifestations of behavioral automatisms and experiential symptoms are strongly associated with temporal lobe lesions. Tonic movements and
head turning are, if appearing at seizure start, strongly
associated with frontal tumor location. Adversive
seizures due to frontal lobe lesions are characterized by
turning of the neck, head and eyes to one side. Another
typical manifestation of frontal lobe epilepsy is motor
agitation, characteristically occurring at night or upon
awakening in early morning, occurring in patients with
deep frontal lobe tumors. Focal Jacksonian motor or
somatosensory seizures occur with tumor location in
perirolandic areas.
Patients with temporal lobe tumors may report characteristic focal complex seizures with auditory hallucinations, language or memory disturbances, dj vu
phenomena, epigastric or gustatory or olfactory auras.
They can also experience isolated auras without other
signs of epileptic manifestations. In general, complex
automatisms are strongly associated with temporal
lobe lesions.
Seizures in patients with occipital lobe tumor location typically consist of positive visual symptoms as
hallucinations (flickering or blinking lights) and other
visual disturbances but may also include visual field
defects or blurred vision. Thus, a differential diagnosis
of migraine is sometimes difficult to exclude even in
patients with known lesions of the occipital lobe.
Status Epilepticus
Considering the frequency of epilepsy in the glioma
population, studies on status epilepticus in this patient
group are surprisingly sparse. In line with the general pattern of tumor-related seizures, status epilepticus
seems to occur most often at disease onset or at
the time of tumor recurrence. Data from the literature suggest that status epilepticus is more common
in patients with low-grade gliomas with an oligodendroglial tumor component (Cavaliere et al., 2006).
402
Seizures and Tumor Progression

Recurrent seizures after an initial seizure-free period
or late-onset seizures should raise the suspicion of
increased growth rate of the tumor in patients with a
history of glioma, although the exact correlation with
tumor progression is not known. As mentioned, the
onset of epileptic seizures in glioma patients is usually
an early symptom of disease. However, in 1015% of
all patients with glioma-related epilepsy, seizures manifest at a later stage of disease. Half of all patients with
late-onset seizures are diagnosed with a tumor recurrence or progressive tumor growth (Hildebrand et al.,
2005).
In case neuroimaging does not reveal progressive
tumor growth, seizure breakthrough due to insufficient therapeutic serum levels of antiepileptic drugs
should always be ruled out. Likewise, interactions
of antiepileptic drugs with other ongoing medication have to be considered. However, frequent breakthrough seizures may antedate radiological evidence
of tumor recurrence. Therefore, radiological investigations should be repeated over time or other imaging
modalities should be considered in patients with deterioration of seizure control.
Differential Diagnoses
The wide availability of MRI-technology has greatly
improved the efficiency of glioma diagnosis in patients
presenting with epileptic seizures. The two main
causes for delaying tumor diagnosis are clinical misinterpretation of symptoms as non-epileptogenic and
misinterpretation of radiological findings.
As a rule, a brain tumor diagnosis should be
excluded by neuroimaging in all adults presenting with
a first unprovoked epileptic seizure. Symptomatic focal
seizures are typical for tumor-related epilepsy, but
when the focal onset is of short duration and rapidly
followed by secondary generalization the seizures are
often interpreted as primary generalized. In clinical
practice, a tumor should therefore always be considered as a differential diagnosis in adult-onset epilepsy,
regardless of seizure semiology.
Another important clinical issue is that focal
epileptic seizures are quite commonly not recognized
as such and misdiagnosed by both patients and health
care professionals. The presence of focal seizures,

with or without secondary generalization, can easily
be confused with other diagnoses, resulting in delayed
tumor diagnosis. Some of the differential diagnoses,
together with their typical symptoms and signs, are
described here.
Transient Ischemic Attack (TIA)

or Ischemic Infarction
Vascular disease etiology is the main differential diagnosis in elderly individuals presenting with intermittent or persistent neurological symptoms. In vascular
disease, there is typically an abrupt loss of neurological function that, in contrast to simple focal motor or
somatosensory seizures, does not show the characteristic spread (March) to adjacent parts of the body. The
duration of symptoms induced by transient ischemic
attacks (TIA) is usually minutes hours, whereas focal
seizures are typically shorter (seconds minutes).
Syncope
In case of syncope, there is a transient self-limiting
loss of consciousness, usually preceded by symptoms
and signs of light-headedness, nausea and pallor and
followed by rapid recovery without prolonged confusion or drowsiness. Provoking factors like orthostatic
hypotension and emotional distress are often present.
Migraine Aura
A migraine aura has a typical gradual onset of neurological symptoms and lasts approximately 530 min,
often migrating from one part of the body to another.
Migraine auras may involve several neurological functions (visual, sensory, speech) and several vascular
domains. The aura is usually, but not always, followed
by headache and by symptoms such as nausea and
photophobia.
Drop Attacks
Drop attacks are characterized by abrupt loss of tonus
causing sudden falls, without any precipitating factors and without affecting consciousness. Such attacks
403
Disturbances in cardiac conduction may clinically

resemble epileptic seizures, especially when not typically provoked by physical exercise. This underlines
the importance of a thorough clinical history and physical examination together with ambulatory electrocardiographic (ECG) and electroencephalographic (EEG)
recording in these patients.
activities of daily life (Taphoorn and Klein, 2004). In

particular elderly patients or patients with prior cognitive impairment are prone to disabling sedative and
cognitive side effects of antiepileptic drugs. Finally,
elderly patients with aggressive high-grade gliomas
may experience a prolonged post-ictal loss of neurological function following generalized seizures and are
at risk of not or very slowly regaining their pre-ictal
level of functioning.
It is important to realize that in many cases optimal seizure control is not achieved by antiepileptic
drugs only but requires additional surgical resection
and/or radiotherapy and chemotherapy. Thus, a multidisciplinary team of dedicated neurologists, neurooncologists and neurosurgeons is required for optimal
clinical management of glioma-related seizures.
Psychogenic Symptoms
Antiepileptic Drug Treatment
Non-convulsive temporal lobe seizures or frontal

seizures without generalization may be misdiagnosed
as panic or anxiety attacks, hyperventilation spells or
other emotional and behavioral disorders, especially in
patients with psychiatric co-morbidity.
The choice of which antiepileptic drug to prescribe to

a patient with glioma should be made based upon drug
efficacy, pharmacodynamic profiles and side effects
of the individual substances. Antiepileptic drugs that
induce components of the hepatic cytochrome P450
enzyme system include carbamazepine, phenytoin,
phenobarbital, and to a lesser degree lamotrigine and
topiramate. Enzymatic induction may result in an
increased metabolism of chemotherapeutic agents as
well as of other antiepileptic drugs and dexamethasone, which is frequently used in patients with highgrade gliomas.
Although carbamazepine is one of the most effective and well-studied antiepileptic drugs for focal
epilepsy, its use in patients with tumor-associated
epilepsy is therefore hampered. The other drawback
of carbamazepine in patients with glioma receiving
chemotherapy is the potential risk of bone marrow
depression. However, carbamazepine is still one of the
drugs of choice for patients who are not to receive
chemotherapy within the nearest future.
Phenytoin, in spite of its potent antiepileptic effect,
is not used that often any more because of its serious side effects on cognitive function and its complex
pharmacokinetics with a high risk of toxicity. The prodrug phosphenytoin, on the other hand, is a useful drug
when administered intravenously in the emergency
situation for treating status epilepticus.
can be of purely benign nature, typically occurring

in middle-aged women. However, drop attacks may
also be caused by insufficient brain perfusion during episodes of increased intracranial pressure, due to
space occupying lesions.
Cardiac Arrythmia
Treatment
Although epileptic seizures are common and distressing symptoms of disease in patients with glioma, the
clinical management of seizures by antiepileptic treatment is not always optimal. A possible explanation for
this is that tumor treatment often has higher priority,
consuming most of the time and resources of health
care professionals, especially in the case of high-grade
gliomas. However, optimal seizure control during the
entire course of disease is of great importance for all
patients with glioma.
It has been confirmed by several studies that persistent seizures have a profound negative impact on the
quality of life for these patients (Klein et al., 2003).
Furthermore, poor seizure control often leads to multiple antiepileptic drug strategies, at the price of serious side effects such as drowsiness and disturbances
of memory and attention. Impairment of cognitive
function induced by polypharmacy may further deteriorate the capability to maintain social and professional
404
Valproate is effective in treating focal as well as

primary generalized seizures. Furthermore, the putative inherent antitumor properties of the drug, through
its inhibition of histone deacetylase, are of interest in treating patients with glioma-related seizures.
Valproate has relatively mild side effects, apart from
rare but potentially fatal liver damage. However, valproate is an enzyme-inhibiting drug. Thus, decreased
hepatic metabolism can cause increased risk of toxicity by concomitantly administered drugs such as
chemotherapeutics. Furthermore, valproate may prolong bleeding time and is therefore avoided by some
neurosurgeons.
Several of the more modern antiepileptic drugs
that are used for focal epilepsies have established an
evident role in the treatment of gliomas-associated
seizures, either in monotherapy or as add-on drug
(van Breemen et al., 2007). Commonly used and
well-tolerated antiepileptic drugs are levetiracetam,
gabapentin and lamotrigin. These drugs do not have the
strong enzyme-influencing effects of some of the older
antiepileptic drugs, and the risk of interactions with
other drugs is therefore probably smaller. On the other
hand, psychotropic side effects of these antiepileptic drugs have been reported in some patients, which
should be remembered in the clinical follow-up.
Status epilepticus in glioma patients should be
treated according to standard guidelines. Small patient
series suggest that this subgroup responds well to conventional therapy with benzodiazepines or phosphenytoin (Lowenstein and Alldredge, 1993; Cavaliere et al.,
2006).
In spite of the frequent occurrence of epilepsy
in patients with glioma, routine prophylactic use of
antiepileptic drugs for patients who have not experienced seizures is not recommended (Glantz et al.,
2000).
Refractory Seizures
Refractory seizures are defined as seizures so frequent
or sincere that they limit daily life, despite the use
of antiepileptic drugs in adequate serum concentrations (Devinsky, 1999). According to this definition,
around 35% of the overall patient population with
focal epilepsy suffer from refractory seizures. The state
of refractory seizures is commonly associated with a
structural lesion in the brain, including brain tumors.
Whereas approximately half of all adults with lowgrade gliomas presenting with seizures at disease onset
become seizure-free on antiepileptic treatment, the
other half of this population continues to have seizures
in spite of optimal antiepileptic drug treatment (Chang
et al., 2008). Some specific tumor locations in the brain
are highly epileptogenic and thus frequently associated with refractory seizures, such as the insular and
peri-insular region and the mesiotemporal lobe. These
areas constitute common localizations for low-grade
gliomas, in contrast to high-grade gliomas. Epileptic
seizures in high-grade gliomas are less frequent but
may be more difficult to control (Chaichana et al.,
2009).
The frequent lack of efficacy of antiepileptic drugs
in glioma-associated epilepsy suggests that the pharmacological mechanisms of the anticonvulsant drug do
not interfere with the critical epileptogenic pathways of
these lesions. There are a number of alternative causes
behind drug failure of glioma-related seizures, such as
insufficient concentrations of the active drug metabolite in the tumor or blood due to active defense mechanisms by multi-drug resistance proteins. This hypothesis is supported by studies showing the expression of
multi-drug resistance proteins in tumor types associated with refractory seizures (Aronica et al., 2003).
Other possible mechanisms of drug failure include
restricted penetration of lipophilic substances into the
brain and loss of receptor sensitivity for antiepileptic drugs in tumor cells. In addition, interactions with
other drugs may result in increased metabolism of
the antiepileptic drug with insufficient serum levels of active drug metabolite (van Breemen et al.,
2007).
Antitumor Treatment
Alternative treatment modalities should be considered
for patients who continue to have ongoing seizures in
spite of optimal treatment with antiepileptic drugs. The
effectiveness of surgery on seizure control has been
shown in a number of studies, in particular for patients
with low-grade gliomas (Brogna et al., 2008). Maximal
tumor resection is a valuable strategy for improving
seizure control in these patients but resection of epileptic foci needs to be included for optimal effect. Longterm video-electroencephalographic (EEG) monitoring is recommended for complicated cases where
seizure start and clinical seizure manifestations need

to be explored. Also, the use of intraoperative electrocorticography monitoring has improved seizure control
through identifying separate seizure foci surrounding
the tumor (Brogna et al., 2008). By intraoperative
electrostimulation mapping, the extent of the surgical
resection can be increased while preserving the eloquent areas. This method has increased the indication
for surgery of brain tumors that previously have been
considered as inoperable, such as tumors located
in the insular lobe and the rolandic area (Duffau and
Capelle, 2004). Valuable lessons can be learned from
the specialized medical staff taking care of patients
admitted for epilepsy surgery. In fact, slowly growing gliomas such as gangliogliomas are frequently
represented in this patient group and there is a considerable overlap between the two diagnostic patient
groups (Duffau, 2009; Lombardi et al., 1997).
In addition, radiotherapy, as single therapy or in
combination with surgery, has shown beneficial effects
in terms of seizure reduction. Studies have been performed mostly on patients with low-grade gliomas.
In the randomized EORTC (European Organization
for Research and Treatment of Cancer) 22845 trial,
designed to compare survival in adult low-grade
gliomas treated by early versus delayed radiotherapy, seizure outcome by radiotherapy was included
as secondary endpoint in patients still progressionfree 2 years from randomization (van den Bent et al.,
2005). Twenty-five percent of the irradiated patients
had seizures compared to 41% of those not irradiated. Since there was no difference in seizure control
at baseline between the two groups, it was concluded
that seizures were better controlled in the group receiving radiotherapy. Consistent findings have been found
in small retrospective reviews, showing a significant
seizure reduction in irradiated patients.
Small clinical studies are available on the
role of chemotherapy in glioma-related seizures.
Chemotherapy with temozolomide has shown to be
effective in improving seizure control in patients
with progressive low-grade gliomas (Brada et al.,
2003). Interestingly, a significantly better seizure
control by temozolomide was found in non-enhancing
tumors compared to enhancing tumors in a study on
progressive low-grade gliomas (Pace et al., 2003).
Other chemotherapeutic agents may also have positive
effects on seizure control, as shown in a study on
paralimbic low-grade gliomas treated by surgery,
405
chemotherapy or by a combination of these two

(Taillandier and Duffau, 2009).
Summary and Discussion

The clinical management of seizures is an important
and challenging issue for all patients with gliomarelated seizure, irrespective of tumor histology. The
specific aspects that need to be considered in optimizing seizure control for individual patients are related
to the delicate balance between the efficacy, pharmacological profiles and side effects (in particular on
cognitive function) of antiepileptic drugs, the impact
of refractory seizures on cognition and the social and
professional consequences of refractory seizures.
In choosing antiepileptic drugs, other ongoing or
planned medication needs to be taken into account
and possible interactions between these drugs and
the antiepileptic drug must be anticipated. When
antiepileptic drug treatment fails, other treatment
modalities for optimizing seizure control may be
realistic options, such as surgery, radiotherapy and
chemotherapy.
For achieving optimal effect of surgery on seizure
control, resection of the epileptic foci needs to be
included in addition to tumor resection. Modern tools
to identify separate epileptic foci such as preoperative video-EEG monitoring and intraoperative electrocorticography monitoring have largely improved the
chances of favorable postoperative seizure control. The
use of intraoperative electrostimulation mapping has
increased the indication for epilepsy surgery, including patients with tumors previously considered to be
inoperable.
For optimal management of patients with gliomarelated seizures, a close collaboration is recommended
between neuro-oncologists and a qualified team of
neurosurgeons, neurophysiologists and neurologists
involved in epilepsy surgery.
References
Aronica E, Gorter JA, van Vliet EA, Spliet WG, van Veelen
CW, van Rijen PC, Leenstra S, Ramkema MD, Scheffer GL,
Scheper RJ, Sisodiya SM, Troost D (2003) Overexpression
of the human major vault protein in gangliogliomas.
Epilepsia 44:11661174
406
Beaumont A, Whittle IR (2000) The pathogenesis of tumour
associated epilepsy. Acta Neurochir (Wien) 142:115
Berntsson SG, Malmer B, Bondy ML, Qu MQ, Smits A (2009)
Tumor-associated epilepsy and glioma; Are there common
genetic pathways? Acta Oncol 48:328342
Brada M, Viviers L, Abson C, Hines F, Britton J, Ashley S,
Sardell S, Traish D, Gonsalves A, Wilkins P, Westbury C
(2003) Phase II study of primary temozolomide chemotherapy in patients with WHO grade II gliomas. Ann Oncol
14:17151721
Brogna C, Gil Robles S, Duffau H (2008) Brain tumors and
epilepsy. Exper Rev Neurother 8:941955
Cavaliere R, Farace E, Schiff D (2006) Clinical implications of
status epilepticus in patients with neoplasms. Arch Neurol
63:17461749
Chaichana KL, Parker SL, Olivi A, Quiones-Hinojosa A (2009)
Long-term seizure outcomes in adult patients undergoing primary resection of malignant brain astrocytomas. J Neurosurg
111:282292
Chang EF, Potts MB, Keles GE, Lamborn KR, Chang SM,
Barbaro NM, Berger MS (2008) Seizure characteristics and
control following resection in 332 patients with low-grade
gliomas. J Neurosurg 108:227235
Cibula JE, Gilmore RL (1997) Secondary epileptogenesis in
humans. J Clin Neurophys 14:111127
Devinsky O (1999) Patients with refractory seizures. N Eng J
Med 340:15651570
Duffau H, Capelle L (2004) Preferential brain locations of lowgrade gliomas. Cancer 100:26222626
Glantz MJ, Cole BF, Forsyth PA, Recht LD, Wen PY,
Chamberlain MC, Grossman SA, Cairncross JG (2000)
Practice parameter: anticonvulsant prophylaxis in patients
with newly diagnosed brain tumors. Report of the Quality
Standards Subcommittee of the American Academy of
Neurology. Neurology 54:18861893
Hildebrand J, Lecaille C, Perennes J, Delattre JY (2005)
Epileptic seizures during follow-up of patients treated for
primary brain tumors. Neurology 65:212215
Kalachikov S, Evgrafov O, Ross B, Winawer M, BarkerCummings C, Martinelli Boneschi F, Choi C, Morozov P,
Das K, Teplitskaya E, Yu A, Cayanis E, Penchaszadeh G,
Kottmann AH, Pedley TA, Hauser WA, Ottman R, Gilliam
TC (2002) Mutations in LGI1 cause autosomal-dominant
partial epilepsy with auditory features. Nat Genet 30:
335341

King MA, Newton MR, Jackson GD, Fitt GJ, Mitchell LA,
Silvapulle MJ, Berkovic SF (1998) Epileptology of the firstseizure presentation: a clinical, electroencephalographic,
and magnetic resonance imaging study of 300 consecutive
patients. Lancet 352:10071011
Klein M, Engelberts NH, van der Ploeg HM, KasteleijnNolst Trenit DG, Aaronson NK, Taphoorn MJ, Baaijen H,
Vandertop WP, Muller M, Postma TJ, Heimans JJ (2003)
Epilepsy in low-grade gliomas: the impact on cognitive
function and quality of life. Ann Neurol 54:514520
Lombardi D, Marsh R, de Tribolet N (1997) Low grade glioma
in intractable epilepsy: lesionectomy versus epilepsy surgery.
Acta Neurochir 68(Suppl):7074
Lowenstein DH, Alldredge BK (1993) Status epilepticus at an
urban public hospital in the 19802. Neurology 3:483488
Manford M, Fish DR, Shorvon SD (1996) An analysis of clinical seizure patterns and their localizing value in frontal and
temporal lobe epilepsies. Brain 119:1740
Olafsson E, Ludvigsson P, Gudmundsson G, Hesdorffer D,
Kjartansson O, Hauser WA (2005) Incidence of unprovoked
seizures and epilepsy in Iceland and assessment of the
epilepsy syndrome classification: a prospective study. Lancet
Neurol 4:627634
Pace A, Vidiri A, Gali E, Carosi M, Telera S, Cianciulli
AM, Canalini P, Giannarelli D, Jandolo B, Carapella CM
(2003) Temozolomide chemotherapy for progressive lowgrade glioma: clinical benefits and radiological response.
Ann Oncol 14:17221726
Schaller B, Regg SJ (2003) Brain tumor and seizures: pathophysiology and its implications for treatment revisited.
Epilepsia 44:12231232
Taillandier L, Duffau H (2009) Epilepsy and insular Grade II
gliomas: an interdisciplinary point of view from a retrospective monocentric series of 46 cases. Neurosurg Focus
27:E8
Taphoorn MJ, Klein M (2004) Cognitive deficits in adult patients
with brain tumours. Lancet Neurol 3:159168
van Breemen MS, Wilms EB, Vecht CJ (2007) Epilepsy in
patients with brain tumours: epidemiology, mechanisms, and
management. Lancet Neurol 6:421430
van den Bent MJ, Afra D, de Witte O, Ben Hassel M, Schraub
S, Hoang-Xuan K, Malmstrm PO, Collette L, Pirart M,
Mirimanoff R, Karim AB, EORTC Radiotherapy and Brain
Tumor Groups and the UK Medical Research Council (2005)
Long-term efficacy of early versus delayed radiotherapy for
low-grade astrocytoma and oligodendroglioma in adults: the
EORTC 22845 randomised trial. Lancet 366:985990
Chapter 40
Brain Tumors Arising in the Setting of Chronic Epilepsy

Richard A. Prayson
Abstract Brain tumors, along with malformations of

cortical development (cortical dysplasia), hippocampal sclerosis, and remote ischemic damage, are among
the more common identifiable causes of pharmacoresistent chronic epilepsy. The majority of tumors arising in this setting present in the first two decades
of life and fortunately most tumors tend to be low
grade neoplasms, with at least half of the tumors representing glioneuronal neoplasms. The spectrum of
tumor types which have been identified in this clinical setting of chronic epilepsy continues to expand
with the recent addition of the angiocentric glioma to
The WHO Classification of Tumours of the Central
Nervous System. This chapter will review some of the
more commonly encountered tumors that arise in the
setting of medically intractable epilepsy.
Keywords Brain tumor Malformation Chronic
epilepsy Ganglioglioma DNET Astrocytic neoplasms
Introduction
Tumors are a well recognized cause of medically
intractable, chronic epilepsy. In cases where pathology is identifiable, neoplasms account for the primary pathology in anywhere from 10 to 30% of
patients going undergoing surgical treatment for
R.A. Prayson ()

Section of Neuropathology, Department of Anatomic
Pathology, Cleveland Clinic, Cleveland, OH 44195, USA
e-mail: PRAYSOR@ccf.org
chronic epilepsy (Britton et al., 1994). Collectively,

these tumors typically present in the first two decades
of life and are generally low grade neoplasms (WHO
grade I or II). In the larger series that have been
reported in the literature, gangliogliomas, dysembryoplastic neuroepithelial tumors (DNET) and low grade
astrocytomas (pilocytic or fibrillary) are the most commonly encountered neoplasms (Bourgeois et al., 1999;
Pasquier et al., 1996; Zentner et al., 1997). In a recently
published large series of pediatric patients with tumors
in chronic epilepsy, the majority of tumors presented
in the temporal lobe (59.7%) and were WHO grade I
tumors (56.6%) (Prayson, 2010). The most common
tumor types encountered in this series of 129 patients
in descending order included ganglioglioma (37.2%),
DNET (13.2%), low grade astrocytoma (11.6%), low
grade mixed glioma (6.2%), and low grade oligodendroglioma (3.9%). A coexistent malformation of cortical development or cortical dysplasia was observed
in 29.8% of evaluable cases, suggesting that many of
these glioneuronal tumors may have a maldevelopmental basis to their origin.
Some of the differences that have been reported
in terms of incidences of various tumors in surgical series may be explained by a number of factors.
Differences in terms of how one interprets definitions of these various entities clearly exist. This is
particularly problematic at times because a number
of these tumors can resemble one another. A microcytic oligodendroglioma, on a limited pathology specimen, can resemble areas of a DNET. If one is not
careful to closely follow the diagnostic criteria for
DNET, one can very easily interpret the oligodendroglioma as representing a DNET, which has prognostic and potential treatment implications. The typical
DNET is a multinodular, cortical based glioneuronal

407
408
neoplasm with associated cortical dysplasia in contrast

to the typical uninodular, white matter based oligodendroglioma, which is not known to be associated
with cortical dysplasia. Definitions of what constitutes a mixed glioma (oligoastrocytoma) also varies
and can result in different interpretations of the pathology. The exact percentage of a minor component in a
mixed glioma that needs to be present in order to make
that diagnosis has not been well established. Mixed
gliomas, which contain areas in which both the astrocytoma and oligodendroglioma component are mixed,
are also problematic. The extent of tissue sampling
at the time of dissection of the pathology specimen
can also impact the diagnosis. Gangliogliomas are
tumors that comprised of areas resembling low grade
glioma, typically astrocytoma, accompanied by areas
which contain atypical neuronal or ganglionic cells.
Unfortunately, both components of the tumor are not
evenly mixed throughout the neoplasm. A limited sampling may not demonstrate the atypical ganglion cell
component of the tumor and may result in an erroneous
diagnosis of a low grade astrocytoma. Extensive sampling of the neoplasm is therefore warranted to ensure
that diagnostic features of the tumor are sampled.
There are a variety of tumors which are less frequently encountered in this setting and are relatively
uncommon neoplasms in general. Unfamiliarity with
such entities as protoplasmic astrocytoma, angiocentric glioma, and meningioangiomatosis may present
additional challenges to the pathologist who may
not have had much experience with diagnosing these
lesions. Also, there are entities which are more complex and do not clearly fit into the current classification
systems for brain tumors. Composite tumors consisting of mixtures of DNET and ganglioglioma or ganglioglioma and pleomorphic xanthoastrocytoma have
been also described.
Ganglioglioma (WHO Grade I)

