Cognition and Cancer

This page intentionally left blank Cognition and Cancer

Edited by Christina A. Meyers Professor and Chief Section of Neuropsychology Department of Neuro-Oncology The University of Texas M. D. Anderson Cancer Center Houston, Texas, USA

James R. Perry Tony Crolla Chair in Brain Tumour Research Head, Division of Neurology Sunnybrook Health Sciences Centre and Odette Cancer Centre University of Toronto Toronto, Ontario, Canada CAMBRIDGE UNIVERSITY PRESS Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo

Cambridge University Press The Edinburgh Building, Cambridge CB2 8RU, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521854825

This publication is in copyright. Subject to statutory exception and to the provision of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published in print format 2008

ISBN-13 978-0-511-45573-5 eBook (EBL)

ISBN-13 978-0-521-85482-5 hardback

Cambridge University Press has no responsibility for the persistence or accuracy of urls for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate. Every effort has been made in preparing this publication to provide accurate and up-to-date information which is in accord with accepted standards and practice at the time of publication. Although case histories are drawn from actual cases, every effort has been made to disguise the identities of the individuals involved. Nevertheless, the authors, editors and publishers can make no warranties that the information contained herein is totally free from error, not least because clinical standards are constantly changing through research and regulation. The authors, editors and publishers therefore disclaim all liability for direct or consequential damages resulting from the use of material contained in this publication. Readers are strongly advised to pay careful attention to information provided by the manufacturer of any drugs or equipment that they plan to use. Contents

List of contributors page vii Preface xi

Section 1: Cognition and the brain: measurement, tools, and interpretation

1 Introduction3 Christina A. Meyers and James R. Perry

2 Clinical neuropsychology6 Jill B. Rich and Angela K. Troyer

3 Brain imaging investigation of chemotherapy-induced neurocognitive changes 19 Brenna C. McDonald, Andrew J. Saykin, and Tim A. Ahles

4 Role of neuropsychological assessment in cancer patients 33 Elana Farace

5 Neuropsychological assessment of adults with cancer 44 Anne E. Kayl, Robert Collins, and Jeffrey S. Wefel

6 Neuropsychological assessment of children with cancer 56 Louise Penkman Fennell and Robert W. Butler

v vi Contents

Section 2: Effects of cancer and cancer 17 Paraneoplastic disorders 239 treatment on cognition Edward Dropcho

7 Biological bases of radiation injury to 18 Symptomatic therapies and the brain 83 supportive care issues 258 Edward G. Shaw and Mike E. Robbins Alan Valentine and Eduardo Bruera

8 Cognitive dysfunction related to 19 Animal models and cancer-related chemotherapy and biological symptoms 270 response modiﬁers 97 Adrian Dunn Jeffrey S. Wefel, Robert Collins, and Anne E. Kayl

9 Effect of hormones and hormonal Section 3: Interventions and treatment on cognition 115 implications for clinical trials Christien Schilder, Sanne Schagen, and Frits van Dam 20 Behavioral strategies and rehabilitation 281 10 Low-grade gliomas 142 Dona E. C. Locke, Jane H. Cerhan, and Martin J. B. Taphoorn and Charles G. Niel¨ James F. Malec 11 High-grade gliomas 156 Michael J. Glantz and John W. Conlee 21 Support services 295 Bebe Guill and Renee H. Raynor 12 Brain metastases 170 Deepak Khuntia, Beela S. Mathew, Christina 22 Pharmacological interventions for A. Meyers, Sterling Johnson, and the treatment of radiation-induced Minesh P. Mehta brain injury 312 Edward G. Shaw, Jerome Butler, L. Douglas 13 Primary central nervous system Case,Ralphd’Agostino,Jr.,JohnGleason,Jr., lymphoma 187 Edward Ip, Mike E. Robbins, Paul Saconn, Denise D. Correa and Stephen R. Rapp 14 Childhood brain tumors 198 H. Stacy Nicholson, Louise Penkman Fennell, 23 Neurocognitive testing in clinical trials 320 and Robert W. Butler Jennifer A. Smith and Jeffrey S. Wefel

15 Neuroﬁbromatosis 211 Index 329 Bartlett D. Moore, III and John M. Slopis

16 Hematological malignancies 228 Color plate section will be found between Melissa Friedman and Mercedes Fernandez page 32 and 33. Contributors

Ralph d’Agostino, Jr., Ph.D. Professor, Department of Biostatistical Sciences Division of Public Health Sciences Wake Forest University School of Medicine Winston-Salem, North Carolina, USA

Tim A. Ahles, Ph.D. Department of Psychiatry and Behavioral Sciences Director, Neurocognitive Research Laboratory Memorial Sloan Kettering Cancer Center New York, New York, USA

Eduardo Bruera, M.D. Professor and Chair Department of Palliative Care & Rehabilitation Medicine The University of Texas M. D. Anderson Cancer Center Houston, Texas, USA

Jerome Butler, M.D. Department of Radiation Oncology Wake Forest University School of Medicine Winston-Salem, North Carolina, USA

Robert W. Butler, Ph.D., ABPP-Cn Department of Pediatrics Hematology/Oncology Oregon Health & Science University Portland, Oregon, USA

vii viii List of contributors

L. Douglas Case, Ph.D. Director of Clinical Research, Department of Professor, Department of Biostatistical Sciences Neurosurgery Division of Public Health Sciences Program Co-Leader, Cancer Prevention and Control, Wake Forest University School of Medicine Penn State Cancer Institute Winston-Salem, North Carolina, USA Pennsylvania State University Hershey, Pennsylvania, USA Jane H. Cerhan, Ph.D., ABPP-Cn Department of Psychiatry and Psychology Mercedes Fernandez, Ph.D. Mayo Clinic College of Medicine Department of Psychology Rochester, Minnesota, USA Carlos Albizu University Miami, Florida, USA Robert Collins, Ph.D. Melissa Friedman, Ph.D. Clinical Neuropsychologist Division of Psychiatry Neurology Care Line Department of Medicine Michael E. DeBakey Veterans Affairs Medical Center Mount Sinai Medical Center Houston, Texas, USA Miami Beach, Florida, USA

John W. Conlee, Ph.D. Michael J. Glantz, M.D. Department of Oncology Professor of Oncology, Neurosurgery and Neurology Huntsman Cancer Institute Huntsman Cancer Institute University of Utah School University of Utah School of Medicine Salt Lake City, Utah, USA Salt Lake City, Utah, USA

Denise D. Correa, PhD, ABPP-Cn John Gleason, Jr., B.A. Department of Neurology Wake Forest University School of Medicine Memorial Sloan Kettering Cancer Center Winston-Salem, North Carolina, USA New York, New York, USA Bebe Guill, M.Div. Director, Survivorship Programs & Services Edward Dropcho, M.D. The Preston Robert Tisch Brain Tumor Center Professor Duke University Medical Center Department of Neurology Durham, North Carolina, USA Indiana University Medical Center Indianapolis, Indiana, USA Edward Ip, Ph.D. Professor Adrian Dunn, Ph.D. Departments of Biostatistical Sciences and Social Professor Sciences and Health Policy Department of Psychology Division of Public Health Sciences Paciﬁc Biosciences Research Center Wake Forest University School of Medicine University of Hawaii at Manoa¯ Winston-Salem, North Carolina, USA Honolulu, Hawaii, USA Sterling Johnson, Ph.D. Elana Farace, Ph.D. Associate Professor of Medicine Associate Professor of Neurosurgery and Public Health University of Wisconsin Sciences Madison, Wisconsin, USA List of contributors ix

Anne E. Kayl, Ph.D. Christina A. Meyers, Ph.D., ABPP-Cn Assistant Professor Professor and Chief Section of Neuropsychology Section of Neuropsychology Department of Neuro-Oncology Department of Neuro-Oncology The University of Texas M. D. Anderson Cancer Center The University of Texas M. D. Anderson Cancer Center Houston, Texas, USA Houston, Texas, USA

Deepak Khuntia, M.D. Bartlett D. Moore, III, Ph.D. Professor and Chief Assistant Professor Section of Behavioral Pediatrics Department of Human Oncology Division of Pediatrics Radiation Oncology University of Wisconsin The University of Texas M. D. Anderson Cancer Center Houston, Texas, USA Madison, Wisconsin, USA

H. Stacy Nicholson, M.D., MPH Dona E. C. Locke, Ph.D., ABPP-Cn Professor of Pediatrics Mayo Clinic Arizona Chief, Division of Pediatric Hematology/Oncology Division of Psychology Oregon Health & Science University Scottsdale, Arizona, USA Portland, Oregon, USA

James F. Malec, PhD, ABPP-Cn,Rp Charles G. Niel,¨ M.D. Professor and Co-chair Reinier de Graaf Groep Division of Tertiary Psychiatry and Psychology Department of Radiation Oncology Department of Psychiatry and Psychology Delft, The Netherlands Mayo Clinic College of Medicine Rochester, Minnesota, USA Louise Penkman Fennell, Ph.D., ABPP-Cn Assistant Professor and Director of Internship Training Argosy University, Hawaii Campus Beela S. Mathew, M.D. Consulting Neuropsychologist Associate Professor Kapiolani Behavioral Health Specialists Regional Cancer Centre Kapiolani Women and Children’s Hospital Trivandrum, India Honolulu, Hawaii, USA

Brenna C. McDonald, Psy.D., M.B.A. James R. Perry M.D., FRCPC Assistant Professor of Radiology and Neurology Tony Crolla Chair in Brain Tumour Research IU Center for Neuroimaging Head, Division of Neurology Division of Imaging Sciences Sunnybrook Health Sciences Centre and Odette Department of Radiology Cancer Centre Indiana University School of Medicine University of Toronto Indianapolis, Indiana, USA Toronto, Ontario, Canada

Minesh P. Mehta, M.D. Stephen R. Rapp, Ph.D. Chairman, Department of Human Oncology Professor and Head, Section of Psychology Radiation Oncology Department of Psychiatry and Behavioral Medicine University of Wisconsin Wake Forest University School of Medicine Madison, Wisconsin, USA Winston-Salem, North Carolina, USA x List of contributors

Renee H. Raynor, Ph.D. Edward G. Shaw, M.D. Clinical Neuropsychologist Professor and Chairman Director of Clinical Operations Department of Radiation Oncology The Preston Robert Tisch Brain Tumor Center Wake Forest University School of Medicine Duke University Medical Center Winston-Salem, North Carolina, USA Durham, North Carolina, USA John M. Slopis, M.D., M.P.H. Associate Professor of Child Neurology and Cancer Jill B. Rich, Ph.D. Genetics Department of Psychology Department of Neuro-Oncology and Division of York University & Baycrest Centre for Geriatric Care Pediatrics Toronto, Ontario, Canada The University of Texas M. D. Anderson Cancer Center Houston, Texas, USA Mike E. Robbins, Ph.D. Professor and Head, Section of Radiation Biology Jennifer A. Smith, Ph.D. Department of Radiation Oncology Biostatistics Wake Forest University School of Medicine Geron Corporation Winston-Salem, North Carolina, USA Menlo Park, California, USA Martin J. B. Taphoorn, M.D. Paul Saconn, M.D. Medical Center Haaglanden Department of Radiation Oncology Department of Neurology/Neuro-Oncology Wake Forest University School of Medicine The Hague, The Netherlands Winston-Salem, North Carolina, USA Angela K. Troyer, Ph.D. Department of Psychology Andrew J. Saykin, PsyD, ABPP-Cn Baycrest Centre for Geriatric Care Professor of Radiology Toronto, Ontario, Canada Director, IU Center for Neuroimaging Division of Imaging Sciences Alan Valentine, M.D. Department of Radiology Professor Indiana University School of Medicine Department of Psychiatry Indianapolis, Indiana, USA University of Texas M. D. Anderson Cancer Center Houston, Texas, USA Sanne Schagen, Ph.D. Frits van Dam, Ph.D. Department of Psychosocial Research and Department of Psychosocial Research and Epidemiology Epidemiology Netherlands Cancer Institute/Antoni van Netherlands Cancer Institute/Antoni van Leeuwenhoek Hospital Leeuwenhoek hospital Amsterdam, The Netherlands Amsterdam, The Netherlands

Christien Schilder, M.Sc. Jeffrey S. Wefel, Ph.D. Department of Psychosocial Research and Assistant Professor Epidemiology Section of Neuropsychology Netherlands Cancer Institute/Antoni van Department of Neuro-Oncology Leeuwenhoek Hospital The University of Texas M.D. Anderson Cancer Center Amsterdam, The Netherlands Houston, Texas, USA Preface

This volume is different from anything that has been published in the fields of oncology and neurosciences. The study of cognitive function in cancer patients is in its infancy, and far behind the research in other diseases. However, cognitive impairment and other adverse symptoms associated with cancer are becoming increasingly important to patients and are identified as a major source of concern for survivors. To date there is no comprehensive text that brings together the basic research and clinical perspectives of the many disciplines involved in understanding the impact of cancer and cancer treatment on brain function. Thus, we felt there was a growing need to address cognitive function across cancers and treatments as a resource for oncologists and other professionals who treat cancer patients and those who are involved in transla- tional research that has an impact on cancer- related symptoms. We are pleased that we have brought together the research and views of the most prominent professionals in the field of cognition and cancer. The book is intended to be acces- sible to a diverse audience: research and clinical neuropsychologists, neuroscientists, medical and neuro-oncologists, surgeons, radiation oncologists, palliative care health teams, nurses, nurse practitioners and physician assistants, and postgraduate trainees and fellows in these disciplines. We would like to acknowledge the contributions made by our colleagues and our patients from whom we learn daily. We also appreciate the help of Lori Bernstein, Ph.D, who was involved in the conceptual development of the book in its early

xi xii Preface

stages, as well as the tireless support and enthusi- their patience and unselﬁsh support of projects that asm of Betty Fulford and Laura Wood of Cambridge take us away from them. University Press. We would also like to acknowl- As the book cover implies, solving the disorders edge the leadership of Sunnybrook Health Sciences of the brain and mind in cancer patients is multi- Centre and the University of Toronto for having faceted, challenging and at times frustrating, but, the foresight and wisdom to embrace an inter- in the end, solvable. We hope that this text serves disciplinary approach to neurosciences, including to enhance the quality of life of cancer patients cross-programmatic collaboration, which this book and stimulates awareness, research, and knowledge embodies. We also acknowledge our families for transfer in the area of cognition and cancer. SECTION 1

Cognition and the brain: measurement, tools, and interpretation

Introduction

Christina A. Meyers and James R. Perry

Cancer patients experience a number of adverse to learn less information, and learning may be less symptoms, including cognitive impairment, fatigue, efficient), and inefficient memory retrieval (e.g., pain, sleep disturbance, and others often in com- spontaneous recall may be somewhat spotty). How- bination rather than alone. Fortunately detailed ever, the ability to consolidate or store new infor- symptom assessment is becoming increasingly rec- mation is generally intact, so that the memory ognized as a part of routine patient care by physi- disturbance observed in cancer patients is vastly cians, allied health care providers, and accrediting different from that observed in neurodegenera- agencies. Cancer treatment may only be consid- tive disorders such as Alzheimer’s disease, and is ered successful if these symptoms are managed, but often subtle and relative to the individual’s pre- successful management is hampered by insufficient illness level of function. Additional common symp- knowledge of mechanisms. toms include periodic lapses of attention, dis- Cognitive dysfunction occurs in the majority of tractibility, and slowed cognitive processing speed. cancer patients on active therapy, and is not infre- In general, reasoning and intellectual functions quently a symptom that heralds the diagnosis. In are not affected, but patients often have difficulty addition, it persists in a substantial number of performing their normal work due to cognitive patients long after treatment is discontinued. In inefficiencies. some situations this type of cognitive dysfunc- The effect of these symptoms on daily life can tion is popularly termed “chemobrain” or “chemo- be quite profound, depending upon the demands fog” although cognitive impairment can be due present in the individual’s work and home life. Many to a large number of factors (Table 1.1), many patients observe that they can no longer multi-task, of which are discussed in detail throughout this and that they may become overwhelmed when too text. much is happening at once. They are often easily The components of cognitive dysfunction will distracted, and find that they may go from project vary as a result of the specific etiology, but there to project without getting them done. Cognitive pro- are several core cognitive domains that appear to cessing speed is generally diminished, so the person be differentially affected. Cancer patients with cog- is slower to perform their usual activities. Finally, nitive dysfunction often present with complaints patients note that it takes increased mental effort of memory disturbance. However, objective testing to perform even routine tasks. This contributes to of memory generally demonstrates a restriction of the fatigue that is often a co-existing symptom. In working memory capacity (e.g., the person is able fact, cognitive impairment generally does not occur

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

3 4 Section 1. Cognition and the brain

Table 1.1. Potential causes of cognitive impairment in Table 1.2. Predictors of cognitive impairment cancer patients r Soil (host-related factors) r r Primary or metastatic cancer in the brain Genetic factors r r Indirect effects of non-brain cancer Immune reactivity r r Neurotoxic effects of treatment Nutrition r r Chemotherapy Cognitive reserve r r Radiation therapy Seed (disease-related factors) r r Immunotherapy Tumor genetic mutations r r Hormonal therapy Paraneoplastic disorders r r Surgery Cytokines r r Effects of adjuvant medications Pesticides (treatment-related factors) r r Co- or pre-existing neurologic and psychiatric illness Cytokines r r Reactive mood and adjustment disorders Poisons r r Sensory impairment and general frailty Specific mechanisms of action r r Secondary gain Interactions between host-, disease-, and treatment- related factors in isolation, but interacts in a negative way with fatigue, pain, sleep disturbance, etc. neurotoxicity of specific agents (Meyers et al., The impact of cognitive dysfunction on cancer 1997; Scheibel et al., 2004). These will guide the patients depends upon their developmental stage of interventions to be offered to minimize the impact life, the type of work they do, and their pre-illness of cognitive dysfunction on patients’ lives. lifestyle. For instance, the symptoms described Cognitive dysfunction in cancer patients can be above may not significantly impact the quality of thus conceptualized as a result of the interaction life of an older retired person who can take things between the seed (cancer), the soil (the individ- at his or her own pace. However, those symptoms ual), and pesticides that are offered as treatment may be disabling to an attorney in a court-room set- (Table 1.2). New intervention strategies are being ting, and may necessitate changing jobs or going on developed, to improve patient function and qual- disability. ity of life as well as to provide valuable informa- Assessment of cognitive function in cancer tion for clinical trials. This is an exciting time for patients is becoming more routine. For many researchers who are interested in the effect of can- patients, addressing cognitive problems that exist cer and cancer treatment on brain function. Under- before treatment begins is important, and the standing the mechanisms of cognitive impairment underlying cause can be proactively addressed. In and the development of efficacious interventions addition, cognitive testing is increasingly becoming will require a multidisciplinary approach, including an endpoint in clinical trials. In this way, the effect oncology, neuropsychology, cognitive neuroscience, of new agents or treatments on brain function genomics, proteonomics, molecular epidemiology, can be evaluated. New studies are incorporating functional neuroimaging, neuroimmunology, ani- advances in neuroimaging and biomarkers to mal models, and drug discovery. help improve understanding of the mechanisms This book represents the first attempt to bring by which cognitive dysfunction and other symp- together clinicians and scientists to address the toms develop. A number of possible mechanisms effect of cancer and cancer treatment on cognitive are being studied, including the inflammatory function, and the intervention strategies that may response (Lee et al., 2004; Meyers et al., 2005), be helpful for patients. We hope that the reader will autoimmune phenomena (Dropcho, 2005), hor- take away our firm belief that optimizing the qual- monal influences (Wefel et al., 2004), and direct ity of life of cancer patients is possible, essential, Chapter 1. Introduction 5

and should be on equal footing with antineoplastic Meyers CA, Albitar M, Estey E (2005). Cognitive impair- therapy. ment, fatigue, and cytokine levels in patients with acute myelogenous leukemia or myelodysplastic syndrome. Cancer 104: 788–793. REFERENCES Scheibel RS, Valentine AD, O’Brien S, Meyers CA (2004). Cognitive dysfunction and depression during Dropcho EJ (2005). Update on paraneoplastic syndromes. treatment with interferon-alpha and chemother- Curr Opin Neurol 18(3): 331–336. apy. J Neuropsychiatry Clin Neurosci 16: 185– Lee BN, Dantzer R, Langley KE et al. (2004). A cytokine- 191. based neuroimmunological mechanism of cancer-rela- Wefel JS, Lenzi R, Theriault RL, Davis RN, Meyers CA ted symptoms. Neuroimmunomodulation 11: 279–292. (2004). The cognitive sequelae of standard dose adjuvant Meyers CA, Kudelka AP, Conrad CA, Gelke CK, Grove W, chemotherapy in women with breast cancer: results of a Pazdur R (1997). Neurotoxicity of CI-980, a novel mitotic prospective, randomized, longitudinal trial. Cancer 100: inhibitor. Clin Cancer Res 3: 419–422. 2292–2299. 2

Clinical neuropsychology

Jill B. Rich and Angela K. Troyer

Neuropsychology is a specialized area of study universities or teaching hospitals where they may within the field of psychology that focuses on develop their own test stimuli and procedures or brain–behavior relations, most particularly involv- administer clinical neuropsychological instruments ing structural–functional connections between either to healthy individuals with presumptively the nervous system and mental behavior. Out- normal cognition or to patients with known or side of psychology, its closest allies are behavioral suspected brain damage. When experimental neu- neurology, functional neuroanatomy, neuropsychi- ropsychologists administer clinical instruments atry, speech and language pathology, and, more to patients, however, it is most often to advance recently, cognitive neuroscience. A distinction understanding of the cognitive processes involved may be made between clinical and experimental in performing a particular task or for the com- neuropsychology, although these branches are parison of cognitive processes in different patient complementary, as evidenced by a large number groups rather than for diagnostic purposes. of neuropsychologists who identify themselves This chapter focuses on the basic principles of as clinical researchers and work as true scientist clinical neuropsychology. Following a brief overview practitioners. For example, neuropsychological of the historical background that gave rise to mod- rehabilitation generally includes diagnosis and ern clinical neuropsychology, we review the primary treatment (both clinical) as well as outcome stud- goals of neuropsychological evaluation and detail ies (research) assessing the efficacy of various the procedures common to most evaluations. The interventions. Clinical neuropsychology refers to remainder of the chapter provides an annotated list the practice of neuropsychological evaluation of of frequently used neuropsychological tests orga- individuals with known or suspected brain damage. nized by the behavioral domain (cognitive, motor, Clinical neuropsychologists typically work in hospi- mood) that they are purported to assess. Our list of tal settings or private clinics where they administer tests and our definitions of cognitive constructs are standardized, clinical neuropsychological mea- necessarily selective, as there are literally hundreds sures to patients referred by physicians, school of published neuropsychological tests now avail- systems, or insurance companies. Experimental able to clinicians. Interested readers are encour- neuropsychology is the descriptive term for the aged to consult some of the excellent compendia academic branch of neuropsychology that focuses that describe test stimuli and administration proce- on research rather than clinical service delivery. dures in detail (e.g., Lezak et al., 2004; Strauss et al., Experimental neuropsychologists typically work in 2006). In contrast to the comprehensiveness of

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

6 Chapter 2. Clinical neuropsychology 7

those texts, which are invaluable to actual prac- rologist Paul Broca’s demonstration in 1861 of the titioners in the field, our intent is merely to importance of the “anterior lobe” to the faculty of introduce health care professionals to the scope articulate speech (Benton, 2000). Although Broca and general purposes of clinical neuropsychol- himself acknowledged the much earlier work of ogy. Specifically, after reading this chapter, one Bouillaud (1825) for the identification of this asso- should have an idea of when it might be appro- ciation, the year 1861 has been heralded as the priate to refer a patient for a neuropsychological beginning of modern neuropsychology as a formal- evaluation and be sufficiently familiar with con- ized field of study (Benton & Joynt, 1960). Further cepts and tests to understand a neuropsychological advances came in short order, including works on report. receptive aphasia by Wernicke in 1874, a model of sensory and perceptual processing (the agnosias) by Lissauer in 1889, Dejerine’s reports of alexia with Historical background and without agraphia in 1891 and 1892, and Liep- mann’s distinction between apraxia and agnosia in Neuropsychology has a long history but only a short 1900. Thus, the neuropsychological disorders with past as a formalized field of study. The Edwin Smith the longest history of systematic observation and Surgical Papyrus, which has been dated to around taxonomic categorization are generally character- 2500 bc, documents 48 cases of individuals suffer- ized by loss of specific functions following cere- ing from traumatic lesions of the head, neck, and brovascular accidents. other parts of the body, contains the first known The Second World War produced an unfortunate record of a word for “brain,” and is the first written boon to neuropsychology as large numbers of head- record demonstrating an awareness of localization injured veterans returned to society. In the inter- of function (Walsh, 1978). Around 2000 years later vening years leading up to the present, neuropsy- in Classical Greece, Hippocrates and other physi- chologists have moved away from discrete mapping cians observed an association between damage on of isolated brain structures with simple or com- one side of the brain and spasms or convulsions plex behaviors in favor of seeking patterns of inter- on the other. By 200 ad, the Greeks and Romans connections in distributed systems or networks. For recorded atheoretical observations about aphasias, example, amnesia has been associated with three alexias, and other types of functional loss following general brain regions: diencephalon [(Korsakoff’s head injuries but with no analysis of the underly- syndrome), mesial temporal lobe damage (as repre- ing cognitive schema. The 1500s brought descrip- sented by the patient HM (Scoville & Milner, 1957) tions of focal symptoms and syndromes involving who developed a permanent anterograde amnesia speech loss following brain damage, unlike previ- following bilateral surgical resection of the medial ous reports, which had been limited to diffuse prob- temporal lobes for intractable epileptic seizures], lems such as dementia, anoxia, or clouding of con- and posterior cerebral artery stroke (which serves sciousness. Building on these observations, clinical the hippocampus). More complex disorders, includ- descriptions of nearly all the major neuropsycho- ing dementia, schizophrenia, closed-head injuries logical syndromes appeared over the next 300 years (i.e., non-penetrating injuries caused by rotational (see Benton, 2000; Gibson, 1969). However, prior to forces of the brain as occur in motor vehicle acci- 1800, there was very little theory and virtually no dents), and those with undetected (or undetectable) attempt to correlate these syndromes with particu- brain damage (which is sometimes the case with lar brain regions. irradiation or chemotherapy), are even less localiz- The late nineteenth century brought significant able. In sum, the history of neuropsychology may advances in brain–behavior relations. Arguably, the be traced from mentioning the brain in an ancient most significant contribution was the French neu- Egyptian papyrus to documentation of associations 8 Section 1. Cognition and the brain

between a localized brain lesion and loss of a Differential diagnosis is one of the most common specific cognitive function, to the identification and goals of neuropsychological evaluation in cases expectation of the involvement of complex brain where the underlying disease is unknown. In mem- systems and networks in complex behaviors and ory clinics and some general hospital settings the syndromes. primary referrals come from neurologists or geria- tricians to assist in differential diagnosis of dementia for elderly patients with reported memory prob- Goals of assessment lems. The evaluation can determine whether the patient has dementia (or mild cognitive impairment The original goal of assessment was localization of or normal aging or amnesia, for example), and, if function for its own sake. Tests were designed to so, what the most likely cause may be (Alzheimer’s assess the functional integrity of specific anatomical disease, frontal lobar degeneration, subcortical vas- regions. Thus, the inability to identify shapes when cular disease, alcoholism, or a potentially reversible palpated with one’s right hand while blindfolded dementia syndrome of depression). A very differ- would lead to a “diagnosis” of left parietal brain ent type of diagnosis may be sought with younger damage, specifically in the “hand” region of the left populations, especially in university or other aca- postcentral gyrus or somatosensory strip. With the demic settings where neuropsychologists may be advent of brain imaging, of course, neuropsychol- called upon to assess individuals with poor aca- ogists have been called upon less and less to iden- demic achievement. In these cases, the objective tify the presence and localization of brain lesions may be to help determine whether the learning dif- that can routinely be obtained by computed tomog- ficulties are primarily attributable to a learning dis- raphy (CT) or magnetic resonance imaging (MRI) ability as opposed to environmental circumstances. scanning. Nevertheless, neuropsychology continues Neuropsychological evaluations are also frequently to play this role when imaging is contraindicated or requested following closed-head injuries sustained otherwise not available and for lesions that are dif- in motor vehicle accidents. Even when there is a ficult to discern with imaging. For example, a neu- documented concussion, it may be unclear whether rosurgeon may refer an epileptic patient for eval- the person has sustained structural brain damage. uation of language and memory functions prior to A thorough neuropsychological evaluation includ- surgical resection of a suspected seizure focus in the ing assessment of both cognitive and personality left mesial temporal lobe. In such cases, the specific variables can help determine whether poor con- seizure focus may be unknown. centration and attention following the accident, for More typically, however, neuropsychological eval- example, are likely due to brain damage, psycholo- uations are requested for other purposes, most of gical factors, or even the medications that the per- which vary according to the clinical setting. The goal son may be taking for physical injuries sustained in for any particular evaluation may be determined on the accident. the basis of a mutual understanding arising from When the diagnosis is not in question, such an established relationship between the physician as with a genetic disorder (Huntington’s disease, and neuropsychologist, by a specified request in the Wilson’s disease), medical disorder [infection written referral, or, when the referral question is with human immunodeficiency virus (HIV), unclear, by contacting the physician to determine epilepsy, multiple sclerosis, Parkinson’s disease], the purpose of the referral. In other words, the neu- documented brain lesion (neoplasm, aneurysm, ropsychologist requires a contextual framework in arteriovenous malformation), or trauma (closed- order to design the evaluation and report the results. or open-head injury, electrical injury), an evalua- Some of the more common evaluation goals are dis- tion may nevertheless be requested to provide a cussed below. descriptive report of the patient’s cognitive strengths Chapter 2. Clinical neuropsychology 9

and weaknesses. This evaluative purpose can serve for neuropsychological evaluations both pre- and a variety of goals, including treatment planning, post-operatively to assess comparative outcomes workers’ compensation and other employment from surgical and cognitive perspectives. This is issues requiring vocational guidance, long-term particularly true for surgeries with high morbid- care planning, or documentation of a baseline ity rates or in vulnerable brain regions, where the against which to gauge future abilities. Importantly, potential for adverse cognitive outcomes would add many diseases have widely varying behavioral to the “cost” in a risk-benefit analysis of whether expressions, much as there are multiple pheno- to perform the surgery (e.g., carotid endarterec- types of a single genotype. Thus, two individuals tomy, pallidotomy for Parkinson’s disease, certain with the same type of brain tumor may have brain tumors, shunt placement for normal pres- completely non-overlapping symptoms or may sure hydrocephalus in elderly patients or those with have similar symptoms with differing magnitudes severe dementia). that leave one person functionally intact and the There are a number of situations in which neu- other compromised. Individual differences in ropsychological evaluation may be particularly use- pre-existing abilities and different occupational or ful or relevant for the care of cancer patients.The social demands also lead to different functional actual cognitive profiles associated with various outcomes among individuals with the same disease. brain cancers and the cognitive effects of var- Neuropsychological evaluation in these cases may ious cancer treatments are described in subse- be helpful in contextualizing the impact of the brain quent chapters. Below, we briefly sketch some of diseaseinthatperson’slife. the circumstances that might lead an oncologist Neuropsychologists are also called upon to con- to refer a cancer patient for a neuropsychological duct serial assessments to document changes over evaluation: r time in response to behavioral or drug interven- When there are subjective complaints from the tions (in the case of clinical trials) or when nat- patient of (1) cognitive declines, such as poor con- uralistic changes may be expected that would centration, slowed thinking, word-finding diffi- affect care needs, such as degenerative demen- culties, trouble making decisions, right-left con- tias, chronic progressive disease (e.g., multiple fusion, short-term memory problems, difficulty sclerosis), or rapidly growing tumors. Many neu- performing calculations, becoming lost in familiar ropsychological measures are highly sensitive to areas; (2) sensory or perceptual changes, such as practice effects. In fact, stable performance (i.e., visual field cuts, anosmia (loss of sense of smell), with no practice-associated improvement) from inability to recognize faces or some other class of initial to repeated administration on some mea- objects (cars, buildings); (3) motor changes, such sures may actually indicate the progression of a as a change in handwriting, difficulties with bal- disease or a decline in function. Practice effects ance, gait, or fine-motor skill; or (4) psychological make the selection of tests and interpretation changes, such as irritability, depression, excessive of performance particularly important for serial anxiety. r assessments. Among the armamentarium of tests When there are external reports from friends or outlined below are a number of measures with family members of any of the above symptoms or alternate versions of equivalent difficulty specif- of any of the following symptoms that may indi- ically designed for serial testing (e.g., Hopkins cate compromised integrity of the frontal lobes Verbal Learning Test-Revised; Brandt & Benedict, (and which may be unnoticed or denied by the 2001). patient): (1) abrupt changes in personality (e.g., Neuropsychological evaluations are also sought lack of empathy, depression, becoming enraged to help clarify the outcome of a surgical intervention. easily); (2) uncharacteristic behaviors, such as Many neurosurgeons routinely refer their patients making inappropriate sexual remarks, spending 10 Section 1. Cognition and the brain

largesumsofmoney,engaginginstrangeorrit- commonly used battery of tests is the Halstead– ualistic eating habits; (3) declines in self-care or Reitan Battery (Reitan & Wolfson, 1993), which was hygiene; or (4) hypersomnolence or insomnia. originally developed as a sensitive diagnostic mea- r In cases where the primary tumor is rapidly sure of patients with frontal lobe or lateralized brain changing. lesions. Currently, many neuropsychologists use a r To determine potential cognitive or mood effects “core” battery of tests to tap functions in several key of radiation treatment. functional domains, including general mental sta- r To determine potential cognitive or mood effects tus or intelligence, attention, visual perception, con- of chemotherapy. struction, language, memory, executive function, r To document possible neuropsychological seque- and mood, personality, or emotional status. This lae that may have been incurred from destruc- core battery is then supplemented by additional tion of healthy tissue during neurosurgical tumor measures as warranted by the referral question, resection. the patient’s capacity for testing, the patient’s abilities as ascertained throughout the assessment, and the clinical setting. For example, even “Halstead– Standard neuropsychological Reitanners” who use the current version of the orig- evaluation procedures inal Halstead–Reitan Battery generally supplement their assessment with one of the Wechsler scales of The clinical neuropsychological evaluation com- intelligence as well as tests of memory and other prises several discrete sections, including history specific functional domains. Some brief evaluations taking, test selection, the clinical interview, test of very elderly patients or those with severe brain administration (also called the assessment), inter- damage may include only a single mental status pretation of results, and the dissemination of find- examination that encompasses a minimal sample ings and conclusions. The history taking begins with of several of the domains listed above (e.g., Mattis the referral question. In many cases, the referral Dementia Rating Scale; Mattis, 2001). Other general- itself may include several reports, such as a neuro- purpose batteries (e.g., Kaplan–Baycrest Neurocog- logical examination, radiological reports from brain nitive Assessment; Leach et al., 2000; The Repeat- imaging, bloodwork results, other medical reports, able Battery for the Assessment of Neuropsycho- or even an entire hospital record. This part of the logical Status; Randolph et al., 1998) have been history review takes place before the assessment developed for use as contained measures when and may be done several days or weeks in advance. a brief assessment is appropriate. However, these If the patient has undergone a previous evaluation, same measures, as well as brief screening instru- as is often the case for patients referred by insurance ments such as the Mini-Mental State Examina- companies and/or involving a legal claim, reports tion (Folstein et al., 1975), may be used in longer from those evaluations are typically reviewed as part assessments as a preliminary measure to guide the of the history. This is particularly important when selection of tests for subsequent evaluation of spe- the prior assessment was conducted within the past cific functional domains (many of which are listed year (or even within the past month, such as when below). evaluating a patient pre- and post-surgically), as it The clinical interview is typically conducted with will likely affect the selection of tests for the current the patient alone, especially for inpatients, although evaluation. permission may be requested to contact a spouse Historically, clinical neuropsychologists could be or caregiver separately by phone. When secondary divided into two camps in terms of test selection: sources are present, as is often the case with out- those who used fixed batteries and those who opted patients, they may be interviewed together with for a flexible approach. The most widely known and the patient, especially when the patient may be Chapter 2. Clinical neuropsychology 11

an unreliable historian, or separately (with the Published tests include strict guidelines for stan- patient’s permission in most settings). When a child dardized test administration, and this is critical for is being tested, he or she is typically interviewed subsequent interpretation of results. For example, briefly, though the parents are often asked to report inexperienced examiners may “coach” patients or such things as the timing of developmental mile- give extra cues in an effort to help them get the stones in addition to current symptoms or prob- right answer. Alternatively, the overly rigid exam- lems. Depending on the setting and availability of iner may refuse to repeat a question that the patient ancillary records, the clinical interview can take didn’t hear because of poor auditory acuity, a com- from 10 min to an hour or more. At a minimum, the peting public address announcement, or a sneeze. neuropsychologist ascertains critical demographic Either of these approaches could yield unrepre- variables that may affect test interpretation, such as sentative test results. Among the data to be inter- the patient’s age, education, native language, and preted are the summary scores obtained on the var- handedness, as well as social variables, such as high- ious tests administered, the qualitative responses est and most recent occupational attainment, cur- that led to those scores, the consistency of perfor- rent living situation (including language spoken at mance on multiple measures of the same domain, home), medical history and current medical status, relative strengths and weaknesses observed across and the patient’s understanding of the reason for the domains, the degree to which the test environ- referral. ment conformed to or deviated from optimal condi- In general, even when the patient suffers from tions, the patient’s co-operation with the test proce- dementia, the interview will provide useful infor- dures, normative expectations for individuals with mation, such as whether the patient is aware of similar demographic and social backgrounds, and his or her deficits and other aspects of insight. It any motor, visual, auditory, comprehension, or ver- also provides an opportunity to assess spontaneous bal expressive difficulties that may have impacted or conversational speech, including length and the patient’s ability to perform the presented tasks. appropriateness of responses to open-ended ques- Qualitative interpretation of quantitative data is tions. Behavioral observations, such as eye contact, essential. Consider, for example, the many ways in impulsivity, distractibility, and inattention are made which a score of 0 may be obtained on a single item, during the interview and throughout the assess- such as the identification of a line drawing: (1) no ment. Although there are specific tests to assess for response; (2) identification of the item at the super- malingering (used frequently in medicolegal con- ordinate level of taxonomic categorization (animal texts), the examiner also tries to gauge the patient’s for rhinoceros); (3) identification of an exemplar motivation level and fatigue to determine whether from the same class (hippopotamus for rhinoceros); the results obtained represent the patient’s true (4) phonemic paraphasic response (rhinosteros); abilities. (5) neologism (pinder); (6) misperception of gestalt, The actual neuropsychological assessment entails with focus on a single detail (horn for rhinoceros); the test administration component of the evalua- or (7) correct response after the time limit. These tion. This may be done by the neuropsychologist, responses have differing interpretive significance, but more typically the assessment is carried out which is why a test score in isolation may be by a psychometrist (trained technician) or clinical misleading. trainee (such as a predoctoral intern or postdoc- Following test selection appropriate to the spec- toral fellow). In the latter cases, the neuropsychol- ified purpose, accurate test administration and ogist makes or approves of the test selection and scoring, and interpretation of the obtained results supervises the test administration and scoring accu- in the context of the history, presenting symptoms, racy. Commonly used tests for various functional and testing circumstances, the final step in the domains are listed in the next section. evaluation is dissemination of the findings.This 12 Section 1. Cognition and the brain

most often takes the form of a written report to the Intelligence Scale – III (WAIS-III; Wechsler, 1997a) referring physician and/or a verbal report given to and the Stanford-Binet Intelligence Scale (Roid, multidisciplinary teams in hospital settings. The 2003), contain multiple subtests tapping a variety of length of the report varies widely across settings and abilities. Tests of verbal intelligence assess vocabu- depends primarily on the familiarity and interest lary, general factual knowledge, verbal abstraction, of the referring physician. Reports to insurance or social judgment. Tests of non-verbal (or per- companies for head-injured patients involved in formance) intelligence assess visual abstraction, litigation may be 15–20 pages long. At the other end visual construction, detection of visual details, or of the continuum, a two-paragraph chart note may arrangement of pictures to tell a story. Often, the be made in an inpatient record at a nursing home purpose of administering intelligence tests is to for a patient with severe dementia who could not obtain an estimate of a person’s overall level of cog- undergo much testing. Between these extremes is nitive ability, rather than to determine performance a standard report of three to six pages, which is a on the various subcomponents of intelligence. To typical length for outpatient evaluations. In many achieve the former, only a subset of intelligence clinic and hospital settings, the neuropsychologist tests need be administered (Wechsler, 1999). On the will schedule a verbal feedback session with the WAIS-III, Verbal IQ is most highly correlated with patient and/or the patient’s family or caregivers. performance on the Vocabulary subtest, and Per- Although patients are entitled to receive a copy of formance IQ is most highly correlated with perfor- their full written report, it is undesirable to merely mance on the Matrix Reasoning (visual abstraction) send them a copy in the mail with no supplemental subtest. explanations. Instead, many neuropsychologists offer to provide the patient with a summary of the Attention and processing speed results after they review it together in the feedback session. Attention is the ability to focus or concentrate on specific stimuli. It comprises selective, sustained, divided, and alternating attention and plays a large role in working memory. Selective attention is the Commonly used neuropsychological tests by ability to focus on information relevant to the task domain assessed at hand and to filter out irrelevant information. Most people automatically display auditory selec- Most of the tests described below have alternative tive attention for the sound of their name being spo- versions, modifications, or other corollary measures ken in a crowded room with a lot of background developed specifically for children. In some cases, noise. Tests of visual selective attention may require modified instruments have been developed for use the search for target stimuli while scanning arrays with other populations, such as the elderly or indi- of target and non-target stimuli (e.g., Ruff 2 & 7; Ruff viduals with severe impairments who are not capa- & Allen, 1996). Sustained attention is the ability to ble of completing many of the measures described maintain attention on a task over an extended time, here. such as watching a long movie or reading a book. It is tested by having the patient perform a relatively simple task for several minutes at a time, such as hit- Intelligence ting a key every time a particular number appears Intelligence is a multidimensional construct on a screen or a letter is spoken on an audiotape that comprises many cognitive abilities. Clinical (Test of Everyday Attention; Robertson et al., 1994). tests of intelligence, such as the Wechsler Adult Divided attention is the ability to focus on multiple Chapter 2. Clinical neuropsychology 13

tasks simultaneously, such as watching a child and Processing speed refers to the ability to quickly talking on the telephone. It is tested by having the process and respond to new information, such as patient perform two concurrent tasks, such as scan- slamming on one’s brakes when the car ahead stops ning a stimulus array and counting tones. suddenly. It can be measured by tests requiring Alternating attention is the ability to switch atten- patients to rapidly transcribe symbols paired with tion between two or more sources of information, numbers or to scan series of symbols for the pres- such as going back and forth between recipes ence of target symbols (Digit-Symbol Substitution when preparing different parts of a meal. It is most subtest and Symbol Search subtest of the WAIS-III; frequently tested by the Trail Making Test (Reitan & Wechsler, 1997a). Wolfson, 1993; Delis et al., 2001), which requires one to alternate between number and letter sequences Visual ability (e.g., 1–A–2–B–3–C). It is important to keep in mind that most attention tests tap more than one Assessment of visual ability typically involves test- component of attention. For example, sustained ing object and spatial perception as well as visual attention almost always requires selective attention. construction. Object perception is an ability used The Trail Making Test includes two subtests: the one in everyday life to recognize such things as house- described above and a simple numeric sequencing hold items and people’s faces. It may be mea- subtest (1–2–3–4). Both tasks require visual scan- sured by the identification of missing visual details ning, sequencing, and psychomotor control of a from a line drawing (Picture Completion subtest of pencil, but only the alphanumeric subtest requires the WAIS-III; Wechsler, 1997a), matching a photo- cognitive set shifting from one sequence to the graph of a face from an array containing the target other. Thus, the ability to alternate attention can face along with different faces or of the same face be assessed by comparing performance on the two photographed from different views (Benton Facial tasks. Recognition Test; Benton et al., 1983), the identifi- Working memory refers to the ability to manipu- cation of pictures of objects (Boston Naming Test; late information being held in memory for a short Kaplan et al., 1983, although this task also mea- period, such as calculating the tip on a restau- sures naming ability), discrimination of real from rant or taxi bill. Working memory certainly requires nonsense figures, or the identification of incomplete attention, and some neuropsychologists character- objects, rotated objects, objects embedded in com- ize it as a type of attention, although others report plex arrays, or overlapping figures (see Lezak et al., on this function in the memory section of their 2004; Warrington & James, 1991). Spatial perception reports. Working memory may be tested by ask- is the ability to appreciate the physical location of ing the patient to perform mental arithmetic (Arith- objects either alone or in relation to other objects. metic subtest of the WAIS-III; Wechsler, 1997a), to This ability underlies many everyday motor activi- listen to strings of digits and repeat them in back- ties, such as walking through a doorway rather than ward sequence (Digit Span subtest of the WAIS-III), into the door frame, reaching for a desired object in to listen to strings of randomized digits and letters the refrigerator, and buttoning one’s clothes. Spa- and repeat back the numbers first in numeric order tial perception may be measured by the compari- followed by the letters in alphabetic order (Letter- son and matching of line segments drawn at vary- Number Sequencing subtest of the Wechsler Mem- ing angles (Judgment of Line Orientation; Benton ory Scale-III; WMS-III; Wechsler, 1997b), or to listen et al., 1983), the discrimination of relative spatial to a series of digits and add each digit to the previ- positions of objects (Visual Object and Space Per- ously presented one (Paced Auditory Serial Addition ception Test; Warrington & James, 1991), and mental Test; Gronwall, 1977). rotations (Standardised Road Map; Money, 1976). 14 Section 1. Cognition and the brain