The most common tumor encountered in the setting of
chronic epilepsy is ganglioglioma (Miller et al., 1993;
Prayson et al., 1995; Wolf et al., 1994). The majority of these tumors arise in the first three decades of
life. The tumor has been described to arise throughout
the central nervous system, although the majority of
tumors are localized to the temporal lobe. On imaging,
R.A. Prayson
the tumor appears as a circumscribed solid or cystic

mass. Calcifications are frequently present. Contrast
enhancement is common but may be absent in some
neoplasms.
Histologically, the tumor is marked by a glioma
component that generally resembles a low grade fibrillary astrocytoma, pilocytic astrocytoma, or less commonly an oligodendroglioma (Fig. 40.1a). Intermixed
with this component is a population of atypical neuronal or ganglionic cells marked by a cytologic alterations including an abnormal distribution of Nissl
substance, binucleation, and occasional neurofibrillary
tangles (Fig. 40.1b). Vascular proliferative changes
may be focally evident but do not carry the same poor
prognosis that they do when seen in the setting of
a fibrillary astrocytoma or oligodendroglioma. Focal
leptomeningeal extension of the tumor may be evident
but does not appear to carry with it a poor prognosis
(Fig. 40.1c). Other common features include the presence of perivascular chronic inflammation consisting
of benign appearing lymphocytes, microcalcifications,
and eosinophilic granular bodies. Necrosis and prominent mitotic activity are generally not features of the
typical ganglioglioma, and if present in any appreciable amount, may be indicative of a higher grade lesion,
so called anaplastic ganglioglioma (WHO grade III).
Not surprisingly, markers of cell proliferation, such as
Ki-67, generally show low rates of cell proliferation;
Ki-67 labeling indices are 1%.
Immunohistochemistry may be useful in confirming that the ganglionic-like cells are neuronal in
the derivation. Tumors with large or giant astrocytic
cells (such as gemistocytic astrocytoma) may mimic
a ganglioglioma, if the large cells are interpreted as
being neuronal in derivation rather than astrocytic
(Fig. 40.1d). Care should be used when interpreting
immunohistochemical staining for neuronal markers.
Normal entrapped resident neurons will also stain with
neuronal antibodies and should not be misinterpreted
as being part of the tumor. Care should be taken in
using cell proliferation labeling indices to assist in differential diagnosis. Although a higher labeling index
in excess of 5% would be quite unusual in a ganglioglioma and may be indicative of a grade II or
III neoplasm, a low labeling index does not necessarily make a diagnosis of ganglioglioma; some grade II
gliomas will focally demonstrate similarly low labeling indices. As briefly mentioned in the introduction,
these tumors need to be extensively sampled, since
40 Brain Tumors Arising in the Setting of Chronic Epilepsy
409
Fig. 40.1 (a) The components of ganglioglioma are often not

evenly distributed throughout the neoplasm. This area shows the
glioma component of the tumor resembling a low grade diffuse
astrocytoma with moderate hypercellularity and atypia (original magnification 200); (b) This area of a ganglioglioma is
marked by atypical neuronal cells with cytoplasmic neurofibrillary tangles. Microcalcification is observed in the lower right
corner (original magnification 400); (c) Focal leptomeningeal
extension can be observed in some gangliogliomas and does
not appear to adversely influence outcome (original magnification 100); (d) The large cells seen in this tumor turned out
to be GFAP positive and synaptophysin negative and represent

a component of a gemistocytic astrocytoma. Immunostaining
can be useful in delineating the lineage of cellular components in ganglioglioma (original magnification 200); (e) Low
magnification appearance of a DNET highlighting the intracortical location and multinodular architecture (original magnification 50); (f) DNET marked by a proliferation of generally
rounded cells with scant cytoplasm arranged against a microcystic background. Occasional normal appearing neuronal cells
are observed floating in the cystic spaces (original magnification
200)
the neuronal or ganglionic component of the tumor

may be only focally evident. A failure to sample the
neuronal component may potentially result in the erroneous diagnosis of a grade II low grade glioma, which
has the potential of recurring and progressing to a
higher grade lesion in a shorter interval of time.
The prognosis of a ganglioglioma is excellent.
These tumors are presumably cured with gross total
resection. A significant relief of epilepsy has been documented in many of these tumors following resection.
Interestingly, the tumors themselves are frequently
electrically silent and the seizures appear to originate
from the surrounding tissue. Studies have documented

that the majority of these tumors are accompanied
by adjacent malformations of cortical development
or cortical dysplasia. The malformations of cortical
development represent a group of abnormalities of cortical architectural disorganization thought to be related
to some abnormality that occurs during the development of the brain. The patients, who have recurrent or persistent seizures following excision of their
ganglioglioma, are thought to have residual cortical
dysplasia.
410
Dysembryoplastic Neuroepithelial
Tumor (DNET) (WHO Grade I)
DNETs are another glioneuronal tumor that have a particularly good prognosis (Daumas-Duport et al., 1988;
Daumas-Duport, 1993; Prayson et al., 2002). Like gangliogliomas, these tumors are most commonly encountered in younger age patients, with the temporal lobe
being the most common, albeit not the only, location
for these neoplasms. On imaging, the tumor is predominately cortical-based and may appear hyperintense on
T2-weighted MRI images or hypointense or isointense
on T1-weighted images. The lesion often has a pseudopolycystic appearance and may show evidence of
calcification.
Histologically, the classic lesion is marked by
a cortical location and a multinodular architecture
(Fig. 40.1e). The predominate tumor cell resembles
an oligodendroglial-like cell with generally rounded
nucleus and scant cytoplasm (Fig. 40.1f). These
cells are often arranged against a microcystic background. Floating in the microcystic pools are normal
appearing neuronal cells. There is minimal cytologic
atypia in either component of this glioneuronal neoplasm. Prominent mitotic activity and necrosis are
unusual; occasional tumors may show vascular proliferative changes. Calcification may be evident. Like
gangliogliomas, careful inspection of the adjacent
parenchyma shows architectural disorganization consistent with a malformative or developmental origin for
the neoplasm (Fig. 40.2a). Proliferation markers show
low levels of staining, on par with ganglioglioma.
Like gangliogliomas, excision of the lesion results
in good seizure outcome. The tumor is generally
considered benign. Occasional case reports of more
aggressive or malignant behavior in these tumors
have been reported; these cases are often based on criteria which deviate from the WHO definition and raise
questions regarding the proper classification of these
unusual cases.
From a practical differential diagnostic standpoint, differentiation of this lesion from a low grade
oligodendroglioma or mixed glioma can be problematic, particularly in a less intact surgical specimen. A small sampling of an infiltrative low grade
oligodendroglioma with a microcystic background
and entrapped cortical neurons can closely resemble areas of a DNET and the two entities may be
R.A. Prayson
indistinguishable. Characteristic features or clues to

resolve the differential may not always be evident i.e.
things such as predominate cortical location, multinodularity, and adjacent cortical dysplasia, all of which
favor a DNET versus subpial aggregation of cells,
predominate white matter location, uninodularity, and
the presence of minigemistocytes which favor oligodendroglioma. In cases where these distinguishing
features are not evident, one can employ molecular studies looking for evidence of large deletions on
chromosomes 1p and 19q, features with have been
associated with oligodendroglioma. In the presence
of large deletions, a diagnosis of oligodendroglioma
or mixed glioma would be favored. In cases where
the deletions are not evident, the diagnostic dilemma
remains; deletions have not been reported in DNETs
and are not present in approximately 2040% of oligodendrogliomas.
Astrocytic Neoplasms
A variety of astrocytic neoplasms have been associated with chronic epilepsy. Low grade diffuse or
fibrillary astrocytomas (WHO grade II) may occasionally present with chronic seizures and may arise
anywhere throughout neuroaxis (Louis et al., 2007).
These neoplasms are characterized by mild hypercellularity and a proliferation of atypical appearing
fibrillary-type astrocytes marked by nuclear enlargement, irregularities of the nuclear contour, and nuclear
hyperchromasia (Fig. 40.2b). A rare mitotic figure
may also be observed. Focally, calcifications or microcystic changes may be evident. Vascular proliferative
changes and necrosis are absent. In contrast to the
gangliogliomas and DNETs, these tumors tend to be
more infiltrative in nature and are not as amenable
to surgical resection. With time, these tumors will
progress to a higher grade lesion and may require
adjuvant radiation therapy or chemotherapy. Cell proliferation indices are generally low and may overlap
with indices observed in grade I glioneuronal tumors.
Although astrocytomas may present in the first two
decades of life, they are typically more commonly
encountered in a slightly older population of patients,
with a peak incidence in young adults between the
ages of 3040 years. Imaging studies show an illdefined, homogeneous lesion of low signal intensity.
411
Fig. 40.2 (a) The area adjacent to a DNET showing an abnormal cortical architecture marked by an absence of cortical layer
2 consistent with a malformation of cortical development or
cortical dysplasia (original magnification 200); (b) This low
grade diffuse or fibrillary astrocytoma is characterized by mild
hypercellularity due to a proliferation of atypical cells with
enlarged nuclei and irregular nuclear contours (original magnification 400); (c) The looser area of a pilocytic astrocytoma
characterized by multiple eosinophilic granular bodies (original magnification 400); (d) A microcystic area of a pilocytic
astrocytoma showing focal vascular proliferative changes; this

finding does not carry with it the same poor prognosis as it does
when encountered in the setting of a diffuse astrocytoma (original magnification 200); (e) This protoplasmic astrocytoma
is comprised of generally rounded cells with scant cytoplasm
arranged against a microcystic background (original magnification 200); (f) Pleomorphic xanthoastrocytomas are typically
marked hypercellular and are comprised of very pleomorphic
cells. Lipidized astrocytes, as seen here, are present in only a
subset of cases (original magnification 200)
Occasional tumors may show appreciable numbers of

large gemistocytes. These gemistocytes are marked
by abundant eosinophilic cytoplasm and laterally displaced nuclei, which are slightly enlarged and may
have a slightly irregular nuclear contour. Care should
be taken in not confusing gemistocytes with a neuronal or ganglion cells proliferation, resulting in an
erroneous diagnosis or ganglioglioma. If one is uncertain about the linage of these cells, they are readily
highlighted with a GFAP (glial fibrillary acidic protein)
immunostain in contrast to neuronal or ganglionic cells
which will stain with markers of neural differentiation
such as synaptophysin.
Pilocytic astrocytomas (WHO grade I) represent a

special variant of astrocytoma that have a particularly
good prognosis (Louis et al., 2007). They comprise
approximately 56% of all gliomas and arise most
commonly in children. Most tumors originate in the
cerebellum, brain stem, thalamic/basal ganglia, and
optic chiasm regions. Occasional tumors arise in the
region of the third ventricle and may present with
epilepsy. The classic imaging appearance of the lesion
is that of a cystic tumor with enhancing mural nodule
or nodules.
Histologically, the classic pilocystic astrocytoma
has a biphasic appearance. Areas of the tumor are
412
marked by cells that are elongated and arranged against

a fibrillary background. Eosinophilic Rosenthal fibers
are usually identifiable in these regions if present
(about 75% of cases). These fibrillary zones alternate with areas in which the cells are more rounded
and arranged against a microcystic background. The
eosinophilic granular bodies, if present, are usually more readily observable in the looser regions
(Fig. 40.2c). Pleomorphic nuclei are not uncommon in
these tumors. Vascular proliferative changes and vascular sclerosis may be also evident (Fig. 40.2d); the
vascular proliferation does not carry the same ominous prognosis it does when observed with a fibrillary
astrocytoma. Infiltration of the leptomeninges may on
occasion be noted. Areas of the tumor may consist of
more rounded cells and resemble a low grade oligodendroglioma, creating the overall impression of a mixed
glioma. Cell proliferation indices are generally low,
typically on the order of approximately 1%; however,
occasional cases may show labeling indices as high as
34%.
The challenge from a diagnostic standpoint is usually in cases where the sampling is limited and the
characteristic features of the tumor are not evident.
This can result in an erroneous diagnosis of a fibrillary astrocytoma or low grade mixed glioma. The
glioma component of a ganglioglioma can also resemble a pilocytic astrocytoma and a failure to identify
the atypical ganglion cell component may result in
an incorrect diagnosis of pilocytic astrocytoma. The
pilocytic tumors are generally amenable to surgical
resection and have an excellent prognosis.
The rare protoplasmic astrocytoma (WHO grade II)
has also been associated with chronic epilepsy
(Prayson and Estes 1995, 1996). Protoplasmic astrocytomas show a preference for temporal and frontal
lobes and seem to be predominately cortical based.
Microscopically, they are marked by a proliferation of
astrocytic cells with generally round or slightly oval
nuclear contours and a scant amount of eosinophilic
cytoplasm (Fig. 40.2e). The tumor cells are generally arranged against a microcystic background. The
histologic appearance warrants a differential diagnosis with microcystic oligodendroglioma and DNET. In
contrast to the DNET, the protoplasmic astrocytoma
does not show evidence of adjacent cortical dysplasia and is not a multinodular neoplasm. Chromosomal
studies have shown no evidence of the signature 1p
deletions that are associated with oligodendroglioma
R.A. Prayson
in the protoplasmic astrocytoma. The number of cases

that have been reported to date are rather small and the
literature suggests that a subset of tumors may progress
or degenerate into a higher grade lesion over time.
The pleomorphic xanthoastrocytoma (WHO grade
II) is a rare variant of astrocytoma which typically
presents in children and young adults, often with a
long history of seizures (Giannini et al., 1999; Kepes
et al., 1979). The vast majority of tumors arise in the
temporal or parietal lobes. Like DNETs, they are often
cortical based. They may have a cystic component with
a mural nodule.
Histologically, they are markedly hypercellular and
as its name suggests marked by prominent nuclear
pleomorphism (Fig. 40.2f). Prior to its recognition
as a distinct entity, many of them were mislabeled as representing glioblastoma. In contrast to
glioblastoma, these tumors show little in the way of
mitotic activity and lack necrosis. Vacular proliferative changes may be evident. The xanthoastrocytes
(lipidized astrocytes) are a variable feature of this
tumor but are helpful when present. Perivascular lymphocytes, eosinophilic granular bodies, focal extension
into the leptomeninges, and microcalcifications may
all be present. Reticulin staining highlights the presence of increased staining surrounding individual cells
or small groups of cells; this is in contrast to the typical
glioblastoma lesion where reticulin staining is usually
observed only in association with vascular proliferative
changes.
Most of the tumors cells stain positively with antibodies to GFAP and S-100 protein. Interestingly, a
subset of cells in the tumor has been reported to variably stain with markers of neuronal differentiation,
such as synaptophysin, neurofilament and class III
beta-tubulin. This finding raises interesting questions
regarding the lineage of this tumor and suggests the
possibility of a glioneuronal neoplasm. A report of
adjacent malformation of cortical development (cortical dysplasia) has been published. Cell proliferation
labeling indices may be variable and range from low
to intermediate levels. Rare cases of anaplastic pleomorphic xanthoastrocytoma (WHO grade III) have
been described and are typically marked by increased
mitotic activity (5 or more mitoses per 10 high power
fields), necrosis and high cell proliferation labeling
indices. Progression of low grade tumors to this higher
grade lesion has also been described. The prognosis
is generally favorable with a 70% 10 year survival
(much better than its histologic appearance would

suggest).
Low Grade Oligodendroglioma

and Mixed Glioma (Oligoastrocytoma)
(WHO Grade II)
Low grade oligodendrogliomas are a relatively infrequent case of chronic epilepsy in young patients (Louis
et al., 2007). Overall, oligodendrogliomas are much
less common than their fibrillary astrocytoma counterparts. The peak incidence is in the 5th decade of life.
Tumors arise in the white matter with the frontal and
temporal lobes being the most common sites of origin. On imaging studies, calcifications and evidence of
hemorrhage are fairly common.
Microscopically, the tumor is marked by a fairly
monomorphic proliferation of cells with generally rounded nuclear contours and scant cytoplasm
(Fig. 40.3a). Pericellular haloes (fried egg appearance) represent an artifact of delayed formalin fixation; therefore, the haloes will not be evident on
a frozen section slides or smear cytologic preparations. An arcuate or chicken wire vascular pattern is
evident due to accentuation of small vessels resulting from a paucity of cell processes. Microcystic
changes may be evident and may result in an appearance resembling the DNET. Some tumors contain
cells which resemble small gemistocytes (minigemistocytes) (Fig. 40.3b). Cells comprising the tumor have
a propensity to satellite around preexisting structures,
such as neurons or blood vessels. Subpial aggregation of infiltrating tumor cells may also be evident. Similar to fibrillary astrocytoma, higher grade
lesions (anaplastic oligodendrogliomas- WHO grade
III) have been described and are generally characterized by increased cellularity and nuclear pleomorphism, increased mitotic activity (typically 5 or more
mitoses per 10 high power fields), and necrosis. These
higher grade tumors usually are too rapidly growing
to result in a presentation of chronic epilepsy. Markers
of cell proliferation show higher indices in the higher
grade tumors.
Oligodendrogliomas are infiltrative tumors and not
always amenable to complete resection. A majority of oligodendrogliomas are chemoresponsive and
grade for grade, they have a better prognosis than
413
fibrillary astrocytomas. With time, many of the low

grade tumors will progress to higher grade neoplasms.
Of particular note is the association of chemoresponsiveness in oligodendrogliomas with large deletions on chromosomes 1p and 19q. The large deletions afford tumors (especially the higher grade
lesions) a better prognosis because of the association with chemoresponsiveness. These molecular findings can also be useful in terms of sorting
out differential diagnoses; oligodendroglioma lookalike tumors like DNET and protoplasmic astrocytomas do not demonstrate the large chromosomal
deletions.
The mixed gliomas (oligoastrocytomas) represent a
group of tumors which are marked by geographically
distinct areas resembling a fibrillary astrocytoma and
an oligodendroglioma. The diagnosis is not terribly
reproducible, since the criteria are a bit nebulous and
identification of patterns is not always agreed upon.
Guidelines suggest that somewhere between 20 and
35% of a minor component (depending on who you
read) should be recognized in order to make the diagnosis. It is not unusual in many of these tumors for
the two phenotypes to be intermixed in a more diffuse pattern. The diagnosis ends up being reflective
of the biases of the pathologist; some pathologists
may use the diagnosis fairly regularly; whereas, others may use it sparingly. Consequently, interpretation
of the literature on the subject is difficult. Many studies simply lump mixed gliomas together with pure
oligodendrogliomas for purpose of analysis, further
confusing the matter. Like fibrillary astrocytomas and
low grade oligodendrogliomas, mixed gliomas generally arise in the white matter and are infiltrative neoplasms. The same parameters that are used for grading
astrocytomas and oligodendrogliomas are also used to
evaluate and grade these tumors as well. Anaplastic
mixed gliomas are generally more cellular, more pleomorphic, marked by increased mitotic activity and may
demonstrate vascular proliferative changes. The presence of necrosis now warrants classification of the
tumor as a glioblastoma with an oligodendroglioma
component. Chromosome 1p and 19q analysis shows
a deletion prevalence intermediate between pure astrocytoma and oligodendroglioma. Prognosis is thought
to be intermediate between the low grade astrocytoma
and low grade oligodendroglioma, although 1p/19q
deleted mixed gliomas tend to behave more like oligodendrogliomas.
414
R.A. Prayson
Fig. 40.3 (a) The classic appearance of a low grade oligodendroglioma is marked by a monomorphic proliferation of rounded
cells; the pericellular clearing is an artifact of delayed formalin
fixation (original magnification 400); (b) Occasional oligodendrogliomas may contain a subpopulation of cells with increased
eosinophilic cytoplasm and eccentrically placed nucleus, socalled minigemistocytes (original magnification 400); (c)
Meningioangiomatosis is characterized by a proliferation of
blood vessels surrounded by collars of spindled meningothelial
cells (original magnification 200); (d) Sometimes, prominent

sclerosis of blood vessels in meningioangiomatosis can result
in an appearance suggestive of a vascular malformation original magnification 400); (e) Perivascular pseudorosettes are a
characteristic feature of the angiocentric glioma (original magnification 200); (f) The interface between the angiocentric
glioma and adjacent brain shows an infiltrative border (original
magnification 400)
Meningioangiomatosis (No WHO Grade)
of a vascular malformation. Microscopically, the lesion

is marked by a proliferation of blood vessels surrounded by collars of spindled meningothelial cells
(Fig. 40.3c). Calcifications are frequently seen in
the intervening parenchyma. Neurons in the cortex may show dysmorphic changes including neurofibrillary tangles and granulovacuolar degeneration.
In some cases, prominent sclerotic vessel changes
are observed and recognition of the meningothelial collars may be difficult (Fig. 40.3d). Excision
of the lesion usually results in dramatic seizure
improvement.
Meningioangiomatosis is a rare developmental or

hamartomatous abnormality which may present like
a neoplasm (Halper et al., 1986; Perry et al., 2005;
Prayson, 1995). There is a known association with
neurofibromatosis type 2. The lesion appears as a
somewhat discrete, plaque-like cortical mass (although
microscopically, the lesion extends to involve the white
matter). Overlying meningioma may be present in
some cases. Large blood vessels may be situated on
the surface of the lesion, giving the gross impression
Angiocentric Glioma (WHO Grade I)

Angiocentric glioma is a fairly recently recognized
entity which presents primarily in children or young
adults (Lellouch-Tubiana et al., 2005; Preusser et al.,
2007; Wang et al., 2005). The tumor appears fairly
well circumscribed and solid on imaging studies,
although it is typically infiltrative microscopically.
Histologically, the neoplasm is marked by a proliferation of spindled cells focally arranged around blood
vessels (predominantly in the cortex) forming perivascular pseudorosette structures (Fig. 40.3e). Subpial
aggregation of tumor cells is also evident in some
cases. The tumor has an infiltrative pattern at the
perimeter (Fig. 40.3f). Prominent mitotic activity, vascular proliferative changes and necrosis are not seen.
The prominence of the pseudorosette pattern raises a
differential diagnosis with ependymoma and astroblastoma. Interestingly, a subset of these tumors demonstrate dot-like EMA (epithelial membrane antigen)
positivity on immunostaining as well as microvilli,
cilia and desmosomes, all features typically associated
with ependymal differentiation. The current hypothesis
regarding origin is that the tumor arises from bipolar radial glia that have the capability to differentiate
along ependymal lines. Cell proliferation indices are
generally very low. The tumor has a generally excellent
prognosis and is thought to be cured with total surgical
resection.
Composite Tumors
There are still occasional tumors that one encounters in
the setting of chronic epilepsy that do not neatly fit into
the current WHO classification. Rare examples of socalled composite tumors have been described, tumors
with combined features of pleomorphic xanthoastrocytoma/ganglioglioma or tumors with features of
ganglioglioma/DNET (Hirose and Scheithauer, 1998;
Napekoski and Prayson, 2010; Perry et al., 1997;
Vajtai et al., 1997). One recent small series of the
later group found them to be more common in children or young adults and most commonly arising in
the temporal lobe. The tumors were all multinodular
with some nodules showing typical features of either
ganglioglioma or DNET, with some nodules showing
415
intermixed features of both tumors intermixed. Cell

proliferation indices were low in the tumors evaluated
and adjacent malformations of cortical development
were observed in a subset of cases.
References
Bourgeosis M, Sainte-Rose C, Lellouch-Tubiana A, Malucci C,
Brunelle F, Maixner W, Cinalli G, Pierre-Kahn A, Renier D,
Zerah M, Hirsch J-F, Goutieres F, Aicardi J (1999) Surgery
of epilepsy associated with focal lesions in childhood. J
Neurosurg 90:833842
Britton JW, Cascino GD, Sharbrough FW, Kelly PH (1994)
Low-grade glial neoplasms and intractable partial epilepsy:
efficacy of surgical treatment. Epilepsia 35:11301135
Daumas-Duport C (1993) Dysembryoplastic neuroepithelial
tumors. Brain Pathol 3:283295
Daumas-Duport C, Scheithauer BW, Chodkiewicz JP, Laws ER,
Vedrenne C (1988) Dysembryoplastic neuroepithelial tumor:
a surgically curable tumour of young patients with intractable
partial seizures. Neurosurgery 23:545556
Giannini C, Scheithauer BW, Burger PC, Brat DJ, Wollan PC,
Lach B, ONeill BP (1999) Pleomorphic xanthoastrocytoma.
What do we really know about it? Cancer 85:20332045
Halper J, Scheithauer BW, Okasaki H, Laws ER (1986)
Meningoangiomatosis: a report of six cases with special
reference to the occurrence of neurofibrillary tangles. J
Hirose T, Scheithauer BW (1998) Mixed dysembryoplastic neuroepithelial tumor and ganglioglioma. Acta Neuropathol
95:649654
Kepes JJ, Rubinstein LJ, Eng LF (1979) Pleomorphic xanthoastrocytoma: a distinctive meningocerebral glioma of young
subjects with relatively favorable prognosis. A study of 12
cases. Cancer 44:18391852
Lellouch-Tubiana A, Boddaert N, Bourgeois M, Fohlen M,
Jouvet A, Delalande O, Seidenwurm D, Brunelle F, SainteRose C (2005) Angiocentric neuroepithelial tumor (ANET):
a new epilepsy-related clinicopathological entity with distinctive MRI. Brain Pathol 15:281286
IARC Press, Lyon, France
Miller DC, Lang FF, Epstein FJ (1993) Central nervous system
gangliogliomas. Part 1. pathology. J Neurosurg 75:859866
Napekoski KM, Prayson RA (2010) Composite ganglioglioma/dysembryoplastic neuroepithelial tumor: a clinicopathologic study of 6 cases. Mod Pathol 23(Suppl 1):
378379A
Pasquier B, Bost F, Peoch M, Barnoud R, Pasquier D (1996)
Neuropathologic data in drug-resistant partial epilepsy.
Report of a series of 195 cases. Ann Pathol (Paris) 16:
174181
Perry A, Giannini C, Scheithauer BW, Rojiani AM, Yachnes AT,
Seo IS, Johnson PC, Kho J, Shapiro S (1997) Composite
pleomorphic xanthoastrocytoma and ganglioglioma report of
four cases and review of the literature. Am J Surg Pathol
21:763771
416
Perry A, Kurtkaya-Yapicier O, Scheithauer BW, Robinson
S, Prayson RA, Kleinschmidt-DeMasters BK, StemmerRachaminov AO, Gutmann DH (2005) Insights into meningioangiomatosis with and without meningioma: a clinicopathologic and genetic series of 24 cases with review of the
literature. Brain Pathol 15:5565
Prayson RA (1995) Meningioangiomatosis: a clinicopathologic
study including MIB-1 immunoreactivity. Arch Pathol Lab
Med 119:10611064
Prayson RA, Castilla EA, Hartke M, Pettay J, Tubbs RR,
Barnett GH (2002) Chromosome 1p allelic loss by fluorescence in situ hybridization (FISH) not observed in dysembryoplastic neuroepithelial tumors. Am J Clin Pathol 118:
512517
Prayson RA (2010) Tumors arising in the setting of pediatric
chronic epilepsy. Pathology 42:426431
Prayson RA, Estes ML (1995) Protoplasmic astrocytoma. A
clinicopathologic study of 16 tumors. Am J Clin Pathol
10:705709
Prayson RA, Estes ML (1996) MIB-1 and p53 immunoreactivity in protoplasmic astrocytomas. Pathol Int 46:
862866
R.A. Prayson
Prayson RA, Khajavi K, Comair YG (1995) Cortical architectural abnormalities and MIB-1 Immunoreactivity in gangliogliomas: a study of 60 patients with intracranial tumors. J
Neuropathol Exp Neurol 1995 54:513520
Preusser M, Hoischen AM, Novak K (2007) Angiocentric
glioma: report of clinico-pathologic and genetic findings in
8 cases. Am J Surg Pathol 31:17091718
Vajtai I, Varga Z, Aguzzi A (1997) Pleomorphic xanthoastrocytoma with gangliogliomatous component. Pathol Res Pract
193:617621
Wang M, Tihan T, Rojiani AM (2005) Monomorphous angiocentric glioma: a distinctive epileptogenic neoplasm with
features of infiltrating astrocytoma and ependymoma. J
Wolf HK, Mller MB, Spnle M, Zentner J, Schramm J,
Wiestler OD (1994) Ganglioglioma: a detailed histopathological and immunohistochemical analysis of 61 cases. Acta
Zentner J, Hufnagel A, Wolf HK, Ostertan B, Behrens E,
Campos MG, Elger CE, Wiestler OD, Schramm J (1997)
Surgical treatment of neoplasms associated with medically
intractable epilepsy. Neurosurgery 41:378387
Chapter 41
Low-Grade Gliomas: Role of Relative Cerebral Blood Volume

in Malignant Transformation
Neil Upadhyay and Adam D. Waldman
Abstract Accurate characterisation of gliomas is

important for appropriate risk stratification, which will
inform decisions on potentially invasive investigation
and treatment that have an associated risk of morbidity
and mortality. In many institutions aggressive therapy
is only commenced when low grade gliomas (LGGs)
show signs of malignant transformation. Limitations
with conventional contrast enhanced CT and MRI
for glioma evaluation are well-recognised. Relative
Cerebral Blood Volume (rCBV) measurements offer a
non-invasive method of characterising and monitoring
LGGs which has been validated by histopathological
correlations and clinical endpoints. The widespread
use of rCBV in clinical practice will require consensus
on evidence-based rCBV values which discriminate
aggressive gliomas from more indolent tumors.
The future role of rCBV is likely to be within the
framework of a multimodality approach to glioma
evaluation.
Keywords Gliomas LGG rCBV WHO
Astrocytomas Oligodendrogliomas
Introduction
The management of low grade diffuse gliomas (WHO
grade II, LGGs) is one of the most controversial areas
N. Upadhyay ()
Institute of Clinical Science, Imperial College London, Charing
Cross Hospital, W68RF, London, UK
e-mail: neil.upadhyay@imperial.ac.uk
A.D. Waldman ()
Institute of Clinical Science, Imperial College London, Charing
Cross Hospital, W68RF, London, UK
e-mail: adam.waldman@csc.mrc.ac.uk
of neuro-oncology. These tumors frequently present

with readily-controlled seizures in patients who are
often young, and have no fixed neurological deficit.
They behave indolently for an unpredictable period,
which may be many years, before undergoing malignant transformation into high grade lesions which are
rapidly fatal. Prospective evidence for survival advantage from early intervention is weak, and although
total or subtotal resection may confer benefit, this is
frequently not possible due to eloquent anatomical
location and the infiltrative nature of the tumor. The
potential benefit of early aggressive treatment must
therefore be balanced against the morbidity associated
with investigation and treatment in patients who are
frequently clinically well, and may not develop significant problems for several years. In many institutions,
patients with LGG are therefore monitored on a watch
and wait policy until there are signs of malignant
transformation, at which point radiotherapy and/or
surgery is commenced. Conventional techniques for
assessing gliomas include contrast enhanced MRI
and histopathological assessment based on stereotactic biopsy or bulk resection, although limitations with
these methods are well recognised in the literature
(Knopp et al., 1999; Coons et al., 1997). The development of more reliable methods for characterisation and
risk stratification of LGG is important, to aid clinical
decision making.
Perfusion Magnetic Resonance (MR) imaging techniques allow calculation of cerebral blood volume
(CBV), which is a measure of localised cerebral blood
volume within tissue. Changes in the CBV of gliomas
can be explained by tumor pathophysiology; neoangiogenesis is associated with an increase in vessel
density, tortuosity and permeability, and are a hallmark
of aggressive elements in histopathological analysis of

417
418
gliomas. CBV therefore provides a non-invasive quantitative imaging biomarker that reflects angiogenesis in
gliomas, and hence can aid characterisation of gliomas.
In particular, it may be helpful in grading gliomas
and identifying malignant transformation at an earlier
stage.
Acquisition of rCBV
Dynamic susceptibility contrast enhanced MRI (DSCMRI) is the most common perfusion method used in
clinical practice, and the technique on which most
published relative cerebral blood volume (rCBV) data
in gliomas has been based. The method involves
acquiring time-resolved images during the transit of
an exogenous gadolinium chelate (Gd) contrast agent
through the brain. Local magnetic field inhomogeneities caused by the Gd leads to a transient reduction in the transverse relaxation time of water protons
in adjacent tissues, and a consequent reduction in tissue signal intensity (T2 effect). The concentration of
agent is proportional to the change in relaxation rate,
and analysis of the kinetics of signal intensity change
allows perfusion parameters to be derived. In order to
sample the signal profile from the first pass of contrast agent, the compound is injected intravenously as a
rapid bolus, and rapid echo planar imaging (EPI) based
spin echo or gradient echo sequences allow the whole
brain to be imaged with sufficient time-resolution.
CBV is proportional to the area under the deconvolved time intensity curve, assuming there has been
no recirculation or contrast leakage. Because the
blood-brain barrier of contrast-enhancing tumors is
leaky, contrast leakage needs to be accounted for
when CBV values are calculated. A significant correlation between glioma grade and CBV values corrected
for contrast extravasation has been previously demonstrated, but no correlation for uncorrected values was
found (Boxerman et al., 2006). CBV is the proportion of the volume of interest which is comprised
of vessels. It is measured as millilitres of blood per
grams of tissue. Because of difficulties in reproducible
absolute quantification of CBV, relative CBV is usually expressed as a ratio of CBV within the tissue
of interest against normal contralateral white matter;
this facilitates stable longitudinal perfusion measurement and inter-subject comparisons. Comparison of
quantitative data across institutions and different MRI
manufacturer platforms, however, remain a challenge.
N. Upadhyay and A.D. Waldman
Perfusion parameters can also be derived from the

first pass phase of dynamic contrast enhanced MRI,
which is based on T1-dependent signal abnormality, and also yields permeability data. These methods have been used to study gliomas, but are less
widely validated. Arterial Spin Labelling is an alternative perfusion technique which uses magneticallylabelled water as an endogenous tracer. Water is
labelled by inversion or saturation pulses and imaging is performed downstream from the site of labelled
water to allow flow to the region of interest. The
technique offers the advantage of being completely
non-invasive and allows absolute quantification of
blood flow, but its clinical application has been limited by long acquisition times and requirements for
complex off-line processing. Technical advances with
the current generation of commercial 3T MRI systems and more user-friendly processing software are
likely to allow wider clinical application of ASL in
future.
Comparison of rCBV and Conventional

MRI for Glioma Evaluation
Several studies have demonstrated that rCBV can
give useful information about tumor grade and
patient prognosis (Table 41.1). Accurate characterisation of gliomas is important to allow appropriate
risk stratification, which will inform difficult decisions on potentially invasive investigation and treatment. Conventional assessment of gliomas includes
evaluation with contrast enhanced MRI sequences
and histopathological evaluation following biopsy or
tumor resection. Limitations of conventional methods
of assessment are well recognised in the literature
(Knopp et al., 1999; Coons et al., 1997). Relative CBV
gives supplementary information on tumor physiology which complements that from conventional contrast enhanced MRI and can potentially detect signs
of transformation earlier than conventional methods
(Figs. 41.1, 41.2, and 41.3). Relative CBV is related
to perfusion, and is felt to reflect angiogenesis, which
is an important marker of tumor aggressiveness in the
pathological grading of gliomas. Relative CBV can
therefore differentiate biologically aggressive tumors
from more indolent tumors and guide the most appropriate management. Contrast enhancement is often
used as an indicator of progression on conventional
41 Low-Grade Gliomas: Role of Relative Cerebral Blood Volume in Malignant Transformation
419
Table 41.1 rCBV values suggested by different authors for characterising gliomas
Study
Tumor type
Low grade
High grade
Aronen et al. (1994)

Arvinda et al. (2009)
Bisdas et al. (2009)
Danchaivijitr et al.
(2008)
Gliomas
Gliomas
Astrocytomasb
Gliomas
Hakyemez et al. (2005)

Law et al. (2006)
Law et al. (2008)
Gliomas
Gliomas
Gliomas
1.11
3.64
1.21
3.29
/
/
1.31c
1.93c
No statistically significant
difference
1.69
3.32
2.14
5.18
/
/
Lev et al. (2004)

Jackson et al. (2002)
Gliomase
Astrocytomas
Astrocytomas
Jenkinson et al. (2006)
Oligodendrogliomas
Optimal
discrimination
Reference
/
2.91a
3.8
/
Grade
Grade
1 yr survival
Time to
progressiond
2.00
1.75a
1.75
Grade
Grade
Time to
progressiond
Grade and
survival
Grade
1.5
2.9
1.5
1.3
3.1
Specific values not published. rCBV increased
significantly with grade
rCBV did not increase with
1.59f identifies
grade
deletion of 1p/19q
1.44
5.07
/
1.30
3.87
/
2.00
4.91
2.93
1.74
6.10
/
2.01
2.47
1.6
Grade 1p/19q
deletion
Grade
Grade
Grade
Grade
Grade
Knopp et al. (1999)

Astrocytomas
Rollin et al. (2006)
Gliomasg
Shin et al. (2002)
Gliomas
Yang et al. (2002)
Gliomas
Zonari et al. (2007)
Gliomas
P < 0.05 unless otherwise stated
a Minimising C2 error, which is the fraction of misclassified tumors
b Subgroup analysis on astrocytomas; when oligodendrogliomas and oligoastrocytomas were included, results were not statistically
significant
c Longitudinal study following LGG; rCBV of patients with no transformation was 1.31 at baseline whilst rCBV of patients with
transformation was 1.93 at baseline. There was no statistically significant difference at baseline. The group mean rCBV was 5.36 at
transformation
d Progression was defined by clinical and conventional MRI parameters
e Study on gliomas; subgroup analysis performed on astrocytomas and oligodendrogliomas
f Study on oligodendrogliomas; threshold value discriminates between oligodendrogliomas which have 1p/19q deletion and those
which do not have 1p/19q deletion
g Study on intra-axial tumors; subgroup analysis performed on gliomas
Fig. 41.1 Grade 2

astrocytoma. Left:
T1-weighted
contrast-enhanced (T1CE)
study demonstrates minimal
contrast enhancement. Right:
The rCBV map demonstrates
low rCBV values
420
Fig. 41.2 Glioblastoma

multiforme. Left: T1CE study
demonstrates marked contrast
enhancement around the
tumour. Right: The rCBV map
demonstrates elevated rCBV
values
Fig. 41.3 Glioblastoma

multiforme. The rCBV map
(right) shows elevated rCBV
values before any significant
contrast enhancement is
demonstrated on the T1CE
study (left)
MRI studies, but this can lead to an inaccurate assessment of the tumor (Knopp et al., 1999). Enhancement
represents loss of the local blood-brain barrier but does
not necessarily give a good indication of vascularity.
Lack of enhancement does not always equate to a low
grade tumor (Fig. 41.3) and variability in enhancement
patterns between subgroups of low grade and high
grade gliomas is well recognised (White et al., 2005;
Ginsberg et al., 1998). Other characteristics on conventional MRI such as mass effect, oedema, necrosis
and haemorrhage also lack accuracy in differentiating gliomas by grade (Aronen et al., 1994; Law et al.,
2003; Rollin et al., 2006).
Differentiating Gliomas Using rCBV

Many groups have attempted to determine rCBV measurements that will discriminate LGGs from high
grade gliomas (HGGs). Table 41.1 summarises the
results from 15 studies which analysed tumors using
rCBV. The range of mean maximal rCBV values for

LGGs is 1.112.14. The range of mean maximal rCBV
values for HGGs is 2.476.10. A wide range of threshold values were suggested to discriminate between low
grade and high grade (1.53.8).
Challenges of Glioma Evaluation

with rCBV
For use in routine clinical practice, consensus on a precise rCBV value is required to distinguish between low
grade and high grade gliomas. However, Table 41.1
demonstrates a range of values have been suggested in
the literature, which can be attributed to a number of
factors.
Technical differences in acquisition and analytical
methods make comparison and synthesis of data from
these studies difficult. A study found CBV values were
higher using Gradient Echo Echo Planar imaging
(GE EPI) compared with Spin Echo Echo Planar
imaging (SE EPI) (Sugahara et al., 2001). GE EPI has

been preferred to SE EPI by some groups because it
is more sensitive to the contribution of vessels of a
range of sizes whilst SE derived imaging has been
reported as being more sensitive for capillary sized
vessels (Boxerman et al., 2006; Sugahara et al., 2001).
High grade tumors contain larger calibre vessels and
studies have suggested GE rCBV values show significant correlation with tumor grade whilst SE rCBV
values do not (Boxerman et al., 2006). The technique
for selecting regions of interest (ROIs) may also affect
values (Lev et al., 2004). A relatively small hot spot
surrounded by lower rCBVs will be perceived as a
lower value if a larger ROI is used than a smaller
ROI. Inclusion or exclusion of visible intratumoural
vessels also affects the magnitude of calculated rCBV
and hence thresholds for discriminating between tumor
groups (Caseiras et al., 2008).
Variability in the subgroups of gliomas that
were studied may also contribute to the differences.
A difference in tumor architecture between astrocytomas and oligodendrogliomas is likely to cause an
intrinsic difference in the rCBV of tumors of comparable grade. Maximum rCBV of low grade oligodendrogliomas have demonstrated higher values than low
grade astrocytomas, which is thought to be because
the former have a higher density of capillaries than
the latter (Lev et al., 2004; Cha et al., 2005); moreover, deletion of chromosome 1p has also been associated with elevated rCBV values (Jenkinson et al.,
2006). Relative CBV is a sensitive discriminant for
1p 19q deletion status, the genetic hallmark of oligodendroglioma features, in LGGs. This is, in effect,
a confounder of reliable tumor grading in a lesion
of unknown subtype; it may not be possible to distinguish between aggressive elements within a low
grade astrocytoma and a stable low grade oligodendroglioma. Pilocytic astrocytomas have been excluded
from some studies because they can demonstrate high
rCBV despite their classification as grade 1 gliomas
(Arvinda et al., 2009); fortunately they are frequently
morphologically distinctive on structural imaging,
although may occasionally mimic high grade diffuse
gliomas.
Differences in the management of patient cohorts
between studies will also affect rCBV. Treatment with
steroids can decrease tumor permeability and blood
volume measurements. Changes in permeability have
been described at one hour after steroid administration,
421
with a 15% decrease in the rCBV of peritumoral tissue

(Ostergaard et al., 1999).
The level at which to set a threshold value will
depend on what is felt to be the optimum balance
between acceptable false positives and false negative
rates. A high false positive rate will lead to potentially
unnecessary investigation and treatment with associated risks. A high false negative rate will lead to
inappropriate management of patients at high risk. The
judgement of the most appropriate balance is likely to
vary between authors.
The choice of reference point used to determine threshold values also varies between studies.
Arguments have been made for the use of hard clinical
endpoints rather than histology to determine threshold values (Bisdas et al., 2009). A clinical endpoint
may not necessarily correlate well with grade and there
are concerns over the validity of the use of histology
as a gold standard. Histological grading of gliomas
can be difficult, particularly when a tumor demonstrates an atypical pattern; suboptimal reproducibility
and interobserver variation are recognised challenges
(Coons et al., 1997). There is also the risk of sampling
a misrepresentative area of tumor, which will result
in misleading histopathological analysis in heterogeneous lesions; this typically results in an underestimate
of grade or aggressiveness.
As with other methods, characterising tumor
heterogeneity with perfusion presents challenges.
A commonly used technique is to analyse the brightest discrete region, although such an area may not
always be apparent, and the decision on the most
appropriate region of interest for analysis may vary
between observers. Some groups have overcome this
by analysing the whole tumor, although these methods
are more labor-intensive and time consuming.
The Role of rCBV Within a Multimodalilty

Approach to Glioma Evaluation
Opinions on the performance of rCBV compared to
other advanced imaging techniques for the evaluation
of LGGs vary, but there is general consensus that the
future role of rCBV is likely to be within the framework of a multimodal approach which incorporates
both conventional MRI and other advanced MRI techniques. Such an approach is advantageous because it
422
allows interrogation of different aspects of tumor biology, growth rate and tumor metabolism, in addition
to angiogenesis. Amalgamation of this information is
likely to give a more accurate estimation of risk of
transformation than the individual techniques, which
have their own intrinsic weaknesses.
Some comparisons between rCBV measurements,
diffusion imaging, MR Spectroscopy (MRS) and conventional MR imaging have found rCBV to be the best
performing parameter to distinguish low grade from
high grade gliomas (Law et al., 2003; Zonari et al.,
2007), and for distinguishing gliomas for other brain
tumors (Weber et al., 2006). MRS was found to be better than rCBV for regions near the cortex because the
rCBV of grey matter is less distinct from tumor values
than white matter (Henry et al., 2000), although coregistration of rCBV maps with structural images can help
to distinguish tumor signal from that of adjacent cortex. The evidence for rCBV as a marker of tumor
behaviour and grade is best established in diffuse
gliomas of astrocytic lineage. A borderline significant
elevation in choline/creatine ratio on MRS was demonstrated 12 months before rCBV changes (p = 0.06) in
oligodendrogliomas (Hlaihel et al., 2010); other techniques may be of particular use in characterising this
tumor subtype, in which rCBV is known to correlate
less reliably with malignancy. A study of a heterogeneous group of paediatric tumors has also reported
changes in choline adjacent to enhancing tumor bed
without corresponding changes in rCBV (Tzika et al.,
2002). A comparison of tumor volume growth over a
6 month interval with a number of other parameters,
including rCBV but not MRS measurements, found
tumor growth at 6 months was the best overall predictor for time to transformation in LGGs (Caseiras
et al., 2009), although comparative risk stratification
over short and long time periods was not investigated.
Some groups have attempted to overcome the limitations of individual advanced MR techniques by
developing algorithms which use a multimodality
approach. Groups have analysed sensitivity, specificity and predictive value of tumor grading following
addition of rCBV and MRS findings to conventional
MRI, and concluded that accuracy increases (Zonari
et al., 2007; Law et al., 2003; Arvinda et al., 2009).
A multimodal algorithm has been suggested, which
uses different techniques in a systematic manner to narrow the differential for an unknown intra-axial lesion
(Al-Okaili et al., 2006).
Conclusion
Relative CBV measurements offer a validated, noninvasive method of characterising and monitoring
LGGs. Standardisation of techniques for acquisition of
rCBV, consensus on the optimal threshold value to discriminate gliomas and appreciation of the variability in
rCBV between subgroups of gliomas will facilitate its
use in routine clinical practice. Current evidence suggests variable performance, with rCBV demonstrating
strengths in some areas but weaknesses in others when
comparison is made with other advanced MRI techniques. The future role of rCBV is likely to be as part
of a multimodal approach to glioma characterisation
and monitoring.
References
Al-Okaili RN, Krejza J, Wang S, Woo JH, Melhem ER
(2006) Advanced MR imaging techniques in the diagnosis of intraaxial brain tumors in adults. Radiographics 26:
S173S189
Aronen HJ, Gazit IE, Louis DN, Buchbinder BR, Pardo
FS, Weisskoff RM, Harsh GR, Cosgrove GR, Halpern
EF, Hochberg FH (1994) Cerebral blood volume maps of
gliomas: comparison with tumor grade and histologic findings. Radiology 191:4151
Arvinda H, Kesavadas C, Sarma P, Thomas B, Radhakrishnan
V, Gupta A, Kapilamoorthy T, Nair S (2009) Glioma grading: sensitivity, specificity, positive and negative predictive
values of diffusion and perfusion imaging. J Neurooncol 94:
8796
Bisdas S, Kirkpatrick M, Giglio P, Welsh C, Spampinato MV,
Rumboldt Z (2009) Cerebral blood volume measurements by
perfusion-weighted MR imaging in gliomas: ready for prime
time in predicting short-term outcome and recurrent disease?
Boxerman JL, Schmainda KM, Weisskoff RM (2006) Relative
cerebral blood volume maps corrected for contrast agent
extravasation significantly correlate with glioma tumor
grade, whereas uncorrected maps do not. AJNR Am
J Neuroradiol 27:859867
Caseiras G, Ciccarelli O, Altmann DR, Benton CE, Tozer DJ,
Tofts PS, Yousry TA, Rees J, Waldman AD, Jger HR
(2009) Low-grade gliomas: six-month tumor growth predicts
patient outcome better than admission tumor volume, relative
cerebral blood volume, and apparent diffusion coefficient.
Caseiras GB, Thornton JS, Yousry T, Benton C, Rees J,
Waldman AD, Jger HR (2008) Inclusion or exclusion of
intratumoral vessels in relative cerebral blood volume characterization in low-grade gliomas: does it make a difference?