Visual construction is the ability to put together Picture subtest of the Boston Diagnostic Aphasia individual parts to make a coherent whole, such Examination; Goodglass & Kaplan, 1972). as assembling a new appliance from a box of parts (with or without an instruction manual). This Memory skill requires visual perception, integration of visual details, and a motor response. Visual construction Memory is a complex cognitive ability to measure may be measured by having patients arrange col- because it can be broken down into individual ored blocks into designs (Block Design subtest of the components along several dimensions, including WAIS-III; Wechsler, 1997a), assemble puzzle pieces temporal span, sensory modality, and stage or pro- to create an object (Object Assembly subtest of the cess. The majority of clinical memory tests mea- WAIS-III; Wechsler, 1997a), or copy a detailed geo- sure short-term and long-term memory for informa- metric figure (Rey–Osterrieth Complex Figure Test; tion that was presented seconds to hours ago, as Strauss et al., 2006). opposed to remote memory for events from many years ago. In everyday life, these types of memory correspond to remembering a phone number long Language enough to dial it, remembering what you had for Assessment of language functioning typically breakfast this morning, and remembering your high includes evaluation of both receptive and expres- school graduation. In addition, most tests measure sive abilities. Comprehensive batteries of language episodic memory for new information presented tests such as the Boston Diagnostic Aphasia Exami- during the assessment (such as a story or a list or nation (Goodglass & Kaplan, 1972) contain multiple words, designs, or faces) as opposed to semantic tests of language ability and are useful for charac- memory for previously known general facts (e.g., the terizing severe language impairments. Receptive capital of Norway). language refers to the ability to understand orally A thorough memory assessment typically or visually presented verbal information and is includes measurement of both verbal and non- necessary for following a conversation or reading verbal memory. Verbal memory may be assessed a book. It may be assessed with a multiple-choice by asking the patient to remember lists of words, test of vocabulary (e.g., Peabody Picture Vocabulary such as the California Verbal Learning Test (Delis Test – III; Dunn & Dunn, 1997) or a sentence- et al., 2000) or Hopkins Verbal Learning Test- comprehension test requiring the patient to follow Revised (Brandt & Benedict, 2001), series of word simple verbal commands (e.g., Token Test; see pairs (e.g., Verbal Paired Associates subtest of the Strauss et al., 2006). Expressive language refers WMS-III; Wechsler, 1997b), or prose passages (Log- to the ability to generate words or sentences, as ical Memory subtest of the WMS-III). Non-verbal used for speaking and writing in everyday life. It memory may be assessed by asking the patient to maybemeasuredbyaskingpatientstonameline remember geometric figures (e.g., Brief Visuospatial drawings of common and low-frequency objects Memory Test; Benedict, 1997; Rey–Osterrieth Com- (Boston Naming Test; Kaplan et al., 1983), to plex Figure Test; Strauss et al., 2006) or new faces generate words in a limited time (typically 60 s) (e.g., Family Pictures subtest and Faces subtest of according to specified rules, such as words begin- the WMS-III). Memory can be tested with several ning with a given letter or belonging to a specific procedures, including free recall of the information semantic category (Verbal Fluency subtest of the without any hints or cues from the examiner and Delis–Kaplan Executive Function Scale, or D-KEFS; recognition of previously presented items randomly Delis et al., 2001), to define words (Vocabulary sub- dispersed among non-presented items. test of the WAIS-III; Wechsler, 1997a), or to describe The typical procedure of many clinical mem- a complex scene depicted on a card (Cookie Theft ory tests involves: (1) multiple presentations of the Chapter 2. Clinical neuropsychology 15

information to be remembered, with each presenta- ment. For example, patients may be asked to copy tion followed by free recall; (2) an intervening period a detailed geometric figure (e.g., using the Boston of 20–30 min during which the individual engages in Qualitative Scoring System for the Rey–Osterrieth unrelated tasks; (3) delayed free recall of the infor- Complex Figure Test; Stern et al., 1999) or to re- mation; and (4) delayed recognition of the informa- arrange a given structure to match a target structure tion. Using this procedure, memory can be parsed using a minimum number of responses in accor- into several processes. Acquisition is the ability to dance with fixed rules (e.g., Tower subtest of the D- encode new information into memory and is mea- KEFS; Delis et al., 2001). sured by level of recall after the initial presentations. Abstract thinking is the ability to form generalized Learning is the ability to benefit from repeated pres- concepts from discrete instances. Tests of abstract entation of the information and is measured by the thinking require the individual to describe simi- increase in items recalled from the first to the last larities between words (e.g., Similarities subtest of trial of the initial presentations. Generally, recollec- the WAIS-III; Wechsler, 1997a), to select the miss- tion improves with successive presentations. Reten- ing component of visual sequences arranged in sim- tion is the ability to hold newly acquired informa- ple to complex patterns (e.g., Matrix Reasoning sub- tion in memory over a delay. Retention is often test of the WAIS-III; Wechsler, 1997a), to sort cards represented as a proportion of initial acquisition, according to various principles (e.g., Card Sorting calculated as the number of items recalled after subtest of the D-KEFS; Delis et al., 2001; Wisconsin the delay divided by the number of items recalled Card Sorting Test; Heaton et al., 1993), or to interpret after the initial presentations. Retrieval is the abil- proverbs (Proverbs subtest of the D-KEFS). ity to recall information that has been stored in Response inhibition is the ability to inhibit an memory. A retrieval problem is suspected when automatic response in favor of a more unusual free recall is significantly poorer than recognition, response. For example, experienced drivers who are because this pattern indicates that the informa- conditioned to go when a traffic light is green must tion was stored in memory but was not properly instead stop when an ambulance or funeral proces- accessed. sion is coming by. A classic measure of response inhibition is the Stroop (1935) paradigm, which involves presenting color names printed in disso- Executive function nant colors (e.g., the word “red” is written in blue Executive functions are defined as higher order ink) and asking the patient to state the ink color. cognitive abilities that are necessary for appro- Because reading the word is an automatic pro- priate, socially responsible, and effective con- cess for fluent readers, this task requires inhibi- duct (Goodwin, 1989). They encompass many tion. Standardized Stroop tests include Stroop Color different types of cognitive ability, including plan- and Word (Golden, 1978), Victoria Stroop (Strauss ning, abstract thinking, response inhibition, and et al., 2006), and D-KEFS Color–Word Interference switching. (Delis et al., 2001). Inhibition can also be mea- Planning involves the abilities to formulate and sured with a sentence-completion task requiring the weigh alternative approaches to a task and to carry patient to provide a word that is unrelated to the out an effective approach to achieve a goal. Every- sentence, such as the Hayling Sentence Completion day tasks that require planning include packing Test (Burgess & Shallice, 1997). a suitcase, preparing a meal, and mapping out Switching, also known as cognitive set shifting, a transportation route (Shallice, 1982). It can be is the ability to alternate between different types measured by examining the way in which patients of information or different categories of response. accomplish complex tasks in which impulsive, early Switching may be measured by asking the patient steps may slow down the ultimate goal attain- to alternately sequence numbers and letters (as 16 Section 1. Cognition and the brain

in tests of alternating attention; Trail Making Test; (Snaith & Zigmond, 1994), Personality Assessment Delis et al., 2001; Reitan & Wolfson, 1993), to switch Inventory (Morey, 1991), and the Minnesota Mul- between sorting principles on card sorting tasks tiphasic Personality Inventory (MMPI-2; Butcher (D-KEFS Card Sorting; Delis et al., 2001; Wisconsin et al., 1989). Card Sorting Test; Heaton et al., 1993), to generate words from alternating semantic categories on a As can be gleaned from this brief review of the his- verbal fluency task (Verbal Fluency subtest of the D- tory, purpose, procedure, and basic principles of KEFS; Delis et al., 2000), or to switch between color clinical neuropsychology, a number of varying con- naming and word reading on a Stroop task (D-KEFS; ditions will generally determine the exact nature Color–Word Interference subtest; Delis et al., 2001). of a particular neuropsychological evaluation. In addition to the compendia by Lezak et al. (2004) and Strauss et al. (2006) considered critical for the Sensorimotor ability practice of neuropsychology, the interested begin- Gross measures of sensory and motor ability are ner reader is referred to the undergraduate text by often included in neuropsychological assessment, Kolb and Whishaw (2003) or a standard graduate both to provide information about the functional text edited by Heilman and Valenstein (2003). Sev- integrity of specific brain regions and to detect any eral other edited and authored volumes elaborate right/left asymmetries that may indicate lateralized on the topics touched upon here (e.g., Adams et al., brain dysfunction. Measurement of sensory ability 1996; Bradshaw & Mattingley, 1995; Grant & Adams, may include simple tests of detection of visual, audi- 1996; Ogden, 1996). tory, or tactile stimuli on the left and right sides of the body (Bilateral Simultaneous Sensory Stim- ulation; Reitan & Wolfson, 1993). Measurement of REFERENCES manual motor ability may include assessment of fine-motor speed (Finger Tapping, also called Finger Adams RL, Parsons OA, Culbertson JL, Nixon SJ (1996). Oscillation Test; Reitan & Wolfson, 1993), strength Neuropsychology for Clinical Practice: Etiology, Assess- of hand grip (Grip-Strength Dynamometer; Reitan & ment, and Treatment of Common Neurological Disorders. Washington, DC: American Psychological Association. Wolfson, 1993), and fine-motor dexterity (Grooved Beck AT, Steer RA (1990). Beck Anxiety Inventory. San Anto- Pegboard; Kløve, 1963). nio, TX: Psychological Corporation. Beck AT, Steer RA, Brown GK (1996). Beck Depression Mood and personality Inventory-II. San Antonio, TX: Psychological Corpora- tion. Assessment of mood and/or personality is some- Benedict RHB (1997). Brief Visuospatial Memory Test- times included in a neuropsychological evalua- Revised. Lutz, FL: Psychological Assessment Resources. tion because brain dysfunction can cause changes Benton A (2000). Exploring the History of Neuropsychology: in either emotional responsiveness or even longer Selected Papers. New York: Oxford University Press. term temperament. Evaluation can provide infor- Benton AL, Joynt RJ (1960). Early descriptions of aphasia. − mation about symptoms of depression or anxi- Arch Neurol 3: 205 222. Benton AL, Hamsher K de S, Varney NR, Spreen O (1983). ety, level of stress, somatic complaints, paranoid Contributions to Neuropsychological Assessment: A Clini- thoughts, substance abuse, aggressive behaviors, cal Manual. New York: Oxford University Press. and interpersonal styles. Mood and personality Bouillaud JB (1825). Recherches cliniques propres a` are usually measured with self-report question- demontrer´ que la perte de la parole correspond ala` naires, such as the Beck Depression Inventory-II lesion´ des lobules anterieursducerveau,et´ a` confirmer (Beck et al., 1996), Beck Anxiety Inventory (Beck & l’opinion de M. Gall sur le siegedel’organedulangage` Steer, 1990), Hospital Anxiety and Depression Scale articule.´ Arch Generales Med 8:25−45. Chapter 2. Clinical neuropsychology 17

Bradshaw JL, Mattingley JB (1995). Clinical Neuropsychol- Kløve H (1963). Clinical neuropsychology. Med Clin North ogy: Behavioral and Brain Science. San Diego, CA: Aca- Am 47: 1647–1658. demic Press. Kolb B, Whishaw IQ (2003). Fundamentals of Human Neu- Brandt J, Benedict RHB (2001). Hopkins Verbal Learn- ropsychology (5th edn.). New York: Worth Publishers. ing Test-Revised. Lutz, FL: Psychological Assessment Leach L, Kaplan E, Richards B, et al. (2000). Kaplan-Baycrest Resources. Neurocognitive Assessment. San Antonio, TX: Psychologi- Burgess PW, Shallice T (1997). The Hayling and Brixton cal Corporation. Tests. Bury St. Edmunds: Thames Valley Test Company. Lezak MD, Howieson DB, Loring DW (2004). Neuropsycho- Butcher JN, Dahlstrom WG, Graham JR, et al. (1989). logical Assessment (4th edn.). New York: Oxford Univer- Minnesota Multiphasic Personality Inventory-2 sity Press. [Manual]. Minneapolis, MN: University of Minnesota Mattis S (2001). Dementia Rating Scale-2 (Manual). Odessa, Press. FL: Psychological Assessment Resources. Delis DC, Kramer JH, Kaplan E, et al. (2000). California Money J (1976). A Standardised Road-Map Test of Direction Verbal Learning Test-II. New York: Psychological Corpo- Sense Manual. San Rafael, CA: Academic Therapy Publi- ration. cations. Delis DC, Kaplan E, Kramer JH (2001). Delis-Kaplan Morey LC (1991). Personality Assessment Inventory. Odessa, Executive Function System. New York: Psychological Cor- FL: Psychological Assessment Resources. poration. Ogden JA (1996). Fractured Minds: A Case-Study Approach Dunn LM, Dunn LM (1997). Peabody Picture Vocabulary to Clinical Neuropsychology. New York: Oxford University Test – III. Circle Pines, MN: American Guidance Service. Press. Folstein MF, Folstein SE, McHugh PR (1975). “Mini-mental Randolph C, Tierney MC, Mohr E, et al. (1998). The Repeat- State”: a practical method for grading the cognitive able Battery for the Assessment of Neuropsychological state of patients for the clinician. J Psychiatr Res 12: Status (RBANS): preliminary clinical validity. JClinExp 189–198. Neuropsychol 20: 310–319. Gibson WC (1969). The early history of localization in the Reitan RM, Wolfson D (1993). Halstead-Reitan Neu- nervous system. In PJ Vinken & GW Bruyn (eds.) Hand- ropsychological Battery. Tucson, AZ: Neuropsychology book of Clinical Neurology (Vol. 2, pp. 4–14). Amsterdam: Press. North-Holland Publishing Co. Robertson IH, Ward T, Ridgeway V, et al. (1994). The Test Golden CJ (1978). Stroop Color and Word Test: A Manual of Everyday Attention. Bury St Edmunds: Thames Valley for Clinical and Experimental Uses. Chicago, IL: Stoelting Test Company. Company. Roid GH (2003). Stanford-Binet Intelligence Scale (5th edn.) Goodglass H, Kaplan E (1972). The Assessment of Aphasia Itasca, IL: Riverside Publishing. and Related Disorders. Philadelphia, PA: Lea & Febiger. Ruff RM, Allen CC (1996). Ruff 2 & 7 Selective Attention Test. Goodwin DM (1989). A Dictionary of Neuropsychology.New Lutz, FL: Psychological Assessment Resources. York: Springer-Verlag. Scoville WE, Milner B (1957). Loss of recent memory after Grant I, Adams KM (eds.) (1996). Neuropsychological bilateral hippocampal lesions. J Neurol Neurosurg Psy- Assessment of Neuropsychiatric Disorders (2nd edn.). chiatry 20: 11–21. New York: Oxford University Press. Shallice T (1982). Speciﬁc impairments of planning. Gronwall DMA (1977). Paced Auditory Serial-Addition task: Philosoph Trans Soc London B 298: 199–209. a measure of recovery from concussion. Percept Motor Snaith RP, Zigmond AS (1994). The Hospital Anxiety and Skills 44: 367–373. Depression Scale. Windsor: NFER-Nelson. Heaton RK, Chelune GJ, Talley JL, et al. (1993). Wisconsin Stern RA, Javorsky DJ, Singer EA, et al. (1999). The Card Sorting Test (WCST) Manual Revised and Expanded. Boston Qualitative Scoring System for the Rey-Osterrieth Odessa, FL: Psychological Assessment Resources. Complex Figure. Odessa, FL: Psychological Assessment Heilman KM, Valenstein E (eds.) (2003). Clinical Neu- Resources. ropsychology (4th edn.). New York: Oxford University Strauss E, Sherman EM, Spreen O (2006). A Compendium Press. of Neuropsychological Tests: Administration, Norms, and Kaplan EF, Goodglass H, Weintraub S (1983). The Boston Commentary (3rd edn.). New York: Oxford University Naming Test (2nd edn.) Philadelphia, PA: Lea & Febiger. Press. 18 Section 1. Cognition and the brain

Stroop JR (1935). Studies of interference in serial verbal Wechsler D (1997a). The Wechsler Adult Intelligence Scale – reaction. JExpPsychol18: 643–662. III. San Antonio, TX: Psychological Corporation. Walsh KW (1978). Neuropsychology: A Clinical Approach. Wechsler D (1997b). The Wechsler Memory Scale – III.San New York: Churchill Livingstone. Antonio, TX: Psychological Corporation. Warrington EK, James M (1991). Visual Object and Space Wechsler D (1999). The Wechsler Abbreviated Scale Perception Battery. Bury St Edmunds: Thames Valley Test of Intelligence. San Antonio, TX: Psychological Company. Corporation. 3

Brain imaging investigation of chemotherapy-induced neurocognitive changes

Brenna C. McDonald, Andrew J. Saykin, and Tim A. Ahles

Introduction therapies, has been a topic of increasing study for 20 years. While more detailed discussion of cogni- Structural and functional neuroimaging techniques tive studies of changes in function related to cancer provide a unique opportunity to examine the neu- treatment can be found elsewhere in this volume, ral basis for cognitive changes related to cancer and these issues will be briefly summarized here, as they its treatment. While the link between cognitive dys- form a major component from which subsequent function and central nervous system (CNS) can- neuroimaging research has grown. Cross-sectional cers (e.g., primary brain tumors, primary CNS lym- and longitudinal neuropsychological studies of can- phoma, brain metastases of cancer in other organ cer survivors (Ahles et al., 2002; Brezden et al., systems, etc.) or non-CNS cancers treated with pro- 2000; Castellon et al., 2004; Schagen et al., 1999; phylactic whole-brain radiation seems clear, our Shilling et al., 2005; Tchen et al., 2003; Van Dam understanding of the causes for cognitive changes et al., 1998; Wefel et al., 2004b; Wieneke & Dienst, following chemotherapy for other non-CNS cancers 1995) have contributed to a growing body of liter- remains much more limited. Research using a vari- ature suggesting detrimental effects of chemother- ety of neuroimaging modalities has begun to delin- apy on cognitive performance, although some eate the brain mechanisms for cognitive changes studies have not found such an effect (Donovan related to cancer and chemotherapy, across a num- et al., 2005; Jenkins et al., 2006). The cognitive ber of cancer subtypes. This chapter will briefly changes associated with chemotherapy are typically summarize the cognitive domains most likely to be subtle, with patients often showing mildly (though affected following chemotherapy, review the avail- statistically significant) reduced functioning rela- able data relating cognitive performance and structive to control groups, though overall performance tural and functional neuroimaging changes in var- remains within normal limits by clinical stan- ious cancer populations, and suggest avenues for dards. Changes related to chemotherapy have been future work in this area. reported across several cognitive domains, includ- As clinical efficacy of cancer treatment has ing working memory, executive function, and pro- improved survivorship, increased awareness has cessing speed (Ahles & Saykin, 2002; Anderson- arisen of issues critical to the functioning and qual- Hanley et al., 2003; Ferguson & Ahles, 2003; Tannock ity of life of cancer survivors. Specifically, cogni- et al., 2004). Acutely, cognitive symptoms are often tive impairment related to cancer and its treatment, reported during chemotherapy (Ahles & Saykin, including radiation, chemotherapy, and hormone 2002; Ferguson & Ahles, 2003), but appear to persist

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

19 20 Section 1. Cognition and the brain

post-treatment only in a smaller subset of patients, that shared factors interact to increase the risk of with estimates ranging from 17% to 34%. These cancer itself, as well as cognitive decline more gen- effects were observed even after accounting for vari- erally, where the presence of such factors might ables which may be related to negative cognitive account for the observed baseline impairment in outcomes as well as to cancer and its treatment, some cancer patients. In the presence of such risk including psychological factors such as depres- factors, one would perhaps expect to observe exac- sion or anxiety, or side-effects of cancer treat- erbation of cognitive deficits following chemother- ments such as fatigue. Studies have also found a apy. From a neuropsychological perspective, an higher than expected incidence of impaired cog- important avenue of research to pursue is the delin- nitive performance in cancer patients at baseline eation of changes in brain structure and func- (i.e., before exposure to chemotherapy) (Ahles et al., tion associated with the cognitive abnormalities 1998; Meyers et al., 2005; Wagner et al., 2006; Wefel observed following chemotherapy, changes which et al., 2004a), and a history of cancer has been may be related to risk factors for cancer and cogni- suggested as a potential risk factor for cognitive tive decline. impairment and Alzheimer’s disease in the elderly (Heflin et al., 2005). To date, chemotherapy-related cognitive changes (independent of the effects of Structural and functional cranial irradiation) have been studied most exten- neuroimaging methods sively in breast cancer patients; however, there are reports in the literature suggesting that patients Recent advances in neuroimaging technologies per- with other non-CNS cancers (e.g., lung cancer, lym- mit investigation of cognitive changes related to phoma) may also demonstrate cognitive changes cancer and chemotherapy in vivo.Examplesof related to cancer and/or chemotherapy (e.g., Ahles these brain imaging methods are briefly reviewed et al., 2002; Kanard et al., 2004; Komaki et al., here, and are illustrated in Figure 3.1. Struc- 1995; Meyers et al., 1995; Van Oosterhout et al., tural magnetic resonance imaging (MRI) can be 1996). used to provide a high-resolution picture of neu- While studies like those cited above have doc- roanatomical details and atrophy (T1-weighted umented cognitive changes that appear related to scans), and visible pathology such as microvascu- chemotherapy, the neural mechanism underlying lar and inflammatory lesions [T2-weighted scans; thesechangesisasyetpoorlyunderstood,though fluid-attenuated inversion recovery (FLAIR) scans]. several possible biological pathways have been pro- Semi-automated or manual methods can be used posed to account for cognitive changes related to segment or classify the structural images into to both chemotherapy and cancer itself (Ahles & the main brain tissue compartments (Figure 3.1a), Saykin, 2007). These include genetic factors which including gray and white matter (GM and WM) and increase risk for both cancer and cognitive impair- cerebrospinal fluid (CSF), as well as to delineate ment, and the potential interaction of these fac- hyperintense lesions, which can reflect microvas- tors with chemotherapy and hormonal cancer treat- cular changes or areas of demyelination. Volume ments. In summary, research to date suggests that and other characteristics of each tissue type can cognitive changes associated with chemotherapy then be quantitated and compared using imaging need to be examined within the broader context of and statistical software. Prior to the development of genetic and other risk factors and biological pro- MRI, computerized axial tomography (CAT or CT) cesses associated with the development of can- scanning was the predominant method of structural cer. Significant cognitive changes likely occur only brain imaging. CT methodology involves comput- in a subset of chemotherapy-treated patients as a erized integration of multiple X-ray images to gen- result of these risk factors. It may also be the case erate cross-sectional views of the brain. Due to its Chapter 3. Brain imaging investigation of chemotherapy-induced neurocognitive changes 21

(a)

(b) (c)

(d)

Figure 3.1. Neuroimaging methods relevant to cognitive changes. (a) Structural MRI (gray and white matter atrophy); (b) diffusion tensor imaging (white matter connectivity); (c) functional MRI (brain activity), and (d) PET (brain metabolism). Reprinted with permission from Ahles and Saykin (2007). See color version in color plate section. much lower intracerebral tissue contrast and the Voxel-based morphometry (VBM) is a recently exposure to radiation involved in the technique, developed method for analyzing structural MRI however, CT is much less commonly used than data to quantitatively evaluate atrophy and other MRI for most neuroimaging research related to can- changes on a voxel-by-voxel basis throughout the cer chemotherapy, though CT remains the optimal entire brain (Ashburner & Friston, 2000, 2001; Good method for some purposes (e.g., imaging of bone or et al., 2001). Unlike the above-noted morphological acute hemorrhage). methods that involve manual segmentation of 22 Section 1. Cognition and the brain

selected structures, VBM is a fully automated pro- neuroimaging, as yet no published reports have cedure for examining tissue integrity, providing the used this technique to examine the potential effects ability to assess regional volume and density of of chemotherapy independent of radiation ther- brain tissue compartments. VBM utilizes statisti- apy. However, DTI appears to have a great deal of cal parametric mapping procedures similar to those promise for closer study of the WM abnormalities employed for analysis of functional neuroimaging that have been demonstrated after chemotherapy data. Because VBM assesses signal intensities across using other structural neuroimaging methods. every voxel in the brain relative to a user-defined Magnetic resonance spectroscopy (MRS) utilizes a priori statistical threshold, it provides an unbi- the differing magnetic properties of biochemical ased, comprehensive, and highly reliable assess- compounds in the brain to allow graphic rep- ment of tissue volume that is sensitive to local resentation of metabolite concentration, synthe- changes. VBM has been used in a small number sis rates, and relative volumes in neural tissue, in of studies, discussed below, to investigate brain vivo and without exposure to radioactivity. Recent changes related to chemotherapy. developments in MRS allow measurement of spe- Diffusion tensor imaging (DTI) is a recently cific brain neurotransmitters, including glutamate, developed technique that capitalizes on variation glycine, and gamma-amino butyric acid (GABA), in the degree and directionality of diffusion of among others, in addition to other metabolite water molecules in different brain tissue types as markers of neuronal status and integrity such as an indicator of tissue integrity. Diffusion of water N-acetyl-aspartate (NAA), creatine, choline, and molecules in GM and CSF is largely isotropic (ran- myo-inositol. dom), but is directionally restricted by axonal mem- Functional MRI (fMRI) employs detection of branes and myelin. Therefore, in aggregate, WM increases in local signal intensity (Belliveau et al., fiber bundles are normally highly non-random in 1992; Kwong et al., 1992) to assess the activation diffusion characteristics. By measuring the degree of cortical and subcortical regions during the per- and orientation of anisotropy of diffusion, DTI can formance of cognitive or sensorimotor tasks in the demonstrate the directionality of fiber tracts (Figure scanner [Figure 3.1c; for details of key fMRI meth- 3.1b) (Le Bihan et al., 2001), and can demon- ods, see Bandettini & Wong (1997), Rosen et al. strate neuroanatomic connectivity of fiber path- (1998), and Moonen and Bandettini (2000)]. The dif- ways among brain regions involved in a particular fering magnetic susceptibilities of oxyhemoglobin network (Basser et al., 1994; Pierpaoli et al., 1996), (diamagnetic) and deoxyhemoglobin (paramag- thus showing great promise for investigating sub- netic) permit deoxyhemoglobin to act as an endoge- types of WM pathology. The integrity of WM path- nous contrast agent sensitive to blood oxygenation ways is also quantitatively indicated by the degree (Ogawa et al., 1998). This has become known as of anisotropy, while pathological changes in GM blood oxygen level dependent (BOLD) contrast. It can be detected by examining differences in mean has been hypothesized that microvascular changes diffusivity of tissue. Further, DTI can be combined are specific to the dynamic activity of local neural with functional MRI (see below) to relate anatomi- circuits, and recent evidence confirms this model cal connectivity to functional brain activation pat- (Logothetis et al., 2001). The change in signal inten- terns related to cognitive or motor tasks, and to sity is induced by local field gradients which result recovery from brain injury or insult (Werring et al., from the intravascular compartmentalization of 1998, 1999). Therefore, DTI shows great promise for the endogenous contrast agent (Lai et al., 1993). investigating subtypes of structural brain pathol- The onset of detectable change in signal inten- ogy, particularly in WM. Although DTI has been sity is time dependent and occurs maximally at used to document abnormalities in other clinical 5–8 s post-activation, with drop off approximately populations where WM disease is a hallmark on 5–9 s post cessation of stimulation (Bandettini Chapter 3. Brain imaging investigation of chemotherapy-induced neurocognitive changes 23

et al., 1993). Fast acquisition capability can take have not yet been examined in systematic, prospec- advantage of this rapid change in blood flow. Neu- tive studies of chemotherapy-induced or cancer- ronal activity and local cerebral blood flow, volume, associated cognitive changes, but these approaches and parenchymal oxygenation are normally tightly hold promise for identifying the neural bases of such coupled, but this may be altered by disease pro- changes. While the available neuroimaging litera- cesses and possibly aging (D’Esposito et al., 1999; ture examining chemotherapy effects is somewhat Johnson et al., 2001; Ross et al., 1997). limited, some relevant clinical observations have In contrast to the MRI-based measures described been published. above, molecular imaging methods such as positron emission tomography (PET) use radiotracers to pro- MRI studies of childhood leukemia vide data on cerebral blood flow or specific neurotransmitter/receptor systems (Figure 3.1d). PET Structural MRI has been demonstrated to be sen- utilizes short-lived radioisotopes to examine brain sitive to abnormalities in children treated with function either at rest or during task performance, chemotherapy for acute lymphoblastic leukemia most commonly by measuring levels of blood flow (ALL) (for review, see Reddick et al., 2007). While (e.g., with 15O-labeled water) or glucose metabolism considerable work has examined the effects of (e.g., using 18F-fluorodeoxyglucose, or FDG). More chemotherapy in conjunction with RT, a grow- recent applications allow targeted examination of ing number of neuroimaging studies have also specific neurotransmitter systems, including the examined the independent effects of chemother- dopaminergic, cholinergic, and serotonergic sys- apy by studying children treated for ALL who tems. Tracers have likewise been developed to study did not receive RT. Ciesielski, Lesnik and col- receptor binding for opioids and benzodiazepines. leagues (Ciesielski et al., 1999; Lesnik et al., 1998) A highly promising area is the development of used MRI morphometry to study 10 children who targeted PET probes for molecular pathology of underwent neurotoxic intrathecal chemotherapy disease, such as amyloid imaging for Alzheimer’s with methotrexate but not RT for ALL before age disease and other dementias (e.g., Klunk et al., five. In a study focused on the development of 2004). brain regions important for memory, Ciesielski et al. (1999) found significant volume reductions in the mammillary bodies and prefrontal cortices Neuroimaging of cognitive changes related (PFC), and non-significant reductions in the cau- to cancer and chemotherapy date nuclei in ALL patients relative to healthy controls. In another study using the same patient Toga and Mazziotta (1996) provide an overall review and control groups, this team found smaller vol- of structural and functional neuroimaging method- umes in the PFC and cerebellar lobuli VI–VII, and ologies and their application to brain disorders neuropsychological deficits in visuospatial atten- (Mazziotta et al., 2000). Very limited systematic tion, short-term memory, and visuomotor organi- neuroimaging research has been conducted with zation and co-ordination, implicating cerebellar- patients undergoing chemotherapy for non-brain frontal system alterations in the cognitive changes cancers in the absence of radiation therapy (RT), following chemotherapy in this population (Lesnik with much of the available literature reflecting et al., 1998). case series, convenience samples, or retrospec- Harila-Saari et al. (1998) found MRI abnormal- tive studies. In addition, the literature is largely ities in 8 of 32 patients 5 years after RT and/or restricted to structural neuroimaging studies, with chemotherapy for ALL. Only 2 of 15 patients who only a few studies utilizing functional neuroimag- received chemotherapy without RT demonstrated ing techniques. Functional neuroimaging methods abnormalities (WM changes) however, and one of 24 Section 1. Cognition and the brain

these subjects had been born prematurely, perhaps Reddick et al. (2006) also examined the relation- indicating a predisposing factor for WM change. ship between WM volume and neurocognitive per- Kingma et al. (2001) studied 17 ALL patients treated formance within the ALL patients, and found sig- with chemotherapy only with neuropsychological nificant inverse correlations for nearly all measures assessment and MRI, and compared this group to examined, including estimated intellect, academic children who also received RT and to healthy con- achievement, and attentional functioning. These trols. While structural MRI abnormalities were less findings suggest that chemotherapy and RT likely frequent in the chemotherapy-only group relative to have both independent and interacting negative patients who had also received RT (38% vs. 63%), effects in terms of structural brain changes and cog- definite abnormalities were noted in 3 of 16 chil- nitive performance following treatment. dren with available MRI scans who were treated with chemotherapy without RT, with probable abnormal- Neuroimaging in osteosarcoma ities noted in another 3 cases. These abnormalities reportedly included WM changes and/or atrophy. In Early case reports of radiologic abnormalities fol- this small sample, the presence of an MRI abnor- lowing chemotherapy for osteosarcoma utilized mality was unrelated to other cognitive or academic CT scanning. Packer et al. (1983) reported the first variables for either ALL group. case of abnormal CT findings in a child treated with Most recently, Reddick et al. (2006) studied a high-dose methotrexate followed by citrovorum large cohort of children who had been treated for rescue in 1983. Another early case study also noted ALL with chemotherapy alone (n = 84) or with transient encephalopathy and CNS toxicity related chemotherapy and RT (n = 28), in comparison to to high-dose methotrexate treatment for metastatic healthy sibling controls (n = 33). Neurocognitive osteosarcoma (Fritsch & Urban, 1984), with accom- testing demonstrated significant differences rela- panying CT abnormalities including periventricular tive to normative performance expectations in both hypodensity, particularly around the frontal horns, cancer groups. Patients treated with RT showed the as well as a hypodense left temporal lesion. These most significant declines, with cognitive impair- abnormalities were noted to be persistent even 14 ment consistent with prior reports. Patients treated months after the acute symptom onset, though with chemotherapy alone also demonstrated stat- cognitive and neurological function was reportedly istically significant declines relative to age norms normal 5 years post-illness, aside from absent deep across most neurocognitive tasks, though level of tendon reflexes. Another case study similarly noted impairment was less severe than for patients treated the onset of leukoencephalopathy during high- with RT. For most measures, performance in the dose methotrexate treatment with calcium leuco- chemotherapy-only group was within broad nor- vorin rescue for osteosarcoma (Glass et al., 1986), mal limits; however, clinically meaningful impair- with head CT showing bilateral non-enhancing ment (>1.0 SD lower than normative mean) was symmetric hypodensities in the periventricular WM seen on measures of attentional functioning. A and centrum semiovale. In earlier related work, similar pattern was observed for brain WM vol- Allen et al. (1980) noted diffuse WM hypodensity ume on MRI. Patients treated with RT showed (five patients) and atrophic changes (five patients) the greatest reduction in WM volume, with signif- in individuals treated with high-dose methotrexate icantly smaller volumes than both other groups. with leucovorin rescue without RT for bone or The chemotherapy-only group also showed a signif- soft-tissue sarcomas who developed leukoen- icant reduction in WM volume relative to controls cephalopathy. however, consistent with the studies noted above Later studies utilized MRI to detect abnor- showing reduced brain volumes in other struc- malities in osteosarcoma patients with greater tures following chemotherapy without RT for ALL. anatomic resolution. In a sample of eight patients, Chapter 3. Brain imaging investigation of chemotherapy-induced neurocognitive changes 25

mostly adolescents, treated with single-agent three patients improved with corticosteroid therapy high-dose methotrexate without RT, Ebner et al. after chemotherapy ended. This study suggested (1989) observed brain abnormalities on CT or MRI that inflammatory mechanisms may play an impor- in four, including chronic edema, multifocal WM tant role in the pathophysiology of 5-fluorouracil necrosis, and deep atrophy. Interestingly, while two neurotoxicity, though the authors noted that they of these four became encephalopathic, the other could not rule out an effect of levamisole. Subse- two did not manifest clinically evident cognitive quently, this group of investigators (Kimmel et al., abnormalities concurrent with the noted brain 1995) reported similar MRI and clinical findings in abnormalities. Lien et al. (1991) reported MRI a patient receiving adjuvant levamisole therapy for findings in 22 patients treated for osteosarcoma malignant melanoma, which again improved after with a chemotherapy regimen which included high- discontinuation followed by corticosteroid therapy. dose methotrexate with citrovorum factor rescue, This suggests that levamisole may have been a key cisplatin, doxorubicin HCl, bleomycin, cyclophos- factor in the previously reported cases. phamide, and actinomycin D. Ten patients who received cisplatin-based treatment for testicular Structural and functional MRI studies of cancer were used as a control group. In the osteosar- chemotherapy-related cognitive changes coma group, 14 patients showed WM lesions on in breast cancer T2-weighted imaging, which appeared related to time since treatment, as they were observed in 12 of Studies that have incorporated neuroimaging tech- 14 subjects whose MRI scan occurred within 2 years niques have reported structural and functional of chemotherapy, but in only 2 of 8 patients who changes in the brain associated with chemother- had later MRI studies. Lesions were bilateral in all apy (Saykin et al., 2003b; Stemmer et al., 1994). A but one patient, and were most commonly adjacent reduction in the volume of brain structures impor- to the lateral ventricle, though they were also noted tant for cognitive functioning (such as the frontal in the centrum semiovale and corpus callosum. No cortex) and changes in the integrity of WM tracts MRI abnormalities were observed in the testicular that connect brain structures have been associated cancer group, suggesting that cisplatin was unlikely with changes in cognitive functioning, and have to be the major agent responsible for such changes, been seen using structural MRI in patients after and arguing that high-dose methotrexate was the chemotherapy. Stemmer et al. (1994) reported WM most probable cause, though an interactive effect changes in 9 of 13 breast cancer patients (stage II– with cisplatin could not be ruled out. In this sample, IV) following treatment with high-dose cyclophos- no CNS symptoms were apparent, though detailed phamide, cisplatin, carmustine, and autologous neurological and psychiatric examinations were not bone marrow support. Based on neuroradiologi- conducted. cal ratings, four of these patients had severe WM changes, four had moderate changes, and one had mild changes. It is important to recognize that – Other MRI reports of chemotherapy-induced despite at times extensive structural brain abnor- leukoencephalopathy in adults malities, particularly of the WM – many of the stud- Development of multifocal gadolinium-enhancing ies cited in this chapter note that these changes inflammatory brain WM lesions has been reported appear to be “clinically silent,” with no noted after treatment with adjuvant 5-fluorouracil and lev- deficits on neurological examination or cognitive, amisole for adenocarcinoma of the colon (Hook psychological, or functional complaints reported by et al., 1992). Biopsy in two of three cases was con- their study participants. sistent with active demyelination, with axonal spar- In a very recent study using VBM in female breast ing and perivascular lymphocytic inflammation. All cancer survivors, Inagaki et al. (2007) compared 26 Section 1. Cognition and the brain

patients who had been treated with chemotherapy C+ than the C− group) in both studies, and the (C+)tothosewhohadnot(C−), and to a healthy possible presence of baseline cognitive deficits control group. Groups were compared at two time in the cancer patients was not taken into con- points, labeled 1 and 3 years post-chemotherapy, sideration. The rationale behind some of the though it should be noted that the mean post- comparisons presented by Inagaki et al. (2007) was treatment interval for the “1 year” sample was also unclear. It does not appear that the C+ and about4months,whilethemeanforthe“3-year” C− groups were independently compared to the sample was about 3.25 years. The samples were par- control group, which would have been inform- tially overlapping, but the study was not longitu- ative. It is also unclear whether correlational anal- dinal as such. At the 1-year time point, the C+ yses between cognitive performance and brain group was found to demonstrate decreased GM volume were done across groups, or just within the andWMvolumerelativetotheC− group, includ- C+ group. Attention to such issues will be important ing prefrontal, parahippocampal, cingulate gyrus, in future research. and precuneus regions. No significant intergroup In contrast to the findings from Inagaki et al. (C+ vs. C−) differences were apparent at the 3-year (2007), VBM studies in our laboratory have demon- time point. Despite the finding of smaller regional strated smaller regional brain volumes in breast brain volumes in the C+ relative to the C− group cancer patients who received chemotherapy com- at the 1-year time point, when all cancer patients pared to those who did not and to healthy con- were compared to healthy controls, no significant trol participants. In a study of long-term (>5years volume differences were apparent at either time post-diagnosis) breast cancer and lymphoma sur- point. Correlational analyses demonstrated a sig- vivors, we utilized VBM to study 12 women treated nificant relationship between prefrontal, parahip- with chemotherapy compared to 12 demographi- pocampal, and precuneus volumes and indices cally matched healthy control subjects (Saykin et al., of attention-concentration and/or visual memory 2003a, 2003b). Relative to controls, the C+ group from the Wechsler Memory Scale, Revised (WMS-R) showed local bilateral reduction of neocortical GM at the 1-year time point. and cortical and subcortical WM volume in several The lack of significant differences between C+ regions. There were no brain regions in which con- and C− groups at the 3-year time point was consis- trol subjects demonstrated volume reduction rela- tent with prior work by this group (Yoshikawa et al., tive to cancer patients. These results provided pre- 2005), in which they compared hippocampal vol- liminary evidence of distributed structural changes umes obtained using a manual tracing technique in GM and WM many years after treatment with between C+ and C− patients who were all more chemotherapy, in a relatively diffuse neuroanatomic than 3 years post-surgery, and found no between- pattern generally consistent with the relatively dif- group differences. They likewise found no group diffuse profile of neuropsychological declines noted ferences in WMS-R memory performance, though above and described elsewhere in this volume. attention-concentration performance was signifi- Our laboratory is currently undertaking a longitu- cantly weaker in the C+ group. dinal, prospective study of the neural mechanisms As noted by the authors themselves and in of cognitive changes related to breast cancer and other commentary, there are methodological its treatment, utilizing a comprehensive assessment issues that affect interpretation of these results. battery including neurocognitive evaluation, struc- As noted by Eichbaum et al. (2007), the majority tural and functional MRI, and blood biomarker and of patients in the Inagaki et al. (2007) study did genotyping analysis. Breast cancer patients treated not receive a recommended standard chemother- with (C+)andwithout(C−) chemotherapy are apy regimen. Groups were also confounded by being compared to healthy control subjects prior hormonal treatment status (significantly more to adjuvant treatment and 1 and 12 months after patients received hormonal treatment in the completion of chemotherapy (or yoked intervals). Chapter 3. Brain imaging investigation of chemotherapy-induced neurocognitive changes 27