Cha S, Tihan T, Crawford F, Fischbein NJ, Chang S, Bollen
A, Nelson SJ, Prados M, Berger MS, Dillon WP (2005)
Differentiation of low-grade oligodendrogliomas from lowgrade astrocytomas by using quantitative blood-volume measurements derived from dynamic susceptibility contrastenhanced MR imaging.. AJNR Am J Neuroradiol 26:
266273
Coons SW, Johnson PC, Scheithauer BW, Yates AJ, Pearl
DK (1997) Improving diagnostic accuracy and interobserver
concordance in the classification and grading of primary
gliomas. Cancer 79:13811393
Danchaivijitr N, Waldman AD, Tozer DJ, Benton CE, Caseiras
GB, Tofts PS, Rees JH, Jger HR (2008) Low-grade gliomas:
do changes in rCBV measurements at longitudinal perfusionweighted MR imaging predict malignant transformation?
Ginsberg LE, Fuller GN, Hashmi M, Leeds NE, Schomer DF
(1998) The significance of lack of MR contrast enhancement
of supratentorial brain tumors in adults: histopathological
evaluation of a series. Surg Neurol 49:436440
Hakyemez B, Erdogan C, Ercan I, Ergin N, Uysal S, Atahan S
(2005) High-grade and low-grade gliomas: differentiation by
using perfusion MR imaging. Clin Radiol 60:493502
Henry RG, Vigneron DB, Fischbein NJ, Grant PE, Day MR,
Noworolski SM, Star-Lack JM, Wald LL, Dillon WP, Chang
SM, Nelson SJ (2000) Comparison of relative cerebral blood
volume and proton spectroscopy in patients with treated
gliomas.. AJNR Am J Neuroradiol 21:357366
Hlaihel C, Guilloton L, Guyotat J, Streichenberger N, Honnorat
J, Cotton F (2010) Predictive value of multimodality MRI
using conventional, perfusion, and spectroscopy MR in
anaplastic transformation of low-grade oligodendrogliomas.
Jackson A, Kassner A, Annesley-Williams D, Reid H, Zhu XP,
Li KL (2002) Abnormalities in the recirculation phase of
contrast agent bolus passage in cerebral gliomas: comparison with relative blood volume and tumor grade. AJNR Am
J Neuroradiol 23:714
Jenkinson MD, Smith TS, Joyce KA, Fildes D, Broome J, Du
Plessis DG, Haylock B, Husband DJ, Warnke PC, Walker
C (2006) Cerebral blood volume, genotype and chemosensitivity in oligodendroglial tumors. Neuroradiology 48:
703713
Knopp EA, Cha S, Johnson G, Mazumdar A, Golfinos JG,
Zagzag D, Miller DC, Kelly PJ, Kricheff II (1999) Glial
neoplasms: dynamic contrast-enhanced T2 -weighted MR
Law M, Oh S, Babb JS, Wang E, Inglese M, Zagzag
D, Knopp EA, Johnson G (2006) Low-grade gliomas:
dynamic susceptibility-weighted contrast-enhanced perfusion MR imaging prediction of patient clinical response.
Law M, Yang S, Wang H, Babb JS, Johnson G, Cha S, Knopp
EA, Zagzag D (2003) Glioma grading: sensitivity, specificity,
and predictive values of perfusion MR imaging and proton
MR spectroscopic imaging compared with conventional MR
imaging. AJNR Am J Neuroradiol 24:19891998
423
Law M, Young RJ, Babb JS, Peccerelli N, Chheang S,

Gruber ML, Miller DC, Golfinos JG, Zagzag D, Johnson
G (2008) Gliomas: predicting time to progression or survival with cerebral blood volume measurements at dynamic
susceptibility-weighted contrast-enhanced perfusion MR
Lev MH, Ozsunar Y, Henson JW, Rasheed AA, Barest GD,
Harsh GRIV, Fitzek MM, Chiocca EA, Rabinov JD, Csavoy
AN, Rosen BR, Hochberg FH, Schaefer PW, Gonzalez
RG (2004) Glial tumor grading and outcome prediction
using dynamic spin-echo MR susceptibility mapping compared with conventional contrast-enhanced MR: confounding effect of elevated rCBV of oligodendrogliomas. AJNR
Ostergaard L, Hochberg FH, Rabinov JD, Sorensen AG, Lev
M, Kim L, Weisskoff RM, Gonzalez RG, Gyldensted C,
Rosen BR (1999) Early changes measured by magnetic resonance imaging in cerebral blood flow, blood volume, and
blood-brain barrier permeability following dexamethasone
treatment in patients with brain tumors. J Neurosurgery
90:300305
Rollin N, Guyotat J, Streichenberger N, Honnorat J, Tran Minh
VA, Cotton F (2006) Clinical relevance of diffusion and perfusion magnetic resonance imaging in assessing intra-axial
brain tumors. Neuroradiology 48:150159
Shin JH, Lee HK, Kwun BD, Kim JS, Kang W, Choi CG, Suh
DC (2002) Using relative cerebral blood flow and volume
to evaluate the histopathologic grade of cerebral gliomas:
preliminary results. Am J Roentgenol 179:783789
Sugahara T, Korogi Y, Kochi M, Ushio Y, Takahashi M (2001)
Perfusion-sensitive MR imaging of gliomas: comparison
between gradient-echo and spin-echo echo-planar imaging
techniques. AJNR Am J Neuroradiol 22:13061315
Tzika AA, Zarifi MK, Goumnerova L, Astrakas LG, Zurakowski
D, Young-Poussaint T, Anthony DC, Scott RM, Black
PM (2002) Neuroimaging in Pediatric Brain Tumors: GdDTPA-enhanced, hemodynamic, and diffusion MR imaging
compared with MR spectroscopic imaging. AJNR Am J
Weber MA, Zoubaa S, Schlieter M, Juttler E, Huttner HB,
Geletneky K, Ittrich C, Lichy MP, Kroll A, Debus J, Giesel
FL, Hartmann M, Essig M (2006) Diagnostic performance
of spectroscopic and perfusion MRI for distinction of brain
tumors. Neurology 66:18991906
White ML, Zhang Y, Kirby P, Ryken TC (2005) Can tumor
contrast enhancement be used as a criterion for differentiating tumor grades of oligodendrogliomas? AJNR Am J
Yang D, Korogi Y, Sugahara T, Kitajima M, Shigematsu Y,
Liang L, Ushio Y, Takahashi M (2002) Cerebral gliomas:
prospective comparison of multivoxel 2D chemical-shift
imaging proton MR spectroscopy, echoplanar perfusion and
diffusion-weighted MRI. Neuroradiology 44:656666
Zonari P, Baraldi P, Crisi G (2007) Multimodal MRI in the characterization of glial neoplasms: the combined role of singlevoxel MR spectroscopy, diffusion imaging and echo-planar
perfusion imaging. Neuroradiology 49:795803
Chapter 42
Angiocentric Glioma-Induced Seizures: Lesionectomy

Dave F. Clarke and Timothy M. George
Abstract Glioneuronal tumors, such as dysembryoplastic neuroepithelial tumors and gangliomas, are
known to induce intractable partial onset epilepsy.
A recently described tumor type, angiocentric glioma,
has been found to be extremely epileptogenic with
seizures affecting all diagnosed patients. Individuals
with complete tumor removal achieve freedom from
seizure, and those with partial removal have a significant reduction in seizures. This chapter will review
the complications of intractable symptomatic epilepsy,
discuss medical management, and make the case proceed directly to lesionectomy, however others require
further investigation prior to surgery. Identifying eloquent cortex is essential in surgical planning and
minimizing post-surgical morbidity. Non-invasive and
invasive techniques required to define the epiletogenic
zone and identify areas of critical cortical function will
be described.
Keywords Lesionectomy Angiocentric glioma
Intractable epilepsy Antiepileptic medication
Epilepsy surgery Seizure semiology
Introduction
Thirty to forty percent of patients with epilepsy
become pharmacoresistant after taking 23 medications (Kwan and Brodie, 2000). Many of these
D.F. Clarke ()

Division of Neurology, Department of Pediatrics, Dell
Childrens Medical Center, Austin, TX 78723, USA
e-mail: dfclarke@seton.org
patients are suitable candidates for epilepsy surgery

which may potentially cause seizure cessation or
significant reduction in seizures (Spencer and Huh,
2008). The potential for seizure freedom increases
if a lesion is found and completely removed (Lote
et al., 1998). Sir Victor Horsley, in his initial description of epilepsy surgery in 1886, described performing a lesionectomy on a tubeculoma, specifically for treating intractable epilepsy (Feindel, 2003).
Horsley used the clinical description of the seizure
to localize the area for resection. The field of
epilepsy surgery has since seen significant investigative advances with continuous evolving neuroimaging, electrophysiological techniques to identify seizure
foci, as well as invasive and non-invasive means of
identifying function. Identifying exactly where the
lesions are located in three dimensions and determining the spatial distribution of primary functional
(eloquent) cortex allows for better resection and minimizes postsurgical motor, language and memory
impairments.
Abnormalities of cortical development are the most
frequently described lesions causing seizures, however, tumors and or destructive lesions may also be
causative (Lellouch-Tubiana et al., 2005; Beaumont
and Whittle, 2000; Britton et al., 1994). Seizure prevalence is dependent on the pathological features of the
tumor (van Breemen et al., 2007). Dysembryoblastic
neuroepithelial tumors (DNETS) (100%) or gangliogliomas (90%), are described as having the highest seizure frequency, followed by other low-grade
gliomas (6085%) (Lellouch-Tubiana et al., 2005;
(Beaumont and Whittle, 2000; van Breemen et al.,
2007; Tandon et al., 2001). A recently described
tumor type, angiocentric glioma, likely approximates
the epileptogenicity found in DNETs. All patients

425
426
described in the literature with this tumor type thus

far, manifested seizures and most had pharmacoresistant epilepsy (Lellouch-Tubiana et al., 2005; Wang
et al., 2005; Preusser et al., 2007; Shakur et al., 2009;
Fulton et al., 2009). Most patients became seizure free
or had a significant reduction in seizures after the
removal of their angiocentric glioma (Shakur et al.,
2009). Unfortunately, in some individuals there was a
prolonged period between seizure onset and epilepsy
surgery potentially increasing risk of seizure related
morbidity (Lellouch-Tubiana et al., 2005; Wang et al.,
2005; Preusser et al., 2007).
There is an increase in mortality in patients with
epilepsy (Mohanraj et al., 2006). Patients who respond
to treatment have a much lower mortality risk when
compared to those with intractable seizures (Hitiris
et al., 2007; Lhatoo et al., 2001). Sudden unexpected death in epilepsy (SUPED) increases with
increasing seizure frequency and may be as high as
3.6/1000 in high risk populations (Ryvlin et al., 2006).
Significant psychosocial and neurocognitive pathology
is also found in patients with chronic epilepsy (Elliot
et al., 2005). The potential for significant morbidity and mortality associated with intractable epilepsy
suggests that patients with a lesion, especially those
who fit characteristics described in benign gliomas or
angiocentric gliomas, epilepsy surgery, which includes
lesionectomy, should be expedited. The concurrence
of epileptogenic gliomas with an approximating or
surrounding region of epileptogenic cortical dysplasia further supports the stance for expedited surgical resection (Duffau, 2005). Though this concurrent
phenomenon has not been confirmed in angiocentric
gliomas, it is important that the epiletogenic zone
be adequately delineated as it may extend beyond
the area visualized on imaging studies (Awad et al.,
1991).
Clearly, a multidisciplinary approach is required
to identify suitable surgery candidates. An ideal
team is comprised of an epiletologist, epilepsy surgeon, neuropsychologist, psychiatrist, neuroradiologist, electroencephalographer, specialized epilepsy
nurse, a social worker and, in the case of neoplasia, a neuro-oncologist. These individuals are
vitally important in identifying the epileptogenic
zone, minimizing post surgical impairment and
addressing any cognitive, developmental or psychosocial needs of the patient (Clarke and Boop,
2009).
D.F. Clarke and T.M. George
Patients Demographics
Angiocentric glioma-induced seizures were initially
described by Wang et al. (2005). Patients were presented as early as 3 years of age, and most presented
with seizures starting in childhood. Even earlier onset
of seizures has been described (Fulton et al., 2009).
Time from seizure onset to surgery in Wangs cohort
ranged from several months to as long as 25 years. Two
other studies, of 10 and 8 patients experienced maximal diagnosis to surgery intervals over 15 and 57 years,
respectively (Lellouch-Tubiana et al., 2005; Preusser
et al., 2007). More recent studies, show shorter delays
(Fulton et al., 2009; Shakur et al., 2009). Although
angiocentric glioma is a relatively rare tumor, no gender predilection has been described.
Complications of Intractable Epilepsy

The somewhat subjective definition of intractability
may have a role in the prolonged period from the time
of seizure onset to epilepsy surgery. It is known however, that two to three antiepileptic drug failures, either
as monotherapy or in combination, predict a low probability of seizure freedom with another drug Kwan and
Brodie found that 47% of individuals became seizure
free with the first anti-epileptic medication tried. A
second drug worked in an additional 13%, while only
another 4% gained seizure freedom with a third (1%
with monotherapy, and 3% in combination) (Kwan and
Brodie 2000). With a structural lesion, patients are
even less likely to become seizure free unless the lesion
is removed (Koepp and Woermann, 2005).
As stated earlier, seizures starting in childhood, the
period when most patients with angiocentic gliomas
present, may have the longest and most severe
impact on neurocognitive and psychosocial development (Elliot et al., 2005). Mortality in patients with
intractable epilepsy may be up to 47 time higher than
in the general population. Sudden unexplained death
in epilepsy (SUPED) is the most common epilepsyrelated death in adolescents and adults, occurring 24
times more frequently than death in the general population (Sperling, 2004). Both seizure frequency and
the number of AEDs used have been shown to be
risk factors (Ryvlin et al., 2006). Gliomas may also
be causative of the neurocognitive and behavioral
42 Angiocentric Glioma-Induced Seizures: Lesionectomy
dysfunction seen in patients with symptomatic

epilepsy. Epilepsy surgery leading to effective seizure
control has been shown to potentially improve quality
of life through neurocognitive and IQ score improvements (Freitag and Tuxhorn, 2005). Such changes are
likely due to multiple factors, including better seizure
control, lower dosing antiepileptic medication, which
decreases side effects, and the psychosocial improvements associated with seizure freedom.
Distribution and Semiology of

Angiocentric Glioma Induced Seizures
Studies suggest that most angiocentric gliomas occur
in the frontal lobe or temporal lobe. They are seen less
frequently in the parietal lobe and only rarely in the
occipital region (Lellouch-Tubiana et al., 2005; Wang
et al., 2005; Preusser et al., 2007; Shakur et al., 2009;
Fulton et al., 2009). Clinical presentation depends on
the neuroanatomical location of the lesion. Seizures
arising from the frontal lobe occur predominantly at
night, are brief, frequent, and cause little or no postictal changes (Losey and Ng, 2009). Due to their
atypical symptom, they may be mistaken for psychogenic non-epileptic spells. With a lesion involving
the primary motor area, seizures present with contralateral clonic or tonic motor activity. They begin as
simple partial seizures but cortical spread of the epileptic activity leads to more complex symptoms, including
altered consciousness and generalization. In addition
to spreading throughout the frontal lobe or crossing to
the contralateral hemisphere, the seizure may march
somatatopically along the homunculus to involve the
neuroanatomical area representing the upper extremity, face and tongue. The term Jacksonian march was
coined by Hughlings Jackson to describe this phenomenon (Holms, 1927).
Seizures involving the supplementary motor area
are also brief and nocturnal. They may be preceded
by a poorly defined aura, which is followed by atypical, but stereotyped movements involving one or both
sides. Consciousness may be preserved but vocalization and speech arrest is common, dependant on
the hemisphere involved. Posturing is an inconsistent
phenomenon.
Both lateral and fronto-polar seizures present with
versive head and eye deviation, as well as tonic/clonic
427
activity contralateral to the side of seizure onset.

Speech arrest is common with left hemisphere involvement. Automatisms, witnessed often, affect the side
ipsilateral to seizure onset, and may be accompanied
by paucity of movement on the contralateral side.
Both orbito-frontal and cingulated gyrus epileptic discharges may cause seizures simulating those arising
from the temporal lobe because both areas are involved
in limbic circuitry. Autonomic features maybe prominent in both. Olfactory hallucination and urinary
incontinence, laughing or atypical smiling, and sexual
automatism maybe experienced. Sensation of hunger
or thirst has been described. Pelvic thrusting, often
seen in non-epileptic spells may also be noted primarily in orbito-frontal seizures. Staring and diminished
responsiveness are also common. The rare opercular
lesion may cause excessive chewing and swallowing, increase salivation, speech arrest, and autonomic
features. Seizure spread and an invasive angiocentric glioma involving multiple lobes may trump the
involvement of only one area; producing combinations of the symptoms described above. Seizures are
however often stereotyped at onset in each individual.
Seizures arising from the temporal lobe, another
area frequently affected by angiocentric gliomas, have
been described by Weiser as exhibiting five typical symptom constellations (Wiser, 1983). The temporobasal limbic pattern stemming from hippocampal
and perhaps amygdale involvement causes alteration of
consciousness, symptoms of nausea and a cephalic
aura. Pupillary change, altered temperature sensation,
pallor or flushing (visceromotor symptoms), changes
in emotion (often fear), and focal motor activity are
also prominent. The temperopolar pattern involves
similar visceromotor symptoms and alteration of consciousness, but oroalimentary automatisms, behavioral
change, and psychic phenomenon may also occur.
The fronto-basal-cingulate type involves gesticularexploratory complex automatisms, bilateral lower limb
and trunk automatisms, and change in behavior and
facial expressions. Ictal warning with retained consciousness, psychic and auditory hallucinations and
aphasia is described in seizures of the opercular type.
Lastly, the posterior neocortical pattern is associated
with aphasia, staring and changes in facial expression,
vestibular and visual hallucinations, and unilateral
gestural-exploratory automatisms of the upper limb.
Younger children may not be able to describe their
symptoms. Instead they run towards a parent or other
428
adult as a sign that they are experiencing something

unusual. Parents and physicians therefore should be
more vigilant in evaluating this age group.
Angiocentric gliomas seldom affect the parietal
lobe, but those that do may present with atypical sensations such as numbness or other parathesias. Once
again, the younger patient may be incapable of elaborating and may even suggest pain or grab the area
during the seizure (Fulton et al., 2009). Due to the rapid
spread to the frontal and or temporal lobe, parietial
seizures are the most difficult to localize. Occipital
angiocentric gliomas are also rare, but present with
unusual visual disturbance. Either non-specific visual
hallucinations, such as white or colored light, or
the negative phenomenon, of itcal blindness may be
seen. Since headaches are a feature of migraines and
brain tumors, one must discern visual features associated with migraine headaches from those caused by
seizures. Children may have idiopathic partial or focal
generalized seizures independent of their tumor, therefore one has to always determine a causal relationship
between tumor and seizure.
Medical Management
Most researchers agree that treatment should be initiated with antiepileptic medication after a second
unprovoked seizure. At least two antiepileptic agents
should be tried before surgery is entertained (Kwan and
Brodie, 2000). Since evidence suggests that surgery
is potentially curative in this patient population, an
aggressive approach of simultaneously considering
surgery, while employing medication trials is warranted.
When choosing an antiepileptic medication for
patients with an intracranial mass, one has to take
into account not only the seizure type, but potential
co-morbidities such as headaches, memory loss and
other neurocognitive and neuropsychological deficits.
The tumor type is often unclear prior to removal;
therefore one must also consider the potential need
for chemotherapeutic agents. Expert consensus suggests that, in both children and adults, carbamazepine
and oxcarbazepine are drugs of first choice for partial
focal seizures (Wheless et al., 2005). With either agent
one has to be aware of the potential for worsening
frontal lobe seizures. One must also consider potential liver induction by both medications, which may
alter the distribution of both chemotherapeutic agents.
If comorbid headaches are experienced, valproic acid
or topiramate may be agents of choice; however, as
with carbamazepine, hepatic function may be altered
(Glauser et al., 2006). Both medications may affect
weight, with valproic acid stimulating hunger and topiramate having the opposite effect. Fosphenytoin may
be used for seizures in an acute setting, but long term
use is suboptimal due to induction, significant protein
binding, chronic gum hyperplasia, hirsutism, and bone
loss, to name a few potential pitfalls. Many of these
agents may also suppress the bone marrow, which is
less than ideal in this patient population. Gabapentin
and levatiracetam have little or no effect on both
chemotherapeutic agents. Levatiracetam occasionally
causes significant behavior changers, potentially compounding psychiatrics issues caused by lesions affecting the frontal lobe. Gabapentin is often less effective
than other antiepileptic agents and may cause weight
gain. There are pros and cons of using newer agents
such as locosamide and rufinamide. The metabolism,
volume, distribution and available preparation of all
antiepileptic drugs have to be taken into account. This
far from comprehensive list of antiepileptic drug side
effects offer another motive for considering surgery
sooner rather than later.
Investigative Studies
It has been recommended by both the International
League Against Epilepsy (ILAE) and the American
Epilepsy Society (AES) that an electroencephalogram
(EEG) be completed after the first unprovoked seizure
(Hirtz, 2000). Though the EEG yield is less than 50%,
a finding of focal slowing or discharges may help
to identify a seizure focus. An EEG capturing wakefulness and sleep using a, internationally recognized,
1020 electrode placement system increased yield.
Time locked Video EEG, is useful in distinguishing
nonepileptic from epileptic events.
Brain magnetic resonance imaging (MRI) is also
suggested after the first unprovoked seizure (Fig. 42.1).
T1 and T2 weighted images are essential (Koepp and
Woermann, 2005; Hirtz, 2000). T1 weighted images
define neuroanatomy and T2 is highly sensitive in
No
LESION
429
Anti-epileptic Medications
Subdural
Electrodes
with Cortical
Mapping
+/
Involving Eloquent
Cortex or Function
Focal
Seizure(s)
MRI
LESION
Seizure
Localization:
VEEG +/
MEG, PET,
SPECT
Function :
fMRI
MEG
WADA
Not involving
Eloquent Cortex
LESIONECTOMY
Fig. 42.1 Evaluation of tumor induced intractable epilepsy
detecting and defining pathological brain processes. T1

gradient echo, T2 fast spin echo, fluid attenuation iversion recover (FLAIR) and proton density sequences are
other techniques used in identifying dysplastic cortex
or tumors. Contrast material should always be given in
cases where tumors are suspected. The combination of
tumor and mesial temporal sclerosis is not unusual in
patients with intractable epilepsy. In such cases spoiled
gradient echo sequence with slices perpendicular to
the long access of the temporal lobe are sensitive in
detecting hippocampal atrophy.
Specific to angicentric gliomas, Lellouch-Tubiana
et al., described isointense cortical gyri with an intrinsic rim-like hyperintensity on T1 weighted images. On
T2 weighted image or FLAIR sequences the lesions
appeared hyperintense. The authors also described a
stalk-like extension of the lesion to the ventricle. In
other cases, however, there was neither lesional extension to the ventricular wall, nor ring-like enhancement
(Fulton et al., 2009; Preusser et al., 2007) Contrast
enhancement and calcification is seldom seen with
these lesions. They are described as most often seen
in the frontal and temporal regions. Several parasagittal lesions were described, yet others involved the
highly epiletogenic mesial temporal structures. Several
authors described streaks of grey matter from the surrounding normal cortex dispersed within the tumor
tissue (Fig. 42.2a, b). It is unclear whether these
entrapped neurons or the tumor cells are the epileptogenic culprits.
Positron emission tomography (PET) and single photon emission computed tomography (SPECT)
imaging studies enable identification of potential
epileptogentic zones (PET), which usually employs

the ligand, depicts metabolic differences in the brain.
PET performed done in the inter-ictal (between
seizures) period reveals hypometabolism of the epileptogenic zone (zone of seizure origin) (Knowlton, 2005;
Koepp and Woermann, 2005). It may delineate either
hypometabolism or, rarely hypermetabolism in the
region occupied by the lesion in question. SPECT
measures cerebral blood flow utilizing an injected
radionucleotide. Interictal SPECT detects abnormal
shows hypoprofusion similar in distribution to that
seen in PET, while an ictal (during seizure) SPECT
performed by injecting the radionucleotide at the onset
of the seizure can yield increased signal of the seizure
focus. A subtraction technique in which the ictal
data is removed from the interictal data significantly
improves specificity. Interictal PET and SPECT alone
offer less specificity in that the area of hypometabolism
is often larger in diameter then the actual epiletigenic
zone.
The Magnetoencephalogram (MEG) is a tool used
to detect magnetic fields produced by cortical neurons (Knowlton, 2005). The maximal magnetic field
is then superimposed on a three dimensional magnetic resonance imaging through the technique called
magnetic source imaging (MSI). This technique is
superior to surface electrodes placed on the scalp at
identifying the site of seizure onset. Because a special
magnetically shielded room is necessary to identify
only eliminate extraneous magnetic fields, the study
is usually obtained interictally. It may however better
delineate whether the seizure is coming from within or
around the tumor in question. MEG/MSI is less able
430
Fig. 42.2 (a) T1 Axial

showing an isointense
parasagital lesion with
imbedded depth electrodes
and interhemisperic
electrodes; (b) T2 MRI with a
hyperintense parasagital
lesion and subdural electrodes
visualized. Entrapped cortical
gray matter with finger like
projections are observed with
depth electrodes placed by
stealth to sample these areas;
(c) Resected gross
pathological specimen of
angiocentric gliomas shown in
Fig. 42.2a, b; (d) Histology
reveals the malignant cells
concentrically placed around
blood vessels of varying sizes.
There is diffuse
immunoreactivity of GFAP
and focal epithelial membrane
antigen staining
to evaluate very deep structures. Furthermore its sensitivity and specificity have been questioned in detecting
discharges in mesial temporal structures.
As is true with other seizure types, angiocentric
gliomas cannot always be removed due to proximity to
or involvement of the eloquent cortex. Unfortunately,
many prior studies fail to identify the method of functional analysis employed. A lesion merely displacing
eloquent cortex may allow for a complete resection;
therefore neuroanatomical estimation of primary functional areas is insufficient (Duffau, 2005). Identifying
the eloquent cortex prior to surgery is therefore essential (Fig. 42.3). If deficits are not preventable they may
be minimized and identifying function preoperatively
may enable, better preoperative counseling of patient
and relatives.
The best means of identifying cortical function is
direct cortical stimulation using electrodes placed on
the brain. This involves significant surgical and procedural risk, however, meaning that extra-operative
functional studies might be more advisable. fMRI and
MEG are non-invasive techniques now frequently used
to identify essential areas of primary function in the
pre-operative period and are therefore instrumental

in preoperative planning (Knowlton, 2005). WADA
may be used in cooperative patients with temporal
lobe involvement to lateralize memory and language
function (Clarke and Boop, 2009).
MEG functional studies identify somatosensory
evoked potentials, visual evoked potentials, auditory
evoked potentials, expressive and receptive language,
and primary motor cortex. This is accomplished by
using, magnetic field dipole localization, the same
technique described above for characterizing epileptiform activity. Frequent epileptiform discharges may
be difficult to discern from the discharges generated
by the magnetic field created by evoked potentials.
Functional magnetic resonance imaging (fMRI) will
also identify identical areas i.e., language, motor, sensory, auditory and visual cortices. fMRI samples blood
flow to produces still or moving images of the subjects
brain activity. Because of its dependence on blood
flow, sensitivity and specificity may suffer in vascular
lesions and tumors which distort blood flow.
Motor and language mapping require patient cooperation, which limits the use of MEG and fMRI in
431
Fig. 42.3 Graphic depiction

of subdural grid and strips
placed over the frontal and
temporal lobes. Eloquent
motor and language cortex is
labeled and is displaced by the
tumor shown on the T2 axial
MRI image in the lower right
of the figure
younger patients, due to their lesser ability to cooperate during testing. This disadvantage has however,
been overcome in more recent studies by the utilization of passive movement to identify the sensory
and motor (Ogg et al., 2009) regions. Recent literature has suggested that sedated receptive language
testing may also be carried out successfully during
light sleep (Perkins et al., 2008). Both fMRI and
MEG identify activity on the cortical surface, but are
unable to determine the trajectory of cortical pathways through white matter, which may be grossly
distorted by lesions. Magnetic resonance tractography
is a newer technique growing in popularity and use. It
delineates cerebral white matter tracts with diffusion
tensor imaging, which examines in water molecule
diffusion in tissue. The functional area can then be
identified on the cortical surface by MEG and fMRI
and the cortical tracts followed. Although the sensitivity, specificity, and therefore accuracy of the technique
are continually improving, it presently finds use in limited centers to assist in the Neurosurgical approach to
avoid functional pathways.
The most established extra-operative technique
for lateralizing expressive and receptive language
and memory has been the sodium amobarbital test
(WADA). In this procedure sodium amobarbital is
injected into the carotid artery to temporarily anesthetize one hemisphere, while neuropshycological testing of language and memory is carried out. Once the
patient recovers, testing is repeated by injecting the

cortralateral carotid artery. The technique described
above, however have substituted its use for language
in many centers and memory paradigms are being
attempted for both fMRI and MEG.
Cortical Mapping and Lesionectomy

Direct cortical stimulation is the most sensitive and
specific technique available for defining cortical function, in cases of close tumor proximity to somatosensory, motor, or expressive and receptive language
cortices. Language mapping is dependent on patient
response during current application while the patient
is actively speaking or a question is being asked.
Micro-currents are applied directly to subdural electrodes placed over the cortex just below after-discharge
(a rhythmic discharge caused by electrically activating cortical neurons which, if it persists, may
spread to other neurons and cause a clinical seizure).
After-discharge threshold is higher in young children
and patients with tumors; therefore more current is
required for an adequate response (Chitoku et al.,
2003). Stimulation over the motor area may cause tonic
or clonic movement while that over the sensory area
may cause a change in sensation. Testing can be done
in awake older individuals with an open craniotomy,
432
intraoperatively, or extraoperatively post intracranial

electrodes placement at the bedside. In younger children, motor mapping can be done with anesthesia in
the operating room, but language mapping efforts usually prove futile. Therefore mapping in young children
is best done extra-operatively.
Potential discordance between the seizure focus or
epileptogenic zone and the tumor, necessitates invasive
monitoring. Studies suggest that subdural electrodes
often identify the ictal onset zone or seizure onset zone
in the region immediately surrounding tumors (Tandon
et al., 2001; Awad et al., 1991). A more recent case
study questions whether placement of depth electrodes
in the body of certain tumors may reveal seizures originating from within some tumor types (Fulton et al.,
2009). Although further studies are required to test
this hypothesis, depth electrodes may assist subdural
electrodes may be helpful in obtaining more detailed
sampling of the areas around and within a tumor
(Fulton et al., 2009). This may then prevent unnecessary removal of normal functioning brain tissue.
Practical Surgical Application

A lesionectomy should always be performed if the
tumor does not approximate eloquent cortex and electrophysiological data supports the seizure coming
solely from the tumor. This is concordant with studies that support complete tumor resection as being the
best predictor of good seizure outcome (Awad et al.,
1991). As stated earlier, a two-staged procedure using
intracranial recording electrodes is performed to define
seizure onset or cortical function in uncertain situations. In the first surgical procedure the tumor site and
surrounding tissue is exposed. Pre-made grids or strips
with embedded subdural electrodes are placed over the
area in question, while depth electrodes can be placed
directly in the tumor. A digital picture may be taken
of the electrode array to document placement, identify
neuroanatomical landmarks, and allow later labeling
of seizure focus and functional mapping (Fig. 42.3).
The patient then returns to the video-EEG monitoring
room, with or without overnight ICU stay, and antiepileptic medications are withdrawn to capture typical
seizures. Imaging studies are carried out to define three
dimensional electrode placements. The necessary cortical mapping is then done at the bedside before the
patient returns for the second surgery. A hand drawn

or digital depiction of the region of seizure onset and
functional cortex is presented to the epilepsy team,
which then develops appropriate surgical plan (Rutka
et al., 1999). Removal of the abnormal area is then
carried out during the second surgical stage.
Pathology
Angiocentric gliomas are usually intra-cortical and
comprised of monomorphous, gliofibrillary acid protein positive (GFAP), usually slender cells often with
focal epithelial membrane antigen (EMA) immunoreactivity (Fig. 42.2c, d). Features of both astocytic and
ependymal differentiation are present (Fulton et al.,
2009; Lum et al., 2008). The tendency for tumor
cells to accumulate around blood vessels (Wang et al.,
2005) in an angiocentric manner lends the tumor its
name. Tumors cells are aligned in a radial or circumstantial longitudinal manner along small to medium
size vessels. This pattern suggests a developmental
nature of the tumor. Most lesions infiltrate the cortex
and surrounding subdural white matter and gradations
of minimally invasive or compact tumors displacing
surrounding tissues has been described (Lellouch).
White matter infiltration is accompanied by demyelination in some tumors (Lellouch). Preusser et al.
described finding single neurons interdispersed within
the tumor tissue itself, while Wang described trapping of cortical ganglion cells (Wang et al., 2005;
Preusser et al., 2007). Rosenthal fibers, eosinophil
granular bodies, and components of dysembryoplastic
neuroepithelial tumor are absent. Vascular proliferations, necrosis, mitosis and other features associated
with more aggressive tumor types are also seldom
described (Wang et al., 2005).
Seizure Outcome Post-resection

Only one reported death appears in the literature,
that of a 26 year old male with recurrence of a left
frontal lobe tumor (Wang et al., 2005). Seizure outcome in most cases is classified according to the Engle
proposed seizure classification (class 1-free of disabling seizure; class 2-rare disabling seizures; class
3-worthwhile improvement and class 4-no worthwhile

improvement) (Engle et al., 1993). All patients with
complete tumor resection were stated to be seizure free
or free from disabling seizures. The only patients identified as having rare disabling seizures were those in
whom total resection was not possible and the single
reported fatality. This agrees with a study by Awad
et al., which found postoperative control of partial
epilepsy to be up to 94% effective with a complete
lesioncetomy, 83% with incomplete lesionectomy but
complete seizure focus removal and 52% effective
with incomplete lesionectomy and incomplete resection of seizure focus (Awad et al., 1991). Other studies
failed to reveal similar success rates. One revealed
only 66% seizure freedom and 88% significant reduction in seizure frequency in patients with low-grade
neoplasm induced seizures (Britton et al., 1994). The
literature was unclear if patients remained on antiepileptic agents after angiocentric glioma removal.
However Preusser et al. described two seizure-free
patients who were not on any anti-epileptic agents,
both of whom had a total lesionectomy. While a more
recent study have shown that, post-tumor resection,
individuals may be taken off anti-epileptic agents successfully within 36 months, most physicians continue
patients on antiepileptic agents for 12 years prior to
entertaining weaning.
Conclusion
Angiocentric glioma is a recently described, novel
tumor type with distinct pathological features. It is
highly epiltogenic with most affected patients presenting with seizures. As in other more indolent tumor
types, the true prevalence is unclear. Apart from a
single case report, angiocentic glioma has a clinical pathological characteristic of a low-grade lesion.
Prolonged period of sympotomatic epilepsy of up to
57 years in individuals affected is in concordance with
these lesions being biologically benign. Nevertheless,
there are severe psychosocial and neurocognitive ramifications of poorly controlled epilepsy over so many
years. Antiepileptic agents may cause additional problems. As in other low grade lesions the angiocentric glioma, though not aggressive, is not necessarily
benign. Factors described above suggest that surgical resection should be expedited if possible. It is
433
imperative however, that electrophysiological and neuroimaging techniques be utilized appropriately to identify the region of seizure onset and functional areas,
therefore minimizing post-operative mobidity. Total
lesion resection provides best seizure outcome. It is
unclear, however, how many patents could be successfully taken off anti-epileptic medication after tumor
removal.
Acknowledgements We would like to thank Dr. Jeffrey Kerr for
his editorial expertise and Santa LeSure for the graphic labeling
and design. Radiographic images were completed at Le Bonheur
Childrens Medical Center and used with permission.
References
Awad IA, Rosenfeld J, Ahl J, Hahn JF, Luders H (1991)
Intractable epilepsy and structural lesions of the brain: mapping, resection strategies, and seizure outcome. Epilepsia
32:179186
Beaumont A, Whittle IR (2000) The pathogenesis of tumor
associated epilepsy. Acta Neurochir (Wien) 142:115
Britton JW, Cascino GD, Sharbrough FW, Kelly PJ (1994)
Low-grade glial neoplasms and intractable partial epilepsy:
efficacy of surgical treatment. Epilepsia 35:11301135
Chitoku S, Otsubo H, Harada Y, Jay V, Rutka JT, Weiss
SK, Elliott I, Ochi A, Kitayama M, Snead OC (2003)
Characteristics of prolonged after discharges in children
with malformations of cortical development. Child Neurol
18:247253
Clarke DF, Boop F (2009) Epilepsy surgery. In: Wheless
J, Willmore L, Brumback R (eds) Advanced therapy
in epilepsy, vol 52. Peoples Medical Publishing House,
Shelton, CT, pp 402414
Duffau H (2005) Lessons from brain mapping in surgery for low
grade glioma: insights into associations between tumor and
Elliot IM, Lach L, Smith ML (2005) I just want to be normal:
a qualitative study exploring how children and adolescents
view the impact of intractable epilepsy on their quality of
life. Epilepsy Behav 74:664678
Engle J Jr, Van Ness PC, Rasmussen TB, Ojemann LM (1993)
Outcome with respect to epileptic seizures. In: Engel J Jr (ed)
Surgical treatment of the epilepsies, 2nd edn. Raven, New
York, NY, pp 609621
Feindel WJ (2003) Osler and the medico-chirurgical neurologist: horsley, cushing, and penfield. J Neurosurg 99(1):
188199
Freitag H, Tuxhorn I (2005) Cognitive function in preschool
children after epilepsy surgery: rationale for early intervention. Epilepsia 46:561567
Fulton SP, Clarke DF, Wheless JW, Ellison DW, Ogg R, Boop FA
(2009) Angiocentric glioma-induced seizures in a two-year
old child. J Child Neurol 24(7):852856
Glauser T, Ben-Menachem E, Bourgeois B, Cnaan A, Chadwick
D, Guerreiro C, Kalviainen R, Mattson R, Perucca E,
434
Tomson T (2006) ILAE treatment guidelines: evidence-based
analysis of antiepileptic drug efficacy and effectiveness as
initial monotherapy for epileptic seizures and syndromes.
Epilepsia 47(7):10941120
Hirtz D (2000) Practice parameter: evaluating a first nonfebrile
seizure in children. Neurology 55:616623
Hitiris N, Mohanraj R, Norrie J, Brodie MJ (2007) Mortality in
epilepsy. Epilepsy Behav 10(3):363376
Holms G (1927) On local epilepsy. Lancet 1:957962
Knowlton RC (2005) The role of FDG-positron emission tomography, itcal single photon emission computed tomography,
and magentoencephalography in the epilepsy surgery evaluation. Epilepsy Behav 8:91101
Koepp MJ, Woermann FG (2005) Imaging structure and function
in refractory focal epilepsy. Lancet Neurol 4:4253
Kwan P, Brodie MJ (2000) Early identification of refractory
epilepsy. N Engl J Med 342(5):314319
Lellouch-Tubiana A, Boddaert N, Bourgeois M, Fohlen M,
Anne Jouvet A, Delalande O, Seidenwurm D, Brunelle F,
Sainte-Rose C (2005) Angiocentric neuroepithelial tumor
(ANET): a new epilepsy-related clinicopathological entity
with distinctive MRI. Brain Pathol 15:281286
Lhatoo SD, Johnson AL, Goodridge DM, MacDonald BK,
Sander JW, Shorvon SD (2001) Mortality in epilepsy in the
first 11 to 14 years after diagnosis: multivariate analysis of a
long-term, prospective, population-based cohort. Ann Neurol
49(3):336344
Losey TE, Ng Y-t (2009) Partial-onset seizures and localizationrelated epilepsy syndromes. In: Wheless J, Willmore L,
Brumback R (eds) Advance therapy in epilepsy, vol
23. Peoples Medical Publishing House. Shelton, CT,
pp 165178
Lote K, Stenwig AE, Skullerud K, Hirschberg H (1998)
Prevalence and prognostic significance of epilepsy in patients
with gliomas. Eur J Cancer 34:98102
Lum DJ, Halliday W, Watson M, Smith A (2008) Law,
A. Cortical ependymoma or monomorphous angiocentric
glioma? Neuropathy 28:8186
Mohanraj R, Norrie J, Stephen L, Kelly K, Hitiris N, Brodie MJ
(2006) Mortality in adults with newly diagnosed and chronic
epilepsy: a retrospective comparative study. Lancet Neurol
5(6):484487

Ogg RJ, Laningham FH, Clarke DF, Einhaus S, Zou P, Tobias
ME, Boop FA (2009) Passive range of motion functional
magnetic resonance imaging localizing sensorimotor cortex
in sedated children. J Neurosurg Pediatr 4:317322
Perkins F, Clarke D, Richie S, Brewer V, Wheless J, McGregor
A, Ogg R, McManis M (2008) Passive language mapping
using magnetoencephalograpy. Epilepsia (Abst. 2.056)
Preusser M, Hoischen A, Novak K (2007) Angiocentric glioma:
report of clinico-pathologic and genetic findings in 8 cases.
Am J Surg Pathol 31:17091718
Rutka JT, Otsubo H, Kitano S, Sakamoto H, Shirasawa A, Ochi
A, Snead OC (1999) Utility of digital camera derived intraoperative images in the planning of epilepsy surgery for
children. Neurosurgery 45:11861191
Ryvlin P, Montavont A, Kahane P (2006) Sudden unexpected
death in epilepsy: from mechanisms to prevention. Curr Opin
Neurol 19:194199
Shakur SF, McGrit MJ, Johnson MW, Burger PC, Ahn E,
Carson G, Jallo GI (2009) Angiocentric glioma: a case series.
J Neurosurg Pediatr 3(3):197202
Spencer SS, Huh L (2008) Outcome of epilepsy surgery in adults
and children. Lancet Neurol 7:525537
Sperling MR (2004) The consequences of uncontrolled epilepsy.
CNS Spectr 9:98101, 106109.
Tandon PN, Mahapatra AK, Khosla A (2001) Epileptic seizures
in supratentorial gliomas. Neurol India 49:5559
van Breemen MSM, Wilms EB, Vecht CJ (2007) Epilepsy in
patients with brain tumors: epidemiology, mechanisms, and
management. Lancet Neurol 6:421430
Wang M, Tihan T, Rojiani AM, Bodhireddy SR, Prayson RA,
Iacuone JJ, Alles AJ, Donahue DJ, Hessler RB, Kim JH,
Haas M, Rosenblum MK, Burger PC (2005) Monomorphous
angiocentric glioma: a distinctive epileptogenic neoplasm
with features of infiltrating astrocytoma and ependymoma.
J Neuropathol Exp Neurol 64:875881
Wheless JW, Clarke DF, Carpenter D (2005) Treatment
of pediatrics epilepsy: expert opinion. J Child Neurol
20:5156
Wiser HG (1983) Part 3. Electroclinical features of the psychomotor seizure. Butterworths, London, pp 177235
Index
A
Abernathey, C.D., 374375
Abou-Ghazal, M., 317
Achatz, M.I., 36
Acute myeloid leukemias, 60
Adam, J.F., 350
Adeno-associated viruses (AAV), 338
Adenoviral vectors (Ads), 344
CAR, 338
HSV-tk gene therapy, 345
MHC, 338
suicide gene therapy, 345
Adjuvant chemotherapy, 210
Adult brainstem gliomas
diffuse intrinsic/diffusely infiltrative low grade
clinical presentation, 372
diplopia, axial flair MR, 372
epidemiology, 372
follow-up, 373
pathology, 372
radiological presentation, 372
treatment, 372373
focal tectal brainstem gliomas
clinical and radiological presentation, 375
epidemiology, 375
pathology, 375
treatment and follow-up, 376
malignant brainstem gliomas
differential diagnoses, 374375
epidemiology, 373
follow-up, 375
gadolinium infusion, axial T1-weighted MR, 374
pathology, 373
treatment, 374375
neurofibromatosis type 1, 376
posterior exophytic gliomas, 376
Adult gliomas, see Gliomagenesis
Aghili, M., 18, 51
Aghi, M.K., 252
Agnelli, G., 381
Aguado, T., 283
Albert, F.K., 182
Alberts, B., 304

Alexander, A.L., 114115, 118
Allan, R.S., 144, 146
Alldredge, B.K., 404
Al-Okaili, R.N., 422
Ammirati, M., 5
Anaplastic astrocytoma (AA), 207, 336
adjuvant chemotherapy, 210
lomustine chemotherapy, 209
procarbazine chemotherapy, 209
temozolomide, 209210
vincristine chemotherapy, 209
Anaplastic gliomas, 12
Anaplastic oligodendrogliomas (AO) (WHO grade III), 47, 212
Angiocentric glioma, see Lesionectomy
Angiogenesis, 291
Animal glioma models
cell line and host
C6 cell line, 349
9L cell line, 349
selection of, 349
implantation technique
animals head, 352
burr hole, 352
choice of coordinates, 351352
localization, 353
microinfusion pump, 352
MR and PET, 351
parameters, 354
tumor cell suspension, 352353
standardized model
cell line preparation, 354
Antiepileptic drug treatment in glioma
carbamazepine, 403
enzymatic induction, 403
hepatic cytochrome P450, 403
phenytoin, 403
valproate, 404
Anti-glial fibrillary acidic protein (GFAP), 33
Antisense oligonucleotides (AON), 330
Antitumor treatment in glioma
chemotherapy, 405
intraoperative electrostimulation mapping, 405
radiotherapy, 405