Preliminary data (Saykin et al., 2007) have shown a 60-year-old, right-handed sisters participated pattern of reduced brain activation in frontal areas in a comprehensive battery of cognitive assessment on fMRI during a working memory task 1 month and structural and fMRI measures, and completed after chemotherapy, suggestive of dysfunction in detailed self-report inventories. While cogni- circuitry crucial for normal working memory func- tive functioning was not meaningfully different tioning. Unpublished data from VBM analysis of the between the twins as measured by standardized structural MRI data in this cohort have indicated neuropsychological tests, Twin A, who was treated decreased frontotemporal GM 1 month after breast with adjuvant chemotherapy and hormone therapy cancer chemotherapy relative to pre-chemotherapy for stage II breast cancer, self-reported a much baseline in C+ patients relative to C− patients and greater level of cognitive concerns than Twin B, who to healthy controls. These data suggest that struc- had never had breast cancer. MRI measures showed tural and functional changes in the brain can be notable differences between the two subjects. Twin detected quite soon after systemic chemotherapy; A showed a higher volume of WM hyperintensities analysis of the data from the 12-month assessment on FLAIR imaging (9800.68 mm3 vs. 6241.11 mm3) will help determine the natural course of these than Twin B; these hyperintensities were read as of changes over time. uncertain clinical significance by the study neuro- Another fMRI study of breast cancer patients radiologist. Twin A also demonstrated an expanded is noted in a recent review article (Castellon spatial extent of activation during working memory et al., 2005), though it does not as yet appear to processing on an auditory-verbal “n-back” fMRI have been published independently. In this study paradigm, despite comparable task performance (Wagner et al., 2004), ten cancer patients, most to her sister. These findings illustrate a potential of whom reportedly had breast cancer and who explanation for the commonly seen discrepancy were identified as demonstrating cognitive impair- between self-reported cognitive symptoms follow- ment on formal neuropsychological testing, were ing chemotherapy and lack of objective findings compared to a group of demographically matched of cognitive impairment on formal neuropsycho- healthy control participants. Cancer patients logical assessment. The expanded extent of brain were reportedly within 6 months of completion activation demonstrated in Twin A – which is similar of chemotherapy, and demonstrated deficits in to that we have seen in other clinical populations, three or more of the neuropsychological domains including those with multiple sclerosis, trau- evaluated, including cognitive efficiency, working matic brain injury, and mild cognitive impairment memory, visuospatial skills, and delayed mem- (McAllister et al., 2001; Saykin et al., 2004; Wishart ory. All subjects were right-handed. Participants et al., 2004) – may reflect compensatory recruit- completed a blocked, visual, non-verbal n-back ment of additional brain regions in order to paradigm, with 0-, 1-, and 2-back conditions. perform the task successfully. This may be per- In contrast to other functional neuroimaging ceived by patients as increased task effort, or as studies, this group of patients reportedly did not tasks becoming more difficult than they were demonstrate differences in task performance previously. Data from the ongoing longitudinal or brain activation relative to controls, with study noted above will aid in coming to a more the exception of greater activation in the con- detailed understanding of the neural substrate of trol group in the right cerebellum and dentate cognitive changes related to breast cancer and its nucleus (task condition unspecified in the available treatment. citations). As yet only one published study has examined In a case report (Ferguson et al., 2007), we neural effects of chemotherapy using PET (Silver- describe differences in structural and functional man et al., 2007). In this study, 16 women who MRI data in identical twins discordant for breast were 5–10 years post breast cancer chemother- cancer and chemotherapy treatment. These apy were compared to a concurrent sample of 8 28 Section 1. Cognition and the brain

controls (5 non-cancer healthy controls, 3 women after induction chemotherapy. WM changes were with a history of breast cancer but no chemother- apparent in one of these patients 2 months after apy treatment), and to a previously acquired HCD/AHPCS. At 3 months and beyond, three of sample of 10 healthy controls. 15O-water PET was four patients remaining in the study showed an used to evaluate blood flow related to memory pro- increasing volume of WM changes that stabilized cessing, while 18F-FDG PET was used to examin the 6- to 12-month post-treatment phase. Max- ine resting cerebral metabolism. During short-term imal volumes of abnormal WM ranged from 73 to verbal recall, modulation of blood flow in spe- 166 cm3. Despite the clear WM abnormalities, few cific frontal and cerebellar regions was significantly neurochemical changes were detected by MRS, altered in the chemotherapy-treated group rela- although the ratio of NAA to creatine (Cr) suggested tive to controls, most significantly in the left infe- a transient treatment-related decrease (Brown et al., rior frontal gyrus (LIFG). In the chemotherapy- 1998). In an earlier post-treatment study, this group treated patients only, LIFG resting metabolism (Brown et al., 1995) compared MRI and MRS in 13 was observed to correlate directly with perfor- patients undergoing bone marrow transplant for mance on a short-term memory task previously advanced breast carcinoma relative to 13 controls. found by this group (Castellon et al., 2004) to be Extensive HDC-induced WM changes were meas- impaired in chemotherapy-treated patients. Exam- ured in 10 of 13 patients, with an average volume of ination of the effects of hormone treatment on abnormal WM of 49 cm3. NAA/Cr and NAA/choline cerebral metabolism demonstrated that patients ratios were not abnormal despite these promi- treated with chemotherapy and tamoxifen (11 of nent late-stage structural changes, leading the the 16 studied) showed significantly decreased basal authors to conclude that chemotherapy-induced ganglia metabolism relative to those who received WM disease is predominantly a water space and chemotherapy but did not receive tamoxifen, or possibly an extra-neuronal process rather than a those not treated with chemotherapy. These find- primary neuronal/axonal disease. These authors ings offer further evidence of alterations in brain also emphasized the complementary nature of function related to breast cancer treatment, and information from MRS and MRI in understanding highlight the importance of further study of the the pathophysiology of chemotherapy effects. independent and interactive effects of cytotoxic chemotherapy and hormonal treatment. Proton MRS can reveal neurochemical changes in Conclusion brain cellular metabolism that appear to be highly relevant for understanding pathophysiological At present, the data regarding the cognitive effects changes after chemotherapy. Relevant data from of cancer chemotherapy are somewhat lim- several small studies have been reported. Brown ited. Most studies to date have been conducted et al. (1998) prospectively attempted to determine post-treatment, and many are confounded by con- thetimecoursefordevelopmentofWMchanges comitant cranial radiation therapy. The available induced by high-dose chemotherapy (HDC). evidence suggests that multiple neuroimaging Advanced (stage II–IV) breast cancer patients modalities are sensitive to the effects of cancer (n = 8) were studied with serial MRI and MRS treatment on brain structure and function. There is before chemotherapy throughout the 12 months evidence for both GM and WM structural changes, after treatment [carmustine, cyclophosphamide, as well as altered metabolic and activation profiles and cisplatin, with autologous hematopoietic on functional neuroimaging. Systematic research is progenitor cell support (AHPCS)]. MRI appeared needed to determine which neuroimaging modali- normal in all eight subjects at baseline, and in ties are most sensitive and specific for chemother- all six patients for whom scans were available apy, hormonal therapy, and radiation effects, as well Chapter 3. Brain imaging investigation of chemotherapy-induced neurocognitive changes 29

as interactions among these interventions. Further Bandettini PA, Wong EC (1997). Magnetic resonance imag- integration of genetics and other biomarkers is also ing of human brain function. Principles, practicali- likely to yield important new information. Finally, ties, and possibilities. Neurosurg Clin North Am 8(3): structural and functional neuroimaging can be 345–371. expected to play a role in evaluating interventions Bandettini PA, Jesmanowicz A, et al. (1993). Processing strategies for time-course data sets in functional MRI of such as cognitive rehabilitation and medication for the human brain. Magn Res Med 30(2): 161–173. treatment of cognitive changes related to cancer Basser PJ, Mattiello J, et al. (1994). MR diffusion tensor and its treatment. spectroscopy and imaging. Biophys J 66(1): 259–267. Belliveau JW, Kwong KK, et al. (1992). Magnetic resonance imaging mapping of brain function. Human visual cor- ACKNOWLEDGMENTS tex. Invest Radiol 27(Suppl 2): S59–S65. Brezden CB, Phillips KA, et al. (2000). Cognitive function in This work was supported by grants from the National breast cancer patients receiving adjuvant chemotherapy. Cancer Institute (5R01CA087845 & 5R01CA101318) J Clin Oncol 18(14): 2695–2701. and a National Institutes of Health Roadmap Grant BrownMS,SimonJH,et al. (1995). MR and proton spec- (U54EB005149). troscopy of white matter disease induced by high- dose chemotherapy with bone marrow transplant in advanced breast carcinoma. AJNR Am J Neuroradiol REFERENCES 16(10): 2013–2020. Brown MS, Stemmer SM, et al. (1998). White matter disease Ahles TA, Saykin AJ (2002). Breast cancer chemotherapy- induced by high-dose chemotherapy: longitudinal study related cognitive dysfunction. Clin Breast Cancer 3(3): with MR imaging and proton spectroscopy. AJNR Am J 84–90. Neuroradiol 19(2): 217–221. Ahles TA, Saykin AJ (2007). Candidate mechanisms for Castellon SA, Ganz PA, et al. (2004). Neurocognitive per- chemotherapy-induced cognitive changes. Nature Rev formance in breast cancer survivors exposed to adjuvant 7(3): 192–201. chemotherapy and tamoxifen. J Clin Exp Neuropsychol Ahles TA, Silberfarb PM, et al. (1998). Psychologic and 26(7): 955–969. neuropsychologic functioning of patients with limited Castellon SA, Silverman DHS, et al. (2005). Breast cancer small-cell lung cancer treated with chemotherapy and treatment and cognitive functioning: current status and radiation therapy with or without warfarin: a study by the future challenges in assessment. Breast Cancer Res Treat Cancer and Leukemia Group B. J Clin Oncol 16(5): 1954– 92(3): 199–206. 60. Ciesielski KT, Lesnik PG, et al. (1999). MRI morphometry of Ahles TA, Saykin AJ, et al. (2002). Neuropsychologic impact mammillary bodies, caudate nuclei, and prefrontal cor- of standard-dose systemic chemotherapy in long-term tices after chemotherapy for childhood leukemia: multi- survivors of breast cancer and lymphoma. J Clin Oncol variate models of early and late developing memory sub- 20(2): 485–493. systems. Behav Neurosci 113(3): 439–450. Allen JC, Rosen G, et al. (1980). Leukoencephalopa- D’Esposito M, Zarahn E, et al. (1999). The effect of nor- thy following high-dose iv methotrexate chemotherapy mal aging on the coupling of neural activity to the bold with leucovorin rescue. Cancer Treat Rep 64(12): 1261– hemodynamic response. NeuroImage 10(1): 6–14. 73. Donovan KA, Small BJ, et al. (2005). Cognitive func- Anderson-Hanley C, Sherman ML, et al. (2003). Neuropsy- tioning after adjuvant chemotherapy and/or radiother- chological effects of treatments for adults with cancer: a apy for early-stage breast carcinoma. Cancer 104(11): meta-analysis and review of the literature. J Int Neuropsy- 2499–2507. chol Soc 9(7): 967–982. Ebner F, Ranner G, et al. (1989). MR ﬁndings in Ashburner J, Friston KF (2000). Voxel-based morphometry methotrexate-induced CNS abnormalities. Am J – the methods. NeuroImage 11(6 Pt 1): 805–821. Roentgenol 153(6): 1283–1288. Ashburner J, Friston KJ (2001). Why voxel-based morphom- Eichbaum MHR, Schneeweiss A, et al. (2007). Smaller etry should be used. NeuroImage 14(6): 1238–1243. regional volumes of gray and white matter demonstrated 30 Section 1. Cognition and the brain

in breast cancer survivors exposed to adjuvant Kingma A, van Dommelens RI, et al. (2001). Slight cognitive chemotherapy. Cancer 109(1): 146–156. impairment and magnetic resonance imaging abnor- Ferguson RJ, Ahles TA (2003). Low neuropsychologic malities but normal school levels in children treated for performance among adult cancer survivors treated acute lymphoblastic leukemia with chemotherapy only. with chemotherapy. Curr Neurol Neurosci Rep 3: 215– J Pediatr 139(3): 413–420. 222. Klunk WE, Engler H, et al. (2004). Imaging brain amyloid in Ferguson RJ, McDonald BC, et al. (2007). Brain structure Alzheimer’s disease with Pittsburgh Compound-B. Ann and function differences in monozygotic twins: pos- Neurol 55(3): 306–319. sible effects of breast cancer chemotherapy. J Clin Oncol Komaki R, Meyers CA, et al. (1995). Evaluation of cogni- 25(25): 3866–3870. tive function in patients with limited small cell lung can- Fritsch G, Urban C (1984). Transient encephalopathy cer prior to and shortly following prophylactic cranial during the late course of treatment with high-dose irradiation. Int J Radiat Oncol, Biol Physics 33(1): 179– methotrexate. Cancer 53(9): 1849–1851. 182. Glass JP, Lee YY, et al. (1986). Treatment-related leukoen- Kwong KK, Belliveau JW, et al. (1992). Dynamic magnetic cephalopathy. A study of three cases and literature resonance imaging of human brain activity during pri- review. Medicine 65(3): 154–162. mary sensory stimulation. Proc Nat Acad Sci USA 89(12): Good CD, Johnsrude IS, et al. (2001). A voxel-based mor- 5675–5679. phometric study of ageing in 465 normal adult human Lai S, Hopkins AL, et al. (1993). Identification of vascular brains. NeuroImage 14: 21–36. structures as a major source of signal contrast in high Harila-Saari AH, Paakko EL, et al. (1998). A longitudinal resolution 2D and 3D functional activation imaging of magnetic resonance imaging study of the brain in sur- the motor cortex at 1.5T: preliminary results. Magn Reson vivors in childhood acute lymphoblastic leukemia. Can- Imaging Med 30: 387–392. cer 83(12): 2608–2617. Le Bihan D, Mangin JF, et al. (2001). Diffusion tensor imag- Heflin LH, Meyerowitz BE, et al. (2005). Re: Cancer as a risk ing: concepts and applications. J Magn Reson Imaging factor for dementia: a house built on shifting sand. JNat 13(4): 534–546. Cancer Inst 97(20): 1550–1551. Lesnik PG, Ciesielski KT, et al. (1998). Evidence for Hook CC, Kimmel DW, et al. (1992). Multifocal inflamma- cerebellar-frontal subsystem changes in children treated tory leukoencephalopathy with 5-fluorouracil and lev- with intrathecal chemotherapy for leukemia: enhanced amisole. Ann Neurol 31(3): 262–267. data analysis using an effect size model. Arch Neurol Inagaki M, Yoshikawa E, et al. (2007). Smaller 55(12): 1561–1568. regional volumes of brain gray and white matter Lien HH, Blomlie V, et al. (1991). Osteogenic sarcoma: demonstrated in breast cancer survivors exposed MR signal abnormalities of the brain in asymptomatic to adjuvant chemotherapy. Cancer 109(1): 146– patients treated with high-dose methotrexate. Radiology 156. 179(2): 547–550. Jenkins V, Shilling V, et al. (2006). A 3-year prospective Logothetis NK, Pauls J, et al. (2001). Neurophysiological study of the effects of adjuvant treatments on cognition investigation of the basis of the fMRI signal. Nature 412: in women with early stage breast cancer. Br J Cancer 150–157. 94(6): 828–834. Mazziotta JC, Toga AW, et al. (2000). Brain Mapping: The Johnson SC, Saykin AJ, et al. (2001). Similarities and differ- Disorders. San Diego, CA: Academic Press. ences in semantic and phonological processing with age: McAllister TW, Sparling MB, et al. (2001). Differential work- patterns of functional MRI activation. Aging, Neuropsy- ing memory load effects after mild traumatic brain chol Cogn 8(4): 307–320. injury. NeuroImage 14(5): 1004–1012. Kanard A, Frytak S, et al. (2004). Cognitive dysfunction in Meyers CA, Byrne KS, et al. (1995). Cognitive deficits in patients with small-cell lung cancer: incidence, causes, patients with small cell lung cancer before and after and suggestions on management. J Support Oncol 2(2): chemotherapy. Lung Cancer 12(3): 231–235. 127–132. Meyers CA, Albitar M, et al. (2005). Cognitive impairment, Kimmel DW, Wijdicks EF, et al. (1995). Multifocal inflam- fatigue, and cytokine levels in patients with acute myel- matory leukoencephalopathy associated with levamisole ogenous leukemia or myelodysplastic syndrome. Cancer therapy. Neurology 45(2): 374–376. 104(4): 788–793. Chapter 3. Brain imaging investigation of chemotherapy-induced neurocognitive changes 31

Moonen CTW, Bandettini PA (eds.) (2000). Functional MRI. Silverman DHS, Dy CJ, et al. (2007). Altered fron- Berlin: Springer. tocortical, cerebellar, and basal ganglia activity in Ogawa S, Menon RS, et al. (1998). On the characteristics adjuvant-treated breast cancer survivors 5–10 years after of functional magnetic resonance imaging of the brain. chemotherapy. Breast Cancer Res Treat 103(3): 303– Annu Rev Biophys Biomol Struct 27: 447–474. 311. Packer RJ, Grossman RI, et al. (1983). High dose systemic Stemmer S, Stears J, et al. (1994). White matter changes methotrexate-associated acute neurologic dysfunction. in patients with breast cancer treated with high-dose Med Pediatr Oncol 11(3): 159–161. chemotherapy and autologous bone marrow support. Pierpaoli C, Jezzard P, et al. (1996). Diffusion tensor MR Am J Neuroradiol 15(7): 1267–1273. imaging of the human brain. Radiology 201(3): 637– Tannock IF, Ahles TA, et al. (2004). Cognitive impair- 648. ment associated with chemotherapy for cancer: Reddick WE, Shan ZY, et al. (2006). Smaller white- report of a workshop. J Clin Oncol 22(11): 2233– matter volumes are associated with larger deﬁcits in 2239. attention and learning among long-term survivors of Tchen N, Juffs HG, et al. (2003). Cognitive function, fatigue, acute lymphoblastic leukemia. Cancer 106(4): 941– and menopausal symptoms in women receiving adju- 949. vant chemotherapy for breast cancer. J Clin Oncol 21(22): Reddick WE, Laningham FH, Glass JO, Pui CH (2007). 4175–4183. Quantitative morphologic evaluation of magnetic reso- Toga AW, Mazziotta JC (1996). Brain Mapping: The Meth- nance imaging during and after treatment of childhood ods. San Diego, CA: Academic Press. leukemia. Neuroradiology 49: 889–904. Van Dam FSAM, Schagen SB, et al. (1998). Impairment Rosen BR, Buckner RL, et al. (1998). Event-related func- of cognitive function in women receiving adjuvant tional MRI: past, present, and future. Proc Natl Acad Sci treatment for high-risk breast cancer: high-dose versus USA 95(3): 773–780. standard-dose chemotherapy. J Nat Cancer Inst 90(3): Ross MH, Yurgelun-Todd DA, et al. (1997). Age-related 210–218. reduction in functional MRI response to photic stimula- Van Oosterhout AG, Ganzevles PG, et al. (1996). Sequelae in tion. Neurology 48(1): 173–176. long-term survivors of small cell lung cancer. Int J Radiat Saykin AJ, Ahles TA, et al. (2003a). Mechanisms of Oncol Biol Physics 34(5): 1037–1044. chemotherapy-induced cognitive disorders: neu- Wagner LI, Sweet JJ, et al. (2004). Performance of Cog- ropsychological, pathophysiological, and neuroimaging nitively Impaired Oncology Patients and Normal perspectives. Semin Clin Neuropsychiatry 8(4): 201– Controls on Functional MRI Tasks. American Society 216. of Clinical Oncology Annual Meeting, New Orleans, Saykin AJ, Ahles TA, et al. (2003b). Gray matter reduction LA. on voxel-based morphometry in chemotherapy-treated Wagner L, Sweet J, et al. (2006). Trajectory of cognitive cancer survivors. J Int Neuropsychol Soc 9: 246. impairment during breast cancer treatment: a prospec- Saykin AJ, Wishart HA, et al. (2004). Cholinergic enhance- tive analysis. JClinOncolSuppl24: 8500. ment of frontal lobe activity in Mild Cognitive Impair- Wefel JS, Lenzi R, et al. (2004a). “Chemobrain” in ment. Brain 127(7): 1574–1583. breast carcinoma?: a prologue. Cancer 101(3): 466– Saykin AJ, McDonald BC, et al. (2007). Alterations in Brain 475. Activation during Working Memory in Patients with Wefel JS, Lenzi R, et al. (2004b). The cognitive seque- Breast Cancer: Relation to Treatment Modality, Cancer lae of standard-dose adjuvant chemotherapy in Status and Task Performance. Available on CD-Rom in women with breast carcinoma: results of a prospec- NeuroImage, Chicago, IL. tive, randomized, longitudinal trial. Cancer 100: 2292– Schagen SB, van Dam FSAM, et al. (1999). Cognitive deﬁcits 2299. after postoperative adjuvant chemotherapy for breast Werring DJ, Clark CA, et al. (1998). The structural and carcinoma. Cancer 85(3): 640–650. functional mechanisms of motor recovery: complemen- Shilling V, Jenkins V, et al. (2005). The effects of adjuvant tary use of diffusion tensor and functional magnetic chemotherapy on cognition in women with breast can- resonance imaging in traumatic injury of the inter- cer – preliminary results of an observational longitudinal nal capsule. J Neurol Neurosurg Psychiatry 65: 863– study. The Breast 14: 142–150. 869. 32 Section 1. Cognition and the brain

Werring DJ, Clark CA, et al. (1999). A direct demonstra- Wishart HA, Saykin AJ, et al. (2004). Brain activation pat- tion of both structure and function in the visual sys- terns associated with working memory in relapsing- tem: combining diffusion tensor imaging with functional remitting MS. Neurology 62: 234–238. magnetic resonance imaging. NeuroImage 9(3): 352–361. Yoshikawa E, Matsuoka Y, et al. (2005). No adverse effects Wieneke MH, Dienst ER (1995). Neuropsychological of adjuvant chemotherapy on hippocampal volume in assessment of cognitive functioning following chemo- Japanese breast cancer survivors. Breast Cancer Res Treat therapy for breast cancer. Psycho-Oncology 4: 61–66. 92(1): 81–84. 3.1(a)

3.1(b)

3.1(c) 3.1(d)

Avoidance Region

30 Gy 6 Gy 3 GY

IMRT with tomotherapy achieves significant dose reduction (hippocampus), while delivering 30 Gy to the rest of the brain

Figure 12.2. Hippocampus avoidance with intensity modulated radiotherapy (provided by Drs. Hazim Jaradat and Wolfgang Tome, University of Wisconsin) 4

Role of neuropsychological assessment in cancer patients

Elana Farace

Neurocognitive function is a very important issue during active treatment, and improve during inter- in cancer survivorship. When present, neurocogni- treatment intervals; however, often the opposite tive deficits explain the lion’s share of cancer sur- pattern can be seen. Patients will also differ from vivors’ reported decreased quality of life. However, one another in terms of their objective deficits at scientific study of the neuropsychological seque- any given time, and in terms of the impact of those lae of cancer is just beginning to be undertaken. deficits on their overall quality of life (QOL). A Medline search for 1996–2006 of “neuropsychol- The neuropsychology of cancer is particularly ogy or neurocognitive” and “cancer” results in only complex due to the numerous mediators that affect 86 articles. Once those are selected to include only an individual’s abilities. For example, instead of a those that include information on cancer in adults deficit occurring from a one-time injury, such as in (as opposed to pediatric cancer or adult survivors of a traumatic brain injury, the injury to the brain can pediatric cancers), written in English, only 34 papers come from a solid brain tumor (either a primary remain. However, the relative paucity of research brain tumor or a metastasis), which then varies over is in contrast to the recent attention given to this time with tumor growth and treatment. Neurocog- important topic, most recently in the Institute of nitive change may also result from changes in struc- Medicine Report From Cancer Patient to Cancer Sur- ture and function through radiation, chemotherapy, vivor, in which cognitive dysfunction is listed as one changes in hormonal status, and other factors that of the important concerns of cancer survivors after less directly affect brain function. Neurocogni- treatment (Hewitt et al., 2006). tive impairment can be masked or mimicked by Neurocognitive deficits in cancer patients are psychological phenomena such as depression, variable. When patients report having neuropsy- anxiety, and somatization. Although cancer chological impairments, they may note them as patients’ neurocognitive ability typically declines being very minor (e.g., “I’m in a fog” or “I have at some point in the cancer trajectory, the pattern a lot of ‘senior moments’”) or patients may have of decline is variable. Other patient-oriented fac- significant neurocognitive deficits that impair their tors, such as baseline intelligence, education level, ability to speak, remember, or act appropriately. mood, coping skills, and social support, are also Some patients are not aware of their own deficits likely mediators of neurocognitive ability. and only caregivers have noticed the changes. Thus, the purpose of this chapter is to famil- A cancer patient with neurocognitive dysfunc- iarize the reader with the importance of careful tion may decline during some periods, such as neuropsychological assessment throughout cancer

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

33 34 Section 1. Cognition and the brain

survivorship. Neuropsychological assessment helps patients with metastases to the brain either have or clinicians, patients, and family caregivers under- will develop multiple lesions (Sawaya et al., 1995). stand cancer disease, treatment, and survivorship Neurological and neurocognitive impairment sequelae. Neuropsychological assessment is also resulting from a metastatic lesion is very similar to extremely useful in research, whether in determin- the signs and symptoms of primary brain tumors ing the side-effects of an experimental treatment, (Cairncross et al., 1980; Hirsch et al., 1982). Men- as a primary focus of research aimed to improve tal status changes are one of the most frequent QOL, or in determining future directions in the field. symptoms of a primary brain tumor (Klein et al., Each of these areas will be discussed in turn. Ideally, 2001; Packer et al., 1998). Neurocognitive impair- elucidation of these factors will improve decision- ment in primary brain tumor patients at baseline making by patients, families, and the clinicians who is very frequent, with 91% of patients having at care for them. The ultimate goal is to maximize both least one area of deficit compared to the normal survival and QOL, that is, to help patients live a population, and 71% demonstrating at least three longer life and a better life. deficits (Tucha et al., 2000). The following sections will detail the importance of neuropsychological evaluation in the multidisciplinary care of a cancer Benefit of neuropsychological assessment patient. More complete reviews of low-grade and for clinical medical management of the high-grade gliomas can be found in Chapters 10 cancer patient and 11, respectively.

Neuropsychological changes are very common pre- Neuropsychological dysfunction at senting symptoms of a new brain tumor in a can- presentation cer patient, whether it is a primary brain tumor or a primary cancer elsewhere that has metasta- The presenting neurocognitive symptoms typically sized to the brain. The most frequent cause of depend on where the new tumor or tumors are new neuropsychological impairment in cancer sur- located (Farace et al., 1995; Meyers, 2000; Meyers vivors is metastatic brain tumors, the most com- et al., 2000b, Scheibel et al., 1996). For example, mon form of intracranial tumors in adults (Patchell, left hemisphere lesions tend to produce changes 1995). A complete review of brain metastases can be in language (Hahn et al., 2003) and lesions on the found in Chapter 12 in this volume. As treatment motor strip tend to cause seizures and/or paresis. of the primary cancer improves and length of sur- Third ventricle tumors tend to cause impairments in vival after diagnosis increases, the risk of metastases memory, executive function, and manual dexterity to the brain also increases (Carney, 1999; Chidel (interestingly, independent of the effects of surgery et al., 2000; Vermeulen, 1998). Up to 170 000 new or hydrocephalus) (Friedman et al., 2003). There- patients are diagnosed each year in the USA with fore, the onset of a new neuropsychological symp- brain metastases (Packer et al., 1998). One-half of tom can be a very important sign of new brain dis- all invasive cancers will disseminate to the brain ease, which should encourage clinicians to investi- (Vermeulen, 1998) as shown on autopsy (Cairncross gate further. et al., 1980) and 150 000–170 000 cancer patients The size and location of the brain tumor can develop symptomatic brain metastases annually also impact neurocognitive impairment, given that (Chidel et al., 2000; Vermeulen, 1998). Approxi- healthy brain tissue near the tumor may also be mately 90% of patients with a history of cancer who impacted by surgery and treatment. For example, present with a solitary brain lesion on magnetic in tumors that arise in areas with a difﬁcult sur- resonance imaging (MRI) have a brain metastasis gical approach (e.g., deep tumors) or tumors in (Patchell & Tibbs, 1990). Between 21% and 86% of “eloquent cortex” that are not very amenable to Chapter 4. Role of neuropsychological assessment in cancer patients 35

surgical resection (Laws et al., 2003; Packer et al., Cancer surveillance using neuropsychological 1998), there is a risk for the surgical approach dam- assessment aging healthy brain tissue, and the location of any Collecting “baseline” neuropsychological informa- bleeding will affect future neurocognitive status in a tion from patients either before treatment or very similar way to bleeding following a stroke (Meyers, early in their treatment course can provide an excel- 2000). Radiation fields may adversely influence nor- lent benchmark for later analysis. There is evi- mal brain tissue and corresponding neurocogni- dence that neurocognitive decline may precede a tive function; even if brain tumors are treated with change in the neuroimaging of the tumor, sug- focused radiation such as radiosurgery, the adjacent gesting that neuropsychological surveillance may normal tissue still receives a significant dose of radi- be more sensitive to tumor progression than tradi- ation. Finally, the radiation field directed to struc- tional neuroimaging. In a landmark study, Meyers tures adjacent to the brain, as is common in head and Hess (2003) performed a baseline neuropsy- and neck cancer, may overlap parts of the brain and chological assessment on 80 patients with recur- cause neurocognitive impairment (Meyers et al., rent glioblastoma or anaplastic astrocytoma prior 2000a). The biological bases of radiation-induced to beginning a clinical trial for recurrent/relapsed brain injury can be found in Chapter 7. tumor. The patients were then assessed prior to each Tumor location is not the only predictor of neu- cycle of chemotherapy and an MRI was obtained rocognitive deficit. Rate of tumor growth is a pre- at the same time. The median time for patients to dictor of the degree of neurocognitive impairment – deteriorate cognitively on any of the nine assess- slow tumor growth may displace normal tissue ments was 7.4 weeks (95% confidence interval, 5.0– slowly enough so as not to displace function too 9.6); 61% of patients declined neurocognitively prior drastically (Anderson et al., 1990) whereas rapid to radiographic progression and 25% declined at the tumor growth will cause more severe deficits (Hom same time as radiographic progression (4% declined & Reitan, 1984). In a comparison of neuropsycho- after progression and 11% did not change neurocog- logical deficits between patients with brain tumors nitively). Figure 4.1 shows the time to neurocogni- and patients with strokes, matched for lesion loca- tive progression versus neuroimaging progression tion, stroke patients were shown to have more for each individual patient (Meyers & Hess, 2003). severe neurocognitive deficits (Anderson et al., Neurocognitive decline has also been shown 1990). Often a patient with a new brain tumor will to predict MRI evidence of tumor progression develop brain edema, which is typically treated with in low-grade (Armstrong et al., 2003) along with glucocorticoids (Routh et al., 1994; Vecht et al., high-grade (Meyers, 2000; Meyers & Hess, 2003) 1994). However, steroid therapy has been associated gliomas and patients with brain metastases (Meyers with neurobehavioral changes such as depression, et al., 2004). Thus, neuropsychological surveil- hypomania, mood swings, anxiety, acute psychotic lance can help clinicians to determine when their reactions, and even a case of obsessive-compulsive index of suspicion for radiographic change should behavior disorder (Bick, 1983). Thus, the benefit occur (and concomitant changes in treatment be from glucocorticoids can sometimes be offset by considered). their side-effects. For any new neuropsychological symptom in a cancer patient, neuropsychological assessment can Neuropsychological side-effects of cancer help the medical clinician by diagnosing and quan- treatment tifying the new deficit, and can help to distinguish the cause of the deficit from any mediators. A dis- Neuropsychological assessment can also help the cussion of the utility of neuropsychological assess- oncologist to assess side-effects of treatment and ment beyond initial diagnosis will be explored next. to monitor for toxicity. Standard chemotherapies 36 Section 1. Cognition and the brain

worse neurocognitive impairment in breast cancer patients receiving adjuvant chemotherapy com- 50 pared to controls (Brezden et al., 2000) even after controlling for patient age, education level, and 40 menopausal status. Long-term negative neurocognitive effects of standard chemotherapy for breast 30 1st f ailed test cancer and lymphoma have also been shown (Ahles Prog ression Censored et al., 2002). An article by Ganz and colleagues (Ganz 20 et al., 2002) found that the only predictor of poor QOL 5–10 years after breast cancer diagnosis was 10 having had past systemic adjuvant chemotherapy, suggesting possible long-term effects of neurocog- 0 nitive impairment on QOL. 0 20 40 60 80 100 120 140 Radiation also may induce neurocognitive F ollo w-up time (w eeks) impairment, whether directed at the brain as in Figure 4.1. Event chart comparing time to neurocognitive the case of primary or metastatic brain tumors, failure on any of the tests () and radiographic tumor or when a radiation field overlaps the brain (e.g., progression () for each subject (y axis). “Censored” () head and neck cancers, lymphomas, etc.) (Meyers indicates that the individual had not yet progressed. From et al., 2000a). Patients with small cell lung car- Meyers and Hess (2003), with permission cinoma (SCLC) may receive prophylactic cranial irradiation (PCI) to prevent the development of for a number of different types of cancer have brain metastases. The use of PCI continues to be been related to neuropsychological deficits. For debated, as it has been shown to be associated example, cytokines, interleukin-2, and cortic- with neurocognitive impairment, particularly in the osteroids have all been suggested to be agents that neuropsychological domain of attention, within decrease neurocognitive function (Baile, 1996). 2 years of diagnosis of SCLC (van De Pol et al., 1997; Potential complications of chemotherapy include Van Oosterhout et al., 1996), and, in the presence of acute and chronic encephalopathy, cerebellar syn- concurrent chemotherapy, SCLC patients showed drome, and neuropathy. Chapter 8 of this volume significant neurocognitive impairment within discusses the effects of systemic treatment on neu- 4monthsofPCI(Ahleset al., 1998). rocognitive function, and Chapter 9 describes the Neurocognitive impairment following radiation effects of hormonal treatment on neurocognitive therapy (RT) is likely to be dose dependent, and function. may appear immediately or as much as 30 years Conventional chemotherapy for breast cancer has after completion of treatment (Keime-Guibert et al., been the most highly studied primary cancer in 1998). Radiation-induced progressive neurocogni- regards to the potential neurocognitive sequelae tive dysfunction, dementia, ataxia, and death in of treatment, colloquially referred to as “chemo- the absence of tumor recurrence have all been brain.” Several forms of adjuvant chemotherapy described (DeAngelis, 1994; DeAngelis et al., 1989; and secondary prevention such as tamoxifen have Sheline et al., 1980; Sundaresan et al., 1981). Neu- been shown to cause neurocognitive impairment, rocognitive impairment caused by RT is typified by although there is some debate as to the degree of the well-defined profile of subcortical white mat- baseline neurocognitive impairment in this group ter dysfunction similar to that seen in the subcor- (Bender et al., 2001; Brezden et al., 2000; Ganz, 1998; tical dementias (Cummings, 1990; Roman et al., Olin, 2001; Paganini-Hill & Clark, 2000; Schagen 1993) with deficits in the areas of information pro- et al., 1999; van Dam et al., 1998). One study revealed cessing, executive functioning, memory, attention, Chapter 4. Role of neuropsychological assessment in cancer patients 37

and motor co-ordination (Archibald et al., 1994; population of patients with malignant gliomas Hochberg & Slotnick, 1980; Imperato et al., 1990; (Litofsky et al., 2004). Lieberman et al., 1982; Salander et al., 1995; Scheibel When cancer patients have a structural brain et al., 1996; Surma-Aho et al., 2001; Taphoorn et al., lesion, psychiatric symptoms may depend on the 1994). Radiation necrosis may also result in a location of the tumor (Irle et al., 1994; Lezak & focal lesion, which results in neurocognitive decline O’Brien, 1988). Hecaen (1962) found that 67% of (although necrosis can often be temporarily allevi- patients with frontal lobe tumors exhibited con- ated with steroids) (Packer et al., 1998). Treatment- fused states and dementia as would be expected. related neurocognitive impairment can be discrim- Patients with temporal lobe tumors have been inated from deficits caused by the tumor itself using shown to have personality changes and mood standardized neuropsychological testing, with com- swings (Heilman et al., 1993) and patients with right parison of the pattern of deficits to the normal hemisphere tumors may also show paranoia, hal- population and pre-radiation treatment patterns lucinations, and agitation (Price & Mesulam, 1985). of neurocognitive impairment (Cummings, 1990; If the hypothalamic circuitry is disrupted, whether Meyers et al., 2000b). from the tumor, surgery, or treatment, there can be striking dysregulated behavior such as anxiety, depression, emotional lability, hypersexuality, reduced attention, memory loss, and impaired rea- Discrimination of neurocognitive from soning ability (Mechanick et al., 1986). psychological sequelae Depression is thought to negatively impact Depression is extremely common in cancer sur- patients’ scores on neuropsychological assessment vivors. Often cancer patients with metastatic or (Lezak & O’Brien, 1988), although recently this primary brain tumors develop de novo psychiatric has been debated (Arfken et al., 1999; Rohling symptoms (Lezak et al., 2004). A significant number et al., 2002). It is not uncommon for patients to of patients with low-grade infiltrating tumors and change clinically and appear to neurocognitively meningiomas are referred for psychiatric consul- worsen, but on assessment it is determined that tation before the initial diagnosis or even imaging the only change was an increase in depression. of the tumor (Packer et al., 1998). Depression is a Tumor-related neurocognitive impairment can be risk factor for treatment non-compliance (DiMatteo discriminated from depression using standardized et al., 2000) for which various psychoneuroendo- neuropsychological testing, with comparison of the crinological explanations have been proposed pattern of deficits to the normal population and (Capuron et al., 2001; Spiegel, 1996; Tashiro et al., pre-treatment patterns of neurocognitive impair- 2001). Depression has even been suggested to ment and mood (Meyers et al., 2000b). A careful predict response to chemotherapy in breast cancer neuropsychological examiner can differentiate (Walker et al., 1999). between neurocognitive impairment and depres- Depression in cancer patients has been shown sion, and make suggestions as to the best way to to predict shortened length of survival (Litofsky treat the depression. et al., 2004; Spiegel, 1996). For example, one study of depression following stem-cell transplantation Prognostic value of neuropsychological for malignancies found that depressed patients assessment had a threefold greater risk of dying than non- depressed patients, adjusting for other prognostic There is increasing evidence that baseline neu- factors (Loberiza et al., 2002). A significant relation- ropsychological function is a significant predictor of ship between depression at 3 months post-surgery length of survival in a number of patient popula- and length of survival has also been shown in a tions including those with multiple sclerosis (Peyser 38 Section 1. Cognition and the brain

et al., 1990), dementia (Jelic et al., 2000), and medi- gical phenomena (e.g., undiagnosed depression cally ill older adults (Arfken et al., 1999). Meyers and causing a “pseudo-dementia”). colleagues (2000b) have shown that neurocognitive Neuropsychological assessment should be per- status (verbal memory performance) was strongly formed in the case of any new neuropsychological and independently related to length of survival in symptom in a cancer patient, to help diagnose and patients with recurrent glioma even after account- quantify the new deficits, and to help distinguish the ing for age, Karnofsky Performance Scale (KPS), cause of the deficit from other mediators. Indeed, and time since diagnosis. An analysis of 445 cancer some major brain tumor centers are now moving patients with brain metastases also found that base- toward having a baseline neuropsychological eval- line Mini-Mental Status Exam (MMSE) scores were uation with neurocognitive surveillance as standard asignificantpredictoroflengthofsurvival,inaCox practice for all cancer patients with brain tumors. proportional hazards model adjusting for age, gender and Karnofsky Performance Scale (KPS) (Murray et al., 2000). The benefit of neurocognitive informa- Benefit of neuropsychological assessment tion added to a prognostic model has been shown in for the patient and family caregivers patients with leptomeningeal or parenchymal brain lesions (Meyers et al., 2002). Another study found Neuropsychological assessment is performed with that baseline neurological deficits were an indepen- cancer patients to help determine the differential dent predictor of survival in patients with low-grade contributions of neurological and psychological fac- glioma (Pignatti et al., 2002) although neuropsycho- tors to the patient’s function. From a patient and logical deficits were not directly tested. Therefore, family perspective, neurocognitive impairment has as clinicians are called upon to discuss individual a significant negative impact on QOL. This phe- prognosis with the patient, information on baseline nomenon has most often been shown in cancer neurocognitive status may be helpful. patients with malignant brain tumors, but the construct is very likely similar in other types of cancer when significant neurocognitive impairment is present. Benefit of neuropsychological assessment to We performed a study to better detail the relation- the clinician ships between cognitive impairment and QOL out- Thus, information on neurocognitive function from comes in a population of patients with malignant a neuropsychological assessment of the cancer brain tumor (Farace & Shaffrey, 2000). Neuropsy- patient can be beneficial to the clinician because: chological and QOL measures known to be sensitive r De novo neuropsychological symptoms in cancer to the neurocognitive effects of brain tumors were patients help alert clinicians to the possibility of given to 30 patients with malignant brain tumors, new brain disease. and patients’ caregivers were simultaneously given r Neuropsychological surveillance can be useful to measures of QOL and caregiver burden. Patients’ help predict brain tumor growth, and may be QOL was measured using the EORTC QLQ-C30 with more sensitive to changes in tumor size than stan- Brain Cancer module BCM29 (Aaronson et al., 1993) dard neuroimaging. and caregivers were given the Family QOL Tool, a r Neurocognitive changes are measurable side- 20-item measure of caregiver QOL (Ferrell et al., effects of treatments such as chemotherapy and 1993), and Caregiver Strain, which measures per- radiation which significantly impact patients’ ceived burden by caregivers (Robinson, 1983). The QOL. neuropsychological battery included measures of r A neuropsychological assessment can help dis- pre-morbid IQ, divided attention, problem-solving tinguish brain disease from other neuropsycholo- ability and cognitive flexibility, language, verbal Chapter 4. Role of neuropsychological assessment in cancer patients 39

(a) (b)

10.00 10.00

7.50

0.00 5.00 Caregiver strain Caregiver burden total 2.50 −10.00

0.00

−10.00 0.00 10.00 20.00 30.00 100 150 200 250 300 Stroop test results Grooved pegboard time