DOI 10.1007/978-94-007-0618-7, Springer Science+Business Media B.V. 2011
435
436
Anti-vascular endothelial growth factor in glioma
angiogenic regulators
siRNAs, 364
anti-angiogenic therapy
GBM, 364365
hypoxia, 364
immunhistochemical detection, 364365
malignant glioma cells, 365
PI3 levels, 365
gene profiling studies, model in chick embryo
CAM, 362363
GeneChips, 362363
tumour, 362364
U87 cells, 362363
molecular pathways inhibition
CAM, 367368
pre-clinical evaluation
angiogenic and invasive phenotypes, 367
knockdown of IL6 and/or VEGF, 366367
siRNA, 366367
tumor cells in CAM, 366367
Aoki, H., 283, 321
AON, see Antisense oligonucleotides (AON)
Ardon, H., 313330
Aronen, H.J., 82, 419420
Aronica, E., 404
Arsenic trioxide, 45
Arvinda, H., 419, 421422
Ashery, U., 262
Astner, S.T., 143
Astrocytic tumors, 15, 25
pathogenesis and progression, 55
Astrocytoma, 419
Athanassiou, H., 208
ATP citrate lyase (ACLY), 2
R
, see Bevacizumab
Avastin
Awad, I.A., 426, 432433
B
Bailey, P., 25
Balkwill, F., 317
Balss, J., 18, 4849, 54, 5657, 60
Banissi, C., 329
Banks, L., 44
Bao, S., 19, 271, 317
Barker, F.G.I.I., 124
Barkovich, A.J., 388
Barondes, S.H., 262
Barth, R.F., 234235, 349350
Basser, P.J., 115
Batchelor, T.T., 211, 293
Bauchet, L., 143
Bauman, G.S., 218, 226
Beaulieu, C., 114, 116
Beaumont, A., 399, 425
Beckner, M.E., 2
Been, L.B., 104
Belanich, M., 74
Benard, F., 139
Benhar, M., 67
Index
Benjamin, L.E., 291
Bennet, C.L., 258
Bennett, A., 181
Ben Taarit, C., 375
Beppu, T., 116, 118
Berger, M.S., 174175, 181182
Bergers, G., 291, 298
Bergsneider, M., 185
Bernays, R.L., 185
Bernstein, J.J., 144
Berntsson, S.G., 401
Bertolone, S.J., 256
Bevacizumab, 2, 209, 289
clinical and radiological benefit, 295
clinical effect
corticosteroid use, 295
neurocognetive function and performance status,
295296
clinical response evaluation
steroid effect, 296
HGG and efficacy, 294
phase 2 pilot study, 3
recurrent glioblastoma, 3
treatment with irinotecan, 295
VEGF inhibition, 294
Bhujwalla, Z.M., 83
Bi-exponential decay model, 304
Bigner, D.D., 124
Bigner, S.H., 124
Binello, E., 302
Biological fluids biomarkers, 12
cerebrospinal fluid, 13
diagnosis and tumor heterogeneity, 13
LGG, 13
magnetic resonance imaging (MRI), 13
Biomarkers
cancer biology, new findings in
genomics, 16
glioblastomas and, 17
TIC and TSC, 16
transcriptome, 1617
neurosphere cultures
glioblastoma multiforme, 22
gliomas and, 2122
IDH1 and IDH2, 22
LOH 1p/19q, 22
PTEN deficiency, 22
usefulness of, 14
Biomarkers and HGG management
course
anaplastic oligos, 15
glioneuronal tumors, 15
WHO classification, 15
response to treatments
chemotherapy, 16
grade III tumors and, 1516
radiation therapy, 16
stratifications, 16
temozolomide, 16
Biomarkers and LGG management
benign tumors, 13
Index
response to treatments
chemotherapy, 15
methylation status, MGMT, 15
oligendendroglioma, 15
spontaneous course
alkylating chemotherapy, 14
anaplastic oligodendroglial, 14
anticipated usefulness of, 14
chromosomal material, loss of, 14
Delta Ex2-3 isoform of p73 gene, 1314
tumor growth rate and volume, 13
1,3-Bis(2-chloroethyl)- 1-nitrosourea (BCNU), 64
Bisdas, S., 419, 421
Bitan, M., 382
Black, P.M., 182
Blanchard, J., 356
Blazquez, C., 283
Bleeker, F.E., 26, 49
Bloodbrain barrier (BBB), 97, 108, 314, 316
BNL, see Brookhaven National Laboratory (BNL)
Bogler, O., 34
Bohinski, R.J., 185
Bone morphogenetic proteins (BMPs), 4
Boop, F., 426, 430
Boop, F.A., 426, 430
Borbly, K., 108
Borden, E.C., 269, 273
Borghaei, H., 318
Boron neutron capture protocol, 6
Boron neutron capture therapy (BNCT)
beam facilities
bismuth filter, 236237
clinical trials, 235
cross-section of, 236
research reactors, 235236
spectrum control devices, 236
clinical studies
beams penetrability, 232
BNL, 232
BSH distribution, 233
BSH-mediated, 232
external beam, 233
C-methionine-PET, 231
delivery system and
antibody-antigen reaction, 235
characteristics, 235
DDS, 235
direct intratumoral injection, 235
irradiation room and head positioning, 231
LET and PET, 230
reaction of, 230
treatment planning system
INEEL, 237
radiotherapy, 238
ROI, 237
therapeutic advantage in, 237
Boss, A., 132
Bouillot-Eimer, S., 143144, 146147
Bourgeois, M., 407
Bouzier, A.K., 65
437
Boviatsis, E.J., 374
Boxerman, J.L., 84, 418, 421
Brada, M., 77, 251, 405
Bradbury, M.W., 316
Bradbury, P.A., 319321
Brady, L.W., 124
Brain metastases, 113
Brainstem gliomas
classification
cervicomedullary, 388
diffuse intrinsic, 388
dorsal exophytic, 388
focal tectal, 388
clinical manifestations, 388
symptoms duration, 389
diagnosis
exophytic tumor, 391392
hyperintense FLAIR, 391
hypointense sagital T1, 391
Listeria monocytogenes, 389
MRI, 390
T2-weighted sagital MR sequence, 391
differential diagnosis, 389
fluid attenuation inversion recovery (FLAIR), 390
hemorrhagic or ischemic strokes, 390
infectious agents, 389
inflammatory lesions, 389390
epidemiology, 387
neurofibromatosis 1 (NF1), 388
treatment
chemotherapy, 393
mainstay of, 392393
single or multiple drug regimens, 393
surgical resection, 393
Brain tumor angiogenesis, 81
blood volume and permeability quantification
aspect of, 83
contrast agent leakage, 84
dual echo gradient echo perfusion weighted
imaging, 84
imaging techniques, 83
perfusion techniques, 8384
heterogeneity and perfusion imaging, 87
PCT
CBV map, 89
glioma grade and, 85, 86 87
kinetics theory, 8485
parameters, 85, 87
recurrent tumor from radiation necrosis, 8889
technique, 85
perfusion imaging and histopathology
WHO grading system, 82
perfusion parameters and importance, 82
tumor blood volume
endothelial cells, 82
intratumoral MVD, 82
measurement of, 82
solid tumors, 82
tumor vascular permeability, 83
Brain tumor-derived stem cells (BTSC), 5
438
Brain tumor-polyposis (BTP), 35
syndrome type 1 and 2, 36
Brain tumors, 31, 302
case study, 32
classification of, 336
Cowden syndrome and Lhermitte-Duclos disease, 41
DTI use in classification
apparent diffusion coefficient (ADC), 116
combined metrics, 119
FA role, enhancing region, 116, 118
measurements from, 116
shape based diffusion tensor metrics, 118
tumor infiltration assessing, 118119
gene therapy
AAV, 338
NDV, 338
pro-drug activation and suicide, 339
VPCs, 339
Gorlign syndrome, 41
groups, 336
Li-Fraumeni syndrome, 42
melanoma-astrocytoma syndrome, 42
methodology-diffusion tensor imaging, 114
fractional anisotropy (FA), 115
geometric nature, 115
isotropic diffusion, 115
mismatch deficiencies, 3536
mutation spectrum associated with, 3435
pathophysiology
brain metastases, 114
geometric interpretation of, 114
and stem cells, 18
distinct signaling pathways, 18
integrated genomic analysis, 18
NSC, presence of, 18
progenitors, 18
therapeutics, impact of, 18
tuberous sclerosis complex (TSC), 41
Turcot syndrome (TS), 41
von Hippel-Lindau (VHL) syndrome, 41
Brandes, A.A., 7576, 210211, 256
Brasil Caseiras, G., 168
Brem, H., 175
Brem, S., 175
Brisman, R., 379
Bristol, R.E., 164
Britton, J.W., 407, 425, 433
Brodie, M.J., 425426, 428
Brogna, C., 399, 404405
Broniscer, A., 393
Brookhaven National Laboratory (BNL), 232
Brown, K., 211
Brown, P.D., 177
Bru, A., 64
Bru, I., 64
BTP, see Brain tumor-polyposis (BTP)
BTSC, see Brain tumor-derived stem cells (BTSC)
Index
Buatti, J., 229230, 232
Buller, H.R., 380
Burger, P.C., 35, 147
Burian, J., 232
Busse, P.M., 232
Bynevelt, M., 156
C
Cairncross, G., 290
Cairncross, J.G., 16, 26, 82, 212
Calli, C., 116
Camby, I., 261264, 266
Camelo-Piragua, S., 49
Campora, R.G., 143
R
, see Irinotecan
Campto
R
, see Irinotecan
Camptosar
Cancer, 301
stem cells, type I IFNs, 270
tissue, 303
types and worlds population, 1
Candotti, F., 337
Cannabinoids (anti-cancer agents) in glioblastoma, 277
anti-tumoral action, 280
demand for oxygen and nutrients supply, 283
inhibition of pro-survival pathways, 280281
pro-angiogenic factors, 283
role of ER stress and autophagy in, 281283
applications, 278
clinical use of
pilot study, 284285
risks, constrains and benefits, 285
safety profile, 284
THC capsules, 283284
health and brain tumors
CB1 and CB2 receptors, 278
CB2 immunostaining, 279
central and peripheral effects, 278
endogenous, 278
immunohistochemical analysis, 278279
malignancy grade and CB2 expression, 279280
plant-derived 9THX, chemical structures, 279
Cao, Y., 104, 108
Capelle, L., 190, 399, 405
Capper, D., 49, 58
Capuani, S., 235
CAR, see Coxsackie- and adenovirus receptor (CAR)
Carbamazepine, 403, 428
Carbone, F.R., 316
Carmel, P.W., 185
Carpentier, P.A., 3
Carracedo, A., 280281
Carro, M.S., 31
Carson, K.A., 3
Casanovas, O., 361
Caseiras, G., 168
Caseiras, G.B., 421, 422
Cassoni, P., 143148
Castellot, J.J.Jr., 382
Cavaliere, R., 401, 404
CBTRUS, see Central brain tumor registry of united states
(CBTRUS)
Index
CD133-positive cells in glioblastoma, 4
Cell death mechanisms, 302
Cenic, A., 83
Cental nervous system (CNS) tumors, 143
Central brain tumor registry of united states (CBTRUS), 124
Cerebral gliomas, 4
dynamic scale, 156
macroscopic scale, 156
microscopic scale
histopathological tridimentionnal architectural, 155
isolated tumor cells, 154
microscopic tumor, 154
pilocytic astrocytomas, 156
solid tumor tissue, 154
tumor capsule, 154
WHO tumor grades, 156
Cervicomedullary gliomas, 388
Cervio, A., 143
Ceyssens, S., 104, 108109
CGH, see Comparative genomic hybridization (CGH)
Chaichana, K.L., 404
Chakravarti, A., 27
Chamberlain, M.C., 211, 376
Chanana, A.D., 232
Chandy, M.J., 373375
Chang, C.Y., 366
Chang, E.F., 404
Chang, E.L., 216
Chang, S., 250
Chang, S.M., 217
Chang, Y.S., 95, 97
Chan, J.A., 273
Chan, J.L., 216
Cha, S., 8283, 421
Chemotherapy for GB, 222, 393
left inferior limb arteriography, 257
temozolomide, cisplatin, thalidomide
clinical experience, 257
combination of cisplatin and, 256
mielotoxicity, 258
rationale, 256
resistance to, 256
vascular complications, 258
Chenevert, T.L., 115116, 306
Cheng, S.Y., 361
Chen, R., 22
Chen, W., 104, 110
Chen, X., 165166
Chen, Y., 273
Chen, Y.S., 93100
Chen, Z.P., 93100
Chinot, O.L., 7374
Chiocca, E.A., 340
Chitoku, S., 431
Choi, B.D., 319, 321
Choi, S.J., 104, 110
Cho, K.H., 222
Chorio-allantoic membrane (CAM), 361364, 366368
Choyke, P.L., 83
Christofori, G., 368
Ciafre, S.A., 340
439
Cibula, J.E., 399
Ciechomska, I., 244, 252, 277285
Ciesielski, M.J., 326
Cisplatin
clinical experience
phase II study, 257
malignant gliomas, 256
mielotoxicity
dose limiting toxicity (DLT), 258
dynamic contrast enhanced magnetic resonance
(dMRI), 258
phase III study, 255256
thalidomide and temozolomide, 256
vascular complications
phase II study, 258
venous grade 3-4 toxicity, 258
Clarijs, R., 9798
Clarke, D.F., 425433
Clarke, J., 242
Claus, E.B., 185, 190
Clauss, A., 364
Clausse, N., 264
Clinical target volume (CTV), 133, 216
Cloughesy, T.F., 211, 250
Coakham, H.B., 372
Coate, L., 207212
Coderre, J.A., 234235
Collen, A., 382
Colon cancer, 32
Colvin, D.C., 309
Combs, S.E., 215, 221, 223
Comparative genomic hybridization (CGH), 19
Conditionally replicative adenoviruses (CRADs), 340
Conformal radiotherapy, 215
Connor, S.E., 156
Conturo, T.E., 84
Coons, S.W., 64, 70, 417418, 421
Copier, J., 313
Covarrubias, D.J., 88
Coxsackie- and adenovirus receptor (CAR), 338
CRADs, see Conditionally replicative adenoviruses (CRADs)
Criniere, E., 27
Croteau, D., 157
Crum, B.A., 375
Cserr, H.F., 316
CTV, see Clinical target volume (CTV)
Curran, W.J.Jr., 181, 382
Current therapy
adjuvant radiotherapy and chemotherapy, 337
maximum radiation dose, 336
Curtin, N.J., 73
Cushing, H.W., 25
D
Daglioglu, E., 388
Dahlstrand, J., 270
Dai, C., 362
DAmato, R.J., 256
Danchaivijitr, N., 419
440
Dang, C.V., 67
Dang, L., 18, 50, 5860
DAtri, S., 256
Datta, C.K., 144
Daumas-Duport, C., 12, 154, 157158, 160, 165, 410
3D Conformal/stereotactic radiotherapy
BED, 220
conventional radiotherapy with, 221
fractionated stereotactic radiotherapy with, 221
LINAC based stereotactic radiosurgery with, 221
re-irradiation studies, 220
DDS, see Drug delivery systems (DDS)
Deacetylase inhibitors, 36
Death effectors, 4
De Benedictis, A., 194, 202
Debinski, W., 124
De Carli, E., 57
Deckers, R., 4
Deep vein thromboembolism (DVT), 379, 381383
De Groot, J.F., 366
Deitcher, S.R., 380
Delattre, J.Y., 371376
Delayed-type hypersensitivity (DTH), 321
Dello Russo, C., 252
Del Pulgar, T.G., 280281, 285
De Lussanet, Q.G., 83
Deoxyribonucleic acid (DNA), 337
De Sio, L., 393
Desmurget, M., 199201
Devinsky, O., 404
De Vleeschouwer, S., 316321, 323, 325326,
329330
Dhami, M.S., 258
Dhermain, F., 216
Dhodapkar, K.M., 319
Diaz, A.Z., 232
Dickinson, L.D., 382
Diffuse intrinsic gliomas, 388
Diffuse low-grade gliomas (LGG), 163
history of, 154
image-based monitoring of
MTD, 166
natural course, 167
quantitative radiological growth rates, 166168
radiological growth pattern, 166
spontaneous tumor growth rates, 168
VDE, 166
MRI-based estimation of
dynamic approach, 157
FLAIR sequences, 156
malignancy grade, 157
static approach, 156157
tumor growth rate, 157
MRI reflect, actual spatial extent of, 157
MRI underestimates, spatial extent of
biopsy samples, 158
features, 158
isolated tumor cells, cycling, 158159
MIB-1 positive cells, 158
preoperative MRI, 159
Index
prognostic value, 160
surgical resection of, 160
radiological growth rates
limitation of, 170
prognostic factor, 168170
spontaneous VDE, 169
static parameters, 170
surgery benefits, 170
treatment efficacy, 170
spatial configuration of cerebral gliomas
dynamic scale, 156
macroscopic scale, 156
microscopic scale, 154156
surgical resection of, 153
three-step natural history of
initial silent period, 164
malignant progression, 165166
step-by-step approach, 164
symptomatic period of, 165
WHO grading, 153
Di Maio, S., 389, 392393
DIncalci, M., 256
Dirr, L.Y., 372
Dix, A.R., 318
DNA, see Deoxyribonucleic acid (DNA)
DNET, see Dysembryoblastic neuroepithelial tumors (DNET)
Doherty, L., 250
Dme, B., 97
Donghi, R., 37
Dong, J., 96, 100
Dorsal exophytic gliomas, 388
Double strand break repair pathway, 2
Drug delivery systems (DDS), 235
Drug resistance genes, 4
DTH, see Delayed-type hypersensitivity (DTH)
Dubbink, H.J., 26, 57
Duffau, H., 3, 153154, 165, 189204, 399, 405, 426, 430
Dulbeccos modified Eagle medium (DMEM), 354
Dumartin, L., 362
Dunn, G.P., 313314, 316317, 382
Duntsch, C., 280, 284285
DVT, see Deep vein thromboembolism (DVT)
Dysembryoblastic neuroepithelial tumors (DNET), 398399
E
Earnest, Ft, 157158
EI Hallani, S., 9496
Elgamal, E.A., 372
El Hallani, S., 1, 3944, 9496
Ellert-Miklaszewska, A., 277285
Ellika, S.K., 81, 83, 85
Elliot, I.M., 426
Elliott, L., 318
El Naqa, I., 135
Elola, M.T., 263
Elowitz, E.H., 283
EMA, see Epithelial membrane antigen (EMA)
Endothelial-lined vasculature, 95
Engell-Noerregaard, L., 314, 321
Index
Engle, J.Jr., 432433
EOR, see Extent of resection (EOR)
Epidermal growth factor receptor (EGFR), 211
Epileptic seizures in glioma
clinical manifestations
status epilepticus, 401
TIA, 401
tumor progression, 402
types, 401
differential diagnoses
cardiac arrythmia, 403
drop attacks, 402403
migraine aura, 402
psychogenic symptoms, 403
syncope, 402
TIA/ischemic infarction, 402
epidemiology
epileptogenic lesions, 398
epileptogenesis in
astrocytoma grade II, 400
glioblastoma histopathological sections, 400
pathophysiology, 399
peritumoral brain, 399, 401
SMA, 399
genetic factors, 401
tumor characteristics, 399
treatment
antiepileptic drug, 403404
antitumor, 405
hepatic cytochrome P450 enzyme system, 403
refractory seizures, 404
Epithelial membrane antigen (EMA), 430, 432
Epstein, F., 388
Eramo, A., 4, 19, 271, 317
Ernst-Stecken, A., 221
Esquela-Kerscher, A., 272
Esteller, M., 74
Estes, M.L., 412
European Organization for Research and Treatment of Cancer
(EORTC), 74
Evans, D.G., 42
Evans, S.M., 68
Everhard, S., 27
Extent of resection (EOR), 190
F
Fabi, A., 7377
Fabrini, M.G., 211
Fadok, V., 302
Faivre, S., 70
False-negative and false-positive glioblastoma, 25
Familial gliomas
genetic forms, 4344
genetic linkage analysis, 40
genetic predisposition, 40
melanoma-astrocytoma, 42
neurofibromatosis type 1 and type 2, 4041
441
segregation analysis, 40
TP53 gene, role in
polymorphism in sporadic, 44
role of, 4243
tuberous sclerosis complex (TSC), 41
Turcot syndrome (TS), 41
Fantin, V.R., 66
Fan, X., 5
Farmer, J.P., 372, 388
Fecci, P.E., 318
Feindel, W.J., 425
Ferner, R.E., 388
Ferrara, N., 292, 361
Figueroa, P., 143145, 147
Fine, H.A., 256
Fine-needle aspiration cytology (FNAc), 147
Fisher-F98 Model
cell line preparation
ATCC, 354
DMEM, 354
trypsinization of, 354
clinical evolution, 354355
histopathological characterization
immunocytochemistry, 355
pathologic analysis, 355
results of, 355356
implantation technique
bone wax, 354
Fitzek, M.M., 230
FNAc, see Fine-needle aspiration cytology (FNAc)
Focal tectal gliomas, 388
Folberg, R., 95
Folkman, J., 81, 99, 382
Fortin, D., 349358
Fosphenytoin, 428
Fouladi, M., 376
Fournier, E., 350351
Francis, C.W., 337
Frank, D.A., 273
Freeman, C.R., 372
Freitag, H., 427
French, P.J., 2528
Friedman, H., 210
Friedman, H.S., 7374, 215, 224, 290, 294295, 375
Fuchs, I., 375
Fueger, B.J., 138
Fu, J., 65
Fujimoto, K., 375
Fulton, D., 211
Fulton, S.P., 426429, 432
Functional neuronavigation, 191
Furnari, F.B., 47
Furukawa, K., 342
G
Gabapentin, 428
Gabrusiewicz, K., 241252
Gadeberg, P., 128
Gajjar, A., 393
Galanis, E., 211, 250
442
Galectin-1 (Gal 1)
glioblastoma chemo/radioresistance
anticancer drug, 265
hypoxia, 265
temozolomide, 265266
in glioma biology
angiogenesis, 265
cDNA microarray, 264
progression-associated genetic alterations, 264
stress conditions, 264
in vitro studies, 263264
Galectins family
angiogenesis, 265
carbohydrate-recognition domain (CRD), 262
cDNA microarray, 264
extracellular matrix, 263
progression-associated genetic alterations, 264
properties, 262
stress conditions, 264
in vitro studies, 263264
Galldiks, N., 104, 110
Galli, R., 100
Galons, J.-P., 306
Galve-Roperh, I., 278, 280281, 284285
Ganslandt, O., 157
Gathinji, M., 177
Gaviani, P., 372
GBM, see Glioblastoma multiforme (GBM)
Gemistocytic astrocytoma
case study, 32
Gene therapy
DNA, 337
MLV, 337
NIH Clinical Centre, 337
OTC, 337
transforming principle, 337
Gene therapy-induced apoptosis, 301
assessing response, water diffusion, 305307
BDIX rat brain, 306
biological aspects, 305
cell density, 306
inhomogeneous ADC response with, 307
tumor treatment response, diffusion MRI, 305306
water ADC initial decrease, 306
brain tumors, 302
Brownian motion, 301
cancer, 301
cancerous tissue, 303
cell death mechanisms
DNA fragmentation, 302
membrane-encapsulated apoptotic bodies, 302
nuclear condensation, 302
shrinkage of cell, 302
clinical aspects, 308
definition of, 302
FDA, 303
magnetic resonance methods
Index
diffusion and, 303
diffusion measurement of, 304
gradient parameters, 303
tissue, metabolite diffusion, 309
cell diameter and, 307
diffusion weighted spectroscopic imaging data, 308
MRS, 307
treatment response detection, MR contrast methods, 308
water diffusion
bi-exponential decay model, 304
brain tissue, 304
multi-compartmental diffusion, tissue, 304305
Gene transfer vector, adenoviruses as
anti-angiogenic therapy, 339340
scatter factor (SF), 339340
TIMP-3, 340
tumour development, 339
ZFPs, 339
immune modulation
GMCSF, 341
TGF-, 340
oncolytic viruses
CRADs, 340
Saccharomyces cerevisiae, 340
targeting tumour proliferation and
orthotopic rat glioma model, 341
retinoblastoma (Rb), 341
TIMPs, 342
Genome wide molecular markers, gliomas
gene expression profiling, 27
mesenchymal, 27
1p19q status and, 28
proliferative, 27
proneural, 27
George, T.M., 425
Germano, I., 5
Germano, I.M., 302
Germline mutation TP53 in glioblastoma
brain tumors and mutation spectrum
as carcinogen fingerprints, 34
IARC database, 34
IDH gene mutations, 34
R132C, 34
somatic genetic changes, 3435
wild-type allele and, 34
case 1 report, 32
astrocytoma, grade 2, 33
family pedigree, 33
TP53 immunostaining, 33
case 2 report, 32
adenocarcinoma of colon, 33
anti-glial fibrillary acidic protein (GFAP), 33
colon cancer, TP53 immunostaining of, 33
multiforme, 33
transcription repression, 33
CHEK2- related syndrome (LFS2), 31
genetic testing and
brain tumors, 35
Index
criteria for, 35
DNA, 35
germline TP53 mutations, 35
malignant tumors, 35
single nucleotide polymorphisms (SNPs), 35
Li-Fraumeni-like syndrome (LFLS), 31
Li-Fraumeni syndrome (LFS1), 31
Maffuci syndrome, 31
screening problems and preventions
deacetylase inhibitor, 36
gastrointestinal (GI) tract, 36
histone deacetylase type (HDAC) III, 36
neonatal mass screening, 36
p53-activating drugs and, 36
positron emission tomography (PET), 36
R337H mutation of TP53, 36
syndromes and
brain tumor-polyposis (BTP), 3536
hereditary non-polypotic colorectal cancer (HNPCC), 36
mismatch deficiencies, 3536
Muir-Torre syndrome, 36
TP53 disorders, 36
TP53 and glial tumors
adrenocortical cancers and, 32, 34
germline mutations of, 32, 34
Li-Fraumeni-like syndrome, 32
TP53, germline mutation of, 3132
tuberous sclerosis, 31
von Hippel-Lindau syndrome, 31
Gerson, S.L., 26, 28, 73
Geschwind, D.H., 198
Ghering, K., 202
Ghiringhelli, F., 329
Ghulam Muhammad, A.K., 346
Giannini, C., 412
Giese, A., 113, 154, 159, 165
Gilbert, M.R., 164
Gilmore, R.L., 399
Gil Robles, S., 194, 202
Ginsberg, L.E., 420
Glantz, M.J., 404
Glass, J., 222
Glauser, T., 428
Glial fibrillary acidic protein (GFAP), 12
Glial tumors, 32
Glioblastoma multiforme (GBM), 12, 73, 124, 138, 207, 318,
335, 379383
anaplastic astrocytoma (AA), 336
immunohistochemistry with antielafin
antibody, 365
MGMT, 208
mRNA levels of selected genes, 365
prognosis, 5
risks and uncertainity, 5
surgery, 5
survival, 209
temozolomide, 208
Glioblastomas, 113, 229230
angiogenesis and vasculogenesis, 1
boron neutron capture protocol, 6
443
medical imaging, 5
radiation risk, 6
model of, 49
subtypes, 12
treatments
active migration, 2
CD133 brain tumor cells differentiation, 4
chemoresistance of, 4
chemotherapy, 2
NOTCH signaling, 5
radiation, 2
Gliofibrillary acid protein positive (GFAP),
430, 432
Gliomagenesis
in adults, 11
biological fluids biomarkers, 12
cerebrospinal fluid, 13
diagnosis and tumor heterogeneity, 13
low grade gliomas (LGG), 13
biomarkers, 12
GBM, 12
genetic alterations
BRAF gene, amplification of, 17
Cancer Genome Atlas Research Network, 17
IDH1 and IDH2, mutations of, 1718
MSH6 mismatch, 18
pan-RAF inhibitors, 17
pilocytic astrocytomas, 18
prolylhydroxylases (PHDs), 18
low grade gliomas (LGG, grade II), 12
morphology, 12
oligodendroglial tumors, 12
Sainte Anne Hospital classifications
astrocytic, 12
glial fibrillary acidic protein (GFAP), 12
LOH 1p/19q, 12
oligodendroglial phenotype, 12
predominant cellular type, 12
astrocytic, 12
LOH 1p/19q, 12
Glioma-initiating cells, 269
Glioma models, anti-tumor effects
microglia
density, 247
Glioma stem cells (GSC), 95
regulation of microRNAs
/, 273
biological processes, 272
interference, 272273
quantitative real time polymerase chain reaction
(qRT-PCR), 273
terminal differentiation
STAT3 activation, 271
tyrosine residue of STAT3, 272
444
Glioma stem cells (GSC) (cont.)
type 1 IFNs
CD133+ cells and, 270271
chemotherapy resistance genes, 271
GBM, 270
GSC, 271
nestin, 270
tumor culture-maintaining cells, 270
Glycolysis in glioblastomas, 2
GMCSF, see Granulocyte monocyte-colony stimulating factor
(GMCSF)
Godlee, R., 181
Goh, V., 83
Goi, K., 32
Goldberg, L., 262
Goldman, S., 71
Goli, K.J., 256, 258
Gomez del Pulgar, T., 280281, 285
Gmez-Rio, M., 217
Goncalves-Ferreira, A.J., 374375
Gondi, C.S., 342
Gong, J.P., 278
Gorski, D.H., 224
Grade I glioma, 11
Grade III gliomas, 12
Grange, J.M., 319320
Granulocyte monocyte-colony stimulating factor
(GMCSF), 341
Grauer, O.M., 318321, 326, 329330
Gravendeel, L.A., 2528
Gregoire, V., 132
Griffin, C.A., 26
Griguer, C.E., 68
Grobben, B., 350
Grhn, O., 301310
Grhn, O.H.J., 306
Gross, S., 59
Gross tumor volume (GTV), 133
Grosu, A.L., 137140, 216, 223
GSC, see Glioma stem cells (GSC)
GTV, see Gross tumor volume (GTV)
Guertin, D.A., 249
Guha, C., 139
Guillamo, J.S., 371376, 388389, 393
Gunnersen, J.M., 263
Guo, A.C., 116
Gupta, R.K., 118
Gupta, V., 340
Gutin, P.H., 224225
Guzman, G., 97
Guzman, M., 278, 280, 283285
H
Hadani, M., 182183
Hagedorn, M., 361368
Haginomori, S., 237
Hakumki, J.M., 307308
Hakyemez, B., 419
Hale, A.J., 302
Index
Hallock, A., 215226
Hall, W.A., 185
Halperin, E.C., 216
Halper, J., 414
Hamilton, M.G., 379
Hamstra, D.A., 307, 309
Hanahan, D., 298, 361
Hargrave, D., 392
Hartmann, C., 26, 49, 5455, 60
Hasselbalch, B., 289298
Hatakeyama, T., 104, 108110
Hata, N., 144145
Hattori, Y., 258
Hau, P., 211, 330
Hayat, M.A., 16
Heath, W.R., 316
Hegi, M.E., 16, 27, 7375, 208
Heidorn, S.J., 17
Heimberger, A.B., 317
Held-Feindt, J., 285
Hemmati, H.D., 317
Hendrix, M.J.C., 95
Hendrix, M.J.S.E., 98
Henriksson, R., 234
Hentschel, S.J., 63
Hereditary non-polypotic colorectal cancer (HNPCC), 36
Herholz, K., 104, 108
Hess, A.R., 99, 114115
High-grade astrocytoma, 173
cortical mapping impact, 177178
HGG
resection, impact of, 175
historical considerations, 173174
LGG
multimodal therapy importance of, 175176
neuro-imaging impact, 177178
new post-operative deficits
Kaplan-Meier plot, 176177
WHO grade classification, 173
High-grade gliomas (HGG), 124, 313
biological imaging, 133
blood-brain barrier (BBB), 314
cancer immunoediting
dendritic cells (DC), 314
immunoedited phenotype, 316
MDSC, 314
natural killer (NK) cells, 314
tumor cells growth, 314
tumor vaccination, 315
cancer, types of, 314
central nervous system (CNS)
antigen-presenting cells (APC), 316
blood-brain barrier (BBB), 316
cerebrospinal fluid (CSF), 316
cervical lymphatic drainage of brain, 316
gliomas and immunity, 317
lymphocyte-trafficking, 317
clinical studies results, 329
PFS and OS, 328
post-vaccine IFN-, 328
Index
3D-CRT techniques, 132
experimental glioma models for, 319, 326
delayed-type hypersensitivity (DTH), 321
gliomas and immunity
human gliomas
DC, 326
dendritic cells nature, 327328
monitoring, 328
MRI and PET, 328
outcome measures, selection of, 327
RPA and RTOG, 327
immunotherapy
active specific, 319, 320321
adoptive, 318
optimization of, 329330
passive, 318
restorative, 318
TAA, 313
T cells, 313
MRI and PET, 131
NSCLC, 132
PET-guided BTV, 135
amino acid transport, 136
assessment of, 135137
biological tumor volume measurements, 136
CTV, 133
delineation, practice of, 138
18 F-FET PET, 139
18 F-FMISO, 137
gadolinium enhanced T1-weighted MRI, 139
glioblastoma, 138
GTV, 133, 139
hypoxia tracers, 137
LGG, 137
practice of target volume delineation, 138
PTV, 133
radiation delivery techniques, 132
radiation necrosis, 137138
radiotracers, 136
SPECT, 138
SUV, 139
tracer 18 F-FDG, 137
transaxial slices, 139
tumor hypoxia, 137
volumetric patient imaging, 133
radiation delivery techniques, 132133
stereotactic radiation therapy/surgery (SRT/SRS), 132
targeted therapies, 293
tomotherapy, 132
volumetric patient imaging, 133
X-ray CT, 131
High-grade gliomas (HGG) management
biomarkers
course, 15
response to treatments, 1516
in cancer biology, 1617
response to treatments prediction, 1516
spontaneous course, 15
Highsmith, R.F., 379
Hildebrand, J., 25, 401402
445
Hirose, T., 415
Hirschberg, H., 185
Hirtz, D., 428
Hitiris, N., 426
Hlaihel, C., 156, 166, 422
Hlatky, L., 296
HNPCC, see Hereditary non-polypotic colorectal cancer
(HNPCC)
Hochberg, F.H., 216
Hockaday, D., 123129
Holland, E., 362
Holland, E.C., 362
Holms, G., 427
Hong, X., 341
Horbinski, C., 58
Hormigo, A., 13
Hortelano, S., 306
Houston, S.C., 143145
Howlett, A.C., 277278, 280
Hsu, C.Y., 372
Hsu, E., 143
Hsu, P.P., 65
HSV-tk and ganciclovir therapy, 342
bystander effect, 342, 344
current gene therapy, 343
clinical efficacy, 344
mechanism of action, 342
cytomegalovirus, 343
Huang, Y.L., 361
Hudes, R.S., 226
Huff, C.A., 270
Huh, L., 425
Human glioma
I-TM-601 SPECT imaging in, 123
bioelimination, 127
CBTRUS, 124
chlorotoxin (CTX), 124125
co-registration, 127128
decay rate indications, 129
extracellular matrix (ECM), 124
flouro deoxyglucose (FDG), 124
fusion of images, 127
fusion volumes, 129
glioblastoma, 124
HGGs, 124
image analysis, 126127
magnetic resonance, 124
MRI and PET, 124, 128
MRS volumes, formation of, 124
planar images, 125
preparation and preliminary scans, 125
radiolabeled peptide injection, 125
spectroscopy, 126, 128
stereological estimation, 126127
TM-601 use, 125
tumor volumes, 127129
metabolic differences between regions, 6869
tumor nature, 6465
Hussain, S.F., 251
Hussain, S.P., 34
446
Hustinx, R., 138
Hwu, W.J., 99
2-Hydroxyglutarate (2-HG) levels in gliomas, 5051
Hypothetical tumour, 144
Hypoxia
cellular response
HIF- proteins, 291
HIF-1 transcription factor, 291
HIF-1 overexpression and angiogenesis, 290
necrotic regions of glioblastoma, 290
Hypoxia inducible factors (HIF), 66, 290
I
Iberti, T.J., 379
Idaho national engineering and environmental laboratory
(INEEL), 237
Idbaih, A., 1516
Ide, F., 44
IES, see Intraoperative electro-stimulation mapping (IES)
Ignatova, T.N., 270
Image guided radiation therapy (IGRT), 132
Immonen, A., 344
Immunotherapy
active specific
immuno-adjuvant, 319
reported clinical trials, 322325
T cells and, 319
tumor vaccination, 319
adoptive, 319
passive
glioblastoma, 319
PFC, 319
restorative, 318
IMRT, see Intensity modulated radiation therapy (IMRT)
INEEL, see Idaho national engineering and environmental
laboratory (INEEL)
Ingrassia, L., 266
Innate inter individual prognostic variability, 3
Inoue, M., 361
Insug, O., 326
Integrated genomic analysis, molecular classification of tumors
new tools
cellular phenotype, 19
comparative genomic hybridization (CGH), 19
DNA copy number, 19
DNA methylation profiling, 19
gene expression (trancriptome), 19
glioblastoma pathogenesis, 19
glioblastoma subtypes, 20
mesenchymal subtypes, 21
molecular biology, 19
proneural subtype, 20
patients care, transfer to
C/EBP and STAT3, 21
master gene regulators (MR), 21
microRNA (miRs) field, 21
Intensity modulated radiation therapy (IMRT), 132
Interleukin 6 (IL6) therapy in experimental glioma
angiogenic regulators
siRNAs precursors, 364
Index
anti-angiogenic therapy
glioblastoma, 364365
hypoxia, 364
immunhistochemical detection, 364365
malignant glioma cells, 365
PI3 levels, 365
gene profiling studies, model in chick embryo
CAM, 362363
GeneChips, 362363
tumor, 362364
U87 cells, 362363
molecular pathways inhibition
CAM, 367368
pre-clinical evaluation
angiogenic and invasive phenotypes, 367
knockdown of IL6 and/or VEGF, 366367
siRNA, 366367
tumor cells in CAM, 366367
Intracranial disease, 144
Intraoperative electro-stimulation mapping (IES)
cerebral functional anatomy, 194
functional organization, 195
functional organization of Wernickes area, 195
insular lobe, 195
left angular gyrus and calculation, 195196
left dorsolateral cortex, 196
oculomotor behavior, 196
premotor cortex and language, 195
SMA syndrome, 195
spatial awareness, 196
cerebral plastic potential
brain plasticity, 199
intraoperative plasticity, 199200
postoperative plasticity, 200
preoperative plasticity, 199
LGG surgery
Brocas area resection, 201
extensive glioma resection, 201
functional and oncological results in, 202
insular resection, 200
knob of hand resection, 201
non-dominant sensorimotor area resection, 200201
parietal posterior resection, 200
primary somatosensory area resection, 200
SMA resection, 200
temporal language area resection, 201
multiple-stages surgical approach in LGG
effective connectivity, 203
functional reshaping, 202
median time, 203
techniques and concepts, 203204
subcortical connectivity
dorsal phonological root, 197
language pathways, 197
motor tracts, 196197
somatosensory pathways, 197
spatial awareness, 198
ventral semantic root, 198
visual pathways, 197
Intraoperative eletrostimulation, 3
Intratumoral heterogeneity, 64
Index
Irinotecan, 289
carboxylesterase-mediated breakdown, 294
topoisomerase I, 294
toxicitiy, 294
Isocitrate dehydrogenase 1 and 2 (IDH1and IDH2), 47
cellular metabolism, role in, 50