Figure 4.2. a Graph showing relationship between patients’ cognitive impairment and higher caregiver burden (strong correlations with a test of divided attention; Stroop; r =−0.56). b Graph showing the lack of a relationship between patients’ physical ability and caregiver burden, indicating that neurocognitive rather than physical impairment is a greater burden on caregivers memory, visual memory, fine motor skills, divided attention; Stroop; r =−0.56). In Figure 4.2b, psychomotor speed, and psychological distress. the lack of a relationship between patients’ physical Patients included 14 men, 16 women; mean age was ability and caregiver burden can be seen, indicat- 49.5 years (range 20–77); mean years of education ing that neurocognitive rather than physical impair- was 13 (range 6–18); 24/30 were right-handed; ment is a greater burden on caregivers. 28 were Caucasian, 2 African-American; mean In addition to the direct benefit to the medical estimated pre-morbid IQ was 105 (85–117) (Barona clinicians detailed above, the benefit of neuropsy- et al., 1984); and median time since diagnosis chological assessment is clear because the patient was 8 months. Patients’ cognitive impairment was and the family caregivers can receive individualized significantly related to poor patient QOL. Patients’ feedback on the assessment from the neuropsychol- cognitive impairment was also strongly and neg- ogist, in order to help each patient best determ- atively related to caregiver QOL and caregiver ine how to maximize QOL during survivorship. The burden, particularly when the patient’s impairment assessment will help to elucidate the patient’s psy- was in language and executive function. Patient chological strengths and weaknesses related to neu- and caregiver QOL was not correlated to measures rological dysfunction, and the patient, family, and of physical function, such as the Barthel Index (a neuropsychologist work together to understand the measure of activities of daily living), or the grooved results in an ecologically valid “real-world” perspec- pegboard (a test of fine motor skill). The strong tive. This feedback session may be as short as a relationship between cognitive impairment and few minutes for an intact patient with few ques- QOL suggests that neurocognitive outcome is an tions, or it may be as long as 1–2 h in situations important component of overall QOL. where the patient’s neuropsychological deficits are Figure 4.2a shows the significant relationship contributing to causing significant strain on the between patients’ cognitive impairment and higher family system. Therefore, as a result of undergo- caregiver burden (strong correlations with a test of ing a neuropsychological evaluation, information is 40 Section 1. Cognition and the brain

r provided to the patient and caregivers regarding the Caregiver support – does the caregiver need to patient’s: have more assistance (e.g., other family, hired r Neurocognitive profile – to detail a profile of housecleaners, home nurse aid, support groups, strengths and weaknesses and clarify what com- etc.)? r pensatory strategies (“work-around solutions”) Any other questions about patient functioning. might help (e.g., memory aids, changes to home Essentially, the neurocognitive assessment, to the environment, counseling). patient and family caregiver, is a major part of help- r Differential diagnosis – distinguishing between ing cancer patients determine how to live their lives different conditions that have similar symptoms, so as to maximize QOL, rather than how to fight the such as depression versus memory problems ver- cancer per se. sus tumor progression. r Prognosis – how much will the patient improve or decline over time? What to expect, what to be Benefit of neuropsychological information looking for, why changes occur, and how to deal to clinical cancer and survivorship research with changes. r Ability to function independently – how can we A thorough review of the use of neurocognitive maximize patients’ independence? testing in clinical trials can be found in Chap- r Rehabilitation potential – will the patient bene- ter 23. A survey of academically based Cana- fit from a referral to rehabilitation services (e.g., dian oncologists found that the majority identi- speech therapy, physical therapy, occupational fied QOL as a more appropriate endpoint than therapy, vocational therapy, etc.)? survival for future randomized controlled trials in r Ability to return to work or school – what changes cancer (Bezjak et al., 1998). The Food and Drug need to be made to get the patient back to work or Administration (FDA) will also consider neurocog- school and what should be changed in those envi- nitive endpoints in clinical trials. As part of the FDA ronments? Project on Cancer Drug Approval Endpoints, there r Need for specialized school services – does the was a meeting on January 20, 2006, on Primary student need referrals to special education, or Brain Tumor Endpoints wherein public discussion adaptations such as unlimited time on tests? and testimony was given on alternative endpoints r Ability to drive a car or operate other vehicles or in registration trials. The message of the presen- machinery (farm equipment, power tools, wood- tation was that response, freedom from radiologi- stove, oven, microwave, etc.). cal progression and other imaging-based endpoints r Legal issues – is the patient legally competent? Are may not adequately reflect QOL in brain tumor they accurately representing their function? Might patients. As a result, neurocognitive outcomes, and they benefit from an advanced directive and/or to a lesser extent patient-reported outcomes, are power of attorney? increasingly being included in clinical trials (Meyers r Disability – does the patient qualify for disability & Brown, 2006). This serves to broaden the scope of services? Referrals to social work to aid in applica- approvable endpoints and adds valuable information. tion regarding the clinical benefit of new agents. r Safety – for example is the patient safe to stay alone at home, or to care for small children, or to live by themselves? REFERENCES r Psychological follow-up – could the patient benefit from counseling or psychotherapy or support Aaronson N, Ahmedzai S, Bergman V, et al. (1993). The groups? Referrals to physicians for a trial of an European Organization for Research and Treatment of antidepressant or other medications. Cancer QLQ-C30: a quality of life instrument for use in Chapter 4. Role of neuropsychological assessment in cancer patients 41

international trials in oncology. J Natl Cancer Inst 85: symptoms in cancer patients treated with interleukin- 365–376. 2-based therapy. Psychoneuroendocrinology 26: 797– Ahles TA, Silberfarb PM, Herndon J, 2nd, et al. (1998). 808. Psychologic and neuropsychologic functioning of Carney DN (1999). Prophylactic cranial irradiation and patients with limited small-cell lung cancer treated with small-cell lung cancer [editorial; comment]. New Engl J chemotherapy and radiation therapy with or without Med 341: 524–526. warfarin: a study by the Cancer and Leukemia Group B. Chidel MA, Suh JH, Barnett GH (2000). Brain metastases: J Clin Oncol 16: 1954–1960. presentation, evaluation, and management. Clevel Clin J Ahles TA, Saykin AJ, Furstenberg CT, et al. (2002). Med 67: 120–127. Neuropsychologic impact of standard-dose systemic Cummings J (1990). Introduction. In: Cummings J (ed.) chemotherapy in long-term survivors of breast cancer Subcortical Dementia. New York, Oxford University Press. and lymphoma. J Clin Oncol 20: 485–493. DeAngelis LM (1994). Management of brain metastases. Anderson SW, Damasio H, Tranel D (1990). Neuropsycho- Cancer Invest 12: 156–165. logical impairments associated with lesions caused by DeAngelis LM, Delattre JY, Posner JB (1989). Radiation- tumor or stroke. Arch Neurol 47: 397–405. induced dementia in patients cured of brain metastases. Archibald YM, Lunn D, Ruttan LA, et al. (1994). Cognitive Neurology 39: 789–796. functioning in long-term survivors of high-grade glioma. DiMatteo MR, Lepper HS, Croghan TW (2000). Depression J Neurosurg 80: 247–253. is a risk factor for noncompliance with medical treat- Arfken CL, Lichtenberg PA, Tancer ME (1999). Cognitive ment: meta-analysis of the effects of anxiety and depres- impairment and depression predict mortality in medi- sion on patient adherence. Arch Intern Med 160: 2101– cally ill older adults. J Gerontol Series A-Biol Sci Med Sci 2107. 54: M152–M156. Farace E, Shaffrey ME (2000). Relationship of neurocog- Armstrong CL, Goldstein B, Shera D, et al. (2003). The pre- nitive impairment to QOL in malignant brain tumor dictive value of longitudinal neuropsychologic assess- patients [abstract]. J Neuropsychiatry Clin Neurosci ment in the early detection of brain tumor recurrence. 13:1. Cancer 97: 649–656. Farace E, Turkheimer E, Wilkniss S (1995). Utility of ana- Baile WF (1996). Neuropsychiatric disorders in cancer lyzing lesion location in an outcome study of traumatic patients. Curr Opin Oncol 8: 182–187. brain injury [abstract]. International Neuropsychological Barona A, Reynolds CR, Chastain R (1984). A demographi- Society. Seattle, WA. J Int Neuropsychol Soc 1. cally based index of premorbid intelligence for the WAIS- Ferrell B, Rhiner M, Rivera LM (1993). Development and R. J Consult Clin Psychol 52: 885–887. evaluation of the family pain questionnaire. J Psychoso- Bender CM, Paraska KK, Sereika SM, Ryan CM, Berga SL cial Oncol 10: 21–35. (2001). Cognitive function and reproductive hormones Friedman MA, Meyers CA, Sawaya R (2003). Neuropsycho- in adjuvant therapy for breast cancer: a critical review. logical effects of third ventricle tumor surgery. Neuro- J Pain Symptom Manage 21: 407–424. surgery 52: 791–798; discussion 798. Bezjak A, Taylor KM, Ng P, et al. (1998). Quality-of-life infor- Ganz PA (1998). Cognitive dysfunction following adjuvant mation and clinical practice: the oncologist’s perspec- treatment of breast cancer: a new dose-limiting toxic tive. Cancer Prevent Control 2: 230–235. effect? [letter; comment]. J Natl Cancer Inst 90: 182–183. Bick PA (1983). Obsessive-compulsive behavior associated Ganz PA, Desmond KA, Leedham B, et al. (2002). Quality of with dexamethasone treatment. J Nerv Mental Dis 171: life in long-term, disease-free survivors of breast cancer: 253–254. a follow-up study. J Natl Cancer Inst 94: 39–49 [erratum Brezden CB, Phillips, KA, Abdolell M, et al. (2000). Cogni- in: J Natl Cancer Inst 94:463]. tive function in breast cancer patients receiving adjuvant Hahn CA, Dunn RH, Logue PE, et al. (2003). Prospec- chemotherapy. J Clin Oncol 18: 2695–2701. tive study of neuropsychologic testing and quality-of- Cairncross JG, Kim JH, Posner JB (1980). Radiation therapy life assessment of adults with primary malignant brain for brain metastases. Ann Neurol 7: 529–541. tumors. Int J Radiat Oncol Biol Phys 55: 992–999. Capuron L, Ravaud A, Gualde N, et al. (2001). Associa- Hecaen H (1962). Clinical symptomatology in right tion between immune activation and early depressive and left hemisphere lesions. In Mountcastle VB (ed.) 42 Section 1. Cognition and the brain

Interhemispheric Relations and Cerebral Dominance in Glioma Outcomes Project. Neurosurgery 54: 358–366; Man: Baltimore, MD: Johns Hopkins University Press. discussion 366–367. Heilman KM, Bowers D, Valenstein E (1993). Emotional Loberiza FR Jr., Rizzo JD, Bredeson CN, et al. (2002). Asso- disorders associated with neurological diseases. In Heil- ciation of depressive syndrome and early deaths among man KM, Valenstein E (eds.) Clinical Neuropsychology, patients after stem-cell transplantation for malignant 3rd edn. New York: Oxford University Press. diseases. J Clin Oncol 20: 2118–2126. Hewitt M, Greenfield S, Stovall E (eds.) (2006). From Cancer Mechanick JI, Hochberg FH, Larocque A (1986). Hypotha- Patient to Cancer Survivor: Lost in Transition. Washing- lamic dysfunction following whole-brain irradiation. J ton DC: National Academies Press. Neurosurg, 65: 490–494. Hirsch FR, Paulson OB, Hansen HH, et al. (1982). Intracra- Meyers CA (2000). Quality of life of brain tumor patients. nial metastases in small cell carcinoma of the lung: corre- In Bernstein M, Berger, MS (eds.) Neuro-Oncology: The lation of clinical and autopsy findings. Cancer 50: 2433– Essentials. New York: Thieme Medical Publishers. 2437. Meyers CA, Brown PD (2006). The role and relevance of Hochberg FH, Slotnick B (1980). Neuropsychologic impair- neurocognitive assessment in clinical trials of patients ment in astrocytoma survivors. Neurology 30: 172–177. with central nervous system tumors. J Clin Oncol 24: Hom J, Reitan RM (1984). Neuropsychological correlates of 1305–1309. rapidly vs. slowly growing intrinsic cerebral neoplasms. J Meyers CA, Hess KR (2003). Multifaceted end points in Clin Neuropsychol 6: 309–324. brain tumor clinical trials: cognitive deterioration pre- Imperato JP, Paleologos NA, Vick NA (1990). Effects of treat- cedes MRI progression. Neurooncology 5: 89–95. ment on long-term survivors with malignant astrocy- Meyers CA, Geara F, Wong PF, et al. (2000a). Neurocog- tomas. Ann Neurol 28: 818–822. nitive effects of therapeutic irradiation for base of skull Irle E, Peper M, Wowra B, et al. (1994). Mood changes after tumors. Int J Radiat Oncol Biol Phys 46: 51–55. surgery for tumors of the cerebral cortex. Arch Neurol 51: MeyersCA,HessKR,YungWKA,et al. (2000b). Cognitive 164–174. function as a predictor of survival in patients with recur- Jelic V, Johansson SE, Almkvist O, et al. (2000). Quantita- rent malignant glioma. J Clin Oncol 18: 646–650. tive electroencephalography in mild cognitive impair- Meyers CA, Mehta MP, Rodrigus P, et al. (2002). Motexafin ment: longitudinal changes and possible prediction of gadolinium (MGD) delays neurocognitive progression in Alzheimer’s disease. Neurobiol Aging 21: 533–540. patients with brain metastases from lung cancer: results Keime-Guibert F, Napolitano M, Delattre JY (1998). Neu- of a randomized phase III trial. Neurooncology 4: 372. rological complications of radiotherapy and chemother- Meyers CA, Smith JA, Bezjak A, et al. (2004). Neurocognitive apy. J Neurol 245: 695–708. function and progression in patients with brain meta- Klein M, Taphoorn MJB, Heimans JJ, et al. (2001). Neurobe- stases treated with whole-brain radiation and motexafin havioral status and health-related quality of life in newly gadolinium: results of a randomized phase III trial. JClin diagnosed high-grade glioma patients. J Clin Oncol 19: Oncol 22: 157–165. 4037–4047. Murray KJ, Scott C, Zachariah B, et al. (2000). Importance LawsER,ShaffreyME,MorrisA,et al. (2003). Surgical man- of the mini-mental status examination in the treatment agement of intracranial gliomas – does radical resection of patients with brain metastases: a report from the improve outcome? Acta Neurochir Suppl 85: 47–53. Radiation Therapy Oncology Group protocol 91–04. Int Lezak MD, O’Brien KP (1988). Longitudinal study of emo- J Radiat Oncol Biol Phys 48: 59–64. tional, social, and physical changes after traumatic brain Olin JJ (2001). Cognitive function after systemic therapy for injury. J Learn Disabil 21:456–463. breast cancer. Oncology (Williston Park) 15: 613–618; dis- Lezak MD, Howieson DB, Loring DW (2004). Neuropsycho- cussion 618: 621–624. logical Assessment (4th edn.). New York: Oxford Univer- Packer RJ, Miller DC, Shaffrey MS, et al. (1998). Intracranial sity Press. neoplasms. In Rosenberg RN, Pleasure DE (eds.) Com- Lieberman AN, Foo SH, Ransohoff J, et al. (1982). Long term prehensive Neurology (2nd edn.) New York: John Wiley & survival among patients with malignant brain tumors. Sons. Neurosurgery 10: 450–453. Paganini-Hill A, Clark LJ (2000). Preliminary assessment of Litofsky NS, Farace E, Anderson F Jr., et al. (2004). Depres- cognitive function in breast cancer patients treated with sion in patients with high-grade glioma: results of the tamoxifen. Breast Cancer Res Treat 64: 165–176. Chapter 4. Role of neuropsychological assessment in cancer patients 43

Patchell RA (1995). Metastatic brain tumors. Neurol Clin 13: Spiegel D (1996). Cancer and depression. Br J Psychiatry 915–925. Suppl 30: 109–116. Patchell R, Tibbs P (1990). A randomised trial of surgery in Sundaresan N, Galicich JH, Deck MD, et al. (1981). Radi- the treatment of single metastases to the brain. New Engl ation necrosis after treatment of solitary intracranial JMed22: 494–500. metastases. Neurosurgery 8: 329–333. Peyser JM, Rao SM, Larocca NG, et al. (1990). Guidelines for Surma-Aho O, Niemela M, Vilkki J, et al. (2001). Adverse neuropsychological research in multiple sclerosis. Arch long-term effects of brain radiotherapy in adult low- Neurol 47: 94–97. grade glioma patients. Neurology 56: 1285–1290. Pignatti F, Van Den Bent M, Curran D, et al. (2002). Prog- Taphoorn MJ, Schiphorst AK, Snoek FJ, et al. (1994). Cog- nostic factors for survival in adult patients with cerebral nitive functions and quality of life in patients with low- low-grade glioma. J Clin Oncol 20: 2076–2084. grade gliomas: the impact of radiotherapy. Ann Neurol Price BH, Mesulam M (1985). Psychiatric manifestations of 36: 48–54. right hemisphere infarctions. J Nerv Mental Dis 173: 610– Tashiro M, Itoh M, Kubota K, et al. (2001). Relationship 614. between trait anxiety, brain activity and natural killer cell Robinson B (1983). Validation of a caregiver strain index. J activity in cancer patients: a preliminary PET study. Psy- Gerontol 38: 344–388. chooncology 10: 541–546. Rohling ML, Green P, Allen LM, et al. (2002). Depressive Tucha O, Smely C, Preier M, et al. (2000). Cognitive deficits symptoms and neurocognitive test scores in patients before treatment among patients with brain tumors. passing symptoms validity tests. Arch Clin Neuropsychol Neurosurgery 47: 324–333; discussion 333–334. 17: 205–222. Van Dam FS, Schagen SB, Muller MJ, et al. (1998). Impair- Roman GC, Tatemichi TK, Erkinjuntti T, et al. (1993). Vas- ment of cognitive function in women receiving adjuvant cular dementia: diagnostic criteria for research studies. treatment for high-risk breast cancer: high-dose versus Report of the NINDS-AIREN International Workshop. standard-dose chemotherapy. J Natl Cancer Inst 90: 210– Neurology 43: 250–260. 218. Routh A, Khansur T, Hickman BT, et al. (1994). Man- Van De Pol M, Ten Velde GP, Wilmink JT, et al. (1997). agement of brain metastases: past, present, and future. Efficacy and safety of prophylactic cranial irradiation in Southern Med J 87: 1218–1226. patients with small cell lung cancer. J Neurooncol 35: Salander P, Karlsson T, Bergenheim T, et al. (1995). 153–160. Long-term memory deficits in patients with malignant Van Oosterhout AG, Ganzevles PG, Wilmink JT, et al. (1996). gliomas. J Neurooncol 25: 227–238. Sequelae in long-term survivors of small cell lung cancer. Sawaya R, Rambo WM Jr., Hammoud MA, et al. (1995). Int J Radiat Oncol Biol Phys 34: 1037–1044. Advances in surgery for brain tumors. Neurol Clin 13: Vecht CJ, Hovestadt A, Verbiest HB, et al. (1994). Dose- 757–771. effect relationship of dexamethasone on Karnofsky per- Schagen SB, Van Dam FS, Muller MJ, et al. (1999). Cogni- formance in metastatic brain tumors: a randomized tive deficits after postoperative adjuvant chemotherapy study of doses of 4, 8, and 16 mg per day. Neurology 44: for breast carcinoma. Cancer 85: 640–650. 675–680. Scheibel RS, Meyers CA, Levin VA (1996). Cognitive dys- Vermeulen SS (1998). Whole brain radiotherapy in the function following surgery for intracerebral glioma: treatment of metastatic brain tumors. Sem in Surg Oncol influence of histopathology, lesion location, and treat- 14: 64–69. ment. J Neurooncol 30: 61–69. Walker LG, Heys SD, Walker MB, et al. (1999). Psychological Sheline GE, Wara WM, Smith V (1980). Therapeutic irradia- factors can predict the response to primary chemother- tion and brain injury. Int J Radiat Oncol Biol Phy 6: 1215– apy in patients with locally advanced breast cancer. Eur J 1228. Cancer 35: 1783–1788. 5

Neuropsychological assessment of adults with cancer

Anne E. Kayl, Robert Collins, and Jeffrey S. Wefel

Introduction directly on a person’s ability to function in their environment; it also offers a methodology to eval- The survival rate for patients diagnosed with some uate the effectiveness and neurotoxic limitations of types of cancers has increased with advances in therapies at a level not typically accounted for in surgical techniques, radiotherapy, and the develop- most clinical trials. ment of new chemotherapeutic agents. Although most patients continue to face aggressive multi- modality and multi-agent treatment to control or Neuropsychological assessment eradicate disease, cancer is not always regarded as the “terminal” disease of past decades. In fact, some There has been much written about neuropsycho- types of cancer are best conceptualized as a chronic logical assessment and the reader is referred to illness, more akin to diabetes, and are amenable to Chapter 2 in this book and elsewhere for a more long-term management. It continues to be the case, in-depth review of this topic (see Lezak et al., 2004). however, that the majority of cancer patients will Brieﬂy, neuropsychology is the study of brain– require treatment with therapies that are rarely spe- behavior relationships. Through individualized ciﬁc to malignancy and often place normal tissues assessment, practitioners study the impact of injury at risk. The central nervous system (CNS) appears or disease on brain function (Vanderploeg, 2000). particularly vulnerable to therapy-related changes, Neuropsychological testing involves the adminis- and there is ample evidence to suggest that many tration of standardized psychometric instruments treatments are capable of producing cognitive dys- that comprehensively evaluate cognitive aspects of function that can persist well after cessation of cerebral functioning to include attention, the ability treatment. While it is easy to associate such cogni- to acquire new memories, the recall of stored mem- tive changes to observable CNS tissue damage (e.g., ories, expressive speech, language comprehension, post-surgical changes seen on imaging), many cur- visual perception, executive functions, and mood. rent treatments act at a molecular level of obser- The assessment process includes the integration vation and the mechanism that links those various of test results with observations of the patient’s treatments to putative changes in a patient’s cogni- behavior, patient report, and reports of family tive functioning has not been fully elucidated. Neu- members and/or caregivers. Neuropsychological ropsychological assessment is well suited to quan- evaluations have traditionally been utilized to tify such cognitive impairments, which may bear evaluate patients with known CNS injury or disease

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

44 Chapter 5. Neuropsychological assessment of adults with cancer 45

such as head injury, dementia, or stroke but this The importance of cognitive evaluations in method has more recently been applied to a wider patient care and in clinical cancer trials is receiving spectrum of medical disorders (see Tarter et al., greater recognition than previously, but assess- 2001). ment methods remain less than optimal in most Although recognition of the utility of cognitive cases. Cognitive assessment is a complex, mul- assessment for persons with cancer has increased tifaceted undertaking that requires specialized since the early 1990s, the focus of many clinicians, training. Although administration of tests is rela- especially in cases with difficult-to-treat malignan- tively simple, selection of appropriate measures cies, remains on the achievement of disease control and interpretation of results draws on a diverse and/or symptom management. However, patients’ set of skills requiring specialized, advanced train- cognitive and behavioral functioning, which is also ing in the field of neuropsychology, as well as more loosely studied under the rubric “quality of knowledge of the patient population under study. life,” must be given consideration, especially when Case conceptualization relies on information in the course of extending life expectancy there obtained during clinical interviews with patients exists the possibility that added neurotoxicity will and caregivers, an understanding of the patient’s compromise a person’s ability to function in their sociocultural milieu, recognition of the idiosyn- daily environment at a level that is individually cracies of test construction and psychometrics, satisfying. Even subtle cognitive deficits can sig- and knowledge of the human nervous system nificantly limit a patient’s ability to perform their (Wefel et al., 2004b). Finally, the literature is replete usual activities, but they may not be evident on with studies inadequately assessing “cognitive casual observation or detectable via routine med- functioning” in the cancer population. This sug- ical examinations. If unrecognized, these cognitive gests a limited appreciation of the underlying deficits can lead to inaccurate judgment on the part science that makes cognitive assessment a use- of the medical team regarding the patient’s abil- ful, complementary tool in the medical research ity for self-care, requirements for supervision or setting. special safety measures, and reliability in following his/her therapeutic regimen. Neuropsychological assessment in the cancer population provides quan- Principles of neuropsychological assessment titative, objective measurement of potentially sub- in the cancer population tle changes in a patient’s cognitive function, allows for careful evaluation of the costs and benefits of There are both general testing and cancer- a given treatment regimen or supportive therapy, treatment-specific factors that should be given con- has been shown to predict progression of disease sideration in neuropsychological assessment, and prior to progression on imaging (Meyers & Hess, these are not always independent of one another. 2003), and can differentiate cancer-related impair- General testing factors refer to broader principles ment from stroke, dementia, or mood disturbance. of assessment that should be considered across all The results of an appropriate cognitive assessment cancer populations, whereas cancer/treatment fac- may also be used to guide interventions including tors are more specifically concerned with putative compensatory strategy training (cognitive rehabil- mechanisms by which cognitive dysfunction might itation), pharmacotherapy (i.e., psychostimulants, be induced. The following sections are divided into antidepressants, or anxiolytics), or psychotherapy. general and specific considerations but the type of In fact, improvement in cognitive function and questions being asked (e.g., research versus clinical) delayed cognitive progression are now recognized will also bear directly on testing issues. In theory as important study endpoints by the Food and Drug and in practice, we advocate a purpose-driven Administration (FDA). approach, geared toward the goal of obtaining 46 Section 1. Cognition and the brain

maximally relevant data with minimal patient Anderson Cancer Center receive measures tapping burden. multiple cognitive domains (i.e., general intellectual skills, learning and memory, language, attention, visuospatial/visuoperceptual skills, processing Test selection speed, executive functions), motor functions, and Test selection is crucial in both clinical and research mood. In this setting, the evaluation is typically settings and will necessarily vary given the hypothe- completed in 1 day and lasts from 1 h to 4+ h ses being evaluated. Standard assessments of per- depending on the purpose and goals of the assess- formance status such as the Karnofsky Performance ment. For example, our service is involved in a Status Scale (KPS) (Karnofsky & Burchenal, 1949) variety of clinical trials evaluating the efficacy of globally measure the patient’s symptoms or abil- novel and innovative treatments. The clinical trial ity for self-care and ambulation, but do not reli- battery is briefer than a typical neuropsychological ably or validly assess cognition (Hutchinson et al., assessment and is able to be completed by most 1979). For example, a patient who is able to walk patients in 60 min or less. The clinical trial battery and perform basic activities of daily living may be includes measures that have alternative forms that rated as having a good performance status, but on allow for serial testing and are sensitive to changes more careful evaluation there may be clear evidence in learning and memory processes, attention, proof unreliable memory, poor judgment, difficulty cessing speed, executive function, and fine motor managing routine work, or significant personality co-ordination. In addition to the aforementioned change. Brief mental status evaluations such as the domains, a patient referred prior to a planned return Mini-Mental State Examination (MMSE) (Folstein to educational pursuits will require a more thorough et al., 1975) may superficially evaluate aphasia, assessment of intellectual ability and academic apraxia, orientation, and attention, but neglect achievement. Information gleaned from the assess- those functions most susceptible to chemotherapy- ment may be used to develop an individualized related change (learning and memory, processing education plan, ensuring supports that will facili- speed, executive function, and fine motor control) tate a successful transition. An evaluation to assist (Meyers & Wefel, 2003). in the process of returning to competitive employ- The instruments used in the neuropsychologi- ment will certainly include an assessment of those cal evaluation of persons with cancer should be domains sensitive to cancer-related and treatment- sensitive to subtle changes in the aforementioned related cognitive changes, but should also include areas. They should be psychometrically sound, with measures thought to be ecologically valid. That is, established reliability and validity, and appropri- measures that may predict the patient’s success in ate normative studies. As previously noted, the their chosen vocation. An evaluation requested to breadth of the assessment and the choice of instru- assist in ruling out dementia should include a thor- ments will necessarily vary with the purpose of ough assessment of verbal and non-verbal learn- the assessment, be it a clinical referral for diag- ing and memory processes, as well as supporting nosis/documentation of impairment, a standard- cognitive domains that are known to be affected ized research protocol, a capacity assessment, or to across different stages of various dementing dis- assist with transitions back to work or a return to orders. Appropriate measures should, on interpre- school. tation, enable the neuropsychologist to differenti- Although the specific measures selected will vary ate between failures of learning, retrieval and/or given the referral question, the patient’s complaints, consolidation processes. The evaluation should also and the patient’s estimated baseline level of func- include an assessment of apraxia, language (e.g., tioning, most patients referred to the Neuropsy- naming, lexical fluency, semantic fluency, audi- chology Service at The University of Texas M. D. tory comprehension), agnosia, executive function, Chapter 5. Neuropsychological assessment of adults with cancer 47

orientation, and visuoconstruction/visuopercep- ment obvious to their treatment provider or care- tion skills. To summarize, the selection of neuropsy- giver. The timing of the assessment is, therefore, chological instruments cannot be a “one-size-fits- dictated by the referral. These assessments, how- all” proposition. Although many patients diagnosed ever, often occur in the context of on going treat- with and treated for cancer will evidence cognitive ments and the impact of this should be accounted dysfunction, the measures chosen must take into for. In clinical trials assessing the effect of a par- account the purpose and goals of the assessment. ticular treatment, pre-treatment assessments are imperative to adequately discern change over time. From a methodological standpoint, the need for Assessment process pre-treatment assessments has been undervalued Once again, the reader is referred to Chapter 2 and in the cancer literature. In more recent studies pre- other sources (Lezak et al., 2004; Vanderploeg, 2000) treatment cognitive dysfunction has been reported for comprehensive discussions of the assessment in patients with CNS disease (Fleissbach et al., 2003; procedure. Tucha et al., 2000), lung cancer (Meyers et al., 1995), In general, the neuropsychological assessment and breast cancer (Wefel et al., 2004a). Often a should begin with a general introduction explaining patient may undergo several assessments over time why the patient was referred, as well as the nature and in this case it is important to employ measures of the evaluation. The clinical interview that typi- with alternative forms, which are relatively resistant cally precedes test administration may include fam- to practice effects (Wefel et al., 2004b). When inter- ily members and/or caregivers as well as the patient, preting data from repeated assessments it is also but the testing portion of the examination should be imperative to determine if the magnitude of change completed in a quiet, distraction-free environment, across individual, and multiple measures, is clini- with only the patient and examiner being present. cally meaningful. We utilize a reliable change index Establishing a good working relationship with the (Jacobson & Truax, 1991), but there several ways to patient is crucial to the success of the evaluation. account for clinical change over time and there is no Practitioners must be sensitive to the many poten- consensus in the field. tial factors (demographic, psychological, sociocultural, medical, etc.) that may affect the patient’s Cognitive domains of interest willingness or ability to actively participate in the assessment. Tasks included in the assessment In clinical referral cases, the neuropsychologist should vary in difficulty, be sensitive to potential should select tests appropriate to answer the refer- deficits, and should be administered in a stan- ral question (e.g., aphasia evaluation) for that par- dardized fashion. “Testing the limits” is a process ticular patient. For patients with focal CNS disease, that can provide valuable clinical information, but knowledge of brain–behavior relationships should should be attempted only after a measure has been guide initial test selection. As a general rule, the neu- completed in the standardized format (Lezak et al., ropsychological assessment of persons with cancer 2004). should also include measures of learning and memory, processing speed, executive function, and fine motor control, since these are the domains that have Timing of assessment(s) been found to be sensitive to disease-related and The timing of the assessment is an important con- treatment-related change across disease groups. sideration for both general clinical and research The neuropsychological assessment of patients purposes and, again, is in many ways dictated by with cancer should include an evaluation of mood. the hypotheses being evaluated. In most clinical set- In one study, nearly 50% of cancer patients inter- tings, a patient will be referred when there is impair- viewed had some type of psychiatric disorder; the 48 Section 1. Cognition and the brain

majority of these were classified as adjustment There are numerous PROs available and criti- disorders with features of anxiety and/or depres- cal review of a measure’s development, psychomet- sion or major depression (Derogatis et al., 1983). ric properties, and generalizability is necessary to Unfortunately, mood remains under-assessed in the determine its appropriateness for use in a partic- clinical care and research setting. Analysis of data ular setting and with a particular patient popula- from 598 patients enrolled in the Glioma Outcomes tion (Meyers, 1997). The Functional Assessment of Project revealed a remarkable discordance, with Cancer Therapy (FACT) (Cella et al., 1993) and the physicians reporting depression in 15% of patients European Organization for Research and Treatment with high-grade gliomas, but depressive symptom- of Cancer QLQ-C30 (EORTC QLQ-C30) (Aaronson ology was reported by 93% of patients in the early et al., 1993) are two commonly referenced PROs post-operative period. Patient-reported depression in the cancer–cognition literature. Each includes increased throughout the 6-month period after functional scales (physical, emotional, social, func- surgery, but remained underdiagnosed and under- tional) and symptom scales, with diagnosis-specific treated by physicians (Litofsky et al., 2004). modules. Although they provide some insight into patient concerns and perceptions, PRO ratings do not reliably correlate with clinical/functional outcomes (Huang et al., 2001) or cognitive status as Patient-reported outcomes evaluated by objective neuropsychological assess- Though the importance of patient perceptions and ment (Taphoorn et al., 1992). experience in the clinical management of disease has long been recognized, the term “quality of life” has been criticized as being too broad and non- Disease- and treatment-specific specific to be of scientific value. An FDA work considerations group concurred and adopted the term “patient- reported outcomes” (PROs) to include any disease- The assessment of patients with CNS disease or treatment-related study endpoint subjectively reported by the patient (Acquadro et al., 2003). Whether primary or metastatic, CNS tumors nearly Health-related quality of life (HRQOL) measures are always cause cognitive dysfunction. The nature of a subset of PROs, remaining subjective, but also pro- the neuropsychological impairment observed in viding patient evaluations of the impact of disease the individual patient is in part related to the or treatment on their well-being. site of the lesion. For example, tumors located Patient-reported outcomes are now commonly inthelefthemisphereofthebrainmaybeasso- used in clinical trials since they provide a unique ciated with expressive and/or receptive language understanding of treatment outcome from the problems that impede communication, while right patient’s perspective. In one recent review, PROs hemisphere tumors may be associated with per- were reported in 30% of 215 FDA-approved product ception problems, visual-spatial disturbances, or labels reviewed, and were the only type of endpoint attention deficits. In cancer patients, deficiencies in used for 23 products (Willke et al., 2004). Patient- learning efficiency and memory retrieval are com- reported outcomes complement clinical endpoints mon and prevalent among patients with right, as and factor into the evaluation of a treatment’s well as left hemisphere disease. Impairments of impact. Cancer therapy may be associated with sig- frontal lobe function (executive deficits manifested nificant adverse side-effects, so it is important to by impairments of cognitive flexibility, abstraction, weigh the impact of these effects against potentially motivation, planning and organizational skills, abil- small gains in survival (Osoba et al., 2000; Wiklund, ity to benefit from experience, personality changes, 2004). etc.) are also prevalent and may occur in patients Chapter 5. Neuropsychological assessment of adults with cancer 49

without clear evidence of frontal lobe involvement. of co-morbid vascular risk factors (e.g., diabetes) Disruption of afferent and efferent frontal lobe con- (Crossen et al., 1994; Gregor et al., 1996; Lee nections has been suggested as a putative mecha- et al., 2002). Most studies that include neuropsy- nism. chological assessment of patients before and fol- Although research on the neurobehavioral and lowing radiation therapy reveal significant impair- cognitive changes associated with metastatic dis- ments of information-processing speed, executive ease is not that voluminous, some data are avail- functions, memory, sustained attention, and motor able for review. In contrast to the neuropsycho- co-ordination in those with no evidence of disease logical profiles of patients with local CNS disease, recurrence (Archibald et al., 1994; Helfre & Pierga, individuals with metastatic brain involvement fre- 1999; Hochberg & Slotnick, 1980; Imperato et al., quently present with more diffuse cognitive dys- 1990; Lang et al., 2000; Lieberman et al., 1982; Salan- function. For example, we completed neuropsy- der et al., 1995; Scheibel et al., 1996; Taphoorn et al., chological assessments on 55 patients with brain 1994). Even radiation not directed at the brain can metastases (Kayl et al., 2001). The majority of these cause cognitive impairment. For example, a sub- patients carried a primary diagnosis of lung cancer, stantial percentage of patients who receive ther- melanoma, renal cancer, or breast cancer. In most apeutic radiation for tumors of the anterior skull cases, patients had a single metastatic lesion, but base have cognitive deficits. Memory impairment those with multiple metastases were not excluded. was detected in 80% of patients with paranasal sinus Impaired cognitive performance (defined as a score tumors, even though the brain was not the target of greater than 1 standard deviation from the norma- radiation (Meyers et al., 2000). tive mean) was demonstrated on measures of fine motor co-ordination speed (42% of patients), mem- Impact of chemotherapy on cognitive function ory (free recall: 29%), and verbal fluency (20%) prior to treatment for their brain disease. In this group Chemotherapy remains a useful weapon in the of patients, as in others with metastatic brain dis- management of cancer, and has improved sur- ease, the etiology of these impairments is unclear, vival for patients with some types of disease. but microscopic tumor infiltration, diaschisis, and Potential CNS side-effects vary across agents but treatment-related changes may be influencing their may appear as peripheral neuropathy, acute and functioning. reversible encephalopathy, cerebellar syndrome, or persistent cognitive dysfunction (Keime-Guibert et al., 1998). Cognitive and emotional changes The impact of radiation on cognitive function reported during and after chemotherapy include Radiation therapy remains an important therapeu- memory loss, decreased information-processing tic tool in the care and management of patients speed, reduced attention, anxiety, and depression with CNS disease. A comprehensive review is (Meyers & Abbruzzese, 1992). It has been estimated beyond the scope of this chapter, but the interes- that as many as one-third of patients undergoing ted reader is referred to Chapter 7 in this volume, systemic chemotherapy evidence declines in cogni- as well as those articles cited in the paragraphs tive function that interfere with their quality of life that follow, for additional information. Risk fac- (Ferguson & Ahles, 2003). tors for developing cognitive dysfunction and radi- Among patients treated for cancer, the term ation necrosis include patient age >60 years, sched- “chemobrain” is being used with increasing fre- uled dosing >2 Gy per fraction, total dose, volume quency to describe perceived cognitive declines. of brain irradiated, hyperfractionated schedules, In a recently published prospective, randomized, shorter overall treatment time, concomitant or longitudinal trial, Wefel et al. (2004a) found an subsequent use of chemotherapy, and presence association between cognitive dysfunction and 50 Section 1. Cognition and the brain

chemotherapy in a subgroup of women with non- of therapeutic trials for primary brain tumors metastatic breast carcinoma. Patients received and leptomeningeal disease (LMD) (Meyers et al., a baseline assessment prior to the initiation 1991a). These agents are known to have both acute of chemotherapy with 5-fluorouracil, doxoru- and persistent neurotoxic side-effects. Acute toxic- bicin, and cyclophosphamide (FAC), approximately ity is characterized by fever, headache, and myalgia, 3 weeks following chemotherapy, and again at 1 year which generally resolves over several days. Subacute post-treatment. Surprisingly, 33% of these women neurotoxicity, evident within a week of starting ther- had evidence of cognitive dysfunction prior to the apy, is characterized by inattention, slowed think- initiation of treatment. Within-subject analyses ing, and lack of motivation. Patients may develop revealed that 61% of patients evidenced a decline in difficulty with memory, frontal lobe executive func- cognitive function between the baseline assessment tions (e.g., problem-solving, planning, sequencing), and just after the cessation of chemotherapy. Cog- motor co-ordination, and mood as treatment con- nitive decline occurred most often in the domains tinues (Pavol et al., 1995). These neurotoxic side- of attention, learning and memory, and processing effects are not always reversible following treat- speed. Of this subset of patients, 45% remained ment cessation (Meyers et al., 1991b). Please refer to impaired, 45% improved, and 10% had a mixed Chapter 8 in this volume for a more detailed discus- pattern of improvement and persistent symptoms sion of biological response modifier-related cogni- 1 year post-chemotherapy. tive dysfunction. In some cases, chemotherapeutic agents have The brain may also be sensitive to changes in a mechanism of action that is expected to affect its hormonal milieu (Yaffe et al., 1998). Tamoxifen focal brain regions (Meyers et al., 1997). Other (TAM) is a widely used selective estrogen recep- potential mechanisms include direct neurotoxic tor modulator for the treatment of breast can- effects of treatment leading to cortical atrophy or cer. In a prospective, longitudinal trial (Wefel & demyelination and microvascular changes (Saykin Meyers, 2004) completed at The University of Texas et al., 2003). Certain chemotherapeutic agents are M. D. Anderson Cancer Center, adjuvant TAM ther- known to be especially neurotoxic. The incidence apy was found to be associated with significant and severity of the neurotoxicities vary between neurotoxicity in a subgroup of women. This sub- agents and between individual patients, but cogni- group experienced significant dysfunction in the tive changes have been associated with methotrex- domains of memory, executive, and motor function, ate (Madhyastha et al., 2002; Mulhern et al., 1988; as well as increased emotional distress, decreased Ochs et al., 1991; Taphoorn & Klein, 2004), etoposide quality of life, and diminished ability to maintain (Castello et al., 1990; Chamberlain, 1997; Cham- productive activities. Hormonal therapies are also berlain & Kormanik, 1997), high-dose cytarabine among the treatment options available to men diag- (Geller et al., 2001; Hwang et al., 1985; Salinsky nosed with prostate carcinoma. One type of hor- et al., 1983; Schwartz et al., 2000), 5-fluorouracil monal therapy involves treatment with luteinizing- (Lipp, 1999), TNP-470 (Logothetis et al., 2001), hormone-releasing hormone (LHRH) agonists, such CI-980 (Meyers et al., 1997), ifosfamide (Lipp, 1999), as leuprolide (LupronR ) and goserelin (ZoladexR ). and cisplatin (Troy et al., 2000). Please refer to Chap- This treatment prevents the testicles from pro- ter 8 in this volume for a more detailed discussion of ducing testosterone, and also lowers estradiol lev- chemotherapy-related cognitive dysfunction. els (Dawson-Hughes, 2001). There is growing evidence that certain brain structures are susceptible to declines in hormone levels associated with LHRH Impact of immunotherapy and hormonal therapy, but the mechanisms of these changes therapies on cognitive function are not fully understood. The cognitive effects of Cytokines such as interferon alpha (IFN-α)and hormonal agents are reviewed in Chapter 9 of this interleukin-2 (IL-2) have been used in a number volume. Chapter 5. Neuropsychological assessment of adults with cancer 51

Fatigue ing effects of medications. Depending on the clinical situation, steroids may have positive or neg- Fatigue is a common symptom of patients with can- ative effects on neurocognitive function (Lewis & cer and may be related to anemia, insomnia, poor Smith, 1983; Martignoni et al., 1992; Stoudemire nutrition, proinflammatory cytokine activation, or et al., 1996; Varney et al., 1984; Wolkowitz et al., metabolic disturbance. Cancer-related fatigue is not 1990). Antiepileptic medications have also been responsive to increased rest or sleep and can con- associated with cognitive dysfunction (Taphoorn & tribute to cognitive decline (Iop et al., 2004; Nerenz Klein, 2004), the severity of which tends to increase et al., 1982; Tierney et al., 1991). An assessment of with the use of multiple agents and with ele- fatigue should include an assessment of severity and vated serum levels (Ortinski & Meador, 2004). With its functional impact on the patient’s daily activi- the exception of topiramate, the newer antiepilep- ties and perceived quality of life. In one study of tic agents (including lamotrigine, oxcarbazepine, older patients (over 60 years), fatigue was nearly and gabapentin) seem to have more favorable universal, significantly interfered with subjects’ gen- side-effect profiles, with less impact on cognition eral activity level, and was positively correlated (Beyenburg et al., 2004; Loring & Meador, 2004; with severity of depression (Respini et al., 2003). Meador & Baker, 1997). Potential treatment avenues might include correction of underlying metabolic or endocrine disorders, addressing depression or insomnia, institution of Affective distress light/moderate exercise, cognitive therapy, or phar- Depression and anxiety may also affect cognitive macotherapy with steroids, stimulants, or epoetin performance. While depression and anxiety are not alfa to correct anemia (Iop et al., 2004; Lesage & uncommon in the cancer population, the diagno- Portenoy, 2002). sis of depression in cancer patients is complicated by the difficulty in distinguishing vegetative symp- Anemia toms attributable to a mood disorder from symp- The majority of chemotherapy patients experience toms caused by the primary disease or its treatment mild anemia (Hb <12 g/dl) and severe anemia may (Valentine et al., 2002). Assessment and diagno- affect up to 80% of patients (Groopman & Itri, sis are important however, as depression has been 1999). The causes of cancer-related anemia include shown to affect cognitive function, causing impair- infiltration of bone marrow by malignant cells, ments in attention and other cognitive skills, includ- decreased hemoglobin (Hb) production secondary ing memory (Christensen et al., 1997; Tarbuck & to treatment, iron deficiency, or low erythropoi- Paykel, 1995). Interestingly, cognitive impairment etin (EPO) levels. Anemia levels below 12 g/dl in may vary as a function of the depressive dis- patients undergoing chemotherapy have been asso- orders. In a 2004 study, Airaksinen et al.found ciated with both fatigue and declines on measures that severely depressed individuals and patients of attention, speed of cognitive processing, verbal with mixed anxiety-depression evidenced signifi- fluency, and verbal memory (Brown et al., 1991; cant memory dysfunction, while dysthymic patients Jacobsen et al., 2004; Marsh et al., 1991; Temple tended to show impairments of mental flexibility. In et al., 1995). Treatment of anemia results in a 1996 study, Cull et al. failed to detect differences in improved cognitive function (Grimm et al., 1990; cognitive test performance between patients com- Littlewood et al., 2002; Straus, 2002). plaining of concentration and memory difficulties and non-complainers. However, those reporting cognitive difficulties had significantly high scores Adjunctive medications on measures of anxiety, depression, and fatigue. Interpretation of the results of a neuropsychological Anxiety has been associated with reduced cognitive assessment may be complicated by the confound- efficiency, memory problems, and distractibility. 52 Section 1. Cognition and the brain

Neurobiological evidence suggests that anxiety can function often experienced by patients with can- affect medial temporal lobe structures, includ- cer. Attention, processing speed, learning/memory ing the hippocampus, amygdala, and frontal- functions, executive function, and motor skills are subcortical circuits, subserving some memory func- particularly vulnerable and should be carefully eval- tions. In some cases, imaging studies have found uated for signs of dysfunction. These are areas associations between depression and decreased of cognition that cannot be adequately assessed hippocampal volume (Bremner et al., 2000; Steffens using screening measures such as the oft-employed et al., 2000; Videbech & Ravnkilde, 2004). Relative MMSE. to normal, healthy controls, persons diagnosed with anxiety disorders evidence impairment on measures of memory and executive function, but verbal ﬂuency and processing speed were statistically REFERENCES unaffected (Airaksinen et al., 2005). Aaronson NK, Ahmedzai S, Bergman B et al. (1993). The European Organization for Research and Treatment of Conclusion Cancer QLQ-C30: a quality-of-life instrument for use in international cancer trials in oncology. J Natl Cancer Inst 85(5): 365–376. Neuropsychological testing (the act of test admin- Acquadro C, Berzon R, Dubois D et al. for the PRO Harmo- istration) is a relatively simple procedure, but neu- nization Group. (2003). Incorporating the patient’s per- ropsychological assessment is a process requiring spective into drug development and communication: an an understanding of test construction and psycho- ad hoc force report of the Patient-reported Outcomes metrics, functional neuroanatomy, behavioral neu- (PRO) Harmonization Group meeting at the Food and rology, and the patient’s sociocultural milieu, and Drug Administration, February 16, 2001. Value Health should only be undertaken by professionals with 6(5): 522–531. training in these domains. Airaksinen E, Larsson M, Lundberg I et al. (2004). Cogni- The utility and value of neuropsychological tive functions in depressive disorders: evidence from a assessment in the clinical management of cancer population-based study. Psychol Med 34(1): 83–91. Airaksinen E, Larsson M, Forsell Y (2005). Neuropsycholog- patients and in the evaluation of new treatments are ical functions in anxiety disorders in population-based becoming better recognized, but the quality of the samples: evidence of episodic memory dysfunction. J research that has been published to date is less than Psychiatr Res 39(2): 207–214. optimal. The contemporary scientiﬁc literature is Archibald YM, Lunn D, Ruttan LA et al. (1994). Cognitive cluttered with poorly designed studies that may lead functioning in long-term survivors of high-grade glioma. investigators and the readership to incorrect con- J Neurosurg 80: 247–253. clusions. Clinicians and researchers must keep a few Beyenburg S, Bauer J, Reuber M (2004). New drugs for basic principles in mind when developing a plan for the treatment of epilepsy: a practical approach. Postgrad assessment. First, test selection will vary depend- Med J 80: 581–587. ing on the question under consideration. Second, Bremner JD, Narayan M, Anderson ER et al. (2000). Hip- the measures chosen should have alternative forms pocampal volume reduction in major depression. Am J Psychiatry 157: 115–117. or be relatively resistant to practice effects, charac- Brown WE, Marsh J, Wolcott D et al. (1991). Cognitive func- teristics that are especially important if one plans tion, mood and P3 latency: effects of the amelioration of to test patients repeatedly. Third, selected measures anemia in dialysis patients. Neuropsychologia 29: 35–45. should be psychometrically sound, with established Castello MA, Clerico A, Deb G et al. (1990). High-dose car- reliability and validity, and appropriate normative boplatin in combination with etoposide (JET regimen) studies. Finally, it is important to select measures for childhood brain tumors. Am J Pediatr Hematol Oncol that are sensitive to subtle changes in cognitive 12(3): 297–300. Chapter 5. Neuropsychological assessment of adults with cancer 53

Cella DF, Tulsky DS, Gray G et al. (1993). The Functional Groopman JE, Itri LM (1999). Chemotherapy-induced ane- Assessment of Cancer Therapy Scale: development and mia in adults: incidence and treatment. J Natl Cancer Inst validation of the general measure. J Clin Oncol 11: 570– 91: 1616–1634. 579. Helfre S, Pierga J (1999). Cerebral metastases: radiotherapy Chamberlain MC (1997). Recurrent supratentorial malig- and chemotherapy. Neuro-Chirurgie 45(5): 382–392. nant gliomas in children. Long-term salvage therapy Hochberg FH, Slotnick B (1980). Neuropsychologic impair- with oral etoposide. Arch Neurol 54(5): 554–558. ment in astrocytoma survivors. Neurology 30: 172–177. Chamberlain MC, Kormanik PA (1997). Chronic oral VP- Huang ME, Wartella JE, Kreutzer JS (2001). Functional out- 16 for recurrent medulloblastoma. Pediatr Neurol 17(3): comes and quality of life in patients with brain tumors: 230–234. a preliminary report. Arch Phys Med Rehabil 82: 1540– Christensen H, Griffiths K, MacKinnon A et al. (1997). A 1546. quantitative review of cognitive deficits in depression Hutchinson TA, Boyd NF, Feinstein AR et al. (1979). Sci- and Alzheimer-type dementia. J Int Neuropsychol Soc 3: entific problems in clinical scales, as demonstrated in 631–651. the Karnofsky Index of performance status. JChronicDis Crossen JR, Garwood D, Glatstein E et al. (1994). Neurobe- 32(9–10): 661–666. havioral sequelae of cranial irradiation in adults: a review Hwang TL, Yung WK, Estey EH et al. (1985). Central ner- of radiation-induced encephalopathy. J Clin Oncol 12: vous system toxicity with high dose Ara-C. Neurology 35: 627–642. 1475–1479. Cull A, Hay C, Love SB et al. (1996). What do cancer patients Imperato JP, Paleologos NE, Vich NA (1990). Effects of treat- mean when they complain of memory problems? Br J ment on long-term survivors with malignant astrocy- Cancer 74(10): 1674–1679. tomas. Ann Neurol 28: 818–822. Dawson-Hughes B (2001). Bone loss accompanying medi- Iop A, Manfredi AM, Bonura S (2004). Fatigue in cancer cal therapies. NEnglJMed354(13): 989–991. patients receiving chemotherapy: an analysis of pub- Derogatis LR, Morrow GR, Fetting J et al. (1983). The preva- lished studies. Ann Oncol 15: 712–720. lence of psychiatric disorders among cancer patients. J Jacobsen L, Garland M, Booth-Jones K et al. (2004). Am Med Assoc 249: 751–757. Relationship of hemoglobin levels to fatigue and cog- Ferguson RJ, Ahles TA (2003). Low neuropsychologic per- nitive functioning among cancer patients receiving formance among adult cancer survivors treated with chemotherapy. J Pain Symptom Manage 28(1): 7–18 P. chemotherapy. Curr Neurol Neurosci Rep 3(3): 215–222. Jacobson NS, Truax P (1991). Clinical significance: a sta- Fliessbach K, Urbach H, Helmstaedter C et al. (2003). Cog- tistical approach to defining meaningful change in psy- nitive performance and magnetic resonance imaging chotherapy research. J Consult Clin Psychol 59(1): 12–19. findings after high-dose systemic and intraventricular Karnofsky DA, Burchenal JH (1949). The clinical evaluation chemotherapy for primary central nervous system lym- of chemotherapeutic agents in cancer. In Macleod CM phoma. Arch Neurol 60: 563–568. (ed.) Evaluation of Chemotherapeutic Agents (pp. 191– Folstein MF, Folstein SE, McHugh PR (1975). A mini men- 205). New York: Columbia University Press. tal state: a practical method for grading the cognitive Kayl AE, Seabrooke LF, Meyers CA (2001). Neurobehav- status of patients for the clinician. J Psychiatr Res 12: ioral changes due to metastatic brain tumors [abstract]. 189–198. Neurooncology 3(4): 339. Geller HM, Cheng KY, Goldsmith NK et al. (2001). Oxida- Keime-Guibert F, Napolitano M, Delattre J-Y (1998). Neu- tive stress mediates neuronal DNA damage and apop- rological complications of radiotherapy and chemother- tosis in response to cytosine arabinoside. J Neurochem apy. J Neurol 245: 695–708. 78(2): 265–275. Lang FF, Wildrick DM, Sawaya R (2000). Metastatic Gregor A, Cull A, Traynor E et al. (1996). Neuropsy- brain tumors. In Bernstein M, Berger MS (eds.) Neuro- chometric evaluation of long-term survivors of adult Oncology: The Essentials (pp. 329–337). New York: brain tumours: relationship with tumour and treatment Thieme Medical Publishers. parameters. Radiother Oncol 41: 55–59. LeeAW,KwongDLW,LeungS-Fet al. (2002). Factors affect- Grimm G, Stockenhuber F, Schneeweiss B et al. (1990). ing risk of symptomatic temporal lobe necrosis: Signifi- Improvement in brain function in hemodialysis patients cance of fractional dose and treatment time. Int J Radiat treated with erythropoietin. Kidney Int 38: 480–486. OncolBiolPhys53: 75–82. 54 Section 1. Cognition and the brain

Lesage P, Portenoy RK (2002). Management of fatigue in the Meyers CA, Hess KR (2003). Multifaceted end points in cancer patient. Oncology 16(3): 373–378. brain tumor clinical trials: cognitive deterioration pre- Lewis DA, Smith RE (1983). Steroid induced psychiatric cedes MRI progression. Neurooncology 5(2): 89–95. syndromes. J Affect Disord 5: 319. Meyers CA, Wefel JS (2003). The use of the Mini-Mental Lezak MD, Howieson DB, Loring DW (2004). Neuropsycho- State Examination to assess cognitive functioning in can- logical Assessment, (4th edn.). New York: Oxford Univer- cer trials: no ifs ands, buts, or sensitivity. J Clin Oncol sity Press. 21(19): 3557–3558. Lieberman AN, Foo SH, Ransohoff J et al. (1982). Long term Meyers CA, Obbens EAMT, Scheibel RS et al. (1991a). survival among patients with malignant brain tumors. Neurotoxicity of intraventricularly administered alpha- Neurosurgery 10: 450–453. interferon for leptomeningeal disease. Cancer 68: Lipp HP (1999). Neurotoxicity (including sensory toxicity) 88–92. induced by cytostatics. In Lipp HP (ed.) Anticancer Drug- Meyers CA, Scheibel RS, Forman AD (1991b). Persistent Toxicity. Prevention, Management, and Clinical Pharma- neurotoxicity of systemically administered interferon- cokinetics (pp. 431–454). New York: Marcel Dekker Inc. alpha. Neurology 41: 672–676. Litofsky NS, Farace E, Anderson F, Meyers CA, Huang W, Meyers CA, Byrne KS, Komaki R (1995). Cognitive deficits Laws ER (2004). Depression in patients with high-grade in patients with small cell lung cancer before and after glioma: results of the Glioma Outcomes Project. Neuro- chemotherapy. Lung Cancer 12: 231–235. surgery 54(2): 358–366. Meyers CA, Kudelka AP, Conrad CA et al. (1997). Neurotox- Littlewood TJ, Bajetta E, Nrtier JWR et al. (2002). Effects icity of CI-980, a novel mitotic inhibitor. Clin Cancer Res of epoetin alfa on hematologic parameters and qual- 3: 419–422. ity of life in cancer patients receiving nonplatinum Meyers CA, Geara F, Wong P-F et al. (2000). Neurocognitive chemotherapy: results of a randomized, double-blind, effects of therapeutic irradiation for base of skull tumors. placebo-controlled trial. J Clin Oncol 19: 2865–2874. Int J Radiat Oncol Biol Phys 46: 51–55. Logothetis CJ, Wu KK, Finn LD et al. (2001). Phase I trial Mulhern RK, Wasserman Al, Fairclough D et al. (1988). of the angiogenesis inhibitor TNP-470 for progressive Memory function in disease-free survivors of childhood androgen-independent prostate cancer. Clin Cancer Res acute lymphoblastic leukemia given central nervous sys- 7: 1198–1203. tem prophylaxis with or without 1800 cGy cranial irradi- Loring DW, Meador KJ (2004). Cognitive side effects of ation. J Clin Oncol 6: 315–320. antiepileptic drugs in children. Neurology 62: 872–877. Nerenz DR, Leventhal H, Love RR (1982). Factors contribut- Madhyastha S, Somayaji SN, Raoe MS et al. (2002). Hip- ing to emotional distress during chemotherapy. Cancer pocampal brain amines in methotrexate-induced learn- 50: 1020–1027. ing and memory deficit. Can J Physiol Pharmacol 80(11): Ochs J, Mulhern RK, Fairclough D et al. (1991). Compari- 1076–1084. son of neuropsychologic functioning and clinical indi- MarshJ,BrownWE,WolcottDet al. (1991). RhuEPO treat- cators of neurotoxicity in long-term survivors of child- ment improves brain and cognitive function of anemic hood leukemia given cranial irradiation or parenteral dialysis patients. Kidney Int 39: 155–163. methotrexate: a prospective study. J Clin Oncol 9(1): 145– Martignoni E, Costa A, Sinforiani E et al. (1992). The brain 151. as a target for adrenocortical steroids: cognitive implica- Ortinski P, Meador KL (2004). Cognitive side effects of tions. Psychoneuroendocrinology 17: 343. antiepileptic drugs. Epilepsy Behav 5: S60–S65. Meador KJ, Baker GA (1997). Behavioral and cognitive Osoba D, Brada M, Yung WKA et al. (2000). Health-related effects of lamotrigine. J Child Neurol 12 [Suppl 1]: S44– quality of life patients treated with temozolomide ver- 47. sus procarbazine for recurrent glioblastoma multiforme. Meyers CA (1997). Issues of quality of life in neuro- J Clin Oncol 18(7): 1481–1491. oncology. In Vecht CJ (ed.) Handbook of Clinical Neu- Pavol MA, Meyers CA, Rexer JL et al. (1995). Pattern of rology, Neuro-Oncology (Part 1) (pp. 389–409). New York: neurobehavioral deficits associated with interferon-alfa Elsevier Science. therapy for leukemia. Neurology 45: 947–950. Meyers CA, Abbruzzese JL (1992). Cognitive functioning in Respini D, Jacobsen PB, Thors C et al. (2003). The preva- cancer patients: effect of previous treatment. Neurology lence and correlates of fatigue in older cancer patients. 42: 434–436. Crit Rev Oncol Hematol 47: 273–279. Chapter 5. Neuropsychological assessment of adults with cancer 55

Salander P, Karlsson T, Bergenheim T et al. (1995). Troy L, McFarland K, Littman-Power S et al. (2000). Long-term memory deficits in patients with malignant Cisplatin-based therapy: a neurological and neuropsy- gliomas. J Neurooncol 25: 227–238. chological review. Psychooncology 9: 29–39. Salinsky MC, Levine RL, Aubuchon JP et al. (1983). Acute Tucha O, Smely C, Preier M et al. (2000). Cognitive deficits cerebellar dysfunction with high-dose Ara-C therapy. before treatment among patients with brain tumors. Cancer 51: 426–429. Neurosurgery 47: 324–333. Saykin AJ, Ahles TA, McDonald BC (2003). Mechanisms of Valentine AD, Passik SD, Massie MJ (2002). Psychiatric and chemotherapy-induced cognitive disorders: neuropsy- psychosocial issues. In Levin VA (ed.) Cancer in the Ner- chological, pathophysiological, and neuroimaging per- vous System (2nd edn.) (pp. 572–589). New York: Oxford spectives. Semin Clin Neuropsychiatry 8(4): 201–216. University Press. Scheibel RS, Meyers CA, Levin VA (1996). Cognitive dys- Vanderploeg RD (2000). Interview and testing: the data function following surgery for intracerebral glioma: collection phase of neuropsychological evaluations. In influence of histopathology, lesion location, and treat- Vanderploeg RD (ed.) Clinician’s Guide to Neuropsycho- ment. J Neurooncol 30: 61–69. logical Assessment (2nd edn.) (pp. 3–38). New Jersey: Schwartz J, Alster Y, Ben-Tal O et al. (2000). Visual loss Lawrence Erlbaum Associates. following high-dose cytosine-arabinoside (ARA-C). Eur J Varney NR, Alexander B, Macindoe JH (1984). Reversible Haematol 64(3): 208–209. steroid dementia in patients without steroid psychosis. Steffens DC, Byrum CE, McQuoid DR et al. (2000). Hip- Am J Psychiatry 141: 369–372. pocampal volume in geriatric depression. Biol Psychiatry Videbech P, Ravnkilde B (2004). Hippocampal volume and 48: 301–309. depression: a meta-analysis of MRI studies. Am J Psychi- Stoudemire A, Anfinson T, Edwards J (1996). atry 161(11): 1957–1966. Corticosteroid-induced delirium and dependency. Wefel JS, Lenzi R, Theriault RL et al. (2004a). The cognitive Gen Hosp Psychiatry 18: 196–202. sequelae of standard-dose adjuvant chemotherapy in Straus DJ (2002). Epoetin alfa as a supportive measure in women with breast carcinoma. Results of a prospective, hematologic malignancies. Semin Hematol 39 [4 Suppl randomized, longitudinal trial. Cancer 100(11): 2292– 3]: 25–31. 2299. Taphoorn MJB, Klein M (2004). Cognitive deficits in adult Wefel JS, Kayl AE, Meyers CA (2004b). Neuropsychologi- patients with brain tumors. Lancet 3: 159–168. cal dysfunction associated with cancer and cancer thera- Taphoorn MJB, Heimans JJ, Snoek FJ et al. (1992). Assess- pies: a conceptual review of an emerging target. Br J Can- ment of quality of life in patients treated for low grade cer 90(9): 1691–1696. glioma: a preliminary report. J Neurol Neurosurg Psychi- Wefel JS, Meyers CA (2004). Tamoxifen associated neuro- atry 55: 372–376. toxicity in women with breast cancer [abstract]. Final Taphoorn MJB, Schiphorst CP, Snoek JF et al. (1994). Cog- program 32nd Annual International Neuropsychological nitive functions and quality of life in patients with low- Society Conference: 185–186. grade gliomas: the impact of radiotherapy. Ann Neurol Wiklund I (2004). Assessment of patient-reported out- 36: 48–54. comes in clinical trials: the example of health-related Tarbuck AF, Paykel ES (1995). Effects of major depression quality of life. Fundam Clin Pharmacol 18: 351– on the cognitive function of younger and older patients. 363. Psychol Med 25: 285–296. Willke RJ, Burke LB, Erickson P (2004). Measuring treat- Tarter RE, Butters M, Beers SR (2001). Medical Neuropsy- ment impact: a review of patient-reported outcomes chology (2nd edn.). New York: Kluwer Academic/Plenum and other efficacy endpoints in approved product labels. Publishers. Control Clin Trials 25: 535–552. Temple RM, Deary IJ, Winney RJ (1995). Recombinant Wolkowitz OM, Reus VI, Weingartner H et al. (1990). Cogni- erythropoietin improves cognitive function in patients tive effects of corticosteroids. Am J Psychiatry 147: 1297– maintained on chronic ambulatory peritoneal dialysis. 1303. NephrolDialTransplant10: 1733–1738. Yaffe K, Grady D, Pressman A et al. (1998). Serum estro- Tierney A, Leonard R, Taylor J et al. (1991). Side effects gen levels, cognitive performance, and risk of cognitive expected and experienced by women receiving decline in older community women. J Am Geriatr Soc chemotherapy for breast cancer. Br Med J 302:272. 46(7): 918–920. 6

Neuropsychological assessment of children with cancer

Louise Penkman Fennell and Robert W. Butler

Introduction At birth, infants have more than 100 billion neurons (Berger, 2005). In the first 2 years of life the Neuropsychology, broadly defined, is the study of brain undergoes a period termed transient exuber- brain–behavior relationships. The term was coined ance when as many as 15 000 new connections are by William Osler in the early 1900s and gained wider established per neuron (Thompson, 2000). Follow- appeal in the 1960s. The field was influenced by ing this period of rapid growth, there is a period pioneers in neuroanatomy, neurology, and physiol- of rapid elimination of synapses called “pruning” ogy, who began to explore the brain’s functional- that peaks in adolescence and is variable across dif- ity (Broca, 1865; Hughlings-Jackson, 1931; Lashley, ferent brain regions (Kolb & Wishaw, 2003). Fre- 1950; Wernicke, 1874). Modern neuropsychology quently used synaptic connections are strengthened represents a blend of careful clinical observation and unused connections are pruned. It is through grounded in the pioneering work of Alexandr Luria this process of pruning that distributed neural net- (1973), and a more actuarial approach that utilizes works that subserve functions such as language psychometric instruments to describe and quantify are developed. There are certain critical periods an individual’s functioning (Halstead, 1947; Reitan, during development when the brain “expects” cer- 1974). Neuropsychology has become a science of tain experiences from the environment (e.g., lan- human behavior as it is influenced by brain func- guage stimulation) in order to develop. Studies of tioning and by social, psychological, and cultural children who were deprived of appropriate envi- contexts. ronmental stimulation showed that they failed to Pediatric neuropsychologists are concerned with develop expected skills (Ames, 1997; Gunnar, 2001; developmental issues and take into account the Rutter, 1998). However, it appears that these chil- genetic, medical, environmental, behavioral, and dren can demonstrate accelerated development and sociocultural influences that impact the maturation the brain can recover to reach age-appropriate lev- of a child (Baron, 2004). The human nervous system els if the deprivation is brief in duration (Rutter, is never static and development occurs across the 1998). This supports the concept of critical periods lifespan. However, the rapidity of development in for development. childhood and adolescence calls for a specific devel- In addition to synaptogenesis, dendritic arboriza- opmental focus when conducting evaluations with tion, and pruning, myelination is an important this age group. neurodevelopmental process that occurs across the

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

56 Chapter 6. Neuropsychological assessment of children with cancer 57

lifespan to varying degrees. Myelin is the fatty possess less well-established, acquired knowledge sheath that surrounds the axon and is made from than adults. Also, their foundational skill base is less. oligodendrocytes. It serves to insulate the axon So when an injury disrupts the ability to acquire new and to prevent leakage of the electrical potential, information, a child will be more compromised than and is responsible for speeded conduction of neu- an adult whose lifetime of education has created a ral impulses. Myelination begins in the pre-natal broad base of knowledge and skills. period and continues into the adult years. Like synaptogenesis and pruning, myelination occurs in different cortical regions at different rates and is Childhood cancer considered a gross index of cerebral maturation (Anderson, 2001; Kolb & Wishaw, 2003). Dramatic progress in pediatric oncology has It is important to understand the dynamic nature resulted in increased survival rates for certain of the child’s nervous system in order to adequately types of cancer since the late 1980s to the time understand how development may be interrupted of writing. In the 1960s fewer than 5% of children or changed by an illness, injury, or toxic exposure. diagnosed with acute lymphoblastic leukemia (ALL) The plasticity of the child’s nervous system may survived for more than 5 years. In 2004 the 5-year be most dramatically illustrated in cases of hemi- survival rate for children with standard-risk ALL spherectomy. When one hemisphere is removed was approximately 85% (Ries et al., 2005). Likewise, (e.g., to control refractory epilepsy), if the child is children with medulloblastoma were almost certain young enough some of the functions of the resected to die in the mid 1990s. In 2005, medulloblas- hemisphere can be taken over by the remaining, tomas had a 5-year survival rate of 70% (Reddick healthy hemisphere. In contrast, the mature ner- et al., 2005). These dramatic increases in cure rates vous system is based on well-established connec- are attributable in great part to large-scale, well tions and is less malleable. Although the adult ner- co-ordinated clinical trials carried out by groups vous system does retain some degree of plasticity, such as Children’s Oncology Group (COG) and it is not likely to reorganize in response to a catas- its predecessors Pediatric Oncology Group (POG) trophic neurological event (e.g., stroke or severe and Children’s Cancer Group (CCG). However, traumatic brain injury) in the same way that a devel- these medical advances have yielded a cohort of oping nervous system may. childhood cancer survivors with mild to severe Having said this, it is now understood that suf- neuropsychological impairments that are the result fering brain damage early in life does not always of life-saving treatments. result in fewer deﬁcits. The premise that brain dam- Acute lymphoblastic leukemia is the most com- age early in life is associated with a better outcome mon form of childhood malignancy. There are than brain damage later in life is known as the Ken- approximately 2400 individuals under the age of 20 nard Principle (Finger, 1999; Finger & Wolf, 1988). diagnosed with ALL each year in the United States This is a general statement and does not apply in (Smith et al., 1999). The peak incidence is during the many cases. The type and extent of injury determine pre-school years. Brain tumors are the second most whether a mature or developing nervous system is common neoplasm of childhood (Ries et al., 2005). more resilient: greater recovery of function is typi- Of these tumors, approximately 60% occur in the cally seen following focal rather than diffuse injuries posterior fossa. The most common are gliomas such and many childhood injuries are diffuse in nature. as astrocytoma, followed by medulloblastoma and Although compared to adults a child’s nervous sys- ependymoma. Less commonly diagnosed are high- tem may show more plasticity, deﬁned as the capac- grade supratentorial gliomas and pineal tumors ity of the brain to continuously change its structure (Strother et al., 2002). Although ALL and brain and ultimately its function (Kolb, 1995), children tumors are very different diseases, they have both 58 Section 1. Cognition and the brain

been recognized as posing significant risks to neu- for prophylaxis and has contributed to the high rodevelopment and ultimately academic and voca- cure rate. Intrathecal chemotherapy consists of tional functioning. methotrexate (MTX) delivered in combination with other agents or alone, administered directly into the cerebrospinal fluid via lumbar puncture. When Brain tumors CRT and IT-MTX are administered concurrently, the risk for neuropsychological dysfunction increases The diagnosis of a brain tumor carries with it a (Bleyer et al., 1990). risk to neuropsychological functioning. A tumor Young children are at the highest risk for neu- is a space-occupying lesion that can cause deficits ropsychological dysfunction and subsequent aca- through compression of brain structures. An demic difficulties due to radiation and/or chemo- obstructive brain tumor can cause hydrocephalus therapy treatment (Hopewell, 1998; Mulhern et al., and an increase in intracranial pressure, which 1998; Ris et al., 2001; Walter et al., 1999). The imma- can ultimately be fatal if brainstem structures are ture brain is highly vulnerable to the diffuse effects compressed. Deficits can be focal or more diffuse of radiation and IT chemotherapy: young children depending upon tumor location and associated have fewer well-established skills than older chil- complications (Ris & Noll, 1994). dren when they receive treatment with its poten- Surgery is an important component of both diag- tially disruptive influence on their development. nosis and treatment unless the tumor is deemed It is generally thought that skills that are in a inoperable. Medulloblastomas are the most com- critical stage of development when a neurological mon malignant brain tumor of childhood and most insult occurs may be most susceptible to disruption cases are diagnosed in the first decade of life (Dennis et al., 1998). The rapid growth of the ner- (Strother et al., 2002). They are aggressive tumors vous system during early childhood is the reason and treatment usually includes surgery, chemother- for its significant vulnerability to treatment-related apy, and cranial-spinal radiation therapy (CRT) brain injury. It is important to also note that research with a boost of focal radiation to the posterior indicates that treatment-related brain injury is pro- fossa. Ependymomas are also treated with surgery, gressive and that cognitive and academic deficits cranial-spinal or focal radiation, and chemother- may not become evident until 12 months to several apy. Low-grade gliomas are generally treated with years post-treatment (Hoppe-Hirsch et al., 1995; surgery alone. See Chapter 14 for a more complete Langer et al., 2002; Moleski, 2000). review of childhood brain tumors.

Pattern of neuropsychological deficits Acute lymphoblastic leukemia As stated, children treated with IT chemotherapy Children with ALL are also at risk for neuropsycho- and/or CRT are at risk for neurodevelopmental logical dysfunction related to their treatment. Cur- delay and subsequent neuropsychological deficits. rently, standard treatment for ALL does not include These deficits can lead to academic struggles and CRT. However, children who present with ALL that ultimately lower vocational success and health- has spread to the central nervous system (CNS) related quality of life (HRQOL). Research examin- or whose disease relapses in the CNS do receive ing the HRQOL of childhood cancer survivors docu- CRT at a lower dose than that typically adminis- ments the lowest level for children with CNS tumors. tered for treatment of a malignant brain tumor. Children with leukemia and lymphoma have the Because the CNS is a site of recurrence for this next poorest HRQOL (Speechley et al., 2002). For disease, intrathecal (IT) chemotherapy is used a discussion of the pattern of neuropsychological Chapter 6. Neuropsychological assessment of children with cancer 59

deficits observed following treatment for a brain the prevalence of some degree of neuropsycholog- tumor in childhood, please refer to Chapter 14 of ical impairment in this group of children to be at this text. least 30%. In a number of studies examining chil- The neuropsychological outcome of CRT was first dren treated for ALL with CNS-directed chemother- studied in survivors of childhood ALL (Rowland apy only, Brown and colleagues showed that chil- et al., 1984). Similar to brain tumor survivors, intel- dren who received CNS prophylactic chemotherapy lectual functioning has been shown to be com- were more impaired than their sibling controls on promised in children with ALL treated with CRT measures of dominant hemisphere skills (Brown & (Cousens et al., 1988; Fletcher & Copeland, 1988; Maden-Swain, 1993; Brown et al., 1992, 1998). Other Rowland et al., 1984; Spiegler et al., 2006). However, authors have shown negligible impact on overall IQ scores have also been shown to remain within the intellectual functioning when children with ALL normal range (Langer et al., 2002). In general, the are treated with IT chemotherapy alone (Langer level of global impairment tends not to be as severe et al., 2002; Von der Weid et al., 2003). Spiegler as in the brain tumor population. These children and colleagues (2006) assessed neurocognitive out- are treated with a relatively low dose of CRT and a come in survivors of early childhood ALL 5–20 years high dosage has been shown to be a risk factor for post-treatment. They compared children treated poor intellectual outcome (Picard & Rourke, 1995). with high-dose and very high-dose intravenous The pattern of deficits documented for children MTX and IT-MTX to a group of children treated treated for ALL is similar to those seen in child- with CRT and the same chemotherapy backbone. hood brain tumors. Mulhern and Butler (2005) Their results showed no difference from population reviewed this literature and underscored the fact norms for the chemotherapy group, with survivors that non-dominant hemisphere functions seem to treated with CRT showing deficits. However, Moleski be differentially impacted. This includes non-verbal (2000) cautions that IT-MTX cannot be considered memory, visual motor integration, visual spatial a benign agent. In her comprehensive review, she reasoning, and visual perceptual abilities. Simi- summarizes the mixed research findings available. larly, in an earlier comprehensive review Moleski Her overview indicated that in all studies using sib- (2000) documented attention and non-verbal mem- ling controls instead of the general norms, there was ory problems as the most robust findings. In addi- a significant difference in overall intellectual function, these children show difficulties with fine motor tioning for patients treated with IT chemotherapy. and tactile-perceptual functioning due to the use of In addition, children treated with CRT alone fare vincristine as a chemotherapy agent (Moleski, 2000). better than children treated with IT-MTX in com- Lehtinen and colleagues (2002) studied motor- bination with CRT, suggesting a possible synergistic evoked potentials in children treated for ALL. Their neurotoxic effect (Bleyer et al., 1990). findings support the persistence of peripheral, but Researchers interested in the neuropsychological not central, motor abnormalities in children up impact of treatment for ALL are focused on inves- to 5 years post-treatment. These difficulties mani- tigating core cognitive processes such as attention, fest themselves as gross and fine motor problems processing speed, and working memory. Deficits and children struggle with co-ordination for playing in these core processes are seen as underlying the sports and with handwriting. In general, ALL sur- overall intellectual deterioration and academic dif- vivors demonstrate more difficulties with arithmetic ficulty that has been observed. Processing speed, than with language based academic skills (Moleski, working memory, vigilance, and cognitive flexibility 2000). deficits have been reported in ALL survivors (But- The literature is mixed with respect to neuropsy- ler & Copeland, 2002; Langer et al., 2002; Schatz chological outcome in children treated with IT-MTX et al., 2000). Buzier and colleagues (2005) showed and not CRT. Mulhern and Butler (2004) estimate that IT chemotherapy alone has been associated 60 Section 1. Cognition and the brain

with attention dysfunction. They identified intensi- ropsychological assessment when treatment ends. fied treatment and young age as being associated This provides information for the school about any with worse performance. special accommodations that need to be made to In sum, deficits have been reported in a num- optimize a child’s re-integration into school. For ber of different areas. However, attention, process- children with ALL, treatment duration is approxi- ing speed, executive functioning, non-verbal prob- mately 30 months and many children are able to lem solving, and non-verbal memory appear to continue at school during treatment. For these chil- be most impacted for children with brain tumors dren, it may be necessary to conduct a neuropsy- and ALL. Verbal skills tend to be better preserved chological assessment during treatment, particu- except in very young children. However, the presen- larly if teachers are noticing changes in the child’s tation of each individual child will be mediated by performance at school. both risk and protective factors such as gender (see Re-testing should not be carried out exces- Moore, 2005 for a brief overview), age, treatment, sively. Children can be “over-tested.” Some children pre-existing complications, and parental social class remember the test materials and this can contribute (see Moleski, 2000, for a brief discussion). to significant practice effects. Sometimes neuropsychologists will use alternative test forms or different tests to attempt to control for this problem. How- Implications for neuropsychological ever, only a handful of tests are available in alterna- assessment tive forms. In addition, this is not a perfect solution since alternative forms have been shown to Children treated with CRT and/or IT chemotherapy evidence practice effects when administered seri- with MTX are at substantial risk of neuropsy- ally (Beglinger et al., 2005; Franzen et al., 1996). chological dysfunction with implications for When different tests are used, it is sometimes dif- academic, vocational and social functioning, and, ficult to compare scores for the purposes of eval- ultimately, quality of life. Neuropsychological uating the child’s progression of learning because follow-up of these survivors is now described as the test norms may be very different. In an optimal standard care (Children’s Oncology Group, 2004; situation a child should be administered the same Duffner, 2004). measures when possible, at an appropriate inter- test interval. A general practice of clinicians is to Serial assessments use 6–12 months as the minimum inter-test interval. However, test manuals provide reliability estimates Children treated for brain tumors or ALL evidence for much shorter durations of time. Currently, data a delayed onset of neuropsychological deficits that are lacking for inter-test intervals that more closely may progress over time. As such, it is important to approximate clinical practice. For a thorough dis- take a serial approach to testing this group of chil- cussion of approaches to detecting clinically sig- dren. It does not suffice to test a child once at the nificant change over time in children the reader is end of treatment. Serial assessment presents par- referred to Baron (2004). ticular challenges in the selection of instruments, Decisions made about when to test a child may be detection of meaningful change, and determining dependent on the child’s age when the first assess- when and how often to test. More often now, serial ment is conducted. Obviously, younger children will testing is built into treatment protocols. However, need to be seen more often. Some neuropsychol- for many children, planning when testing should ogists make it their practice to leave it up to fam- take place is left up to the neuropsychologist. ilies and educators to contact them if the child Depending upon the age of the child and treat- encounters more problems over time. This can be ment protocol, it may be helpful to conduct a neu- a way to manage heavy demands on one’s time. Chapter 6. Neuropsychological assessment of children with cancer 61

However, some parents find it more difficult to be the child cancer survivor continues beyond the test- proactive than others, and some children risk not ing office, and that communication is established receiving adequate follow-up if educators or parents between the neuropsychologist and the educational do not initiate the assessment. Other practitioners professionals at the child’s school. The neuropsy- set up a more structured plan where children are chologist’s role is to teach these individuals about booked in advance to come back for a re-assessment the nature of the child’s injury and subsequent in 12–36 months following the first assessment. The deficits and how they translate into practical chal- time interval may depend on the age and grade lenges to learning and social development in the placement of the child and the severity of neuropsy- classroom. Furthermore, the neuropsychologist is chological dysfunction. It may be helpful to see instrumental in ensuring that the child receives ade- children at important educational transition points. quate follow-up over time and that new teachers For example, a child who is assessed in the third each year are educated about the child’s specific grade (at about 9 years of age) may receive a follow- needs. up neuropsychological assessment when he or she transitions to junior high school (sixth or seventh grade, i.e., aged 12–14 years) and then a final assess- Approaches to assessment ment at the beginning of high school. This updated assessment information can be taken with the child Neuropsychological assessments for child cancer into their new educational setting so that educators survivorsshouldbecomprehensivebecauseamul- can make the most well-informed decisions about titude of deficits have been described. Similarly, appropriate placement and educational focus. research studies reflect information about groups of children whereas neuropsychological assessments in clinical settings are carried out within the idio- Liaison with the school graphic context, and generalized statements drawn from research findings may or may not apply. Unfortunately, most very busy educators and school Research findings serve as a guideline to suggest psychologists do not have the time or the resources which areas of cognitive function are likely to be to be particularly well informed about the issues impacted and help to direct clinical assessment facing child cancer survivors when they return to planning. school. Educators may assume that once the can- Traditionally, neuropsychological assessment has cer treatment is over that the child should return been a fairly lengthy process. This is due to the num- to their pre-morbid level of functioning. Similarly, ber of areas of cognitive function that are evalu- they may assume that an end-of-treatment neu- ated and the need for rest breaks to ensure adequate ropsychological assessment that revealed no areas effort on the part of the examinee. However, in more of concern may signal a positive educational future recent years various pressures have forced neuro- for a child. However, this is clearly not always the psychologists to find ways to work more quickly case because these children often sustain a pro- yet still provide a comprehensive description of the gressive injury with deficits that manifest over time. strengths and weaknesses of their patient and pro- Many times a neuropsychologist is not available vide relevant, useful recommendations. to evaluate child cancer survivors and they are We present an overview of a thorough neuropsy- seen by a school psychologist who is not trained chological assessment that addresses the areas of in brain development issues or the underlying deficit experienced by child cancer survivors (see neuropathology of brain injury acquired through Table 6.1 for a discussion of individual tests). The cancer treatment. When a neuropsychologist is approach to assessment considered here is in the available, it is important that their involvement with spirit of the flexible battery. With the fixed battery 62 Section 1. Cognition and the brain

Table 6.1. Areas to evaluate in a comprehensive neuropsychological assessment and sample measures for school-age child cancer survivors: neurocognitive measures

Area of function Sample testsa Comments

Overall intellectual Wechsler tests (Wechsler, 2003, The Wechsler tests are the most widely used of the functioning 2002, 1999, 1997), i.e.: intelligence scales in North America and have strong psychometric properties. – Wechsler Preschool and Primary Intelligence Scale (3rd edn.) (WPPSI-III) – Wechsler Intelligence Scale for Children (4th edn.) (WISC-IV) – Wechsler Adult Intelligence Scale (3rd edn.) (WAIS-III) – Wechsler Abbreviated Scale of Intelligence (WASI) Differential Ability Scales (DAS) The DAS is an alternative to the Wechsler tests and (Elliott, 1990) provides a General Cognitive Ability (CGA) score. It has a good normative base and strong psychometric properties. However, the DAS lacks a comparable battery across the age ranges because subtests change as do the factors that subtests load on Stanford-Binet Intelligence Scale The Stanford-Binet Scales provide large ceilings and (5th edn.) (SB5) ﬂoors and may provide better coverage when (Roid, 2003) evaluating individuals at the extreme ends of the distribution Learning and memory California Verbal Learning Test for The CVLT-C provides numerous useful indices Children (CVLT-C) (Delis et al., including learning slope which can be compared 1994) against normative data Children’s Memory Scale (CMS) Story recall tests such as those on the Wide Range (Cohen, 1997) Assessment of Memory and Learning (WRAML) Faces and the CMS provide useful complementary Picture memory information about the ability to learn and Stories remember meaningful information. Used together, a comprehensive picture of verbal memory abilities can be obtained Wide Range Assessment of Both the WRAML and the CMS provide tests of visual MemoryandLearning(2ndedn.) memory that do not require a motor response. This (WRAML-II) (Sheslow & Adams, can be very useful when evaluating children who 2004) may not be able to demonstrate what they Story Memory remember because of problems with visual motor Verbal Learning output Picture Memory Chapter 6. Neuropsychological assessment of children with cancer 63

Table 6.1. (cont.)

Area of function Sample tests Comments

Design memory Rey–Osterrieth Complex Figure Test The RCFT has a number of administration and scoring (RCFT) (Corwin & Bylsma, 1993) systems. It provides considerable information and is used as an indicator of visual spatial abilities, executive functions (particularly planning and organizational abilities), and visual motor integration in additional to incidental visual memory Attention and Freedom from Distractibility Index FDI provides information about a child’s short-term working memory (FDI) of the WISC-IV attention and working memory abilities. Other widely available batteries contain subtests sensitive to working memory dysfunction [e.g., WRAML-II, CMS, Woodcock–Johnson Tests of Cognitive Ability – (3rd edn., TEA-Ch)] Test of Everyday Attention for The TEA-Ch is a unique test battery designed to Children (TEA-Ch) (Manly et al., comprehensively evaluate attention. It is very long 1999) when given in its entirety, but it does not provide a global score so giving the subtests individually does not violate the psychometrics of the test. It provides measures of sustained attention in the auditory modality. Children enjoy the interesting activities and materials and it is available in alternative forms Continuous Performance Tasks (CPTs) CPTs evaluate sustained attention or vigilance. They (various) provide numerous indices such as omission and commission errors and response speed. They are not specific enough to be diagnostic of attention deficit hyperactivity disorder alone but provide useful information about inattentiveness and impulsivity. They are tedious and children tend to dislike them and compliance can be an issue Rating scales – e.g., Child Behavior Parent, teacher and self-rating scales are helpful to Checklist (CBCL) (Achenbach, obtain information about attention performance in 2001); Behavior Assessment Scale real-world settings. Rating scales such as the CBCL for Children (BASC) (Reynolds & and BASC provide coverage of a wide range of Kamphaus, 2004); Conners’ Rating problem areas but do not examine attention Scales – Revised (Conners, 2001) functioning. A measure such as the Conner’s is more specific to attention and behavioral regulation difficulties Processing speed WISC-IV Processing Speed Index (PSI) The Perceptual Organization factor of the WISC-IV is influenced by processing speed but PSI provides a more direct measure. Care must be taken to rule out confounds of slow processing speed such as fine motor or visual tracking difficulties Academic fluency measures from the Academic fluency measures provide estimates of Woodcock–Johnson (3rd edn.) Tests processing speed in a relevant context that is of Achievement (WJ3) (McGrew applicable to school functioning et al., 2001) (cont.) 64 Section 1. Cognition and the brain

Table 6.1. (cont.)

Area of function Sample tests Comments

Executive function Wisconsin Card Sorting Test (WCST) The WCST and CT are less sensitive to anterior (Heaton, 1981) lesions than originally believed but they remain useful measures of planning, shifting, flexible Category Test (CT) thinking, and concept formation. Both (Knights & Tymchuk, 1968) demonstrate significant practice effects because of the novel learning component and their use in serial assessments should be carefully considered Delis-Kaplan Tests of Executive The D-KEFS provides a compilation of executive Function (D-KEFS) (Delis et al., 2001) function measures that are normed on the same national sample Tower of London (Shallice, 1982) Tower tasks have grown in popularity and may be Tower of Hanoi (Simon, 1975) less sensitive to practice effects than the WCST or CT. For a thorough review of the strengths and weaknesses and psychometric properties of Tower tasks see Baron (2004) Behavior Rating Inventory of Executive The BRIEF was developed in an effort to address Function (BRIEF) (Gioia et al., 2000) issues of ecological validity and the insensitivity of laboratory-based measures of EF.The BRIEF is a parent and teacher report instrument with a large normative database. There is now a pre-school version. A strength of this measure is that it is a measure of EF developed specifically for children and is not a downward extension of an adult measure Language Comprehensive Test of Phonological The CTOPP may be used as a follow-up measure for Processing (CTOPP) (Wagner et al., in-depth assessment when a child shows 1999) difficulties with the development of language-based skills, particularly reading Receptive One-Word Picture Vocabulary The other measures listed will provide estimates of Test (ROWPVT) (Brownell, 2000) expressive and receptive vocabulary development Expressive One-Word Picture as well as language comprehension skills. Taken Vocabulary Test (EOWPVT) (Gardner, together with the verbal subtests of the overall 2000) intellectual measures, a screening of language function can be accomplished Peabody Picture Vocabulary Test (PPVT) (Dunn & Dunn, 1997) Token Test (De Renzi & Faglioni, 1978) Visual perceptual, Beery Developmental Test of Visual Screening for adequate visual perceptual and motor, spatial, Motor Integration (VMI) (Beery & visuomotor abilities using tasks such as the VMI and Beery, 2004) can help in the interpretation of more complex constructional tasks such as the RCFT abilities Block Design, Matrices subtests of the Wechsler tests Judge of line orientation (Benton et al., 1983) Chapter 6. Neuropsychological assessment of children with cancer 65

Table 6.1. (cont.)