genome-wide mutational analysis, 47
single molecular markers, gliomas, 26
Isocitrate dehydrogenases 1 and 2 (IDHI and IDH2) mutations
detection, 4849
as favorable prognostic factor, 50
genetic alterations with
changes, 56
gene expression analysis, 56
R132, 55
TP53, 55
transcriptional and epigenetic profiles, 56
gliomagenesis, 4950
in gliomas, 48
grades of, 49
2-HG levels, 5051
patient survival and, 48
properties
early timing, 58
frequency, 58
neomorphic activity, 59
specificity, 5859
wild-type formation, 5859
timing, genetic alterations, 5657
types
acute myeloid leukemias, 60
codons spectrum, 5960
function, impact on, 60
patients, differences between, 60
Isolated tumor cells, 154
Iwaizumi, M., 3136
Iwama, T., 157
J
Jachimczak, P., 355
Jack, C.R., 177
Jackson, A., 419
Jacks, T., 36
Jacobs, A., 104, 110, 139
Jadhav, U., 340
Jain, N., 144145
Jain, R., 8189
Jain, R.K., 291, 294, 297
Jallo, G.I., 388, 391, 393
Javerzat, S., 361368
Jendreyko, N., 361
Jenkinson, M.D., 157, 419, 421
Jenkins, R., 28
Jenkins, R.B., 14, 26
Jensen, R.L., 290, 364
Jeong, M., 341
Jeremic, B., 175
Joensuu, H., 232
Johannesen, T.B., 174
447
Johnson, P.C., 64, 70
Jones, D.T., 17
Jung, T.Y., 263264
Jurcic, V., 375
K
Kaba, S.E., 63
Kageyama, S., 3136
Kakimi, K., 320321
Kalachikov, S., 401
Kaminska, B., 241252, 277285
Kang, M.K., 65, 271
Kang, M.R., 49, 54, 60
Kang, S.K., 65
Kang, X., 340
Kang, Y.A., 339
Kankaanranta, L., 232, 237
Kapoor, G.S., 281
Karayan-Tapon, L., 1122
Kargiotis, O., 340
Karnofsky performance score (KPS), 3, 25
Karpala, A.J., 272
Kaschten, B., 104, 108109
Kato, Y., 49, 58
Katze, M.G., 273
Kawabata, S., 234
Kawai, N., 103111
Kayser-Gatchalian, M.C., 379
Kayser, K., 379
Keles, G.E., 175
Kelly, P.J., 154, 157158, 160
Kepes, J.J., 412
Kerbel, R.S., 97
Kerr, J.F.R., 302
Kesari, S., 211, 222, 224, 373, 388, 393
Kettunen, M.I., 302, 306, 308
Kihlstrom, L., 392
Kikuchi, T., 322323
Kim, B.M., 341
Kim, C.J., 108, 110
Kim, C.Y., 341
Kim, D.G., 108, 110
Kim, E.H., 4
Kim, H., 21
Kim, J.H., 108, 110
Kim, J.S., 108, 110
Kim, M., 341
Kim, M.H., 341
Kim, M.S., 58
Kim, S., 108, 110
Kim, T.H., 341
Kim, W.B., 341
King, A., 51
King, G.D., 341
King, M.A., 398
Kinloch, R.A., 302
Kinoshita, M., 116, 118119
Kiss, R., 261266
Kjaergaard, J., 326
Kleihues, P., 2, 34, 47, 50, 242, 336
448
Klein, M., 397, 403
Kloog, Y., 262
Kloosterhof, N.K., 26
Knisely, J.P., 364
Knopf, P.M., 316
Knopp, E.A., 417420
Knowlton, R.C., 429430
Koeller, K.K., 118
Koepp, M.J., 426, 428429
Kohn, D.B., 337
Ko, L., 349
Kormanik, P., 211
Kormanik, P.A., 211
Korshunov, A., 49
Koumenis, C., 265
Kouwenhoven, M.C., 26
Kracht, L.W., 108
Krainik, A., 195, 200, 202
Kral, T., 192
Kratimenos, G.P., 374375
Kreisl, T.N., 93, 210, 250251, 259, 294295
Krex, D., 75
Kroemer, G., 6566, 68
Kros, J.M., 25
Krown, S.E., 258
Kumabe, T., 4751
Kumada, H., 229238
Kumar, A.J., 88
Kumar, S., 258
Kunkel, P., 361
Kurachi, K., 3136
Kwan, P., 425426, 428
Kyritsis, A.P., 34, 43, 63
L
Lacroix, M., 175
Lactate dehydrogenase A (LDH-A), 66
Laerum, O.D., 63
Lahm, H., 263
Lai, A., 3
Laigle-Donadey, F., 371376
Lakka, S.S., 342
Laks, D.R., 22
Lamborn, K.R., 63
Lamfers, M.L., 346
Lamour, V., 366
Lampson, L.A., 349
Landolfi, J.C., 371372
Lang, F.F., 164
Langleben, D.D., 88
Laperriere, N.J., 216
Laraqui, L., 143
Larsen, C.J., 11
Lassen, U., 289298
Laster, D.W., 156
Law, M., 8184, 87, 119, 419420, 422
Laws, E.R., 181
Lecchi, M., 139
Leclercq, D., 191, 203
Lecomte, R., 357
Index
Ledford, H., 34
Le Duc, G., 64
Lee, H.C., 70
Lee, J., 270271
Lee, T.Y., 84
Lee, A.Y., 380
Leffler, H., 263
Lefranc, F., 3, 261266
Lehtimki, K., 301310
Lellouch-Tubiana, A., 415, 425427, 429
Le Mercier, M., 262266
Lenard, H.G., 389
Leon, S.P., 82
Le, Q.T., 261, 264265
Lesionectomy
angiocentric glioma, distribution and semiology, 427428
cortical mapping, 431432
and cortical mapping, 431432
demographics, patients
angiocentric glioma, 426
distribution and semiology
angiocentric glioma, 427428
seizures arising, 427428
intractable epilepsy, complications, 426427
anti-epileptic medication, 426427
SUPED, 426
investigative studies
eloquent cortex, tumour, 430431
fluid attenuation iversion recover (FLAIR), 429
functional magnetic resonance imaging (fMRI), 429431
magnetic resonance tractography, 431
MEG, 429431
parasagital lesion, 429430
PET and SPECT, 429
tumor induced intractable epilepsy, 429
medical management, 428
antiepileptic medication, 428
carbamazepine, 428
fosphenytoin, 428
gabapentin, 428
levatiracetam, 428
oxcarbazepine, 428
pathology, 432
EMA, 430, 432
GFAP, 430, 432
post-resection, seizure outcome, 432433
patients demographics, 432433
seizure outcome, 432433
post-resection, 432433
studies, 428431
surgical application, 432
Leukoencephalopathy, 218
Levatiracetam, 428
Lev, M.H., 87, 419, 421
Lewis, J., 280
LGG management, see Low grade gliomas (LGG) management
Lhatoo, S.D., 426
Li, A., 27
Liau, L.M., 322323, 327, 329
case study, 32
Index
characteristic tumors of, 34
diagnosis of, 32
glial tumors, 32, 34
missense variant TP53, 32
Liimatainen, T., 301310
Li, K.W., 392
Lima, M.A., 387394
Lim, E.C., 372, 389
Ling, C.C., 132133
Link, M., 375
Li, F.P., 32, 42
Litofsky, N.S., 177
Liu, G., 4, 19, 65, 271, 317
Liu, H.K., 36
Liu, A.K., 393
Liu, F.T., 261263
Liu, Y., 2
Li, V.W., 82
Li, Y.J., 35
LMWH, see Low Molecular Weight Heparin (LMWH)
Loeffler, S., 366
Loffler, D., 273, 275
Lombardi, D., 405
Lombardi, G., 255259
Lomustine chemotherapy, 209
Losey, T.E., 427
Loss of heterozygosity (LOH) of 1p19q, 64
Lote, K., 425
Louis, D.N., 12, 25, 47, 54, 93, 144, 262, 264, 289,
410411, 413
Lowenstein, D.H., 404
Low grade and anaplastic astrocytoma
adjuvant chemoradiation, 210
radiotherapy, 210
Low-grade gliomas (LGG), 137, 417, 419422
intraoperative electrical stimulations, 189
benefit-to-risk ratio, 190
complete resection, 190
EOR, 190
functional mapping, principles of, 191192
intraoperative electrical stimulations method, 192194
partial resection, 190
presurgical functional neuroimaging, 190191
presurgical neurocognitive evaluation, 190
subtotal resection, 190
rCBV role in malignant transformation
acquisition of, 418
conventional MRI, comparison of, 418, 420
differentiating gliomas, 420
evaluation, challenges of, 420421
multimodalilty approach, 421422
Low grade gliomas (LGG) management, 13
biomarkers
benign tumors, 13
response to treatments, 15
spontaneous course, 1314
tumor growth rate, 13
tumor volume, 13
449
Low Molecular Weight Heparin (LMWH), 379381, 383
Lubbe, J., 34, 42
Lucignani, G., 6
Lucio-Eterovic, A.K., 296297
Lum, D.J., 432
Lu, S., 113, 116, 118
Lu, W., 342
Lyman, G.H., 380
M
MacDonald, B.K., 295296
Macdonald, D.R., 211, 295296, 327
Mackie, K., 277
Maeda, M., 84
Maekawa, M., 3136
Maes, W., 326, 329
Maggio, W.W., 351
Magnetic resonance methods
diffusion and, 303
diffusion measurement of, 304
gradient parameters, 303
Major histocompatibility complex (MHC), 338
Maldonado-Lopez, R., 316
Malignant brainstem gliomas
differential diagnoses, 374375
epidemiology, 373
follow-up, 375
gadolinium infusion, axial T1-weighted MR, 374
pathology, 373
treatment, 374375
Malignant gliomas, 47, 261
clinical and experimental data, 262
clinical testing
diagnostic utility, 58
monoclonal antibodies, 58
PCR-based assays, 58
specificity of, 58
Gal1-expression, 261262
Gal1 thereby, 262
glioblastoma, 207
IDH1 and IDH2 mutations, see Isocitrate dehydrogenases 1
and 2 (IDHI and IDH2) mutations
molecular biomarkers, 211
anaplastic oligodendroglioma, 212
p53 mutations, 212
mutation types
acute myeloid leukemias, 60
codons spectrum, 5960
function, impact on, 60
patients, differences between, 60
patient age
adult studies, 57
older median, 57
pediatric, 57
primary glioblastomas, misclassification of, 57
patient survival
Cox regression multivariate analysis, 5758
450
Malignant gliomas (cont.)
multivariate analyses, 57
study on, 57
Ras signaling and oncogenesis, 262
recurrence management, 210
PFS, 211
therapeutic drugs
bevacizumab, 209
EGFR, 211
initial therapy, 208
PDGFR, 211
temozolomide, 208
timing, genetic alterations, 5657
tumor failure, patterns of
CTV, 216
magnetic resonance spectroscopy, 217
radiation treatment, 216
RTOG, 216
SFRT, 216
surgery, 215
tumor type distribution
VEGFR, 211
WHO grading system classification, 262
whole-genome analyses, 53
Malkin, D., 42
Malmberg, K.J., 314
Malmer, B., 40
Mamelak, A.N., 123129
Mancini, M., 242
Mandell, L.R., 392
Mandonnet, E., 153, 156157, 163171, 189, 194
Manford, M., 401
Maniotis, A.J., 93, 9598
Mardis, E.R., 49, 54, 60
Marin-Sanabria, E.A., 372
Markovic, D.S., 247, 251
Marras, L.C., 379
Martino, J., 203204
Marx, G.M., 256
Mason, W., 207212
Massager, N., 374375
Massi, P., 280, 284285
Mathiesen, T., 248
Mathieu, D., 350, 355356
Mathieu, V., 2, 264266
Matsuyama, J., 144145
Mauer, M.E., 177, 190
Mayer, R., 217218, 220
McAllister, S.D., 284285
McCubrey, J.A., 242
McDermott, D.F., 318, 329
McDonald, D.M., 83
McGirt, M.J., 176177, 181, 190
McKnight, S.L., 342
McMahon, E.J., 316
MDSC, see Myeloid-derived suppressor cells (MDSC)
Mean Tumor Diameter (MTD), 166
Melanoma-astrocytoma syndrome, 42
Melhem, E.R., 113119
Mellinghoff, I.K., 224
Index
Mendell, J., 379
Meng, F., 273
Menger, M.D., 83
Mentrikoski, M., 143144, 146147
Metabolic differences in gliomas
brain gliomas, treatment for, 63
energy metabolism and hallmarks
and Akt, 67
cancer cells, 67
glucose 6-phosphate, avaibility, 65
glycolytic pathway, 65
HIF-1, 66
LDH-A, 66
molecular mechanisms, 66
NADPH, 66
PEP, conversion of, 65
phosphatidylinositol 3-kinase (PI3K), 67
ROS, 67
tricarboxylic acid (TCA) cycle, 65
tumor microenvironment/oncogene activation, 6667
heterogeneous nature, tumors of
BCNU, 64
cancer stem cells, 65
histological analyses, 64
intratumoral heterogeneity, 64
LOH, 64
magnetic resonance imaging studies, 64
MGMT, 65
regional heterogeneity, 64
stem cell population, 65
metabolic alterations in, 6466
metabolic reprogramming, 66
tumor altered metabolism, 67
Metabolic evaluation, 358
positron emission tomography (PET), 357
Sherbrooke LabTEP software, 357
Metro, G., 7377
Mettler, F.A.Jr., 6
Meyer, G., 380
Meyer, J., 58
MHC, see Major histocompatibility complex (MHC)
Microscopic tumor, 154
Micro-vascular density (MVD), 97
Miliaras, G., 144, 146147
Millac, P., 379
Miller, D.C., 408
Mismatch repair genes, 4
Mittler, M.A., 12
Miwa, K., 110
Mixed oligoastrocytic (MOA), 25
MLV, see Murine leukaemia virus (MLV) vector
Moehler, T.M., 258
Moffat, B.A., 307
Mohanraj, R., 426
Moi, seeva, E.P., 263264
Molecular biomarkers, 211
anaplastic oligodendroglioma, 212
p53 mutations, 212
Molecular subtypes of gliomas
astrocytic, 25
false-negative GBM, 25
Index
false-positive GBM, 25
genome wide markers
gene expression profiling, 27
mesenchymal, 27
1p19q status and, 28
proliferative, 27
proneural, 27
glioblastoma, 25
incidence rate, 25
KPS, 25
MOA, 25
oligodendrocytic, 25
single markers
IDH1, 26
LOH 1p19q, 26
MGMT, 2627
Moon, K.S., 146
Morabito, A., 256
Mori, H., 3136
Morita, K., 116, 118119
Moserle, L., 270
Moser, M., 316
Mosieniak, G., 242243, 245
Motomura, K., 269275
Mottet, D., 366
Mourad, P.D., 144
MTD, see Mean Tumor Diameter (MTD)
Muir-Torre syndrome, 36
Mujic, A., 143
Mukherjee, D., 173178
Mukherjee, P., 70, 115
Mu, L.J., 319
Mulkern, R.W., 305
Munson, A.E., 280
Muragaki, Y., 185
Murakami, R., 113, 119
Murine leukaemia virus (MLV) vector, 337
Murphy, M., 25
Murphy, S., 256
Mursch, K., 373
Muzi, M., 111
Myeloid-derived suppressor cells (MDSC), 314
N
Nakai, K., 229238
Nakamura, T., 3136
Nalabothula, N., 340
Napekoski, K.M., 415
Nariai, T., 104, 230
Natali, P.G., 124
Nathanson, M., 379
National Cancer Institute of Canada (NCIC) trial, 74
Natsume, A., 269275
Natural killer T (NKT) cells, 314
Nebeling, L.C., 70
Neil, J.J., 305
Nestin glial fibrillary acidic protein, 4
Nestle, U., 132133
Neuron-specific enolase markers, 4
451
New Castles Disease viruses (NDV), 338
Newly diagnosed glioma
BTB, 108
contrast enhancement, 104
endothelial cell damage, 104
FLT-PET study, 110111
fluorothymidine (FLT), 104
L-methyl C-methionine (MET), 104
materials and methods
patients, 104105
PET examinations and data analysis, 105
statistical analysis, 105
WHO grading system, 105
MRI, 103
PET, 104, 108
pilocytic astrocytoma., 110
protein synthesis, 104
radiation, 108
stereotactic radiosurgery, 108109
tracer uptake and tumor grade, 106
brain parenchyma, 107
MET and FLT uptake, 107
proliferation activity, 108
tumor detection of HGG and LLG
MET-PET, 106
PET tracer uptake, 106
Newton, H.B., 143
R
, see Sorafenib
Nexavar
Nghiemphu, P.L., 210
Ng, Y-t, 427
Niclou, S.P., 9495
Nicolay, D.J., 275
Nieder, C., 3
Nielsen, F.H., 234
Nigg, D.W., 237
Ni, H.T., 321
Nimsky, C., 2, 182, 185
Niola, F., 361
Nishiyama, Y., 103111
Nobusawa, S., 56, 58
Nojiri, T., 108109
Non-polypotic colorectal cancer (HNPCC), 36
Non-small cell lung cancer (NSCLC), 132
Norden, A.D., 293294, 296297, 318
Norrby, K., 382
Norris, D.G., 304305
NOTCH signaling pathway, 5
Noushmehr, H., 56
NSCLC, see Non-small cell lung cancer
(NSCLC)
Ntoukas, V., 183
Nurmohamed, M.T., 381
Nutt, C.L., 27
Nuutinen, J., 104
O
Ogawa, T., 108, 110
Ogg, R., 431
Ohgaki, H., 12, 25, 47, 50, 242, 336
Oh, J., 116, 119
Ohno, M., 269275
452
Ojemann, G., 177
Ojemann, J., 177
Okada, H., 324, 326, 329
Okubo, S., 108, 110
Olafsson, E., 398
Oligodendrogliomas (O) (WHO grade II), 47, 56, 207, 210, 420
pathogenesis and progression, 55
Oliver, J., 6371
Olsson, Y., 114, 118
O6 -methyl-guanyl-methyl transferase gene promoter (MGMT),
15, 65
glioblastoma
anaplastic gliomas, prognostic, 76
clinical outcome, 76
depleting strategies, 7677
in grade III anaplastic gliomas, 76
informative role, 7576
inhibitors, 77
methylation, role in, 7475
optimal depletion of, 77
prognostic role, 7576
quantitative assays, 75
recurrent, clinical outcome of, 7576
role of mgmtmethylation, 7576
status assessing, 74
in tumor tissue, 74
Oppitz, U., 216
Ornithine transcarbamylase (OTC), 337
Osborne, R.H., 31
Ostergaard, L., 421
Oudard, S., 67
Oxcarbazepine, 428
P
Pace, A., 405
Packer, R.J., 371, 392
Padhani, A.R., 305306
Padma, M.V., 358
Pallud, J., 13, 153160, 163171, 189
Palmero, E.I., 3536
Palmer, T.D., 3
Pan, D., 341
Paraf, F., 3536
Park, D.M., 36
Park, J.K., 5
Park, J.W., 58
Park, K.Y., 389
Park, S.Y., 58
Parsa, A.T., 350351
Parsons, D.W., 18, 26, 34, 4748, 5457
Pasquier, B., 407
Patronas, N.J., 138
Paulis, Y.W., 99
Paunu, N., 35, 40, 43
PDGFR, see Platelet-derived growth factor receptor (PDGFR)
Pearson, A.D., 32
Peca, C., 217
Pedersen, I.M., 273
Pellegatta, S., 326
Pelloski, C.E., 147, 364
Index
PEP, see Phosphoenolpyruvate (PEP)
Perillo, N.L., 263
Perkins, F., 431
Perry, A., 414415
Perry, J.R., 77, 208, 211, 381
Petersen, G., 280
Peterson, D.L., 349351
Petitjean, A., 32, 42
Peyre, M., 156, 166, 169170
PFS, see Progression-free survival (PFS)
PHD, see Prolylhydroxylase (PHD)
Phenytoin, 403
Phillips, H.S., 27, 56, 147
Phosphoenolpyruvate (PEP), 65
Piccirillo, S., 4
Pichler, B.J., 139
Pichler, J., 139
Pichler, R., 139
Pierpaoli, C., 115
Pignatti, F., 168, 170
Pilocytic astrocytomas, 11, 110
Pim, D., 44
Pirzkall, A., 139
Placenta-like growth factor (PlGF), 291
Planning target volume (PTV), 133, 226
Platelet-derived growth factor receptor (PDGFR), 2, 211
Plate, K.H., 83
Plaza, M., 196
Poduslo, J.F., 356
Pollack, I.F., 376, 394
R
intraoperative OR Setup, 183
PoleStar
Polymerase chain reaction (PCR), 58
Pope, W.B., 175
Poptani, H., 113119, 306, 308
Portwine, C., 35
Posner, J.B., 379
Post-treatment radiation effect (PTRE), 137
Potgens, A.J.G., 97
Poulsen, H.S., 289298
Pouratian, N., 175
Poussaint, T.Y., 393
Pouyssegur, J., 6566, 68
Prayson, R.A., 407415
Prestegarden, L., 4
Presurgical neurocognitive evaluation, 190
Preusser, M., 415, 426427, 429, 432433
Price, S.J., 156157
Primary brain tumors, 39
Primary glioblastomas, 2
Procarbazine chemotherapy, 209
Progression-free survival (PFS), 211, 318
Prolylhydroxylase (PHD), 18
Prominin-1, marker for brain tumor-initiating cells, 4
Provenzale, J.M., 83
Pruitt, A., 216
PTRE, see Post-treatment radiation effect (PTRE)
PTV, see Planning target volume (PTV)
Puchades, M., 261, 265
Purdie, T.G., 83, 85
Pyrzynska, B., 243244
Index
Q
Qaddoumi, I., 388
Quevedo, J.F., 379
Quinn, J.A., 77, 176
Quinones-Hinojosa, A., 173178
R
Raatschen, H.J., 83
Rabinovich, G.A., 261263
Radiation therapy, 392
brain response, 217
late effects
cerebrovascular, 218
histopathological analyses, 217
leukoencephalopathy, 218
white matter injury, 218
Radiation Therapy Oncology Group (RTOG), 216, 327
Radiation therapy (RT), 132
Radiotherapy, 210, 238
Raila, F.A., 351
Rainov, N.G., 344
Raja, A., 375
Rajagopalan, V., 143
Rajapakse, J.C., 125
Rajasekhar, V.K., 249
Rajshekhar, V., 373375
Ramina, R., 185
Ram, Z., 182
Ransohoff, R.M., 316
Rao, R.D., 250
Ratbi, I., 3944
Reactive oxygen species (ROS), 67
Reardon, D.A., 250
Rech, A.J., 329
Recurrent HGG, 289
angiogenesis
high micro vessel density (MVD), 291
tumor vasculature, 291
anti-angiogenic drugs
anti-angiogenic treatment, 293
bevacizumab, 293, 294295
irinotecan, 294
thalidomide and analog lenalidomide, 293294
biomarkers
heterogeneity and insufficiency of tumor vasculature,
296297
immunohistochemical methods, 297
hypoxia
cellular response, 291
HIF-1 overexpression and angiogenesis, 290
necrotic regions of glioblastoma, 290
targeted therapies, 293
temozolomide, 290
treatment for, 289
VEGF pathway inhibitors, resistance to
glioma cells, 297298
traditional concept, 297
VEGF/VEGFR signaling pathway, 291
453
endothelial and tumor cells, 292293
increased vessel permeability, 291
tumor cells, 292
WHO grades, 289
Recursive partitioning analysis (RPA), 327, 382
Rees, J.H., 114, 118
Reiche, W., 118
Reifenberger, G., 14, 47, 355
Reifenberger, J., 14, 47
Reilly, K.M., 362
Reinhardt, M.J., 138
Reirradiation
and chemotherapy, 223
malignant glioma
management approach, 219
salvage option, 219
with molecular targeted agents, 225
occult radiation injury recovery, 218
Reitman, Z.J., 5360
Relative cerebral blood volume (rCBV), 417422
acquisition of
EPI, 418
gadolinium chelate, 418
challenges evaluation
GE EPI technique, 420421
SE EPI technique, 420421
conventional MRI, comparison
astrocytoma, 419
oligodendrogliomas, 420
tumor grade, 418419
differentiating gliomas
HGGs, 420
multimodalilty approach, 421
paediatric tumors, study, 422
Remy, C., 307
Reynolds, B.A., 270
Rezvan, A., 190
Ribom, D., 104
Ricard, D., 1415, 156, 166, 168170
Rickhey, M., 132
Rieber, J., 35
Rivera, A.L., 27
Roberts, D.W., 182
Roberts, H.C., 81, 8384
Roberts, N., 126, 128
Roberts, T.P., 81, 8384
Robins, H.I., 379383
Roca, P., 6371
Rockwell, S., 364
Rollin, N., 419420
Rong, Y., 382
Rorive, S., 263
Roszman, T., 318
Roujeau, T., 390
RPA, see Recursive partitioning analysis (RPA)
RTOG, see Radiation Therapy Oncology Group (RTOG)
Rudin, M., 305
Regg, S.J., 399
Ruff, R.L., 379
Rustin, P., 67
454
Rutka, J.T., 432
Rutkowski, S., 323, 326, 328
Ryvlin, P., 426
S
Saad, A.G., 144, 146
Sabatini, D.M., 65, 249
Sadik, A.R., 144
Saga, T., 104, 110
Saidi, A., 361, 364
Saini, M., 351
Sainte Anne Hospital classifications and gliomagenesis
astrocytic, 12
LOH 1p/19q, 12
Saito, R., 4751, 326
Saka, E., 372
Salazar, M., 280, 282283, 285
Salmaggi, A., 271, 372373, 375
Salvatori, M., 6
Samaranayake, H., 335346
Sameshima, Y., 3233
Sanai, N., 13, 18, 154, 160, 174175, 177, 181,
192194, 202
Sanchez, C., 278281, 284285
Sandberg-Wollheim, M., 175
Sandmair, A.M., 2, 344
Sanson, M., 50, 5658
Santandreu, F.M., 6371
Santhosh, K., 118
Santibanez-Koref, M.F., 32
Santos, A.V., 144145
Sarfaraz, S., 280
Sathornsumetee, S., 292, 297
Sato, N., 108
Savill, J., 302
Savitsky, J.P., 379
Sawaya, R., 63, 175, 379
Schadendorf, D., 321
Schafer, U., 223
Schaller, B., 399
Scheithauer, B.W., 35, 415
Scherer, A., 258
Schickel, R., 272
Schiffbauer, H., 178
Schiff, D., 251
Schley, M., 278279
Schmidt, F., 211
Schneider, J.P., 185
Schneider, T., 330
Schonekaes, K., 223
Schulder, M., 182183, 185
Schultheiss, T.E., 217218
Schultz, S., 144, 146148
Schumacher, M., 390
Schwartzbaum, J.A., 31
Schwartz, J.L., 111
Schwarzbraun, T., 35
Index
Schwer, A.L., 224226
Scott, C.B., 25
Secondary glioblastomas, 2
Seftor, E.A.M.P., 98
Seftor, R.E., 98
Segall, G.M., 88
Seifert, V., 185
Seitz, R.J., 355
Selch, M.T., 221
Selivanoc, V., 357
Selvapandian, S., 371372, 388
Semenza, G.L., 67, 290291
Senetta, R., 143148
Senft, C., 181186
Senthamizhchelvan, S., 131140
Seyfried, T.N., 70
Shad, A., 374
Shakur, S.F., 426427
Shapiro, J.R., 64
Shapiro, W.R., 64
Shenoy, S.N., 375
Shen, S., 126128
Shepherd, F.A., 319321
Shibahara, I., 4751
Shiels, A.F., 104
Shin, J.H., 419
Shinmura, K., 3136
Shirakawa, K., 95, 97
Shlebak, A.A., 379
Short, S.C., 256
Shweiki, D., 83
Sielska, M., 241252
Sierra, A., 308
Signorelli, F., 178
Signorelli, C.D., 178
Simos, P.G., 177
Singh, S.K., 65, 100, 270, 317
Single molecular markers in gliomas
IDH1, 26
LOH 1p19q, 26
MGMT, 2627
Siragusa, S.C.B., 382
Sjoblom, T., 54, 60
Skin metastases
CNS tumours, 143
diagnostic and therapeutic approaches
FNAc, 147
patient neoplastic history, 147148
treatment of, 148
extra-axial tumor spread, 143
factors, 144
facts and hypotheses
GFAP and EGFR, 147
histopathology and immunophenotype, 147
hypothetical tumour, 144
IHC, 147
intracranial cases, 145146
intracranial disease, 144
mesenchymal and angiogenesis, 147
mesenchymal cell clones, 147
temozolomide, 144, 147
Index
GBM, 143
lung and pleura, 143
Skld, K.-H., 234
Slack, F.J., 272
Sliwa, M., 247248, 251
SMA, see Supplementary motor area (SMA)
Sminia, P., 217, 220
Smith, D.B., 379
Smith, J.S., 175, 190
Smith, M.A., 388
Smits, A., 397405
Soffietti, R., 113, 154, 163, 165166, 168, 170
Solid tumors, 82, 93
Solid tumor tissue, 154
Sonoda, Y., 4751, 54
Sonveaux, P., 68
Sorafenib, 293
Sorg, R.V., 319
Soroceanu, L., 124125
Spencer, S.S., 425
Sperling, M.R., 426
Squires, L.A., 390
SRT/SRS, see Stereotactic radiation therapy/surgery (SRT/SRS)
Standardized uptake value (SUV), 139
Stark, A.M., 388, 393
Stark-Vance, V., 294
Stegman, L.D., 306
Stejskal, E.O., 301
Stella, N., 278
Stenning, S.P., 209
Stereotactic radiation therapy/surgery (SRT/SRS), 132
Stereotactic radiosurgery, 108109
Stereotactic radiotherapy with systemic therapies
chemotherapy, 222
cisplatin (CDDP), 222
concurrent chemoradiation, 222
EGFR, 224
malignant gliomas, 222, 224
molecular targeted, 224
planning process, 226
PTV, 226
temozolomide, 222
VEGF-receptor blocking, 224
Stewart, L.A., 209
Stillman, B.N., 263
St. Lawrence, K.S., 84
Stojiljkovic, M., 350351
Storstein, A., 397405
Stratmann, A., 339
Strieth, S., 361
Strik, H.M., 264
Strojnik, T., 248
Struys, E.A., 50
Stummer, W., 2, 185186
Stupp, R., 2, 16, 7475, 174, 176, 208209, 222, 229, 255,
261262, 265, 289, 295, 336, 382
Sudden unexplained death in epilepsy (SUPED), 426
Sugahara, T., 420421
Sugimura, H., 3136
Sun, B., 97
Sunitinib, 293
455
Sun, T., 97
Sun, X., 133, 136
Supplementary motor area (SMA), 195
Surgery for glioma
intraoperative magnetic resonance imaging (iMRI), 182183
development, 186
influence on, 185
safety concerns in, 183, 185
left temporal glioblast, 184
low field magnetic resonance imaging (iMRI)
example of, 182183
neuronavigation and neurophysiological monitoring,
181182
R
intraoperative, 183184
PoleStar
rationale of, 181
surgical workflow, 183
R
, see Sunitinib
Sutent
SUV, see Standardized uptake value (SUV)
Su, Y.B., 329
Swaroop, G.R., 376
Swinson, B.M., 375
Szelenyi, A., 183
Szikla, G., 154, 157158, 165
T
TAA, see Tumor-associated antigens (TAA)
Taal, W., 76
Tabernero, J., 292
Tachibana, I., 40, 43
Taillandier, L., 405
Takahashi, J.A., 256
Tamber, M.S., 11
Tamiya, T., 103111
Tanaka, M., 230
Tanaka, S., 373, 376
Tandon, P.N., 425, 432
Tanner, J.E., 301, 303
Taphoorn, M.J., 403
Tate, M.C., 252
Tchirkov, A., 366
Teicher, B.A., 382
Teixidor, P., 190, 202
Telozolomide, 2
R
, see Temozolomide
Temodal
Temozolomide, 209210, 265266, 290
clinical experience
phase II study, 257
mielotoxicity
(dMRI), 258
phase III study, 255256
thalidomide and cisplatin, 255256
phase II study, 258
venous grade 3-4 toxicity, 258
Terando, A.M., 319321
Ternier, J., 388
Thakur, S.B., 392
456
Thalidomide
cisplatin and temozolomide, 256
clinical experience
phase II study, 257
mielotoxicity
(dMRI), 258
phase II study, 258
venous grade 34 toxicity, 258
Thiebaut de Schotten, M., 196, 198
Thijssen, V.L., 263265
Thomas, D.G., 373375
Thompson, T.P., 124
Thorsen, F., 351
Three-dimensional conformal radiation therapy (3D-CRT), 132
Three-step natural history of LGG
initial silent period
radiological tumour growth, 164
tumor natural course, 164
malignant progression
brain plasticity mechanisms, 166
heterogeneous group, 165
radiological growth pattern, 166
step-by-step approach, 164
symptomatic period of
mild neurological disorders, 165
MRI, 165
Thromboembolic disease in glioma patients
LMWH anti-neoplastic effects
dalteparin, 380382
DVT, 379383
EGFR and RPA, 382
enoxaparin, 380382
glioblastoma, 379383
heparin-induced thrombocytopenia, 380
RTOG/ECOG, role, 380
vitamin K antagonist, 380
VTE, 379383
prophylaxis for nadroparin, 381
Thrombospondin-1 (TSP-1), 34
Thurner, B., 319
TIA, see Transient ischemic attack (TIA)
TIC, see Tumor initiating cells (TIC)
Tmr, J., 95
Tissue inhibitor of metalloproteinase-3 (TIMP-3), 340, 342
Tofts, P.S., 84
Toh, C.H., 116, 118
Toker, A., 242
Tominaga, T., 4751
Torcuator, R., 373, 393
Tortosa, A., 181
Tth, J., 95
TP53 tumor suppressor gene, mutations in, 2
Transforming growth factor- (TGF-), 340
Index
Transient ischemic attack (TIA), 401402
Treatment
fordiffuse intrinsic/diffusely infiltrative low grade
chemotherapy, 373
neurosurgery, 372373
radiotherapy, 373
steroids and gastrostomy, 373
for malignant brainstem gliomas
chemotherapy, 375
radiotherapy, 375
surgery, 374375
Trdan, O., 265
Tronnier, V.M., 182185
Tropine, A., 118
TSC, see Tumors stem cells (TSC)
Tsien, C.I., 136, 139140
Tsuboi, M., 3136
Tsuchiya, K., 113, 116, 118
Tsuyuguchi, N., 104, 108109
Tuettenberg, J., 211
Tu, H., 368
Tumor-associated antigens (TAA), 313
Tumor initiating cells (TIC), 16
Tumors
altered metabolism, 67
classification and diffusion tensor imaging (DTI), 113, 115
apparent diffusion coefficient (ADC) correlation
with, 116
FA, role of, 116118
metrics for, 119
shape based, 118
tumor infiltration assessing in, 118119
infiltration index, 118119
macrophages, role in, 45
type distribution
gliomas, 54
IDH mutations, 54
secondary glioblastomas, 54
WHO grades classification, 54
vascularization, 1
vasculature, 291
Tumors stem cells (TSC), 16
Turcot syndrome (TS), 32, 36, 41
Tuxhorn, I., 427
Tuyaerts, S., 319, 327
Tyburczy, M., 241252
Type I interferons (IFNs), 269
GSC
CD133+ cells and, 270271
chemotherapy resistance genes, 271
glioblastoma multiforme (GBM), 270
nestin, 270
tumor culture-maintaining cells, 270
miR-21 overexpression in glioma-initiating cells, 274
regulation of microRNAs in GSCs
/, 273
biological processes, 272
interference, 272273
quantitative real timepolymerase chain reaction
(qRT-PCR), 273
Index
STAT3 and miR-21, 275
terminal differentiation of GSC
STAT3 activation, 271
tyrosine residue of STAT3, 272
thyroid hormones (THs), 275
Tzika, A.A., 422
U
Uematsu, H., 84
Ueoka, D.I., 389
Ullrich, R.T., 104
UNBS1450 drug, 3
Upadhyay, N., 417422
Ushio, Y., 175
V
Vahteristo, P., 42
Vajkoczy, P., 83
Vajtai, I., 415
Valenzano, K.J., 278
Valonen, P.K., 307
Valproate, 404
Valtonen, S., 175
Van Breemen, M.S., 398, 404
Van Breemen, M.S.M., 425
Van Calenbergh, F., 313330
Van den Bent, M., 26
Van den Bent, M.J., 76, 211, 290, 405
VanderSpek, L., 215226
Van Dongen, C.J., 380
Van Eekelen, M., 3
Van Gool, S.W., 313330
Van Laere, K., 104, 108, 110
Van Meir, E., 366
Van Meyel, D.J., 43
Van Poppel, H., 319321
Van Sickle, M.D., 278
Van Westen, D., 119
Van Zijl, P.C., 305
Varlet, P., 12
Varley, J.M., 34
Vascular endothelial growth factor receptors
(VEGFR), 211
Vascular endothelial growth factor (VEGF), 291, 361368
elevated expression of, 292
pathway inhibitors, resistance to HGG
glioma cells, 297298
traditional concept, 297
VEGF/VEGFR signaling pathway, 291
tumor cells, 292
tumor induced release, 292
VEGF gene, 292
VEGF/VEGFR signaling pathway
endothelial and tumor cells, 292293
increased vessel permeability, 291
Vasculogenesis, 95
Vasculogenic mimicry (VM) in glioma, 94
advances and challenges, 99
457
biological and clinical significance
BBB ultra-structure observation, 97
clinical prognosis, 97
extra cellular matrix, 97
Kaplan-Meier survival analysis, 98
microcirculation of, 9798
mouse xenograft glioma model, 96
PAS-positive patterned networks, 9697
retrospective analysis, 97
clinical therapeutic value
COL-3, 99
EphA2 gene, 99
Folkman report, 99
treatment modalities, 99
tyrosine kinase activity, 99
concept and characteristic
CD34 staining, 95
endothelial cell (EC) phenotype, 95
endothelial-lined vasculature and, 9596
GSC, 95
mimicking, 95
mosaic vessels, 95
PAS-positive channels, 95
RBCs, 95
vessel-like networks, 96
extracellular matrix (ECM)-rich, 93
human glioblastoma tissues, 94
molecular mechanisms
C81-61 cells, 98
galectin-3 (Gal-3), role in, 99
signaling molecules, 98
studies, 99
VE-cadherin, 99
solid tumors, 93
vasculogenesis, 95
VDE, see Velocity of Diametric Expansion (VDE)
Vectors producing cells (VPCs), 339
Vees, H., 139
VEGFR, see Vascular endothelial growth factor receptors
(VEGFR)
Velasco, G., 281
Velocity of Diametric Expansion (VDE), 163
Veninga, T., 221
Venous thromboembolic events (VTE), 379383
Verhaak, R.G.W., 1920, 27, 34, 56
Verhoeff, J.J., 318
Vescovi, A.L., 19
Vigneau, M., 190, 195, 201
Vincristine chemotherapy, 209
Vital, A., 34, 42
Vlassenbroeck, I., 74
Vonderheide, R.H., 329
Vordermark, D., 221
von Hippel-Lindau (VHL) syndrome, 41
VPCs, see Vectors producing cells (VPCs)
Vredenburgh, J.J., 93, 294296
W
Wachsberger, P., 382
Wager, M., 1122
Wakabayashi, T., 269275
458
Waldman, A.D., 417422
Walenta, S., 68
Walker, C., 212
Walker, D.A., 387, 389
Walker, D.G., 324
Walker, M.D., 208
Wallace, C.J., 143145
Wang, J., 5
Wang, M., 415, 426427, 432
Wang, S., 113119
Wang, W., 118
Warburg effect, 51
Ward, P.S., 54, 5960
Warnick, R.E., 211
Wasser, K., 258
Watanabe, M., 157
Watanabe, T., 18, 34, 4850, 5557, 5960
Watters, J.J., 247, 251
Weber, D.C., 134135
Weber, M.A., 422
Wechsler, W., 350
Weidner, N., 82
Wei, L.H., 249250
Weinberg, R.A., 361
Weissenberger, J., 366
Weissleder, R., 131
Weiss, S., 270
Weller, M., 16, 7374, 7677
Wen, P.Y., 211, 222, 224
Wesolowska, A., 247
Wessels, P.H., 164
Westin, C.F., 115, 118
Westphal, M., 113
Wheeler, C.J., 322, 324, 327329
Wheless, J.W., 428
White, H.H., 371
White, M.L., 420
Whittle, I.R., 356, 376, 399, 425
Wick, A., 7677
Wicki, A., 368
Wick, W., 4950, 58, 76, 77, 210
Willems, P.W., 191
Willett, C.G., 361
Wilson, T.A., 84
Wirth, T., 303, 335346
Wirtz, C.R., 182
Wiser, H.G., 427
Wittig, A., 232
Woermann, F.G., 426, 428429
Wolf, H.K., 408
Wong, E.T., 290, 294
World Health Organization (WHO)
classification of gliomagenesis, 11
astrocytic, 12
LOH 1p/19q, 12
grading system, 105
Wree, A., 356
Index
Wu, J.S., 178
Wurm, R.E., 223
X
Xu, X., 4748
Xu, Y., 45
Xu, Z., 4748
Y
Yaacov, R.L., 217218
Yamada, H., 3136
Yamamoto, T., 6, 229238
Yamamura, Y., 3136
Yamanaka, R., 317318, 322, 324, 327329
Yamane, T., 138
Yamaoka, K., 263
Yamasaki, F., 113, 116
Yamini, B., 340
Yang, D., 419
Yang, I., 251
Yang, W., 235
Yan, H., 34, 4750, 5360, 212
Yankeelov, T.E., 84
Yap, J.T., 132
Yasuhara, T., 143
Yip, S., 73, 76
Yl-Herttuala, S., 335346
Yokota, N., 3136
Yoshida, F., 234235
Young Poussaint, T., 393394
Yousry, I., 375
Yousry, T.A., 375
Yuan, X., 100
Yue, W.Y., 9495, 97
Yu, J.S., 322323, 328
Yuki, K., 269275
Yung, W.K., 77, 256
Z
Zaidi, H., 131140
Zamecnik, J., 116, 118
Zangari, M., 258
Zentner, J., 407
Zha, J., 281
Zhang, D., 98
Zhang, J., 303
Zhang, M., 114, 118
Zhang, S., 98, 115, 118
Zhang, W., 98
Zhao, M., 305
Zhao, S., 18, 50, 58
Zhou, J., 265
Zhou, S., 67
Zhou, X., 21
Zhu, S., 273
Zinc-finger proteins (ZFPs), 339
Zitvogel, L., 314
Zonari, P., 419, 422
Zupanska, A., 245247
Zustovich, F., 255259