Area of function Sample tests Comments

Rey–Osterreith Complex Figure Test (RCFT) Motor and sensory Grooved Pegboard (Klove, 1963) Tests listed represent commonly used measures of perceptual Purdue Pegboard (Tiffin & Asher, 1948) fine motor skills, motor speed, and sensory Astereognosis (Benton et al., 1994) perceptual ability Finger Tapping (Halstead, 1947) Academic Wechsler Individual Achievement Test The WIAT-II or WJ3 are preferred measures when achievement (2nd edn.) (WIAT-II) (Wechsler, 2001) in-depth information about a child’s academic Woodcock–Johnson (3rd edn.) Tests of functioning is required. The various subtests can Achievement (WJ3) help to pinpoint which specific academic areas are Wide Range Achievement Test (3rd problematic for a particular child. The WRAT-3 is a edn.) (WRAT-3) (Wilkinson, 1993) screener only and should be used as such Psychosocial BASC-II; CBCL Rating scales that provide coverage of a range of both functioning Adaptive Behavior Assessment System problem and adaptive behaviors are good starting (2nd edn.) (ABAS-II) (Harrison & points to screen for psychological and social issues Oakland, 2003) that may impact a child’s functioning. These may be followed up with a more in-depth assessment of a child’s psychological functioning if necessary The ABAS-II is a measure of adaptive functioning. It is useful for determining how a child may be functioning in their home, school and community settings aComplete references for all tests are found in the test list at the end of the chapter. approach, a pre-determined battery of tests is given the CNS. The child’s performance will be influenced to every child. The battery is generally designed by the degree of impairment of the CNS and by the to evaluate all necessary areas. With the flexible complexity of the task. It is an evaluation of a child’s battery approach, a core battery of tests is given ability to take in or receive information through in order to generate hypotheses about a child’s their senses, to integrate and process this informa- areas of strengths and weakness. These hypotheses tion, and to interact with their environment through are followed-up by the selection of additional tests the production of a response. Cognitive processes based on the child’s performance on the core bat- impacting a child’s ability to adequately take in tery. This approach is also influenced by the process information in addition to basic sensation and per- approach (Kaplan, 1988), which underscores the ceptual ability include attention, processing speed, importance of attending to the individual’s process and working memory. Memory abilities also impact or test-taking style to guide the neuropsychological the ability to process information. Higher level pro- exploration. The end result is that the assessment is cessing of information includes visual spatial rea- tailored to the individual child and more tests are soning, language processing and problem solving, given in an area where a child is having difficulty to and the integration of information across modali- further tease out the nature of that difficulty. ties in association areas of the brain. Skills impact- Itmaybehelpfultothinkofaneuropsychologi- ing a child’s ability to produce meaningful output cal assessment as a challenge to the functioning of in order to interact with his or her environment 66 Section 1. Cognition and the brain

Intake of information - Sensation - Perception (visual, auditory, somasthetic)

Processing information Executive functions Attention, speed - visual spatial reasoning - organization and of processing, - language inhibition of behavior working memory - problem solving - goal-directed behavior - integration across modalities

Output of information

Memory - language - motor skills

Figure 6.1. Neuropsychological Assessment Heuristic

include fine and gross motor skills and lan- sive assessment of specific areas of weakness in guage. A child’s ability to produce meaningful out- order to make useful recommendations for home put is moderated by executive functions, which and school. help to organize responses and inhibit behavior when necessary. The dynamic interplay of input- Areas to evaluate processing/integration-output is a useful heuristic to consider when trying to ensure that there is ade- Overall intellectual functioning quate coverage of domains of cognitive function in a comprehensive neuropsychological assessment of Both research and clinical reports typically use the child cancer patient or survivor (see Figure 6.1). overall intellectual level as a global index of a In addition, psychological functioning is a key part child’s functioning. The subtest scores for each of the neuropsychological assessment and will be individual are compared to each other (ipsative discussed further at a later point in this chapter. comparison), as well as to population norms, thus Following our overview of the key areas to eval- discerning areas of relative strength and weakness uate, we will present two approaches to neuropsy- within an individual child’s neuropsychological pro- chological assessment of the child cancer patient file. The IQ score will place the child in perspective that represent attempts to provide rich, meaningful relative to others of similar age within the general information in a short period of time through the population (Baron, 2004) and it is a good predictor use of a streamlined core battery of tests. These of school performance (see Sternberg et al., 2001, for assessments might best be thought of as screen- an overview of the predictive value of the IQ score). ing assessments. In particular situations, neuropsy- It is extremely important to measure this domain chologists may need to follow-up with more inten- of functioning in more severely compromised Chapter 6. Neuropsychological assessment of children with cancer 67

children. Intellectual functioning is one of the diag- Learning and memory nostic criteria, together with adaptive functioning deficits, for a diagnosis of mental retardation. Some The assessment of learning and memory is nec- children treated for malignant brain tumors with essary to the comprehensive neuropsychological CRT at a young age meet criteria for this diagno- assessment and nowhere is it more important than sis. In addition, some of the more commonly used in the assessment of children and adolescents. Fail- intellectual tests provide a large normative base ures to learn and remember new information and and opportunities for detailed statistical analysis of to retrieve information at a later time will ulti- strengths and weaknesses. Another benefit of using mately lead to impaired functioning because of a well-known and well-normed instrument such as the importance of acquiring new knowledge to the Wechsler scales is that it is possible to track future learning and development. The assessment rate of learning over time by examining normative of learning and memory can be conceptualized by and raw scores in an individual child over different modality (visual and auditory-verbal), as well as assessments. This is not always possible with other within an information-processing model of mem- instruments where ceiling effects, learning effects, ory functioning (i.e., encoding, consolidation, stor- small normative samples or changes in items across age, retrieval, and recognition). A thorough discus- age ranges can limit the ability to track learning sion of the many models of memory functioning rates over time. is beyond the scope of this chapter. The interested The IQ tests are global instruments and should reader is referred to Lezak et al. (2004) for a review be interpreted as such. Oftentimes, detailed con- of memory models. clusions about brain function are generated based Memory batteries are often lengthy and time con- on poor performance on one subtest of a broader suming and contain a number of subtests that may intellectual measure by inexperienced clinicians. or may not add useful clinical information. For this According to Baron (2004) specific neuropsycholog- reason, Baron (2004) recommends a needs-based ical instruments are much better suited to mak- selection of individual subtests. This may compro- ing detailed inferences about brain function. Often- mise the statistical properties of the test to some times the instrument used will be dictated by the degree, but the experienced clinician can remain school board or institution in which an individual cognizant of these limitations and integrate the psychologist works. information in a meaningful way. In a different viewpoint, Lezak et al. (2004) advo- A child’s capacity and motivation to adequately cate strongly against the use of overall IQ indices. attend to information should be sufficiently An in-depth discussion of the argument is beyond screened because of the importance of intact atten- the scope of this chapter. Briefly, they caution that tion to support memory functioning. The ability with any derived score, important information is to retain information over the short term needs lost and the overall score is not always representa- to be evaluated by tests of immediate memory. tive or meaningful. They also describe unfortunate The duration for “long-term” memory assessment uses of an IQ “cut off” score that limits access to in a day-long neuropsychological assessment is services for individuals. It is important to consider obviously limited and most memory tests include these issues and the well-trained clinician should a delay that is about 20 min in duration. More examine test scores from every vantage point before information about a child’s ability to learn, retain, making inferences about brain function. We believe and retrieve information over the long term can that an overall IQ score can be useful to neuropsy- be gleaned from background information about chological assessment. However, a complete under- the child’s functioning in their home and school standing of the limitations of this overall score is environment. This contextual information should crucial. be considered together with the results of memory 68 Section 1. Cognition and the brain

testing to develop an overall conceptualization of a support the “taking in” of information. If a child child’s memory abilities. cannot attend to information, or processes informa- Learning should be assessed by both repeated and tion extremely slowly, he or she will have limited one-time exposure to determine a child’s capacity opportunity to optimally utilize cognitive resources. for single trial learning as well as their rate of learn- The neuropsychological late-effects research base ing (e.g., learning curve) over time. Memory for rote has pointed to deficits in these areas as driving information (e.g., a list of words), as well as mem- many of the observed declines in other areas of ory for information with more salience (e.g., stories) functioning and subsequent academic failure. should be assessed. When assessing for new learning, both free recall and recognition formats should Attention be used. This allows for delineation of whether difficulties in remembering are related to problems with Attention is not a unitary construct, despite com- encoding or retrieval. mon reference to a general skill called “attention.” Non-verbal memory has been implicated as an There are several models of attention that influence area of deficit in child cancer survivors. Therefore, clinicians’ approaches to its evaluation and a thor- it is an area that should not be omitted. Non-verbal ough discussion is beyond the scope of this chap- memory is usually evaluated through a reproduc- ter (e.g., Mirsky et al., 1991; Posner & Petersen, 1990; tion format (e.g., drawing) or through a recognition Sohlberg & Mateer, 1987). Drawing from these mod- format (e.g., pointing). Obviously, these two formats els, the subset of skills that fall under the overall are not equivalent and may provide additive infor- rubric of attention is broad and includes: sustained, mation. This is particularly true in the case where a focused, selective, divided, alternating, shifting, and child has visuomotor difficulty and is challenged by resistance to distraction. Sohlberg and Mateer’s drawing. It will be difficult for the clinician to ascribe (1987) model is helpful in a clinical setting because failure on a test of non-verbal memory to memory it conceptualizes attention in a hierarchical manner dysfunction if the child is poor at drawing. In this going from the most basic form of attention, arousal case, a task without visuomotor demands may be and responding to the environment, to higher lev- helpful. els of attentional complexity that demand more cog- Apparent problems in memory can be produced nitive capacity (such as alternating attention). It is by a variety of other difficulties such as prob- helpful for the comprehensive neuropsychological lems with attention, motivation, working memory, assessment to attempt to evaluate attentional func- executive functions, and impaired language com- tioninginasmanyofthesedomainsaspossible.Ata prehension. As well, emotional dysfunction can minimum, clinicians should evaluate span of atten- disrupt learning and memory. It is important to tion, sustained attention, and more complex atten- carefully evaluate for the presence of difficulties tion such as the divided or alternating types. Visual in these areas before determining that a child has scanning is also important to assess because of its memory dysfunction. relevance to schoolwork (particularly reading and test taking). Attention is demanded by many different tasks and conclusions about attention func- Core cognitive processes: attention, tioning can be drawn from subtests of various test processing speed, working memory batteries. Assessment of what are now commonly referred It is important to remain aware that attention to as “core” cognitive processes is critical to any is closely related to motivation (Baron, 2004) and neuropsychological assessment. Intact neuropsy- can fluctuate widely over the course of a day. Test chological functioning is dependent on adequate order and breaks should be considered in order to performance of these important processes as they manage fatigue and motivation, which can impact Chapter 6. Neuropsychological assessment of children with cancer 69

performance on tests of attention. For this reason, Processing speed it is also important to obtain estimates of atten- Given the importance of myelin in the speeded con- tion functioning in daily life through the use of par- duction of neural impulses, it is not surprising that ent and teacher report measures. These estimates processing speed deﬁcits have been noted in child can be obtained through relatively broad parent and cancer survivors with CNS impact. The importance teacher report instruments such as the Behavior of adequate processing speed to success at school Assessment Scale for Children – 2nd Edition (BASC- is evident. If a child cannot keep up with the pace II; Reynolds & Kamphaus, 2004) or through the use of presentation of information or is unable to de- of speciﬁc attention rating scales such as the Con- monstrate knowledge because testing time frames ners’ Rating Scales – Revised (Conners, 2001). are too short, he or she will struggle academically.

Working memory Executive function

Workingmemoryisconsideredbymanytofall The construct of executive function (EF) is very under the broad rubric of executive functions. In broad. It has been described as an umbrella term fact, Baddeley and Hitch’s model (1974) concep- encompassing a number of subdomains (Baron, tualizes working memory as being comprised of 2004). Several useful models of EFs have been devel- a Central Executive which manages two slave sys- oped (Fuster, 1980; Shallice, 1988; Stuss & Alexander, tems: the visual spatial scratchpad and the phono- 2000; Stuss & Anderson, 2004). Executive functions logical or articulatory loop. In addition, working have been conceptualized as encompassing the memory tasks have been shown to activate the skills required for purposeful, goal-directed behav- prefrontal cortex (Goldman-Rakic, 1992). Working ior (Anderson, 2001). For this reason, EF skills can memory is also considered a form of attention. In be considered to play a very important role in the fact, Baddeley himself has written a chapter entitled quality of an individual’s output and therefore their “Working memory or working attention?” (Baddeley interaction with the environment. The assessment 1993) which captures the difficulty in differentiating of EF in the pediatric oncology population will be between the two concepts. Some may think of it as discussed here in general terms, recognizing that a form of memory given its label. However it is con- considerable discussion could be devoted to the ceptualized or labeled, working memory is consid- assessment of individual areas of EF alone, such as ered a core cognitive process and is closely related planning or response inhibition. For a more thor- to attentional ability because of its role in support- ough discussion of the assessment of EF in children ing both the input and processing of information. As the interested reader is referred to Anderson et al. discussed above, this is a potential area of deficit for (2001) and Baron (2004). the child cancer survivor. There has been some debate about whether it is Working memory is the attentional store that we appropriate to assess executive functioning in chil- use to hold information in mind for short peri- dren. Given that the prefrontal cortex is the puta- ods of time while we do something with it. For ex- tive seat of EF, that this cortical area is the last ample, while making a phone call the number is to reach maturation (Fuster, 1993), and the find- held in working memory until dialing is finished. ing that children do not perform as well as adults Simple span tasks such as Digit Span, which is a sub- on tests of executive functioning (Chelune & Baer, test of the Wechsler tests of intelligence from ages 6 1986), it may seem inappropriate to assess for this and up, are one way of assessing working memory; domain in young children. Recently, research has however, more complex tasks with greater demands provided evidence for EF skill development in chil- on working memory are more sensitive. dren (Levin et al., 1991; Passler et al., 1985; Welsh 70 Section 1. Cognition and the brain

et al., 1991) and for deficits in these areas fol- ing, spelling, and praxis. To this list we would add lowing brain injury relative to normally develop- both receptive and expressive vocabulary developing children (Anderson & Moore, 1995). However, ment. Basic language skills develop relatively early behavioral descriptors of executive dysfunction in in life. Therefore, they are usually intact in children adults such as poor self-control, impulsivity, and treated for malignant brain tumors and ALL unless poor planning may not be warranted for children, they have intense neurotoxic treatment delivered given the developmental appropriateness of chil- in infancy, prior to language development. How- dren being unable to plan and organize a busy day ever, more subtle difficulties with language may be or demonstrate adequate self-control in some sit- evident. It is important to be aware that children uations. An example of the complexity of assessing treated with cisplatin for a brain tumor may suffer EFs in children and adolescents is provided by Todd hearing loss. This may have an obvious impact on et al., (1996). They found evidence of planning diffi- language development. culties in both brain-injured and normally develop- When children evidence language difficulties it ing adolescents. This underscores the importance is helpful to utilize the specialized expertise of our of basing our assessment of EF in children within a speech and language colleagues. However, neurodevelopmental context and using instruments that psychological assessment should include measures provide an adequate normative database reflecting of language functioning to put the entire assess- the normal development of EF in childhood and ment in context and to document areas requir- adolescence. ing further follow-up by a speech and language An additional challenge in the assessment of EFs pathologist. in any age group is the lack of sensitivity of many laboratory-based tasks to deficits in these areas. Deficits in these skills are often difficult to detect in Non-verbal skills: visual perceptual, visual structured clinical settings with structured assess- motor, visual spatial, and visual constructional ment instruments. Therefore, clinicians often rely abilities on anecdotal evidence from the patient or from family members. Because EF is such a broad term, The non-verbal domain of cognitive functioning is many tests are needed to assess for deficits in probably one of the most complex areas to evalu- each of its subdomains. In her review, Anderson ate. A wide array of skills comprise this domain. It (2001) noted that most test batteries designed for may be helpful to think of visually mediated skills children do not include measures of EF. Despite for the input (i.e., visual perceptual functioning) and these challenges it is important to include the for the output (i.e., visual motor) of information, as assessment of EF when evaluating the child cancer well as skills for higher level processing of visual survivor. Deficits in this area will manifest in daily information (e.g., mental rotation and spatial rea- life as problems with organization, time manage- soning), and complex integrative skills such as the ment, behavioral control, and will negatively impact recognition of facial expression and object recogni- not only learning but also social development. tion. One of the reasons why this is such a difficult area to assess adequately is that it is extremely difficult to create tests that measure discrete areas of Language functioning within this domain. Difficulties in other According to Baron (2004), the speech and lan- areas such as visual acuity, speed, or constructional guage portion of a neuropsychological assessment problems may inadvertently lead a clinician to con- should include screening of: conversational flu- clude that a child has difficulties with non-verbal ency, phonological processing, generative fluency, reasoning. It is very important to rule out deficits in comprehension, repetition, naming, reading, writ- other domains that contribute to task performance. Chapter 6. Neuropsychological assessment of children with cancer 71

Deficits in non-verbal skills can contribute to aca- pertinent and up-to-date school information may demic struggles. For example, a child with spatial be available. However, it can be very useful when difficulties may struggle with arithmetic because he determining where a child is functioning academi- or she cannot line up numbers for mathematical cally when planning for return to school after a long operations. Spatial difficulties can also contribute absence. Tests that provide a more in-depth analysis to difficulties with writing. As the research reviewed of a child’s academic strengths and weaknesses are earlier indicates, this functional domain is impacted preferred over a screener when the goal is to make by CRT and IT chemotherapy. relevant recommendations for school planning. Tests of visual recognition, discrimination and matching can aid in the evaluation of visual per- Psychosocial and adaptive functioning ceptual abilities. More complex perceptual tasks of mental rotation and the understanding of spatial The terms neurocognitive and neuropsychological relationships can tap into more integrative, higher- have been used interchangeably in the literature in level non-verbal reasoning. Difficulties with earlier recent years. However, we would like to propose levels of perception must be ruled out before it can an important distinction. The term “neurocogni- be discerned that a child has a higher-level deficit in tive” implies the assessment of cognition or think- visual reasoning. ing skills and of the integrity of underlying neurological substrates. The term “neuropsychological” implies assessment of an individual’s brain func- Motor and sensory perceptual function tioning, and therefore their various thinking skills or Assessment of basic motor and sensory function is cognition, within the context of their psychological important to rule out deficits in these input/output functioning. To return to our earlier definition, neu- pathways, which may confound assessment ropsychology is a science of human behavior as it is findings. Childhood cancer survivors often have dif- influenced by brain functioning and by social, psy- ficulties with movement and basic sensation related chological, and cultural context. The brain should to their treatment. Vincristine is known to cause not be assessed in isolation but must be consid- peripheral neuropathy. Children who undergo ered within the various contexts of the individual surgery for brain tumors may have cranial nerve person. The impact of anxiety and depression on deficits that interfere with their sensory and motor neuropsychological test performance is well docu- functioning. Children with tumors of the posterior mented (see Lezak et al., 2004, for an overview) and fossa may have problems with movement related to thereforeshouldbeconsideredinaneuropsycho- cerebellar involvement and this should be carefully logical evaluation. observed and considered in the interpretation of For this reason we do not feel that a neuropsycho- test findings. logical assessment is complete without taking into account emotional or intrapsychic and social factors that may impact test performance. The clini- Academic achievement cal interview is a rich source of information about Performance on standardized measures of aca- a child’s emotional functioning. Furthermore, this demic achievement is often included as part of a information can be quantified and compared to neuropsychological assessment although this does normative data through the use of parent- and not fall uniquely within the realm of the neuropsy- teacher-reported behavior-rating scales. Older chil- chologist. The referral question may dictate whether dren can report on their own experience. Specific academic achievement is included in the assess- instruments may be warranted for exploration of ment. It may not be necessary when evaluating a specific concerns about depression or anxiety. This child for a baseline prior to beginning CRT because is not to say that a neuropsychological assessment 72 Section 1. Cognition and the brain

can take the place of a comprehensive and thorough Table 6.2. COG Neuropsychology Test Battery for the evaluation by a clinical psychologist specializing in 6- to 16-year-old age group. Neurocognitive children’s emotional disorders. However, the “psy- measures. WASI Wechlser Abbreviated Scale of chology” in neuropsychology should not be forgot- Intelligence, WRAT-3 Wide Range Assessment Test ten and a child’s emotional state should be screened (3rd edn.), CVLT-C California Verbal Learning Test for as part of the neuropsychological evaluation. Emo- Children, CPT-II Continuous Performance Test (2nd tional status may impact test interpretation and the edn.), D-KEFS Delis–Kaplan Executive Function creation of relevant and useful recommendations. System, BRIEF Behavior Rating Inventory of Executive Function, EOWPVT Expressive One-Word Picture Vocabulary Test, ROWPVT Receptive One-Word Picture The Children’s Oncology Group (COG) Vocabulary Test neuropsychology test battery Area of function Measurea The development of the COG test battery repre- Overall intellectual WASI sents a major accomplishment and should stream- functioning line the collection of data for multisite COG clin- Academic achievement WRAT-3 ical trials. It was put together by members of the Memory CVLT-C neuropsychology subcommittee of COG in an effort Children’s Memory Scale – to standardize neuropsychological data collection selected subtests being conducted within the various COG treatment Attention CPT-II studies. It was difficult to draw robust conclusions Executive function NEPSY – verbal fluency from the data generated because different measures Contingency Naming Test were used across sites and within the same patient D-KEFS – Tower of London Language BRIEF – parent report over time. A further challenge was the time needed EOWPVT to complete these often lengthy batteries. In times of ROWPVT cost containment and managed care, neuropsychological assessments cannot always be completed, aComplete references for all tests are found in the test list particularly for research purposes. at the end of the chapter. The goals of the core battery are to assess the key domains of neurocognitive functioning that are sensitive to the presence of a brain tumor and/or range and the 16 years to adult age range, but these the treatment effects of cancer, and to do so in a are not shown. time-efficient manner (Moore, 2005; P. Brouwers, Although the COG battery represents a major personal communication, 2005). The test battery effort toward standardized assessment for research was designed to represent the minimum standard purposes, some challenges remain. Processing of care and other tests can be added for a particu- speed and visuomotor functioning are not well lar study protocol if necessary. The choice of instru- represented in this battery and problems with ments was driven by a need to select tests that cover both are frequently observed in survivors of ALL the full age range (from infancy to adulthood), are and brain tumors. A further challenge has been psychometrically sound, and are commonly avail- the need to satisfy research and clinical demands able and widely used by psychologists in practice with the same test battery. The COG battery was (seeTables6.2and6.3foralistofthecoreneuropsy- clearly created as a research instrument. However, chological test battery). The battery illustrated cov- because of the comprehensiveness of the assess- ers the age range of 6–16 years only. Modified bat- ment, many of the measures may also need to be teries were also developed to cover the 1–5 years age administered for clinical purposes. In some settings Chapter 6. Neuropsychological assessment of children with cancer 73

Table 6.3. Measures of psychosocial, adaptive responding with areas of cortical and subcortical functioning and quality of life. PedsQL Varni Pediatric brain function. Because the tests comprising the Quality of Life Inventory, BASC-II Behavior Assessment battery were selected on the basis of their good psy- System for Children (2nd edn.), ABAS-II Adaptive chometric properties and sensitivity, they represent Behavior Assessment Scale (2nd edn.) a wide variety of normative data. This is a drawback of this battery and psychometric knowledge and Areas assessed Measurea clinical expertise are essential to inferring appropriate interpretations. Quality of life PedsQL 4.0 Generic Version – parent report (Varni et al., 1999) Psychosocial BASC-II Future directions in pediatric oncology functioning and neuropsychology including parent report of attention The most important future direction in pedi- function atric oncology in addition to increasing cure rates Adaptive functioning ABAS-II is to provide a cure at a smaller price, without such detriment to neurodevelopment. Medi- aComplete references for all tests found in the test list at cal researchers are investigating ways to provide a the end of the chapter. high chance of cure with lower radiation dosage (Packer et al., 1999), and investigations of neuropro- the child’s health insurance pays for the assessment tective agents are also underway (Drachtman et al., or the child is already scheduled to be seen for a 2002). clinical assessment. In these situations it is often In concert with medical attempts to reduce advantageous for the assessment data to be used the neurotoxicity of treatment, neuropsycholo- for both research and clinical purposes, which can gists are turning their focus from assessment create some conflicts. This is an ongoing challenge to treatment. Treatment may include educational that the COG neuropsychology subcommittee will strategies and compensations that can be imple- address. mented once children return to school, or specific targeted intervention strategies that may be delivered prophylactically while a child is on treatment. Oregon Health Sciences University The Cognitive Remediation Program (CRP) test battery (Butler, 1998; Butler & Copeland, 2002) is an approach to the treatment of acquired brain This test battery was assembled by the second injury in child cancer survivors, and has received author to provide a brief, yet comprehensive, core the most systematic study. This is a tripartite screening battery of tests providing information that model and draws from the fields of special educa- is very relevant to the school setting. Like the COG tion/educational psychology, clinical psychology, test battery, it has been designed to cover the full and brain injury rehabilitation. Children meet age range from 1 year of age through to adulthood. individually with a therapist for 2-h sessions for The battery used with the age range 6–16 years is 20 weeks. Massed practice activities are adminis- displayed in Tables 6.4 and 6.5. The test battery tered using the Attention Process Training (APT) is streamlined in that it can be completed in only techniques developed by Sohlberg and Mateer 2–3 h, and it is comprehensive, covering the major (1987). These activities focus on exercising atten- areas of cognitive function discussed above, cor- tional processes and are alternated with more 74 Section 1. Cognition and the brain

Table 6.4. Oregon Health Sciences University Test Battery for the 6- to 16-year-old age group: neurocognitive measures. WASI Wechlser Abbreviated Scale of Intelligence, WISC-III Wechsler Intelligence Scale for Children (3rd edn.), WRAT-3 Wide Range Assessment Test (3rd edn.), WRAML Wide Range Assessment of Memory and Learning, CMS Children’s Memory Scale, CPT-II Continuous Performance Test (2nd edn.), WCST Wisconsin Card Sorting Test, BRIEF = Behavior Rating Inventory of Executive Function

Area of function Measurea

Overall intellectual functioning WASI Academic achievement WRAT-3 Memory Figural Memory (Wechsler, 1945) Memory Cards (Keller et al., 1999) Sentence Memory from the WRAML Story Memory from the CMS Attention and working memory Digit Span from the WISC-III CPT-II (Conners and Staff of Multi-Health Systems, 2000) Processing speed Coding from the WISC-III (Wechsler, 1991) Executive function Color-Word Test (Golden, 1978) WCST TowerofLondon(Kelleret al., 1999) BRIEF – parent report Non-dominant hemisphere function (visual Line Orientation (Benton et al., 1983) spatial processing) Facial Recognition (Benton et al., 1983) Visual Orientation Test (Hooper, 1983) Language Boston Naming Test (Kaplan et al., 1983) Token Test Short Form (De Renzi & Faglioni, 1978)

aComplete references for all tests are found in the test list at the end of the chapter.

intrinsically interesting computer games that also ical trial of the CRP program involving seven institu- exercise attention and problem-solving abilities. tions across the United States of America. Although The participants are also provided instruction in therearenopublishedﬁndingsavailableyet,itis metacognitive strategies designed to help them extremely exciting to see such a large-scale study of with preparing to complete a task, improve their a behavioral intervention. For a review of the pub- on-task performance, and post-task strategies. This lished literature in this area, the interested reader is is all conducted within the context of psychother- referred to Penkman (2004). apeutic support, drawing from clinical psychology. There is nothing in the literature to date looking A cognitive-behavioral approach is used to assist at the efﬁcacy or feasibility of prophylactic interven- the participant in re-framing challenges, acknowl- tions (i.e., remediation programs or academic inter- edging strengths, monitoring internal dialogue, ventions delivered while a child is still on medical stress inoculation, and the development of positive treatment and prior to the documentation of neuro- self-statements. psychological dysfunction). However, Penkman and Results published thus far are promising using Scott-Lane (2007) report a case study where an CRP methods to improve attentional functioning academic intervention was delivered prophylacti- in childhood cancer survivors. Mulhern and Butler cally to a child treated for medulloblastoma. Some (2004) described a recently completed Phase III clin- improvement and maintenance of reading skills in Chapter 6. Neuropsychological assessment of children with cancer 75

Table 6.5. Measures of psychosocial functioning. However, it is time to begin thinking creatively BASC Behavior Assessment System for Children, about our approach to neuropsychological assess- PSI-3 Parenting Stress Index (3rd edn.), SIPA Stress ment and to move beyond the drive for efficiency. Index for Parents of Adolescents Ecological validity is a key issue for neuropsychology. What use are our tests if they are not mean- Areas assessed Measurea ingful in the real world? Parent, teacher, and self- report measures have been developed in recent Psychosocial functioning BASC – parent report PSI-3/SIPA (Abidin, 1995) years to attempt to garner some glimpse into a child’s day-to-day real-world functioning. These aComplete references for all tests are found in the test representanimportantfirststep,howeveritis list at the end of the chapter. our belief that as a profession we need to take a further step forward and create assessment tech- the context of decline in most other areas was docu- niques that are more behaviorally oriented in addi- mented. Mulhern and Butler (2004) underscore the tion to our laboratory based tests. The creation potential usefulness of early, prophylactic interven- of virtual reality technology represents an effort tions for children known to be at increased risk. to link testing more closely to a classroom envi- With survival as a reasonably expected outcome ronment. The ecological validity of our assess- now in childhood cancer and knowledge amassed ment techniques will be further supported through regarding the potential cost of treatment, it is imper- the completion of validity studies that examine ative that clinicians make more concerted efforts to the meaning of our test findings in the classroom communicate with parents about realistic expecta- environment. tions. Not telling parents that their child may experi- In summary, an exciting future awaits neuropsy- ence learning disabilities following CRT is no longer chologists working with child cancer survivors. It is defensible given the large research base that is now an area filled with hope for greater chance of cure available to guide the provision of information. Par- at a smaller price. There remains a vast territory of ents should be given information about research unexplored possibilities in both the treatment and findings and how these findings may relate to their assessment domains. child. There is still much unknown about why certain children experience considerable deficits and others exposed to the same treatment do not. REFERENCES Further delineation of both the treatment and patient characteristics related to higher risk is nec- Ames E (1997). The Development of Romanian Orphan- essary. age Children Adopted to Canada [Final Report to the Although neuropsychology has moved in the National Welfare Grants Program: Human Resources direction of providing behavioral treatments for Development Canada]. Burnaby, British Columbia: acquired brain injury, neuropsychological assess- Simon Fraser University. ment remains the primary task and there is much Anderson DM, Rennier KM, Ziegler RS et al. (2001). Medical work to be done to improve upon current assess- and neurocognitive late effects among survivors of childhood central nervous system tumors. Cancer 92: 2709– ment techniques. The approaches to assessment 2719. discussed in this chapter represent movements Anderson V (2001). Assessing executive functions in chil- toward streamlining the neuropsychological assess- dren: biological, psychological, and developmental con- ment process without losing crucial information siderations. Pediatr Rehabil 4: 119–136. about a child’s strengths and weaknesses. Given the Anderson V, Moore C (1995). Age at injury as a predictor economic conditions of most hospital and clinic of outcome following pediatric head injury. Child Neu- environments this has been an important focus. ropsychol 1: 187–202. 76 Section 1. Cognition and the brain

Baddeley A (1993). Working memory or working attention? Young Adult Cancers, version 1.2. Retrieved March 20, In: Baddeley A, Weiskrantz L (eds.) Attention: Selection, 2008, from http://www.survivorshipguidelines.org. Awareness, and Control (pp. 152–170). New York: Oxford Conners CK (2001). Conners’ Rating Scales – Revised (CRS- University Press. R). Toronto, ON: Multi-Health Systems, Inc. Baddeley AD, Hitch G (1974). Working memory. In Bower CousensP,WatersF,SaidJet al. (1988). Cognitive effects GA (ed.) The Psychology of Learning and Motivation. New of cranial irridation in leukemia: a survey and meta- York: Academic Press. analysis. J Child Psychol Psychiatry 29: 839–852. Baron IS (2004). Neuropsychological Evaluation of the Dennis M, Hetherington R, Spiegler B (1998). Memory Child. Oxford: Oxford University Press. and attention after childhood brain tumors. Med Pediatr Beglinger LJ, Gaydos B, Tangphao-Daniels O et al. (2005). Oncol Suppl 1: 25–33. Practice effects and the use of alternate forms in serial Drachtman RA, Cole PD, Golden CB et al. (2002). Dex- neuropsychological testing. Arch Clin Neuropsychol 20: tromethorphan is effective in the treatment of subacute 517–529. methotrexate toxicity. Pediatr Hematol Oncol 19: 319– Berger K (2005). The Developing Person Through the Life- 327. span (6th edn.). New York: Worth Publishers. Duffner PK (2004). Long-term effects of radiation therapy Bleyer WA, Fallavollita J, Robison L et al. (1990). Influence of on cognitive and endocrine function in children with age, sex, and concurrent intrathecal methotrexate ther- leukemia and brain tumors. Neurologist 10: 293–310. apy on intellectual function after cranial irradiation dur- Finger S (1999). Margaret Kennard on sparing and recov- ing childhood: a report from Children’s Cancer Study ery of function: a tribute on the 100th anniversary of her Group. Pediatric Hematol Oncol 7: 329–338. birth. J Historical Neurosci 8: 269–285. Broca P (1865). Sur le siege de la faculte du langage articule. Finger S, Wolf C (1988). The “Kennard effect” before Ken- Bull Soc Anthropol 6: 377–396. nard. The early history of age and brain lesions. Arch Brown RT, Maden-Swain A (1993). Cognitive, neuropsy- Neurol 45: 1136–1142. chological, and academic sequelae in children with Fletcher JM, Copeland DR (1988). Neurobehavioral effects leukemia. J Learn Disabil 26: 74–90. of central nervous system prophylactic treatment of can- Brown RT, Maden-Swain A, Pais R et al. (1992). Chemother- cer in children. J Clin Exp Neuropsychol 10: 495–538. apy for acute lymphocytic leukemia: cognitive and aca- Franzen MD, Paul D, Iverson GL (1996). Reliability of alter- demic sequelae. J Pediatr 121: 885–889. nate forms of the Trail Making Test. Clin Neuropsychol Brown RT, Maden-Swain A, Walco GA et al. (1998). Cog- 10: 125–129. nitive and academic late effects among children previ- Fuster JM (1980). The Prefrontal Cortex. New York: Raven ously treated for acute lymphocytic leukemia receiving Press. chemotherapy as CNS prophylaxis. J Pediatr Psychol 23: Fuster JM (1993). Frontal lobes. Curr Opin Neurobiol 3: 333–340. 160–165. Butler RW (1998). Attentional processes and their remedia- Goldman-Rakic PS (1992). Working memory and the mind. tion in childhood cancer. Pediatr Oncol Suppl 1: 75–78. Sci Am 267: 111–117. Butler RW, Copeland DR (2002). Attentional processes and Gunnar ME (2001). Effects of early deprivation: findings their remediation in children treated for cancer: a lit- from orphanage reared-infants and children. In Nelson erature review and the development of a therapeutic CA, Luciana M (eds.) Developmental Cognitive Neuro- approach. J Int Neuropsychol Soc 8: 115–124. science (pp. 617–629). Cambridge, MA: MIT Press. Buzier AI, de Sonneville LMJ, Van Den Heuvel-Eibrink MM Halstead WC (1947). Brain and Intelligence: A Quantita- et al. (2005). Chemotherapy and attentional dysfunction tive Study of the Frontal Lobes. Chicago, IL: University of in survivors of childhood acute lymphoblastic leukemia: Chicago Press. effect of treatment intensity. Pediatr Blood Cancer 45: Hopewell JW (1998). Radiation injury to the central nervous 281–290. system. Med Pediatr Oncol Suppl 1:1–9. Chelune GJ, Baer RA (1986). Developmental norms for the Hoppe-Hirsch E, Brunet L, Laroussinie F et al. (1995). Wisconsin Card Sorting Test. J Clin Exp Neuropsychol 8: Intellectual outcome in children with malignant tumors 219–228. of the posterior fossa: influence of the field of irra- Children’s Oncology Group (2004). Long-Term Follow-up diation and quality of surgery. Childs Nerv Syst 11: Guidelines for Survivors of Childhood, Adolescent, and 340–346. Chapter 6. Neuropsychological assessment of children with cancer 77

Hughlings-Jackson J (1931). In Taylor J (ed.) Selected Writ- reduced-dose craniospinal irradiation: a pediatric oncol- ings of John Hughlings-Jackson. London: Hodder. ogy group study. J Clin Oncol 16: 1723–1728. Kaplan E (1988). A process approach to neuropsychological Packer RJ, Goldwein J, Nicholson HS et al. (1999). Treat- assessment. In Boll T, Bryant BK (eds.) Clinical Neuropsy- ment of children with medulloblastoma with reduced- chology and Brain Function: Research, Measurement, and dose cranial spinal radiation and adjuvant chemother- Practice (pp. 143–155). Washington DC: American Psy- apy: A Children’s Cancer Group Study. J Clin Oncol 17: chological Association. 2127–2136. Kolb B (1995). Brain Plasticity and Behavior. New Jersey: Passler MA, Isaac W, Hynd GW (1985). Neuropsychological Lawrence Erlbaum Associates. development of behavior attributed to frontal lobe func- Kolb B, Wishaw IQ (2003). Fundamentals of Human Neu- tioning in children. Dev Neuropsychol 1: 349–370. ropsychology (5th edn.). New York: Worth Publishers. Penkman LC (2004). Remediation of attention deﬁcits in Langer T, Martus P, Ottensmeier H et al. (2002). CNS late- children: a focus on childhood cancer, traumatic brain effects after ALL therapy in childhood. Part III. Neuropsy- injury and attention deﬁcit disorder. Pediatr Rehabil 7: chological performance in long-term survivors of child- 111–123. hood ALL: impairments in concentration, attention, and Penkman LC, Scott-Lane L (2007). Prophylactic academic memory. Med Pediatr Oncol 38: 320–328. intervention for children treated with cranial radiation Lashley KD (1950). In search of the engram. Symp Soc Exp therapy. Dev Neurorehabil 10: 19–26. Biol 4: 454–482. Picard EM, Rourke BP (1995). Neuropsychological conse- Lehtinen SS, Huuskonen UE, Harila-Saari AH et al. (2002). quences of prophylactic treatment for acute lymphocytic Motor nervous system impairment persists in long-term leukemia. In Rourke BP (ed.). Syndrome of Nonverbal survivors of childhood acute lymphoblastic leukemia. Learning Disabilities (pp. 282–330). New York: Guilford Cancer 94: 2466–2473. Press. Levin HS, Culhane KA, Hartmann J et al. (1991). Develop- Posner MI, Petersen SE (1990). The attention system of the mental changes in performance on tests of purported brain. Annu Rev Neurosci 13: 25–42. frontal lobe functioning. Dev Neuropsychol 7: 377–395. Reddick WE, Glass JO, Palmer SL et al. (2005). Atypical Lezak M, Howieson DB, Loring DW (2004). Neuropsycho- white matter volume development in children following logical Assessment (4th edn.). New York: Oxford Univer- craniospinal irradiation. Neurooncology 7: 12–19. sity Press. Reitan RM (1974). Methodological problems in clinical Luria AR (1973). The Working Brain. New York: Basic Books. neuropsychology. In Reitan RM, Davison LA (eds.) Clin- Mirsky AF, Anthony BJ, Duncan CC et al. (1991). Analy- ical Neuropsychology: Current Status and Applications sis of the elements of attention: a neuropsychological (pp. 19–46). Washington, DC: Hemisphere Publishing approach. Neuropsychol Rev 2: 109–145. Corporation. Moleski M (2000). Neuropsychological, neuroanatom- Reynolds CR, Kamphaus RW (2004). Behavior Assessment ical and neurophysiological consequences of CNS System for Children (2nd edn.). Circle Pines, MN: AGS chemotherapy for acute lymphoblastic leukemia. Arch Publishing. Clin Neuropsychol 15: 603–630. Ries LAG, Eisner MP, Kosary CL et al. (eds.) (2005). SEER Moore BD (2005). Neurocognitive outcomes in survivors of Cancer Statistics Review, 1975–2002, NCI. Bethesda, childhood cancer. J Pediatr Psychol 30: 51–63. MD, http://seer.cancer.gov/csr/1975˙2002/, based on Mulhern RK, Butler RW (2004). Neurocognitive sequelae of November 2004 SEER data submission, posted to SEER childhood cancers and their treatment. Pediatr Rehabil website 2005. 7: 1–14. Ris MD, Noll RB (1994). Long-term neurobehavioral out- Mulhern RK, Butler RW (2005). Neuropsychological late come in pediatric brain tumor patients: review and effects. In Brown RT (ed.). Pediatric Hematology/ Oncol- methodological critique. J Clin Exp Neuropsychol 16: 21– ogy: A Biopsychosocial Approach. New York: Oxford Uni- 42. versity Press. Ris MD, Packer R, Goldwein J et al. (2001). Intellectual out- Mulhern RK, Kepner JL, Thomas PR et al. (1998). Neu- come after reduced-dose radiation therapy plus adjuvant ropsychologic functioning of survivors of childhood chemotherapy for medulloblastoma: a children’s cancer medulloblastoma randomized to receive conventional or group study. J Clin Oncol 19: 3470–3476. 78 Section 1. Cognition and the brain

Rowland JH, Glidewell OJ, Sibley RF et al. (1984). Effects of Von der Weid N, Mosimann I, Hirt A et al. (2003). different forms of central nervous system prophylaxis on Intellectual outcome in children and adolescents with neuropsychologic function in childhood leukemia. JClin acute lymphoblastic leukemia treated with chemother- Oncol 12: 1327–1355. apy alone: Age and sex-related differences. Eur J Cancer Rutter M (1998). Developmental catch up, and deﬁcit, 39: 353–365. following adoption after severe global early privation. Walter AW, Mulhern RK, Gajjar A et al. (1999). Survival and J Child Psychol Psychiatry 39: 465–476. neurodevelopmental outcome of young children with Schatz J, Kramer JH, Ablin A et al. (2000). Processing speed, medulloblastoma at St. Jude Children’s Research Hospi- working memory, and IQ: a developmental model of cog- tal. J Clin Oncol 17: 3720–3728. nitive deﬁcits following cranial radiation therapy. Neu- Welsh MC, Pennington BF, Grossier DB (1991). A ropsychology 14: 189–200. normative-developmental study of executive func- Shallice T (1988). From Neuropsychology to Mental Struc- tion: a window on prefrontal function in children. Dev ture. Cambridge, UK: Cambridge University Press. Neuropsychol 7: 131–149. Smith MA, Ries LA, Gurney JG et al. (eds.) (1999). SEER Can- Wernicke C (1874). Der aphasische Symptomenkomplex. cer Statistics Review, 1973–1995. Cancer Incidence and Breslau, Poland: M. Cohn and Weigert. Survival among Children and Adolescents: United States SEER Program 1975–1995. NCI. Bethesda, MD: National Institutes of Health NIH Pub. No. 99–4649. Sohlberg M, Mateer C (1987). Attention Process Training LIST OF TESTS (APT). Washington, DC: Association for Neuropsycho- logical Research and Development. Abidin RR (1995). Parent Stress Index (3rd edn.). Odessa, FL: Speechley KN, Barrera ME, Shaw A et al. (2002). Health Psychological Assessment Resources. related quality of life (HRQL) in childhood cancer sur- Achenbach T (2001). Child Behavior Checklist (CBCL). vivors (<17): survivor-control comparisons and impact Burlington, VT: ASEBA. of age, gender and cancer type. Poster presented at the Beery KE, Beery NA (2004). The Beery-Buktenica Devel- sixth World Congress of Psycho-Oncology; 2002; Banff, opmental Test of Visual-Motor Integration. Minneapolis, Alberta. MN:NCSPearson,Inc. Spiegler BJ, Kennedy K, Maze R et al. (2006). Comparison Benton AL, Hamsher KD, Vanery NR, Spreen O (1983). Con- of long-term neurocognitive outcomes in young children tributions to Neuropsychological Assessment: A Clinical with acute lymphoblastic leukemia treated with cranial Manual. New York: Oxford University Press. radiation or high-dose or very high-dose intravenous Benton AL, Sivan AB, Hamsher KD, Vanery NR, Spreen O methotrexate. J Clin Oncol 24: 3858–3864. (1994). Contributions to Neuropsychological Assessment: Sternberg RJ, Grigorenko EL, Bundy DA (2001). The predic- A Clinical Manual (2nd edn.). New York: Oxford Univer- tive value of IQ. Merrill Palmer Q 47: 1–41. sity Press. Strother D, Pollack I, Fisher PG et al. (2002). Tumors of the Brownell R (2000). Receptive One-Word Picture Vocabulary central nervous system. In Pizzo PA, Poplack DG (eds.) Test–2000 Edition. Novato, CA: Academic Therapy Publi- Principles and Practice of Pediatric Oncology (4th edn.) cations. (pp. 751–824). Philadelphia, PA: Lippincott Williams & Cohen MJ (1997). Children’s Memory Scale (CMS). New Wilkins. York: The Psychological Corporation. Stuss DT, Alexander MP (2000). Executive functions and Conners CK (2001). Conners’ Rating Scales – Revised (CRS- the frontal lobes: a conceptual review. Psychol Res 63: R). Toronto, ON: Multi-Health Systems, Inc. 289–298. Conners CK, Staff of Multi-Health Systems (2000). Con- Stuss DT, Anderson V (2004). The frontal lobes and theory ners’ Continuous Performance Test (CPT-II). North of mind: developmental concepts from adult focal lesion Tonawanda, NY: Multi-Health Systems, Inc. research. Brain Cogn 55: 69–83. Corwin J, Bylsma FW (1993). Translations of excerpts from Thompson RF (2000). The Brain (3rd edn.). New York: Andre Rey’s Psychological Examination of Traumatic Worth Publishers. Encephalopathy and PA Osterreith’s The Complex Figure Todd JA, Anderson VA, Lawrence J (1996). Planning skills in Copy Test. Clin Neuropsychol 7: 3–21. head injured adolescents and their peers. Neuropsychol De Renzi E, Faglioni P (1978). Development of a shortened Rehabil 6: 81–99. version of the Token Test. Cortex 14: 42–49. Chapter 6. Neuropsychological assessment of children with cancer 79

Delis D, Kramer JH, Kaplan E, Ober BA (1994). California McGrew KS, Woodcock RW, Maher N (2001). The Verbal Learning Test – Children’s Version (CVLT-C). San Woodcock-Johnson III Tests of Achievement (WJ3). Antonio, TX: The Psychological Corporation. Itsaca, IL: Riverside Publishing. Delis D, Kaplan E, Kramer J (2001). Delis-Kaplan Executive Reynolds CR, Kamphaus RW (2004). Behavior Assessment Function System (D-KEFS). San Antonio, TX: The Psycho- System for Children (2nd edn.) (BASC-II). Circle Pines, logical Corporation. MN: AGS Publishing. Dunn LM, Dunn LM (1997). Examiner’s Manual for the Roid GH (2003). Stanford-Binet Intelligence Scales (5th Peabody Picture Vocabulary Test (3rd edn.) (PPVT-III). edn.) (SB5). Itsaca, IL: Riverside Publishing. Circle Pines, MN: American Guidance Service. Shallice T (1982). Speciﬁc impairments of planning. Philos Elliott CD (1990). Differential Ability Scales (DAS). San Trans R Soc Lond 298: 199–209. Antonio, TX: The Psychological Corporation. Sheslow D, Adams W (2004). Wide Range Assessment of Gardner MF (2000). Expressive One-Word Picture Vocabu- Memory and Learning (2nd edn.) (WRAML-II). Lutz, FL: lary Test–2000 Edition. Novato, CA: Academic Therapy Psychological Assessment Resources. Publications. Simon HA (1975). The functional equivalence of problem Gioia G, Isquith PK, Guy SC, Kenworthy L (2000). Behavior solving skills. Cogn Psychol 7: 268–288. Rating Inventory of Executive Function. Odessa, FL: Psy- Tifﬁn J, Asher EJ (1948). The Purdue Pegboard: norms and chological Assessment Resources. studies of reliability and validity. J Appl Psychol 32: 234– Golden JC (1978). Stroop Color and Word Test. Chicago, IL: 247. Stoelting Co. Varni JW, Seid M, Rode CA (1999). The PedsQL: measure- Halstead WC (1947). Brain and Intelligence: A Quantita- ment model for the pediatric quality of life inventory. tive Study of the Frontal Lobes. Chicago, IL: University of Med Care 37(2): 126–139. Chicago Press. Wagner RK, Toregensen JK, Rashotte CA (1999). Examiner’s Harrison PL, Oakland T (2003). Adaptive Behavior Assess- Manual: The Comprehensive Test of Phonological Pro- ment System (2nd edn.) (ABAS-II). San Antonio, TX: Har- cessing. Austin, TX: PRO-ED, Inc. court Assessment, Inc. Wechsler D (1945). A standardized memory scale for clini- Heaton RK (1981). Wisconsin Card Sorting Test (WCST). cal use. JPsychol19: 87–95. Odessa, FL: Psychological Assessment Resources. Wechsler D (1991). Wechsler Intelligence Scale for Children Hooper HE (1983). Hooper Visual Orientation Test. Los (3rd edn.) (WISC-III). San Antonio, TX: The Psychological Angeles, CA: Western Psychological Services. Corporation. Kaplan EF, Goodglass H, Weintraub S (1983). The Boston Wechsler D (1997). Wechsler Adult Intelligence Scale (3rd Naming Test. Philadelphia, PA: Lea & Febiger. edn.) (WAIS-III). San Antonio, TX: The Psychological Cor- Keller FR, Davis CM, Davis HP (1999). Memory Cards poration. Test. Colorado Springs, CO: Colorado Assessment Wechsler D (1999). Wechsler Abbreviated Scale of Intelli- Tests. gence (WASI). San Antonio, TX: The Psychological Corpo- Klove H (1963). Clinical neuropsychology. In Foster FM ration. (ed.) The Medical Clinics of North America.NewYork: Wechsler D (2001). Wechsler Individual Achievement Test Saunders. (2nd edn.) (WIAT-II). San Antonio, TX: Harcourt Assess- Knights RM, Tymchuk AJ (1968). An evaluation of the ment, Inc. Halstead-Reitan Category Test for children. Cortex 4: Wechsler D (2002). Wechsler Preschool and Primary Scale of 403–414. Intelligence (3rd edn.) (WPPSI-III). San Antonio, TX: The Korkman M, Kirk U, Kemp S (1997). NEPSY: A Developmen- Psychological Corporation. tal Neuropsychological Assessment. San Antonio, TX: The Wechsler D (2003). Wechsler Intelligence Scale for Children Psychological Corporation. (4th edn.) (WISC-IV). San Antonio, TX: The Psychological Manly T, Robertson IH, Anderson V, Nimmo-Smith I Corporation. (1999). The Test of Everyday Attention for Children: Man- Wilkinson GS (1993). WRAT-3: The Wide Range Achievement ual (TEA-Ch). Bury St. Edmunds: Thames Valley Test Test Administration Manual (3rd edn.). Wilmington, DE: Company, Ltd. Wide Range.

SECTION 2

Effects of cancer and cancer treatment on cognition

Biological bases of radiation injury to the brain

Edward G. Shaw and Mike E. Robbins

Introduction (Calvo et al., 1988), target cell populations leading to white matter necrosis. Neoplasms of the central nervous system (CNS) are a pathologically diverse group of benign and malig- Vascular hypothesis nant tumors for which a variety of management Proponents of the vascular hypothesis argue that strategies, including observation, surgery, radia- vascular damage leads to ischemia with secondary tion therapy, and/or chemotherapy, are employed. white matter necrosis. In support of this hypothe- Regardless of the type of CNS tumor treated, what sis is the large amount of data describing radiation- usually limits the dose of radiation that can be uti- induced vascular changes including blood vessel lized, and therefore what typically determines the (primarily arterial) wall thickening, vessel dilation, local control and cure rate of that tumor, are the tol- and endothelial cell nuclear enlargement (Calvo erance doses of the adjacent or underlying normal et al., 1988; Reinhold et al., 1990; Schultheiss & tissues in and around the CNS. This chapter will out- Stephens, 1992). Quantitative studies in the irradi- line the biological principles of CNS radiation tol- ated rat brain have noted time- and dose-related erance and radiation-induced CNS injury, with an reductions in the number of endothelial cell nuclei emphasis on the brain. and blood vessels prior to the development of necrosis (Reinhold et al., 1990). Further, recent boron neutron capture studies, in which radia- Pathogenesis of radiation-induced tion was delivered essentially to the vasculature CNS injury alone, still led to the development of white matter necrosis (Morris et al., 1996). In contrast, radiation- Classical model of parenchymal or vascular induced white matter necrosis has been reported target cells in the absence of vascular changes (Schultheiss Vascular abnormalities and demyelination are & Stephens, 1992). Moreover, while the vascular the predominant histological changes seen in hypothesis argues that ischemia is responsible for radiation-induced CNS injury. Classically, late white matter necrosis, the most sensitive compo- delayed injury was viewed as due solely to a nent of the CNS to oxygen deprivation, the neuron, reduction in the number of surviving clonogens of is located in the gray matter, a relatively radioresis- either parenchymal, i.e., oligodendrocyte (Van den tant region. Thus, it seems unlikely that radiation Maazen et al., 1993), or vascular, i.e., endothelial injury is due to damage to the vasculature alone.

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

83 84 Section 2. Effects of cancer and cancer treatment on cognition

Parenchymal hypothesis cell kill predominates, such treatments should not negatively impact antitumor efficacy. The parenchymal hypothesis for radiation-induced CNS injury focuses on the oligodendrocyte, required Astrocytes for the formation of myelin sheaths. The key cell for the generation of mature oligodendrocytes is Once considered to be the “brain glue,” provid- the oligodendrocyte type 2 astrocyte (O-2A) pro- ing a supportive role for neuronal distribution and genitor cell (Raff et al., 1983). Ionizing radiation interactions in the normal CNS, astrocytes are now results in the loss of the reproductive capacity of recognized as a heterogeneous class of cells with the O-2A progenitor cells in the rat CNS (van der many important and diverse functions (Volterra Maazen et al., 1991a, 1991b). It is hypothesized & Meldelosi, 2005). Astrocytes secrete a variety of that radiation-induced loss of O-2A progenitor cells cytokines, proteases, and growth factors that regu- leads to a failure to replace oligodendrocytes result- late the response of the vasculature, neurons, and ing in demyelination. However, a mechanistic link oligodendrocyte lineage (Horner & Palmer, 2003). between loss of oligodendrocytes and demyelina- Recent data suggest that hippocampal astrocytes tion has yet to be established. Further, while the are capable of regulating neurogenesis by instruct- kinetics of oligodendrocytes is consistent with the ing the stem cells to adopt a neuronal fate (Muller early transient demyelination seen in so-called early et al., 1995). In addition, astrocytes assume a crit- delayed reactions (discussed later in chapter), it is ical role in the reaction of the CNS to various inconsistent with the late onset of white matter forms of injury, including ionizing radiation, and necrosis (Hornsey et al., 1981). Thus, it is unlikely are vital for the protection of endothelial cells, that loss of O-2A progenitor cells and oligodendro- oligodendrocytes, and neurons from oxidative stress cytes alone can lead to late radiation-induced CNS (Song et al., 2002). In response to injury, astro- injury. cytes exhibit two common reactions: a relatively Recent findings suggest that the classic model of acute cellular swelling and a more chronic reac- parenchymal or vascular target cells is oversimplis- tive gliosis (Pekny & Nilsson, 2005). Of note, time- tic. Pathophysiological data from a variety of late and dose-dependent increases in astrocyte num- responding tissues, including the brain and spinal ber have been observed in the irradiated rat and cord, indicate that the expression of radiation- mouse brain (Calvo et al., 1988; Chiang et al., 1993; induced normal tissue injury involves complex and Reinhold et al., 1990). In addition to increased cell dynamic interactions between several cell types number, an increase in glial fibrillary acidic pro- within the particular organ (Jaenke et al., 1993; tein (GFAP) staining intensity indicative of reac- Moulder et al., 1998; Schultheiss & Stephens, 1992). tive astrocytes has been observed (Chiang et al., In the brain, these include not only the oligodendro- 1993). However, the precise pathogenic mecha- cytes and endothelial cells, but also the astrocytes, nism(s) impacted by the astrocyte in radiation- microglia, and neurons. These now are viewed not induced CNS injury remains unknown. as passive bystanders, merely dying as they attempt to divide, but as active participants in an orches- Microglia trated response to injury (Tofilon & Fike, 2000). This new paradigm offers the exciting possibility that Microglia form approximately 10% of the total radiation-induced CNS injury can be modulated by glial cell population in the adult CNS (Vaughan & the application of therapies directed that alter the Peters, 1974) and are considered the main mediat- steps in the chain of events leading to the clinical ors of neuroinflammation (Van Rossum & Hanisch, expression of damage. Since the cascade of events 2004). Microglia respond to virtually all patholog- does not occur in tumors, where direct clonogenic ical events in the CNS, and in most pathological Chapter 7. Biological bases of radiation injury to the brain 85

settings are assisted by infiltrating macrophages. important component of radiation injury is the rela- Upon activation, they can proliferate, phagocy- tively recent observation that ionizing radiation can tose, and enhance or exacerbate injury through inhibit hippocampal neurogenesis. the production of reactive oxygen species (ROS), lipid metabolites, and hydrolytic enzymes (Stoll & Neural stem cells and neurogenesis Jander, 1999). Irradiation of the rat spinal cord results in a progressive increase in the number of The hippocampus is central to short-term declara- microglial cells 4 and 6 months post-irradiation tive memory and spatial information processing. It (Siegal et al., 1996). Similar increases in activated consists of the dentate gyrus, CA3 and CA1 regions. microglia have been observed in the irradiated rat The dentate gyrus represents a highly dynamic (Mildenberger et al., 1990) and dog (Nakagawa et al., structure and a major site of postnatal and adult 1996) brain. In addition, cranial irradiation studies neurogenesis. Resident in the hippocampus are suggest that radiation-induced microglial activation neural stem cells, self-renewing cells capable of gen- leads to decreased neurogenesis in the rat dentate erating neurons, astrocytes, and oligodendrocytes subgranular zone (DSZ) of the hippocampus and is (Gage et al., 1998; Noel et al., 1998). Neurogen- associated with spatial memory retention memory esis depends on the presence of a specific neu- deficits (Monje et al., 2002; Rola et al., 2004). Thus, rogenic microenvironment; both endothelial cells microglia may play a role in determining the sever- and astrocytes can promote/regulate neurogene- ity of radiation-induced injury in the CNS. sis (Palmer et al., 1997; Song et al., 2002). Exper- imental studies have indicated that brain radiation results in increased apoptosis and neuron loss Neurons (Nakaya et al., 2005; Palmer et al., 2000), with neona- In view of the classic model of radiation-induced tal mice being more susceptible than postnatal normal tissue injury, where DNA damage and loss animals (Nakaya et al., 2005), decreased cell prolif- of slow-turnover stem-cell populations leads to late eration, and a decreased stem/precursor cell differ- effects, the non-proliferating neuron was thought entiation into neurons within the neurogenic region to be radioresistant and a non-participant in of the hippocampus (Bellinzona et al., 1996; Monje radiation-induced CNS injury. Recent reviews of the et al., 2002; Snyder et al., 2003). Rats irradiated with clinical aspects of radiation-induced brain injury asingledoseof10Gyproduceonly3%ofthenew (Armstrong et al., 2004; Crossen et al., 1994; Mulhern hippocampal neurons formed in control animals et al., 2004; Sarkissian, 2005) describe chronic and (Snyder et al., 2003). Of note, these changes were progressive cognitive dysfunction both in children observed after doses of radiation that failed to pro- (Anderson et al., 2000; Roman & Sperduto, 1995) and duce demyelination and/or white matter necrosis in adults (Abayomi, 1996; Moore et al., 1992) follow- the rat brain. ing whole-brain or large-field irradiation, suggest- Further evidence demonstrating the importance ing that neurons are indeed sensitive to radiation. of the microenvironment for successful neuro- Moreover, in vivo and in vitro experimental stud- genesis comes from studies showing that non- ies have shown radiation-induced changes in hip- irradiated stem cells transplanted into the irradi- pocampal cellular activity, synaptic efficiency, and ated hippocampus failed to generate neurons; this spike generation (Bassant & Court, 1978; Surma-aho may reflect a pronounced microglial inflamma- et al., 2001), and in neuronal gene expression tory response, since neuroinflammation is a strong (Pellmar & Lepinski, 1993). Thus, it seems likely inhibitor of neurogenesis (Mizumatsu et al., 2003; that radiation-induced alterations in neuron func- Nakaya et al., 2005). In contrast to the reduction play a role in the development and progression tion in neurogenesis, gliogenesis appears to be of radiation-induced CNS injury. An additional and enhanced after irradiation: microglia and immature 86 Section 2. Effects of cancer and cancer treatment on cognition

oligodendrocytes increase in total and relative num- granular cell layer of the dentate gyrus 2–12 h ber in both in vitro and in vivo conditions (Monje post-irradiation. An oxidative stress, evidenced as a et al., 2002). These results suggest that brain irra- significant increase in lipid peroxidation measured diation does not eradicate hippocampal progen- using malondialdehyde, was noted in the adult itor cells or even alter their intrinsic capability male mouse hippocampus 2 weeks after brain to produce new neurons, but radiation induces irradiation with a single dose of 10 Gy (Limoli currently undefined signals that regulate the pro- et al., 2004). In accompanying in vitro studies liferation, differentiation, and survival of these using isolated multipotent neural precursor cells cells. derived from the rat hippocampus, Limoli et al. (2004) showed that the levels of ROS were significantly elevated when the cells were cultured at Contemporary view on the pathogenesis of low cell density and were associated with elevated radiation-induced CNS injury proliferation and increased metabolic, primar- Based on the assumption that the CNS has a lim- ily mitochondrial, activity. The ROS appeared to ited repertoire of responses to injury, the response result from altered mitochondrial function that of the CNS to other forms of insult has been used ultimately compromised the growth rate of the by Tofilon and Fike (2000) to model a more con- neural precursor cells. At high cell densities, intra- temporary view of the pathogenesis of radiation- cellular ROS and oxidative damage were reduced; induced CNS damage. In this model, radiation not this was associated with a concomitant increase only causes acute cell death, but also induces an in manganese superoxide dismutase (MnSOD) intrinsic recovery/repair response in the form of expression. Irradiation-induced depletion of neural specific cytokines and may initiate secondary reac- precursor cells assessed in the subgranular zone tive processes that result in the generation of a per- also led to increased ROS and altered proliferation, sistent oxidative stress and/or chronic inflamma- confirming the in vitro studies. To further test the tion (Robbins & Zhao, 2004). role of ROS, mice were treated with the antioxidant Data published in the last several years suggest a α-lipoic acid (LA). The administration of LA in primary role for chronic oxidative stress and reac- vivo reduced cell proliferation in both unirradiated tive oxygen/nitrogen oxide species (ROS/RNOS) in and irradiated mice. Indeed, the effect of LA was radiation-induced brain injury. Initial indirect evi- less marked due to the pronounced reduction of dence showed that irradiation of the rat brain inhib- precursor cell numbers observed after irradiation. ited hippocampal neurogenesis, associated with a Of note, LA treatment in irradiated mice lowered marked increase in the number and activation sta- malondialdehyde levels in hippocampal tissue, tus of microglia in the neurogenic zone (Monje et al., supporting the active role of radiation-induced 2002). Subsequent studies showed that inhibiting oxidative stress in radiation-induced brain injury. microglial activation using indomethacin restored More recently, Rola et al. (2005) reported a chronic hippocampal neurogenesis (Monje et al., 2003). inflammatory response in the mouse DSZ 9 months Direct experimental evidence for radiation- following high-LET (linear energy transfer) brain induced oxidative/nitrosative stress has been irradiation; expression of the CCR2 receptor, impor- obtained from studies using neonatal and adult tant in neuroinflammation (Banisadr et al., 2002), rats and mice. Fukuda et al. (2004) treated one increased in the irradiated brains as compared to hemisphere of postnatal-day-8 rats or postnatal- the sham-irradiated control brains. In addition, day-10 mice with a single dose of 4–12 Gy of 4-MV in vitro irradiation of rat neural precursor cells X-rays. Time-dependent increases in nitrosative subjected to chronic exogenous oxidative stress stress, assessed in terms of nitrotyrosine formation, showed increased radiosensitivity (Limoli et al., were observed in the subventricular zone and the 2006). Chapter 7. Biological bases of radiation injury to the brain 87

Laboratory studies of therapeutic from 80% down to 20% (Hopewell et al., 1993). More interventions for radiation-induced CNS injury recently, El-Agamawi et al. (1996) reported that GLA significantly reduced the onset of paralysis follow- As noted earlier, radiation-induced CNS injury has ing spinal cord irradiation in 5-week-old rats. Pro- been well characterized in terms of histological cri- phylactic hyperbaric oxygen (HBO) has also been teria as well as radiobiological parameters. In con- used to try to prevent radiation-induced myelopa- trast, details of the molecular, cellular, and bio- thyinaratmodel.Usingadoseof65Gyin10frac- chemical processes responsible for the expression tions with or without 30 HBO treatments following and progression of radiation-induced CNS injury the irradiation, Sminia et al. (2003) did not demon- currently remain limited. Thus, the rational applica- strate any preventive value to HBO. In fact, there tion of interventional procedures directed at reduc- was a “tendency towards radiosensitization” in the ing the severity of late radiation injury has been HBO-treated rats (Sminia et al., 2003). Administra- problematical. Several pragmatic but unspecific tion of ramipril, an angiotensin converting enzyme approaches have been used. inhibitor, from 2 weeks to 6 months after stereotac- Intrathecal administration of the classic radiopro- tic irradiation with a single dose of 30 Gy was asso- tector WR-2721 (amifostine) before spinal cord irra- ciated with a reduction in the severity of optic neu- diation resulted in a dose-modifying factor of 1.3 ropathy (Kim et al., 2004). and a prolongation of median latency to myelopa- Attempts have been made to rectify the radiation- thy by 63% at the effective dose in 50% of sub- induced decrease in neurogenesis. Rezvani et al. jects (ED50) (Monje et al., 2003). Fike et al. (1994) (2002) transplanted neural stem cells 90 days after observed that the polyamine synthesis inhibitor irradiation of the rat spinal cord with a single dose α-difluoromethylornithine reduced the volume of of 22 Gy. While 100% of irradiated rats treated with radionecrosis and contrast enhancement in the irra- saline exhibited paralysis within 167 days of irradi- diated dog brain; a delayed increase in microglia ation, the paralysis-free survival rate of rats treated was also noted (Nakagawa et al., 1996). Hornsey with neural stem cells was approximately 34% et al. (1990) hypothesized that treating rats with the at 183 days. Conversely, non-irradiated stem cells iron-chelating agent desferrioxamine would reduce transplanted into the irradiated rat hippocamp- hydroxyl-mediated reperfusion-related injury in the us failed to generate neurons, although gliogenesis irradiated spinal cord. Rats were fed a low-iron was spared (Rezvani et al., 2002). Preliminary data diet from 85 days after local spinal cord irradia- suggest that insulin-like growth factor-1 (IGF-1) tion and received desferrioxamine (30 mg in 0.3 ml, may show efficacy in not only preventing radiation s.c., 3 times per week) from day 120, the time myelopathy in adults (Nieder et al., 2000) but also in at which changes in vascular permeability were ameliorating the radiation-induced cognitive dys- noted. The onset of ataxia due to white matter function observed in the rat following whole-brain necrosis was delayed and the incidence of lesions irradiation (Lynch et al., 2002). Other growth fac- was reduced after single doses of 25 and 27 Gy. tors besides IGF-1, such as platelet-derived growth The steroid dexamethasone also delayed the devel- factor, vascular endothelial growth factor, and beta opment of radiation-induced ataxia along with a fibroblast growth factor, may also play a role in reduction in regional capillary permeability. In con- modulating radiation-induced CNS tissue injury trast, indomethacin did not appear to affect any of (Andratschke et al., 2005). these endpoints. In the pig, administration of the As discussed above, recent data suggest a polyunsaturated fatty acids γ -linolenic acid (GLA; role for chronic inflammation in the develop- 18C:3n-6) and eicosapentaenoic acid (EPA; 20C:5n- ment and progression of radiation-induced late 3), starting the day after spinal cord irradiation, was effects, and provide a rationale for the application associated with a reduced incidence of paralysis, of anti-inflammatory interventions to mitigate 88 Section 2. Effects of cancer and cancer treatment on cognition

Table 7.1. Factors associated with radiation tolerance of the normal central nervous system (CNS) tissues (modiﬁed from Leibel & Sheline, 1991; Schultheiss et al., 1995; Sheline et al., 1980)

Factora Factors for increased risk of CNS injury CNS tolerance increased by:

Total dose Higher total dose Decreasing total dose Fractionation Hypofractionation Hyperfractionationb Radiation dose Radiosensitizers Radioprotectors Dose per fraction Dose per fraction >180–200 cGy Dose/fraction to ≤180–200 cGy Volume Increased volume, e.g., whole-organ radiation Decreasing volume, e.g., partial organ radiation Host factors Medical illness, e.g., hypertension, diabetes Unknown Beam quality High-LET radiation beams, e.g., neutrons Low-LET beams, e.g., photons Adjunctive therapy Concomitant use of CNS toxic drugs, e.g., Avoid concomitant use drugs, or use methotrexate sequentially

aTotal time is not a major determinant of normal CNS tissue tolerance. bDeﬁned as multiple daily fractions, usually two with doses per fraction of 180–200 cGy, usually 100–120 cGy, separated by 4–8 h, to total doses higher than those given with “standard” fractionation.

radiation-induced brain injury (Robbins & Zhao, appear to be effective in the treatment of a variety of 2004). Given the anti-inflammatory properties of brain disorders (Bordet et al., 2006). Further, PPARγ the peroxisomal proliferator-activated receptor agonists induce antineoplastic signaling pathways (PPAR)γ agonists in neurological disease (Feinstein, in a variety of cancer cell lines, animal models, 2003), Zhao et al. (2007) investigated the ability and humans, including gliomas (Grommes et al., of the PPARγ agonist pioglitazone (Pio) to mod- 2004). Translating these findings to the clinic offers ulate radiation-induced cognitive impairment the promise of not only improving the quality of using a well-characterized rat model (Brown et life for long-term brain tumor survivors, but also al., 2005). Young adult male F344 rats received increasing their therapeutic window. one of the following: (1) fractionated whole-brain irradiation (WBI); 40 or 45 Gy γ rays in 4 or 4.5 weeks, respectively, 2 fractions/week, and normal Clinical aspects of CNS radiation tolerance diet; (2) sham irradiation and normal diet; (3) WBI plus Pio (120 ppm) prior, during, and for 4 The radiation tolerance of the CNS is dependent or 54 weeks post-irradiation; (4) sham irradiation on a number of factors, including total dose, dose plus Pio; and (5) WBI plus Pio starting 24 h after per fraction, total time, volume, host factors, radi- completion of WBI. Administration of Pio prior to, ation quality (linear energy transfer), and adjunc- during, and for 4 or 54 weeks after WBI prevented tive therapies. Table 7.1 defines the role of these fac- the radiation-induced cognitive impairment. Of tors in radiation tolerance and injury to the brain, interest, administration of Pio for 54 weeks starting as well as ways in which they might be modified after completion of fractionated WBI substantially, to increase tolerance (i.e., reduce injury) (Leibel & but not significantly, reduced the radiation-induced Sheline, 1991; Schultheiss et al., 1995). cognitive impairment (Zhao et al., 2007). The ability Table 7.2 shows partial and whole-organ tolerance of Pio to modulate experimental radiation-induced doses for the brain and spinal cord, and includes cognitive impairment is very significant; Pio (Actos) doses predicted to result in a 5% and 50% prob- has been prescribed for several years as an anti- ability of injury 5 years following treatment with

diabetic agent, and thiazolidinedione drugs (TZDs) radiation (TD5/5 and TD50/5, respectively) (Emami Chapter 7. Biological bases of radiation injury to the brain 89

Table 7.2. Tolerance doses for normal central nervous system tissuesa (modiﬁed from Emami et al., 1991; Rubin & Casarett, 1968)

CNS tissue TD5/5 (Gy) TD50/5 (Gy) Endpoint

Rubin and Casarett (1968) Brain Infarction, necrosis Whole 60 70 Partial (25%) 70 80 Spinal cord Infarction, necrosis Partial (10-cm length) 45 55 Emami et al. (1991) Brain Infarction, necrosis One-third 60 75 Two-thirds 50 65 Whole 40 60 Brainstem Infarction, necrosis One-third 60 – Two-thirds 53 – Whole 50 65 Spinal cord Myelitis, necrosis 5cm 50 70 10 cm 50 70 20 cm 47 – Cauda equina 60 75 Clinically apparent nerve damage Brachial plexus Clinically apparent nerve damage One-third 62 77 Two-thirds 61 76 Whole 60 75 a Assumes 2 Gy per fraction, 5 days per week. et al., 1991; Rubin & Casarett, 1968). These values The linear quadratic model links the response to are derived from mathematical models of brain and fractionated irradiation to the fractional reproduc- spinal cord tolerance based on clinical data describ- tive survival of clonogenic target cells. Fractiona- ing instances of radiation injury and the total doses tion data can be analyzed using the formula shown and fraction sizes at which they occurred. below: None of the mathematical models account for the factors listed in Table 7.1, nor do they adequately E = n (αd + βd2) predict radiation tolerance or injury. The power-law model described by Sheline et al. (1980) represents a modiﬁcation of the Ellis Nominal Standard Dose where the effect (E ) is a linear and quadratic func- formula (Ellis, 1969): tion of the dose per fraction (d) and a function of the fraction number (n). This equation allows determi- − . − . Neuret = (D)(N 0 41)(T 0 03) nation of the α/β ratio, a measure of the “bendiness” of the underlying putative target cell survival curve. where D = total dose, N = number of fractions, and For brain and spinal cord, an average α/β ratio of T = time. 2 Gy appears appropriate (Fowler, 1992). 90 Section 2. Effects of cancer and cancer treatment on cognition

Based on these various models, the TD5/5 for the Table 7.3. Tolerance doses for miscellaneous normal whole brain and for part of the brain is 50 ± 10 Gy tissues of the cranium (modiﬁed from Cooper et al., and 60 ± 10 Gy, respectively. For a 10-cm segment 1995; Emami et al., 1991; Gordon et al., 1995; Sklar & of spinal cord the TD5/5 is 45–50 Gy (Table 7.2). Constine, 1995) Although the TD50/5 value for spinal cord is lower than that for brain, there are no robust data to sup- Normal tissue TD5/5 TD50/5 Manifestations of port this difference. Rather, the sequelae of spinal (Gy) (Gy) severe injury

cord radiation injury are perceived as greater than Ear (middle/ 30–55 40–65 Acute or chronic those of brain injury; therefore, tolerance doses have external) serous otitis been lowered arbitrarily. In clinical practice, TD5/5 Eye and TD1/5 values of 60–65 Gy and 50–55 Gy for par- Retina 45 65 Blindness

tial brain irradiation and TD5/5 and TD1/5 values of Lens 10 18 Cataract formation 55–60 Gy and 45–50 Gy for a limited segment of Optic nerve or 50 65 Blindness spinal cord are commonly used. Clinical data have chiasm borne out these somewhat empiric dose ranges. In a study of 203 adults with supratentorial low-grade glioma, patients were randomized to partial brain for brain and spinal cord tolerance assume a frac- treatment ﬁelds with either 50.4 Gy in 28 fractions tion size of 180–200 cGy per day. For primary CNS of 1.8 Gy each or 64.8 Gy in 36 fractions of 1.8 Gy tumor patients being treated with curative intent, (Shaw et al., 2002). Radiation necrosis developed fraction size should rarely exceed 200 cGy daily, and, in 1% of patients who received 50.4 Gy and 5% of in most situations, should be 180–200 cGy (includ- those who had 64.8 Gy. In a retrospective study of ing areas or volumes of “hot spots”). Fraction sizes 53 head and neck cancer patients undergoing typ- greater than 200 cGy daily (usually 250–300 cGy) are ical posterior cervical treatment in ﬁelds including commonly used for palliation of brain metastases the cervical spinal cord to doses greater than 56 Gy and spinal cord compression, but only because such in fraction sizes of ≤2 Gy, the incidence of radiation patients are not expected to live long enough to myelopathy was 1.9% (McCunniff & Liang, 1989). manifest late radiation-induced brain or spinal cord In a subsequent study of 1048 lung cancer patients injury. treated with thoracic radiation on three Medical Tables 7.2 and 7.3 show the tolerance doses for Research Council Lung Cancer Working Party clini- other normal tissues of the CNS, including the cal trials, the only patients who developed radiation brainstem, eye, ear, optic chiasm, optic nerve, and myelopathy were those treated with 3 Gy fractions pituitary gland. The clinical manifestations of severe or larger. The 2-year risk of radiation myelopathy injury to these structures are listed in the table was 2.2%–2.5% among patients receiving thoracic (Cooper et al., 1995; Gordon et al., 1995; Sklar & Con- spinal cord doses of 17 Gy in 2 fractions or 39 Gy stine, 1995). in 13 fractions. The authors concluded that a total cord dose of 48 Gy given in 2-Gy fractions was safe Quantitative scoring of CNS toxicity (Macbeth et al., 1996). These data emphasize the importance of both total dose and dose per fraction Radiation injury is usually described in terms of in determining CNS tolerance to radiation. These its time course and severity. Acute injury occurs concepts are implied in the neuret model of brain duringthecourseofbrainandspinalcordirradia- tolerance, in which fraction size, which is related to tion, and is extremely uncommon, although acute “N ” (number of fractions), is far more important side-effects of radiation do occur, such as fatigue, than “T ” (time), given that the exponent for N is hair loss, and skin erythema. More common are the

much larger than that for T.TheTD5/5 values given early delayed reactions, which occur several weeks Chapter 7. Biological bases of radiation injury to the brain 91

Table 7.4. RTOG and EORTC central nervous system toxicity tablesa (modiﬁed from Cox et al., 1995)

Acute toxicity grade, brain 1 Fully functional 2 Neurological findings 3 Neurological findings 4 Serious neurological status (i.e., able to sufficient to require home requiring hospitalization impairment that includes work) with minor care; nursing assistance for initial management paralysis, coma, or seizures neurological may be required; >3 per week despite findings; no medications including medication and/or medication needed steroids and anti-seizure hospitalization required agents may be required

Chronic toxicity grade, brain 1 Mild headache; slight 2 Moderate headache; great 3 Severe headaches; severe 4 Seizure or paralysis; coma lethargy lethargy CNS dysfunction (partial loss of power or dyskinesia)

Chronic toxicity grade, spinal cord 1 Mild Lhermitte’s 2 Severe Lhermitte’s 3 Objective neurological 4 Monoplegia, paraplegia, or syndrome syndrome findingsatorbelowcord quadriplegia level treated a Grade 0 toxicity, none; grade 1, mild; grade 2, moderate; grade 3, severe; grade 4, life threatening; grade 5, fatal. to months after radiation has been completed, and oligodendroglial or myelin-producing cells, result- the late delayed reactions, which occur beyond sev- ing in an interruption of myelin synthesis. Myelin eral months (and usually between 1 and 2 years) fol- forms a concentric sheath that surrounds the axons lowing treatment. or nerve fibers. In the brain, this is clinically mani- Clinically, radiation-induced toxicities are usually fest as somnolence, increased irritability, loss of graded as mild, moderate, severe, life-threatening, appetite, and sometimes an exacerbation of under- or fatal, and are defined in an organ-specific lying tumor-associated symptoms or signs. When manner. Table 7.4 shows the toxicity tables used this symptom complex occurs in children follow- for brain tumor clinical research protocols by ing whole brain radiation, it is called the “somno- the Radiation Therapy Oncology Group (RTOG) lence syndrome.” In the spinal cord, symptoms of and its European counterpart, the European demyelination include electric-shock-like paresthe- Organisation for the Research and Treatment of sias radiating into the arms that occur with flexion Cancer (EORTC) (Cox et al., 1995). Alternatively, the of the neck, or Lhermitte’s syndrome. Early delayed National Cancer Institute (NCI) Common Terminol- reactions are nearly always transient, lasting several ogy Criteria for Adverse Events version 3.0 can be weeks to months, and do not predict for subsequent used (http://ctep.cancer.gov/reporting/ctc.html). injury (Esik et al., 2003). Late delayed reactions, To measure quality of life in brain tumor patients in contrast, are usually irreversible. The underlying undergoing combined modality therapy including mechanisms of late delayed reactions are thought to brain radiation, a commonly used and validated include (but are not limited to) injury to the capil- assessment tool is the Functional Assessment of lary endothelium leading to narrowing or oblitera- Cancer Therapy (FACT) that includes the brain tion of the arteries supplying blood to the brain or subscale (Weitzner et al., 1995). spinal cord, and direct damage to all the cells in the Early delayed reactions are thought to occur, at CNS. For both early and late delayed reactions, the least in part, due to the effects of radiation on the result is radiation necrosis, which is tissue damage 92 Section 2. Effects of cancer and cancer treatment on cognition

Ionizing radiation

↓ Oxidative Altered glial/neuronal stress function/phenotype

↓Neurogenesis/neuronal function

Vascular injury/gliosis

Demyelination

Cognitive impairment White matter necrosis

Figure 7.1. Putative pathogenic mechanisms involved in the development and progression of radiation-induced late effects in the brain. Brain irradiation is hypothesized to lead to both acute and chronic oxidative stress/inﬂammation, with resultant alterations in glial and neuronal phenotype that lead to additional and persistent oxidative stress. Accompanying these changes in brain cell phenotype are decreased neurogenesis and neuronal function, vascular injury, gliosis and changes in myelin composition/demyelination. The functional consequence of radiation-induced brain injury is cognitive impairment that may or may not be accompanied by white matter necrosis

to the substance or white matter of the brain and/or cess resulting from acute cell death as well as an spinal cord. The clinical symptoms and signs of intrinsic recovery/repair response inducing specific radiation necrosis are either the direct result of tis- cytokines and secondary reactive processes that sue damage, or indirectly result from swelling of the result in persistent oxidative stress and/or chronic adjacent normal tissues in response to the necrotic inflammation (Figure 7.1). material. Brain necrosis may be asymptomatic if There is evidence that multiple cells are involved, it occurs in non-critical areas, e.g., anterior frontal including the endothelium, oligodendrocytes, as- and temporal lobes, but usually is associated with trocytes, microglia, neurons, and neural stem cells. symptoms that are location specific (e.g., necrosis Neuroprotective and neurotherapeutic approaches in the motor gyrus would result in a contralateral have been utilized both in the laboratory and clini- hemiparesis). Spinal cord necrosis is usually symp- cal settings (Table 7.5) with variable degrees of suc- tomatic, and may include sensory and motor loss in cess. Laboratory research focused on the pathobiol- the legs or arms and legs, depending on the spinal ogy of radiation-induced CNS injury as well as the cord level of the injury, as well as sphincter impair- development of effective preventive and therapeu- ment of the bowel and bladder. tic approaches remain areas of active investigation.

ACKNOWLEDGMENTS Conclusions Supported by NCI grants CA112593, CA122318, and by an Radiation-induced injury to the CNS, including unrestricted educational grant from Elekta Instruments both the brain and spinal cord, is a complex pro- Inc., Norcross, GA. Chapter 7. Biological bases of radiation injury to the brain 93

Table 7.5. Possible preventive and therapeutic adult rats by ionizing irradiation. Neurosci Lett 208: interventions for radiation-induced brain injury. (bFGF 163–166. Basic fibroblast growth factor, IGF-1 insulin-like growth Bordet R, GeleP,DuriezP´ et al. (2006). PPARs: a new tar- factor-1, PDGF platelet derived growth factor, PPARγ get for neuroprotection. J Neurol Neurosurg Psychiatry 77: 285–287. peroxisomal proliferator-activated receptor, gamma, Brown WR, Thore CR, Moody DM et al. (2005). Vascu- VEGF vascular endothelial growth factor) lar damage after fractionated whole-brain irradiation in rats. Radiat Res 164: 662–668. Intervention Reference Calvo W, Hopewell JW, Reinhold HS et al. (1988). Time- Amifostine (WR-2721) Spence et al., 1986 and dose-related changes in the white matter of the rat α-difluoromethylornithine Fike et al., 1994; Nakagawa brain after single doses of X rays. Br J Radiol 61: 1043– et al., 1996 1052. Desferrioxamine Hornsey et al., 1990 Chiang C-S, McBride WH, Withers HR (1993). Radiation- Polyunsaturated fatty acids Hopewell et al., 1993; induced astrocytic and microglial cellular hyperplasia. El-Agamawi et al., 1996 Radiother Oncol 29: 60–68. Hyperbaric oxygen Sminia et al., 2003 Cooper JS, Fu K, Marks J et al. (1995). Late effects of radia- Angiotensin-converting Kim et al., 2004 tionintheheadandneckregion.Int J Radiat Oncol Biol enzyme inhibitors Phys 31: 1141–1164. Indomethacin Monje et al., 2003 Cox JD, Stetz J, Pajak TF (1995). Toxicity criteria of the Neural stem cells Rezvani et al., 2002 Radiation Therapy Oncology Group and the European PDGF,IGF-1, VEGF,bFGF Andratschke et al., 2005 Organization for Research and Treatment of Cancer. Int PPARγ agonists Zhao et al., 2007 J Radiat Oncol Biol Phys 31: 1341–1346. Crossen JR, Garwood D, Glatstein E et al. (1994). Neurobe- havioral sequelae of cranial irradiation in adults: a review of radiation-induced encephalopathy. JClinOncol12: REFERENCES 627–642. El-Agamawi AY, Hopewell JW, Plowman PN et al. (1996). Abayomi OK (1996). Pathogenesis of irradiation-induced Modulation of normal tissue responses to radiation. cognitive dysfunction. Acta Oncol 35: 659–663. Br J Radiol 69: 374–375. Anderson VA, Godber T, Smibert E et al. (2000). Cognitive Ellis F (1969). Dose, time fractionation: a clinical hypothe- and academic outcome following cranial irradiation and sis. Clin Radiol 20:1–7. chemotherapy in children: a longitudinal study. Br J Can- Emami B, Lyman J, Brown A et al. (1991). Tolerance of nor- cer 82: 255–262. mal tissue to therapeutic irradiation. Int J Radiat Oncol Andratschke NH, Nieder C, Price RE et al. (2005). Poten- Biol Phys 21: 109–122. tial role of growth factors in diminishing radiation ther- Esik O, Csere T, Stefantis K et al. (2003). A review on apy neural tissue injury. Semin Oncol 32 [2 Suppl. 3]: radiogenic Lhermitte’s sign. Pathol Oncol Res 9: 115– S67–70. 120. Armstrong CL, Gyato K, Awadalla AW et al. (2004). A criti- Feinstein DL (2003). Therapeutic potential of peroxisomal cal review of the clinical effects of therapeutic irradiation proliferator-activated receptor agonists for neurological damage to the brain: the roots of the controversy. Neu- disease. Diabetes Technol Ther 5: 67–73. ropsychol Rev 14: 65–86. Fike JR, Goebbel GT, Martob J et al. (1994). Radiation Banisadr G, Queraud-Lesaux´ F, Boutterin MC et al. (2002). brain injury is reduced by the polyamine inhibitor alpha- Distribution, cellular localization and functional role of difluoromethylornithine. Radiat Res 138: 99–106. CCR2 chemokine receptors in adult rat brain. J Neu- Fowler JF (1992). Brief summary of radiobiological prin- rochem 81: 257–269. ciples in fractionated radiotherapy. Semin Radiat Oncol Bassant MH, Court L (1978). Effect of whole-body gamma 2: 16–21. irradiation on the activity of rabbit hippocampal neu- Fukuda H, Fukuda A, Zhu C et al. (2004). Irradiation- rons. Radiat Res 75: 595–606. induced progenitor cell death in the developing brain is Bellinzona M, Gobbel GT, Shinohara C et al. (1996). resistant to erythropoietin treatment and caspase inhibi- Apoptosis is induced in the subependyma of young tion. Cell Death Differ 11:1166–1178. 94 Section 2. Effects of cancer and cancer treatment on cognition

Gage FH, Kempermann G, Palmer TD et al. (1998). Multi- three randomized trials of palliative radiotherapy for potent progenitor cells in the adult dentate gyrus. J Neu- non-small cell lung cancer. The Medical Research Coun- robiol 36: 249–266. cil Lung Cancer Working Party. Clin Oncol (R Coll Radiol) Gordon KB, Char DH, Sagerman RH (1995). Late effects 8: 176–181. of radiation on the eye and ocular adnexa. Int J Radiat McCunniff AJ, Liang AJ (1989). Radiation tolerance of the OncolBiolPhys31: 1123–1140. cervical spinal cord. Int J Radiat Oncol Biol Phys 16: 675– Grommes C, Landreth GE, Heneka MT (2004). Antineoplas- 678. tic effects of peroxisome proliferator-activated receptor Mildenberger M, Beach TG, McGeer EG et al. (1990). An γ agonists. Lancet Oncol 5: 419–429. animal model of prophylactic cranial irradiation: his- Hopewell JW, Van Den Aardweg GJMJ, Morris GM et al. tologic effects at acute, early and delayed stages. Int J (1993). Unsaturated lipids as modulators of radiation Radiat Oncol Biol Phys 18: 1051–1060. damage in normal tissues. In Horrobin DF (ed.). New Mizumatsu S, Monje ML, Morhardt DR et al. (2003). Approaches to Cancer Treatment (pp. 88–106). London: Extreme sensitivity of adult neurogenesis to low doses of Churchill Communications Europe. X-irradiation. Cancer Res 63: 4021–4027. Horner PJ, Palmer TD (2003). New roles for astrocytes: the Monje ML, Mizumatsu S, Fike JR et al. (2002). Irradiation nightlife of an “astrocyte”. La vida loca! Trends Neurosci induced neural precursor-cell dysfunction. Nature Med 26: 597–603. 8: 955–961. Hornsey S, Myers R, Coultas PG et al. (1981). Turnover of Monje ML, Toda H, Palmer TD (2003). Inflammatory block- proliferative cells in the spinal cord after X irradiation ade restores adult hippocampal neurogenesis. Science and its relation to time-dependent repair of radiation 302: 1760–1764. damage. Br J Radiol 54: 1081–1085. Moore BD, Copeland DR, Ried H et al. (1992). Neurophysi- Hornsey S, Myers R, Jenkinson T (1990). The reduction of ological basis of cognitive deficits in long-term survivors radiation damage to the spinal cord by postirradiation of childhood cancer. Arch Neurol 49: 809–817. administration of vasoactive drugs. Int J Radiat Oncol Morris GM, Coderre JA, Bywaters A et al. (1996). Boron Biol Phys 18: 1437–1442. neutron capture irradiation of the rat spinal cord: Jaenke RS, Robbins MEC, Bywaters T et al. (1993). Cap- histopathological evidence of a vascular-mediated illary endothelium: target site of renal radiation injury. pathogenesis. Radiat Res 146: 313–320. Lab Invest 57: 551–565. Moulder J, Robbins MEC, Cohen EP et al. (1998). Pharma- Kim JH, Brown SL, Kolozsvary A et al. (2004). Modifica- cologic modification of radiation-induced late normal tion of radiation injury by Ramipril, inhibitor of the tissue injury. In Mittal BB, Purdy JA, Ang KK (eds.) Radia- angiotensin-converting enzyme, on optic neuropathy in tion Therapy (pp. 129–151). Norwell, MA: Kluwer. the rat. Radiat Res 161: 137–142. Mulhern RK, Merchant TE, Gajjar A et al. (2004). Late neu- Leibel SA, Sheline GE (1991). Tolerance of the brain and rocognitive sequelae in survivors of brain tumours in spinal cord to conventional irradiation. In Gutin P, childhood. Lancet Oncol 5: 399–408. Liebel SA, Sheline GE (eds.) Radiation Injury to the Ner- Muller HW, Junghans U, Kappler J (1995). Astroglial neu- vous System (1st edn.) (pp. 211–238). New York: Raven rotrophic and neurite-promoting factors. Pharmacol Press. Ther 65: 1–18. Limoli CL, Rola R, Giedzinski E et al. (2004). Cell-density- Nakagawa M, Bellinzona M, Seilhan TM et al. (1996). dependent regulation of neural precursor cells function. Microglial responses after focal radiation-induced injury Proc Natl Acad Sci USA 101: 16052–16057. are affected by alpha-difluoromethylornithine. Int J Limoli CL, Giedzinski E, Baure J et al. (2006). Altered growth Radiat Oncol Biol Phys 36: 113–123. and radiosensitivity in neural precursor cells subjected Nakaya K, Hasegawa T, Flickinger JC et al. (2005). Sen- to oxidative stress. Int J Radiat Biol 82: 640–647. sitivity to radiation-induced apoptosis and neuron loss Lynch CD, Sonntag WE, Wheeler KT (2002). Radiation- declines rapidly in the postnatal mouse neocortex. Int J induced dementia in aged rats: effects of growth hor- Radiat Biol 81: 545–554. mone and insulin-like growth factor 1 [Abstract]. Neuro- Nieder C, Price RE, Rivera B et al. (2000). Both early and Oncology 4: 354. delayed treatment with growth factors can modulate Macbeth FR, Wheldon TE, Girling DJ et al. (1996). Radi- the development of radiation myelopathy (RM) in rats ation myelopathy: estimates of risk in 1048 patients in [Abstract]. Radiother Oncol 56 [Suppl. 1]: S15. Chapter 7. Biological bases of radiation injury to the brain 95