Tumors of CNS v2 Gliomas Glioblastoma Part 2

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tumors of CNS v2 Gliomas Glioblastoma Part 2

Uploaded by

Copyright:

Available Formats

Tumors of the Central Nervous System

Tumors of the Central Nervous System

For other titles published in this series, go to

Tumors of the Central Nervous

Tumors of the Central

Although touched by technology, surgical pathology always has

Biomarkers and Diagnosis . . . . . . . . . . . . . . . . . . . .

2 Gliomagenesis: Advantages and Limitations of Biomarkers . . . . .

3 Molecular Subtypes of Gliomas . . . . . . . . . . . . . . . . . . . . .

4 Glioblastoma: Germline Mutation of TP53 . . . . . . . . . . . . . . .

5 Familial Gliomas: Role of TP53 Gene . . . . . . . . . . . . . . . . . .

6 The Role of IDH1 and IDH2 Mutations in Malignant Gliomas . . . .

7 Malignant Glioma: Isocitrate Dehydrogenases 1 and 2 Mutations . .

8 Metabolic Differences in Different Regions of Glioma Samples . . . .

Glioblastoma Patients: Role of Methylated MGMT . . . . . . . . . .

10 Brain Tumor Angiogenesis and Glioma Grading:

Vasculogenic Mimicry in Glioma . . . . . . . . . . . . . . . . . . . .

Newly Diagnosed Glioma: Diagnosis Using Positron

Role of Diffusion Tensor Imaging in Differentiation

SPECT imaging of Human Glioma . . . . . . . . . . . .

Assessment of Biological Target Volume Using Positron

Skin Metastases of Glioblastoma . . . . . . . . . . . . . . . . . . . .

Diffuse Low-Grade Gliomas: What Does Complete

Quantitative Approach of the Natural Course of Diffuse

Impact of Extent of Resection on Outcomes in Patients

Glioma Surgery: Intraoperative Low Field Magnetic

Low-Grade Gliomas: Intraoperative Electrical Stimulations . . . . .

Malignant Gliomas: Present and Future Therapeutic Drugs . . . . .

Recurrent Malignant Glioma Patients: Treatment

Glioblastoma: Boron Neutron Capture Therapy . . . . . . . . . . . .

Glioblastoma: Anti-tumor Action of Cyclosporin A

26 Glioblastoma Patients: Chemotherapy with Cisplatin,

27 Glioblastoma: Role of Galectin-1 in Chemoresistance . . . . . . . . .

28 Glioma-Initiating Cells: Interferon Treatment . . . . . . . . . . . . .

29 Glioblastoma: Anti-tumor Action of Natural and Synthetic

30 Patients with Recurrent High-Grade Glioma: Therapy

31 Monitoring Gliomas In Vivo Using Diffusion-Weighted

32 High-Grade Gliomas: Dendritic Cell Therapy . . . . . . . . . . . . .

33 Glioblastoma Multiforme: Use of Adenoviral Vectors . . . . . . . . .

34 Fischer/F98 Glioma Model: Methodology . . . . . . . . . . . . . . .

35 Cellular and Molecular Characterization of Anti-VEGF

37 The Use of Low Molecular Weight Heparin in the Treatment

38 Brainstem Gliomas: An Overview . . . . . . . . . . . . . . . . . . . .

39 Tumor-Associated Epilepsy in Patients with Glioma . . . . . . . . .

40 Brain Tumors Arising in the Setting of Chronic Epilepsy . . . . . . .

Low-Grade Gliomas: Role of Relative Cerebral Blood

Angiocentric Glioma-Induced Seizures: Lesionectomy . . . . . . . .

Glioma Cell Lines: Role of Cancer Stem Cells

Glioblastoma Cancer Stem Cells: Response to Epidermal

Low- and High-Grade Gliomas: Extensive Surgical Resection

Brainstem Gangliogliomas: Total Resection and Close

Glioblastoma: Temozolomide-Based Chemotherapy

Drug-Resistant Glioma: Treatment with Imatinib Mesylate

Glioblastoma Multiforme: Molecular Basis of Resistance to Erlotinib

Enhanced Glioma Chemosensitivity

Malignant Glioma Patients: Anti-Vascular Endothelial

Aggravating Endoplasmic Reticulum Stress by Combined

Targeted Therapy for Malignant Gliomas

Glioblastomas: HER1/EGFR-Targeted Therapeutics

Epidermal Growth Factor Receptor Inhibition as a

Role of Acyl-CoA Synthetases in Glioma Cell Survival and

Malignant Glioma Patients: Combined Treatment with

Malignant Glioma Immunotherapy: A Peptide Vaccine from

Malignant Glioma: Chemovirotherapy