Noel F, Gumin GJ, Raju U et al. (1998). Increased expression Shaw E, Arusell R, Scheithauer B et al. (2002). A prospective of prohormone convertase-2 in the irradiated rat brain. randomized trial of low-versus high-dose radiation ther- FASEB J 12: 1725–1730. apy in adults with supratentorial low-grade glioma: ini- Palmer TD, Takahashi J, Gage FH (1997). The adult rat hip- tial report of a NCCTG-RTOG-ECOG Study. J Clin Oncol pocampus contains primordial neural stem cells. Mol 20: 2267–2276. Cell Neurosci 8: 389–404. Sheline GE, Wara WM, Smith V (1980). Therapeutic irra- Palmer TD, Willhoite AR, Gage FH (2000). Vascular niche diation and brain injury. Int J Radiat Oncol Biol Phys 6: for adult hippocampal neurogenesis. J Comp Neurol 425: 1215–1218. 479–494. Siegal T, Pfeffer MR, Meltzer A et al. (1996). Cellular Pekny M, Nilsson M (2005). Astrocyte activation and reac- and secretory mechanisms related to delayed radiation- tive gliosis. Glia 50: 427–434. induced microvessel dysfunction in the spinal cord of Pellmar TC, Lepinski DL (1993). Gamma radiation (5– rats. Int J Radiat Oncol Biol Phys 36: 649–659. 10 Gy) impairs neuronal function in the guinea pig hip- Sklar CA, Constine LS (1995). Chronic neuroendocrinolog- pocampus. Radiat Res 136: 255–261. ical sequelae of radiation therapy. Int J Radiat Oncol Biol Raff MC, Miller RH, Noble M (1983). A glial progenitor cell Phys 31: 1113–1122. that develops in vitro into an astrocyte or an oligoden- Sminia P, Van Der Kleij AJ, Carl UM et al. (2003). Prophylac- drocyte depending on culture medium. Nature 303: 390– tic hyperbaric oxygen treatment and rat spinal cord irra- 396. diation. Cancer Lett 191: 59–65. Reinhold HS, Calvo W, Hopewell JW et al. (1990). Devel- Snyder JS, Kee N, Wojtowicz JM (2003). Effects of adult neu- opment of blood vessel-related radiation damage in the rogenesis on synaptic plasticity in the rat dentate gyrus. fimbria of the central nervous system. Int J Radiat Oncol J Neurophysiol 85: 2423–2431. Biol Phys 18: 37–42. Song H, Stevens CF, Gage FH (2002). Astroglia induce neu- RezvaniM,BirdsDA,HodgesHet al. (2002). Modification rogenesis from adult neural stem cells. Nature 417: 39– of radiation myelopathy by the transplantation of neural 44. stem cells in the rat. Radiat Res 156: 408–412. Spence AM, Krohn KA, Edmonson SW et al. (1986). Radio- Robbins MEC, Zhao W (2004). Chronic oxidative stress and protection in rat spinal cord with WR-2721 following radiation-induced late normal tissue injury: a review. cerebral lateral intraventricular injection. Int J Radiat Int J Radiat Biol 80: 251–259. OncolBiolPhys12: 1479–1482. Rola R, Raber J, Rizk A et al. (2004). Radiation-induced Stoll G, Jander S (1999). The role of microglia and impairment of hippocampal neurogenesis is associated macrophages in the pathophysiology of the CNS. Prog with cognitive deficits in young mice. Exp Neurol 188: Neurobiol 58: 233–247. 316–330. Surma-aho O, Niemala¨ M, Vilkki J et al. (2001). Adverse Rola R, Sarkissian V, Obenaus A et al. (2005). High-LET radi- long-term effects of brain radiotherapy in adult low- ation induces inflammation and persistent changes in grade glioma patients. Neurology 56: 1285–1290. markers of hippocampal neurogenesis. Radiat Res 164: Tofilon PJ, Fike JR (2000). The radioresponse of the central 556–560. nervous system: a dynamic process. Radiat Res 153: 357– Roman DD, Sperduto PW (1995). Neuropsychological 370. effects of cranial radiation: current knowledge and future Van Der Maazen RWM, Kleiboer BJ, Verhagen I et al. directions. Int J Radiat Oncol Biol Phys 31: 983–998. (1991a). Irradiation in vitro discriminates between differ- Rubin P, Casarett GW (1968). Clinical Radiation Pathology ent O-2A progenitor cell subpopulations in the perinatal (Vols 1 and 2). Philadelphia, PA: WB Saunders. central nervous system of rats. Radiat Res 128: 64–72. Sarkissian V (2005). The sequelae of cranial irradiation on Van Der Maazen RWM, Verhagen I, Kleiboer BJ et al. human cognition. Neurosci Lett 382: 118–123. (1991b). Radiosensitivity of glial progenitor cells of the Schultheiss TE, Stephens LC (1992). Permanent radiation prenatal and adult rat optic nerve studies by an in vitro myelopathy. Br J Radiol 65: 737–753. clonogenic assay. Radiother Oncol 20: 258–264. Schultheiss TE, Kun LE, Ang KK et al. (1995). Radiation Van Den Maazen RWM, Kleiboer BJ, Berhagen I et al. response of the central nervous system. Int J Radiat (1993). Repair capacity of adult rat glial progenitor cells Oncol Biol Phys 31: 1093–1112. determined by an in vitro clonogenic assay after in vitro 96 Section 2. Effects of cancer and cancer treatment on cognition

or in vivo fractionated irradiation. Int J Radiat Biol 63: Weitzner MA, Meyers CA, Gelke CK et al. (1995). The 661–666. Functional Assessment of Cancer Therapy (FACT) scale: Van Rossum D, Hanisch UK (2004). Microglia. Metab Brain development of a brain subscale and revalidation of the Dis 19: 393–411. general version (FACT-G) in patients with primary brain Vaughan DW, Peters A (1974). Neuroglial cells in the cere- tumors. Cancer 75: 1151–1161. bral cortex of rats from young adulthood to old age: an Zhao W, Payne V, Tommasi E et al. (2007). Administration electron microscopic study. J Neurocytol 3: 405–429. of the peroxisomal proliferator-activated receptor γ ago- Volterra A, Meldelosi J (2005). Astrocytes from brain glue to nist pioglitazone during fractionated brain irradiation communication elements: the revolution continues. Nat prevents radiation-induced cognitive impairment. Int J Rev Neurosci 6: 626–640. Radiat Oncol Biol Phys 67:6–9. 8

Cognitive dysfunction related to chemotherapy and biological response modiﬁers

Jeffrey S. Wefel, Robert Collins, and Anne E. Kayl

Chemotherapy-related cognitive Incidence and nature of chemotherapy-related dysfunction cognitive dysfunction

The successful management of many cancers has Adult patients presenting with complaints of been achieved largely through aggressive use of “chemobrain” or “chemofog” typically report cog- therapy, which now generally combines surgery, nitive symptoms arising soon after initiating treat- radiation, chemotherapy, and immunotherapy. ment. For many patients, these symptoms persist Many of these treatment strategies, including even after therapy is complete. It is not uncommon chemotherapy, are not highly specific and therefore for many patients and providers to treat these place normal tissues and organs at risk. While symptoms as an expected, albeit unfortunate, side- the brain is afforded some protection from sys- effect of treatment. Persistent symptoms are also a temic treatments via the blood–brain barrier, cause of considerable distress for individuals who it is increasingly recognized that many agents are unable to return to their previous scholastic, gain access to this environment via direct and/or occupational, or social activities (or are able to do indirect mechanisms, potentially contributing to so only with significant additional mental effort). central nervous system (CNS) toxicity. Further- There has been additional concern that cancer and more, treatment strategies designed to disrupt cancer therapies may increase an individual’s sus- or penetrate the blood–brain barrier are being ceptibility to late emerging cognitive dysfunction explored as treatment options for a number of (Heflin et al., 2005); however, this has yet to be cancers including primary CNS lymphoma and conclusively established (Roe et al., 2005; Wefel & brain metastases (Doolittle et al., 2006). Evidence Meyers, 2005). will be presented supporting the existence of both The most commonly described cognitive prob- chemotherapy-related cognitive dysfunction and lems include difficulties with memory, attention, unique neurobehavioral/psychiatric manifesta- information-processing speed, and organization tions associated with biological response modifiers (i.e., executive dysfunction). In patients with generally, and interferon alpha in particular. chemotherapy-related cognitive dysfunction, the

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

97 98 Section 2. Effects of cancer and cancer treatment on cognition

neuropsychological evaluation frequently uncovers on heterogeneous studies in adult cancer patients difficulties in sustaining focused attention that par- treated with chemotherapy reported statistically allel patient reports of episodes during which they significant deficits in memory, executive, and motor “space out” and lose their concentration. Deficits function compared to normative expectations. in working memory and executive function are also However, when studies without a pre-treatment common and may correspond to patient reports baseline evaluation were excluded, there was no evi- of disorganization and difficulty multi-tasking. dence of a statistically significant effect on domains Measures of information-processing speed and of cognitive function. This may reflect: (1) that fine motor function suggest further inefficiencies the disease itself contributes to the cognitive dys- in cognitive and motor functions. Memory test- function experienced by patients even before they ing is generally consistent with reduced learning receive potentially neurotoxic therapies (Wefel et al., efficiency and memory-retrieval deficits in the 2004a); (2) that a subgroup of patients generally context of relatively better memory consolidation appear to develop these neurotoxicities suggesting processes. This reflects patient reports of forgetful- pharmacogenetic vulnerabilities; and (3) that there ness for recent events and details of conversations, is variability across different chemotherapeutic misplacing items, and repeating themselves in agents in their degree of neurotoxicity. For example, conversations. Notably, it is rare to see syndromes CI-980 was studied in a Phase II trial as a potential of aphasia, agnosia or apraxia that suggest dis- therapy for individuals with ovarian and colorectal turbance in cortical brain areas. This pattern of cancer. This agent is a synthetic mitotic inhibitor cognitive performance is suggestive of preferen- that shares structural and functional similarities tial dysfunction of frontal subcortical networks with colchicine, crosses the blood–brain barrier, (Kayl et al., 2006). and binds to tubulin at the colchicine-binding site. While there is rich anecdotal and clinical evi- Cognitive testing including monitoring of memory dence for the existence of “chemobrain” (Staat & function was a component of this protocol due Segatore, 2005), the cognitive and neurobehavioral to colchicine’s ability to selectively damage cholin- sequelae associated with chemotherapeutic agents ergic neurons in and around the hippocampus have rarely been systematically studied. An emerg- and basal forebrain, structures critical to learning ing body of literature is beginning to character- and memory functions. Serial testing demonstrated ize the cognitive and neurobehavioral sequelae declines in memory function using standard- of chemotherapeutic agents with chemotherapy- ized neuropsychological measures (Meyers et al., related cognitive dysfunction occurring in 15%– 1997). 70% of patients (Bender et al., 2006; Moleski, 2000; In the pediatric literature, there is mixed evidence Shilling et al., 2005; Wefel et al., 2004b). However, for chemotherapy-related cognitive and neurobe- few methodologically rigorous studies exist to guide havioral dysfunction. The majority of these studies clinical practice. Most studies are retrospective, were completed in patients diagnosed with acute fail to incorporate assessments of pre-treatment lymphocytic leukemia (ALL) or brain tumors, and cognitive and neurobehavioral function, consist of were frequently complicated by the administration small and heterogeneous samples of patients who of cerebral radiation. However, after multi-agent often received heterogeneous chemotherapeutic chemotherapy for ALL there are reports of deficits in regimens, lack appropriate control groups, and suf- attention, memory, visuoconstruction, visuomotor fer from poor measurement selection. Further stud- and tactile-perceptual skills, as well as achievement ies are clearly warranted to address both the pos- in arithmetic and, less frequently, spelling and read- sible acute sequelae and the long-term effects ing skills (Moleski, 2000; Moor, 2005). Please refer of these treatments. A recently published meta- to Chapters 6 and 14 for a more detailed review of analysis (Anderson-Hanley et al., 2003) summariz- the cognitive and neurobehavioral sequelae of can- ing the cognitive sequelae of chemotherapy based cer and cancer therapy in children. Chapter 8. Cognitive dysfunction related to chemotherapy and biological response modifiers 99

Patients may also report symptoms of depression, systemic clearance and/or pharmacogenetic mod- anxiety, and fatigue. While there are many studies ulation of drug pharmacokinetics (Shah, 2005); demonstrating the relationship between alterations (2) additive or synergistic effects of multi-agent in mood and cognitive dysfunction in other popu- chemotherapy; (3) additive or synergistic effects lations, studies of chemotherapy-related cognitive of multi-modality therapy that includes admin- dysfunction have consistently failed to demonstrate istration of chemotherapy either concurrently such a link. Moreover, patient-reported cognitive with or subsequent to cerebral radiation (Sheline dysfunction does not correlate with objective evi- et al., 1980; Sul & DeAngelis, 2006); (4) intra-arterial dence of cognitive dysfunction (as assessed through administration with blood–brain barrier disrup- formal, standardized neuropsychological testing). tion; and (5) intrathecal administration (Delattre & Rather, mood disturbance is positively correlated Posner, 1995; Jansen et al., 2005; Keime-Guibert et with self-reported cognitive dysfunction (Castellon al., 1998; Sul & DeAngelis, 2006; Taphoorn & Klein, et al., 2004; Jenkins et al., 2006; Schagen et al., 2002). 2004; Weiss & Vogelzang, 1993). Thus, comprehensive neuropsychological assess- The mechanisms by which these agents impact ments are necessary to assist in the differential diag- CNS function have been reported to occur through nosis of cognitive dysfunction and/or mood disor- both direct and indirect pathways. It is also likely der. that the influence of specific mechanisms varies at Common neurologic toxicities have been charac- different time points in the course of an emerging terized for a host of chemotherapeutic agents (see neurotoxicity. Though not exhaustive, Table 8.2 pro- Table 8.1) and include a variety of non-specific neu- vides an overview of the mechanisms that may be rologic syndromes including: acute encephalopa- involved in the development of cognitive and neu- thy characterized by a confusional state, insomnia, robehavioral toxicities. and often agitation; chronic encephalopathy char- Metabolic abnormalities can be induced by vari- acterized by cognitive dysfunction consistent with ous chemotherapeutic agents and may be respon- a “subcortical dementia,” incontinence, and gait sible for alterations in cognitive and neurobehav- disturbance; leukoencephalopathy;acerebellar syn- ioral function. Both high and intermediate doses drome with symptoms ranging from ataxia to a of 5-fluorouracil (5-FU) have been reported to pancerebellar syndrome; and a variety of peripheral cause encephalopathy in association with hyper- neuropathies. Unfortunately, it is often difficult to ammonemia. With normalization of ammonia lev- discern the specific effect attributable to individual els, the encephalopathy resolved (Kim et al., 2006). agents as they are commonly administered in multi- Methotrexate, a folate antagonist, has been reported agent combinations. to decrease dihydrofolate reductase leading to deficiencies in S-adenosylmethionine (SAM). A deficiency of SAM has been associated with demyelin- Risk factors and mechanisms underlying ation (Shuper et al., 2000), and polymorphisms of cognitive and neurobehavioral toxicity genes involved in methionine metabolism appear Although the relationship between chemotherapy to place individuals at greater risk for this neu- and cognitive dysfunction has not been fully eluci- rotoxicity (Linnebank et al., 2005). Methotrexate dated, there is agreement that cognitive dysfunction has also been reported to lead to hyperhomocys- associated with chemotherapy is a measurable tox- teinemia, high levels of neurotransmitters includ- icity. Several risk factors have been identified that ing homocysteic acid, cysteine sulfinic acid, and appear to increase the risk of developing neuro- homocysteine sulfonic acid, and toxic levels of toxicity associated with chemotherapy, including: adenosine. These abnormalities may lead to min- (1) exposure to higher doses due to planned use of eralizing microangiopathy in the white matter, high-dose regimens or high concentrations of the NMDA-mediated excitotoxicity,andalterations in parent drug and/or its metabolite due to impaired brain monoamines (i.e., norepinephrine, dopamine, 100 Section 2. Effects of cancer and cancer treatment on cognition

Table 8.1. Primary neurotoxicities associated with chemotherapeutic agents. (BCNU Bischloroethyl nitrosurea, carmustine, CHOP cyclophosphamide, adriamycin, vincristine and prednisone, 5-FU 5-ﬂuorouracil, SERMs selective estrogen receptor inhibitors, VM-26 teniposide, VP-16 etoposide)

Syndrome/mechanism Chemotherapy

Acute encephalopathye Ara-C l-Asparaginase Cisplatin 5-FU Fludarabinea Thalidomide Pentostatina Ifosfamide Interferon alpha Interleukin-1 and -2 Methotrexatea,b,c Nitrosureasa,d Procarbazine Vincristine Tamoxifen VP-16a Leukoencephalopathye Ara-C l-Asparaginase BCNUa,d Cisplatin Cyclophosphamide 5-FU Fludarabinea Ifosfamide Methotrexatea Nitrosureas Paclitaxel Vincristine Reversible posterior Ara-C leukoencephalopathy Cisplatin Multi-agent chemo such as CHOP and other combinations (including the following agents: adriamycin, cyclophosphamide, vincristine, ifosfamide, etoposide, and Ara-C) Fludarabinea Ciclosporin Tacrolimus Seizures Ara-C l-Asparaginase BCNUa,d Busulfana Chapter 8. Cognitive dysfunction related to chemotherapy and biological response modiﬁers 101

Table 8.1. (cont.)

Syndrome/mechanism Chemotherapy

Cisplatin Cyclophosphamidea 5-FU Fludarabinea Ifosfamide Interferon alphaa Interleukin-2 Methotrexatea Vincristine VP-16a Chronic encephalopathye Ara-Cb,c BCNUa,c Carmofur Fludarabinea 5-FU (± levamisole) Ifosfamide Methotrexatea,b,c Cerebellar syndrome Ara-C Cytosine arabinosidea 5-FU (± levamisole) Methotrexate Procarbazine Peripheral neuropathy Ara-C Cisplatin Carboplatin Oxaliplatin Procarbazine Suramin Taxanes (paclitaxel, docetaxel) Thalidomide Vinca alkaloids (vincristine, vinorelbine) VM-26 VP-16 Cardiotoxicity Anthracyclines (adriamycin, epirubicin) Ototoxicitye Carboplatin Cisplatin Chemotherapy-related alterations in Alkylating agents hormonal function Carboplatin Cisplatin Doxorubicin 5-FU Methotrexate (cont.) 102 Section 2. Effects of cancer and cancer treatment on cognition

Table 8.1. (cont.)

Syndrome/mechanism Chemotherapy

SERMs Aromatase inhibitors Estrogen antagonists Androgen deprivation therapy Chemotherapy-related anemia Myelosuppressive chemotherapiesf

Note : Adapted from: Delattre & Posner, 1995; Jansen et al., 2005; Keime-Guibert et al., 1998; Rottenberg, 1991; Sul & DeAngelis, 2006; Weiss & Vogelzang, 1993. a Especially if given in high doses. b Especially if administered intrathecally. c Especially if administered intravenously. d Especially if administered intra-arterially. e Especially if therapy includes both chemotherapy and radiation therapy. f See review article by Groopman and Itri (1999).

serotonin) (Haykin et al., 2006; Madhyastha et nitive dysfunction. In women, the incidence of al., 2002; Quinn & Kamen, 1996). Encephalopa- ovarian failure due to chemotherapy is variable thy has also been reported in association with (20%–100%) and depends on the patient’s age, the the syndrome of inappropriate antidiuretic hormone secretion (SIADH) secondary to intravenous vincrisitine or cyclophosphamide treatment (Lipp, Table 8.2. Potential mechanisms of chemotherapy- 1999), hyperglycemia secondary to l-asparaginase, associated cognitive and neurobehavioral dysfunction streptozocin or corticosteroid use, and salt-wasting Acute/subacute nephropathy due to cisplatin (Gilbert & Armstrong, 1996). ∨ Metabolic abnormalities Across all classes of chemotherapeutic agents, ∨ Alterations in excitatory anemia has been estimated to occur in up to 80% neurotransmitters ∨ of patients (Cunningham, 2003) and is a well known Anemia ∨ side-effect of myelosuppressive chemotherapies. Fatigue ∨ Hormonal dysfunction Anemia may cause cerebral hypoxia due to dimin- ∨ Secondary inﬂammatory response ished circulating erythrocyte levels and hemoglobin ∨ Indirect chemical toxicity and oxidative concentration (Birgegard˚ et al., 2005) and is asso- stress ciated with patient reports of reduced quality of ∨ Myeloencephalopathy life, as well as fatigue and cognitive dysfunction ∨ Demyelination (Groopman & Itri, 1999; O’Shaughnessy et al., 2005; ∨ Leukoencephalopathy Wagner et al., 2005). ∨ Microvascular injury Chemotherapy-induced menopause (Knobf, ∨ Direct neurotoxic injury to cerebral 2006; Molina et al., 2005) and treatments targeting parenchyma ∨ sex-hormone systems in both women (Eberling Cerebral atrophy ∨ et al., 2004; Jenkins et al., 2004) and men (Green Non-CNS organ toxicity (e.g., cardiotoxicity) et al., 2002; Jenkins et al., 2005; Mottet et al., 2006) Late/delayed Secondary malignancies have been implicated in the development of cog- Chapter 8. Cognitive dysfunction related to chemotherapy and biological response modiﬁers 103

dose and regimen of chemotherapy, and prior or astating adverse effects. This subacute process concurrent use of radiation therapy (Molina et al., typically occurs within days to weeks after treat- 2005). Estrogen receptors have been found in the ment and can involve progressive limb weakness, hypothalamus, anterior pituitary, amygdala, and cranial nerve palsies, seizures, visual disturbance, CA1 of the hippocampus (McEwen & Alves, 1999). coma, and even death (Moleski, 2000). Methotrex- The mechanisms by which estrogen benefits cog- ate, especially when accompanied by radiation nitive function include: (1) increasing cholinergic therapy, has also been reported to cause microvas- activity through its actions on choline acetyl- cular injury including mineralizing microangi- transferase; (2) maintenance of dendritic spine opathy, vasodilatation, endothelial damage, and density on CA1 pyramidal cells of the hippocampus; stroke-like leukoencephalopathy (Moleski, 2000). (3) facilitating induction of long-term potentiation Whether this represents ischemic events, demyelin- in the hippocampus; (4) increasing serotonergic and ation or edema within the cerebral white matter is cholinergic activity; (5) altering lipoprotein; and (6) a matter of ongoing research (Brown et al., 1998; decreasing the risk of cerebral ischemia (Yaffe et al., Haykin et al., 2006). 1998). It remains unclear whether the cognitive and Gray and white matter volume loss as well as hip- neurobehavioral effects of hormonal manipulations pocampal atrophy have been described in asso- arise through the estrogen receptors (including ciation with some chemotherapeutic treatments aromatization of androgens to estradiol in men) or (Madhyastha et al., 2002; Saykin et al., 2003; via androgen receptors. Please refer to Chapter 9 for Schneiderman, 2004). However, these findings have a more comprehensive review of this literature. not always been replicated (Yoshikawa et al., 2005) Pro-inflammatory cytokine activation may con- suggesting the need for further investigation. Appre- tribute to the cognitive and neurobehavioral seque- ciation of the ability of the adult brain to gener- lae associated with chemotherapy. Several studies ate new neurons has stimulated investigation into have reported increases in interleukins IL-1β,-6, questions regarding the impact of chemotherapies -8, -10, IFN-γ , tumor necrosis factor alpha (TNF-α), on neurogenesis. Crandall et al. (2004) reported that and granulocyte monocyte colony stimulating long-term exposure to 13-cis-retinoic acid was asso- factor associated with different schedules of pacli- ciated with decreased hippocampal neurogenesis taxel, docetaxel, etoposide and carboplatin (Penson and cell proliferation in the hippocampus and sub- et al., 2000; Pusztai et al., 2004; Tsavaris et al., 2002). ventricular zone as well as impaired spatial learning The ability of distinct chemotherapeutic agents to and memory in young adult mice. Other late emerg- induce similar pro-inflammatory cytokine activity ing side-effects that may be associated with cogni- has been hypothesized to occur through the com- tive and neurobehavioral dysfunction include the mon activation of p38 mitogen-activated protein development of cardiotoxicity (e.g., myopericardi- kinase (Wood et al., 2006). Alternatively, the appear- tis, arrhythmia, pericardial effusion, cardiomyopa- ance of cancer-related symptom clusters after thy, congestive heart failure) (Jensen, 2006; Johnson, exposure to different cytokines has been attributed 2006; Lipshultz, 2006; Steinherz & Yahalom, 1993) to activation of nuclear factor-kappa B (Lee et al., and secondary malignancies (e.g., leukemia and 2004). Similarly, adriamycin has been demonstrated myelodysplastic syndrome) (Shapiro & Recht, 2001). to increase oxidative stress in the brain, which may lead to cell dysfunction or cell death and thus Neuroimaging and neurophysiologic contribute to the symptoms of chemobrain (Joshi correlates of chemotherapy-related cognitive et al., 2005). dysfunction Myeloencephalopathy may develop after intrathecal administration of vincristine (Alcaraz et al., 2002) Chemotherapy-related structural imaging changes, or methotrexate (Garcia-Tena et al., 1995) with dev- primarily involving the white matter, have been 104 Section 2. Effects of cancer and cancer treatment on cognition

frequently identified as possible manifestations of metabolism rate have been detected, with evidence a neurotoxic syndrome. However, recent high-dose suggesting a relationship between lower IQ and chemotherapeutic protocols used in the manage- lower thalamus-to-cortex ratio. Cortical hypoper- ment of primary CNS lymphoma have demon- fusion has been demonstrated using single photon strated a dissociation between imaging findings emission computed tomography (SPECT). In addi- and cognitive outcomes with patients developing tion, EEG studies have documented increased treatment-related white matter pathology in the P300 latency and slowed reaction times after context of stable cognitive function (Fliessbach treatment, with significant correlations between et al., 2003). P300 latencies and IQ and achievement test Functional neuroimaging techniques are begin- scores. ning to be utilized to help understand the appearance of cognitive dysfunction, the underlying mech- Pharmacogenetic modulation of anisms of that dysfunction, and the potential chemotherapy-related cognitive dysfunction for recovery of cognitive processes (see Chap- ter 3). Silverman et al. (2007) compared female Clinically, and in most studies to date, cognitive breast cancer survivors who had been treated with and neurobehavioral dysfunction has been found to chemotherapy 5–10 years previously and breast occur only in a subgroup of patients. This finding cancer survivors and non-breast cancer survivors has provoked interest in clarifying the pharma- with no history of chemotherapy exposure. During cogenetic differences that may underlie an indi- a memory-related cognitive activation paradigm, vidual’s vulnerability to these side-effects. Poly- [15O] water positron emission tomography (PET) morphisms that alter the pharmacodynamics of revealed increased activation in the left infe- chemotherapeutic agents may place individuals at rior frontal gyrus and posterior cerebellum near greater risk through increased exposure to poten- the midline for chemotherapy-treated survivors tially toxic agents secondary to reduced detoxi- and in the left parietal region for untreated sur- fication and/or increased permeability of agents vivors. Resting metabolism, as evaluated by [18F]- across the blood–brain barrier (Largillier et al., fluorodeoxyglucose PET, did not differ significantly 2006; McAllister et al., 2004; Okcu et al., 2004). In between the groups. However, survivors treated children with leukemia treated with methotrexate- with both chemotherapy and tamoxifen showed containing regimens with or without cranial radia- decreased metabolism in the lentiform nucleus tion, polymorphisms of genes modulating the folate compared to survivors treated with chemotherapy pathway have been associated with diminished IQ alone. (Krajinovic et al., 2005). The relationships between Electroencephalography (EEG) with event-related polymorphisms in genes responsible for various potentials has also been employed to study aspects repair processes (e.g., apolipoprotein E) and the of information processing in female breast cancer development of cognitive dysfunction are receiving survivors who had been treated with chemother- increased attention (Ahles et al., 2003; Chen et al., apy 2–5 years earlier and with tamoxifen (Kreukels 1997). Recent studies have also examined the rela- et al., 2006). Survivors who received high-dose ther- tionship between brain-derived neurotrophic fac- apy demonstrated decreased P3 amplitude relative tor and memory (Egan et al., 2003), and catechol- to untreated controls, but were not significantly dif- O-methyl transferase and executive function (Egan ferent from survivors treated with standard-dose et al., 2001). There is evidence that polymorphisms chemotherapy. in these genes are related to differences in cognitive A variety of metabolic and evoked potential tech- function amongst healthy individuals. It is unknown niques have been used to evaluate brain function in whether these same polymorphisms confer an addi- pediatric ALL survivors treated with methotrexate tional risk to an individual exposed to a potentially (Moleski, 2000). Abnormalities of cerebral glucose neurotoxic treatment. Chapter 8. Cognitive dysfunction related to chemotherapy and biological response modifiers 105

Figure 8.1. A conceptual model of the relationship between cancer, chemotherapy, and genetics in the development of abnormalities in brain and/or cognitive function

Figure 8.1 provides a model by which disease tion. Layer 5 represents the methods that have been (“seed”), genetics (“soil”), and therapies (“pesti- employed to measure and characterize the altera- cides”) may impact on brain structure and function. tions in brain and/or cognitive function including For ease of discussion this model has been orga- structural and functional neuroimaging techniques, nized into 5 layers. Layer 1 lists three primary poten- neurophysiologic studies, and neuropsychological tial contributing factors to altered brain and/or evaluations. Importantly, the mechanisms repre- cognitive function (layer 4). There are likely direct sented in layer 2 are not believed to be mutually effects of the disease, the therapy, and host genet- exclusive and it is likely that a number of these ics on brain and/or cognitive function, which are mechanisms directly and indirectly influence the represented by the large arrows. Layer 2 shows the expression of one other. For example, inflammatory mechanisms by which chemotherapy is believed to processes such as cytokine activation may directly affect brain and/or cognitive function and is repre- affect brain and/or cognitive function and also con- sented by the light gray ovals. The dark gray bar in tribute to vascular injury, which itself can lead to layer 3 demonstrates the modulatory effect that host altered brain and/or cognitive function. Addition- genetic characteristics are believed to exert on the ally, cancer may invoke a variety of the mechanisms expression of altered brain and/or cognitive func- associated with chemotherapy (i.e., inflammation, 106 Section 2. Effects of cancer and cancer treatment on cognition

metabolic abnormalities) and thereby contribute to been associated with the development of cognitive alterations in brain and/or cognitive function. dysfunction. Some BRMs may also cause psychiatric symptoms such as depression or hallucinations that may require additional treatment. The remainder Biological response modifiers of this chapter will focus on the cytokine interferon alpha (IFN-α), which demonstrates the potential Modification of the cancer patient’s immune neurotoxic effects of these agents in terms of both response, such that a therapeutic advantage is producing cognitive dysfunction and inducing subsequently conferred, is the underlying prin- affective distress. As with the literature examining ciple of treating with biological response modi- the effects of chemotherapy on cognition, there are fiers (BRMs; also referred to as immunotherapy) few well-designed studies in the cancer population (National Institutes of Health, 2003). Given the that fully elucidate the effects of IFN-α, and even various means by which a person’s immune system fewer that assess the association of other BRMs can respond to foreign substances, BRMs repre- (e.g., IFN-β,interleukins,TNF-α, etc.) with cogni- sent a wide range of treatments that often differ tive dysfunction. While it is beyond the scope of this fundamentally in their mechanisms. Cytokines, chapter to review the putative mechanisms for all vaccines, monoclonal antibodies, thymic factors, of the BRMs, the reader is encouraged to use the and colony stimulating factors are all examples of existing literature on IFN-α as a model. BRMs (Clark, 1996). Biological response modifiers can: (1) directly or indirectly augment the patient’s Interferon alpha immunological defenses; (2) modify tumor cells such that the patient’s immunological response Interferon alpha prevents carcinogenesis by induc- is increased; or (3) bolster the patient’s ability ing an innate immunological response from the to manage toxicities secondary to other cancer patient. While the exact mechanism through which treatments (Mihich, 2000). It should also be noted IFN works has not been fully elucidated, the down- that some chemotherapy drugs (e.g., adriamycin, stream effects are natural killer and T cell respon- vincristine, cyclophosphamide), which are typically siveness driving an immunological response, which, thought to be immunosuppressive, can produce in turn, exerts an effect on cellular proliferation and immunoaugmentative effects (depending on dosing differentiation (Krause et al., 2003; Meyers & White- and regimen) and under certain circumstances side, 1996; Tompkins, 1999). That IFN induces a nat- convey a therapeutic benefit through augmentation ural immunological response explains its use in the of a patient’s antitumor host defense (Mihich, treatment of a variety of medical conditions, includ- 2000). ing malignancies, infectious diseases, and neurode- The impact of treatment with BRMs on cogni- generative diseases (Cirelly & Tyring, 1995; Meyers tion is complex, in part due to the wide ranging & Valentine, 1995). For example, IFN has demon- mechanisms by which these agents act. Mono- strated effectiveness, and is approved by the Food clonal antibodies, which recognize specific anti- and Drug Administration (FDA), for the treatment gens, are generally not thought to produce neuro- of chronic myelogenous leukemia, hepatitis C, and toxic effects or secondary cognitive impairment. melanoma. Interferon has also been used for the Colony stimulating factors, such as hematopoietic treatment of multiple sclerosis, amytrophic lateral growth factors, may improve overall quality of life sclerosis, and HIV-1, with limited success. In com- and ameliorate cognitive impairment that occurs bination with chemotherapy (i.e., 13-cis-retinoic secondary to treatment-related anemia (Brown acid, CRA), IFN has received recent attention as et al., 1991; Massa et al., 2006). Other BRMs (e.g., a viable component in the preventative treatment interleukins), like many chemotherapy agents, have of advanced squamous cell carcinoma of the head Chapter 8. Cognitive dysfunction related to chemotherapy and biological response modifiers 107

and neck (Buntzel¨ & Kuttner,¨ 1998; Shin et al., of treatment, although persisting effects of each 2001), although the benefit of CRA and IFN may be have been noted (Malik et al., 2001; Meyers et al., most robust for locally advanced disease rather than 1991; Strite et al., 1997). Antidepressant pharma- metastatic disease (Shin et al., 2002). cotherapies have been found to be effective in a number of trials (Goldman, 1994; Malek-Ahmadi & Ghandour, 2004; Musselman et al., 2001; Valentine & Interferon-related side-effects Meyers, 2005). Many argue that pre-treatment The side-effect profile for IFN is an important con- screening coupled with close serial monitoring of a sideration, as neurotoxicity may prohibit patients patient’s mood is an appropriate strategy to balance from gaining maximum benefit from the ther- the risks of this neurotoxicity against the potential apy (Valentine et al., 1998). Approximately 11% adverse side effects of pharmacotherapies used proof patients treated with IFN elect to discontinue phylactically for symptom prevention (Valentine & treatment because of adverse effects (Spiegel, 1989). Meyers, 2005). Negative effects are seen immediately after initial treatment, as many patients experience flu-like Nature and course of interferon-related symptoms (e.g., fever, chills, body aches, etc.). cognitive dysfunction In most cases this is transient and responsive to non-steroidal anti-inflammatory agents (Meyers & In addition to fatigue and increased depression, Valentine, 1995). In normal, healthy controls, a sin- IFN has been associated with cognitive impair- gle dose of 1.5 million international units (MIU) has ment. However, findings across studies have not been shown to decrease reaction time 6 h and 10 h been consistent, and there has been criticism with after injection (Smith et al., 1988). (As a reference, regard to methodological limitations and/or qual- most cancer patients receive much higher doses for itative differences in the assessment of cognitive longer periods of time.) Fatigue tends to be the most functioning. For example, Pavol et al. (1995) ret- serious side-effect, often persisting for the duration rospectively assessed the cognitive functioning of of treatment, and is most frequently reported across 25 patients with chronic myelogenous leukemia trials. It is estimated that between 70% and 100% undergoing treatment with IFN. There was vari- of patients undergoing IFN treatment experience ability with regard to IFN dosage and chronicity of fatigue, with 10%–40% requiring dose reduction treatment, with an average weekly dose of 51 MIU (Malik et al., 2001). Psychiatric symptoms, usually (range: 17–77) for 26 months (range: 1 week to 84 increased depression, are also reported in 15%–50% months). Patients evidenced impairments in ver- of patients receiving IFN treatment (Dieperink et bal memory, visual motor scanning, and executive al., 2000; Meyers et al., 1991; Valentine & Meyers, functioning, suggesting frontal subcortical network 2005), although the clinical picture may vary to dysfunction. Given the nature of the design (i.e., include mania. no pre-treatment data available), however, it was Increased depression has been observed within not possible to conclude if the pattern of cogni- the first month of treatment and has been linked tive deficits was related to pre-existing depression or to neurovegetative, emotional/affective, and cog- disease-related cognitive dysfunction. In contrast, nitive symptoms (Capuron et al., 2004). Consis- Mayr et al. (1999) found improvements on four tasks tent with this, the effects of IFN for the treatment (e.g., measure of attention, recall of a 10-digit num- of chronic myelogenous leukemia (IFN alone and ber, three simple multiplication problems, and tap- in combination with chemotherapy) were shown ping a pencil for 30 s) in 14 patients with myelo- to increase depressive symptoms independent of proliferative disorders. All patients received IFN fatigue (Scheibel et al., 2004). Both fatigue and treatment for a year, with assessments at 3-month mood alterations tend to remit following cessation intervals post-baseline, with an average weekly dose 108 Section 2. Effects of cancer and cancer treatment on cognition

of 25 MIU (range: 10–35). However, interpretation of were also noted, independent of cognitive dysfunc- their findings is limited by their selection of tasks, tion. Declines over time on measures of divided which have not been validated and may be suscep- attention and graphomotor speed were demonstr- tible to large uncontrolled practice effects. Bender ated for both treatment groups, whereas the com- et al. (2000) assessed information-processing speed bination therapy additionally produced declines on and vigilance in 18 patients with melanoma ran- measures of verbal learning and verbal fluency. The domized to high-dose IFN (100 MIU per week), low- authors suggest the findings are not inconsistent dose IFN (9 MIU per week), or observation control with frontal-subcortical dysfunction (i.e., impaired (6 per arm) and reported no changes in cogni- information processing and executive functioning) tive functioning after a 3-month interval (between- but their data appear to more reliably support group analyses were not made). Unfortunately, the frontal subcortical dysfunction in the combination design itself was likely underpowered (n = 16 and 13 group alone, rather than both IFN and combination at 3 and 6 months, respectively) and demographic treatments. There is, however, some evidence sup- and group data were not presented to facilitate more porting IFN-related frontal subcortical dysfunction, detailed analysis (Bender et al., 2000). as patients exhibiting neurotoxic effects from IFN To date, few methodologically rigorous stud- demonstrated reversible diffuse EEG abnormali- ies have assessed the cognitive effects associated ties with intermittent delta activity (e.g., slowing) in with IFN treatment. Caraceni et al. (1998) assessed the frontal lobes (Farkkila et al., 1984; Honigsberger focused attention, short-term memory, executive et al., 1983; Smedley et al., 1983). Additionally, function, and mood in 64 patients randomized to hypometabolism in the prefrontal cortex has been IFN (dose = 9 MIU per week) or observation for the reported following low-dose treatment with IFN treatment of malignant melanoma. Patients were (Juengling et al., 2000). assessed at baseline, 1, 3, 6, and 12 months. For each patient, the largest negative change on each mea- Mechanisms underlying cognitive and sure from any one of the assessments relative to neurobehavioral toxicity baseline was determined and these were collapsed across groups to form outcome measures. Relative The mechanisms by which IFN causes cognitive to baseline, IFN patients evidenced no changes on effects, including possible alteration of frontal cir- measures of cognition or mood. Group means and cuitry, are not fully understood (Malik et al., 2001; effect sizes were not reported. A significant selec- Schaefer et al., 2002; Valentine et al., 1998). Numer- tion bias appears possible, as the 64 patients agree- ous studies suggest: (1) changes in the endocrine ing to participate were a subset of the 113 enrolled system, (2) dysregulation of neurotransmitter sys- in the actual clinical trial. More recently, Scheibel tems, and (3) activation of secondary cytokine path- et al. (2004) assessed the cognitive effects of IFN for ways as possible mechanisms (Schaefer et al., 2002; the treatment of chronic myelogenous leukemia. In Valentine et al., 1998). all, 30 patients on protocols receiving IFN (n = 13, Interferon is structurally and functionally similar average weekly dose = 40 MIU) or IFN in combi- to adrenocorticotropic hormone (Blalock & Smith, nation with chemotherapy (n = 17, average weekly 1980; Blalock & Stanton, 1980) and is known to dose = 52 MIU) were assessed across domains of stimulate cortisol release into the human blood- verbal memory, verbal fluency, graphomotor speed, stream (Menzies et al., 1996). Increased cortisol visual motor scanning and sequencing, and mood. levels have been associated with mood disorders Evidence for cognitive dysfunction was found in through perturbations of the hypothalamic– both groups, as over half of the patients in the pituitary–adrenal axis (HPA) (Nemeroff et al., 1992). study evidenced a decline of at least 1.5 SD on Endocrinologic disturbance including disrup- at least one or more of the cognitive measures. tion of the hypothalamic–pituitary–thyroid axis Increases in depressive symptoms across groups (HPT) resulting in thyroid abnormalities may also Chapter 8. Cognitive dysfunction related to chemotherapy and biological response modifiers 109

contribute to the mood and cognitive dysfunction Clinical experience and research has informed associated with IFN therapy (Valentine et al., 1998). clinicians about the risks involved with certain Interferon interacts directly with the opioid regimens (e.g., high dose or intrathecal treatment) receptor system, as it is structurally similar to and administration schedules (e.g., concomi- endogenous opioids (Blalock & Smith, 1980), and tant use of radiation therapy and chemother- some of the clinical manifestations of IFN treatment apy) such that many neurotoxicities have been are similar to opioid-like effects (e.g., catatonia and reduced while continuing to achieve adequate analgesia). Moreover, IFN may indirectly modulate cancer control (Keime-Guibert et al., 1998; other neurotransmitter systems (e.g., dopamine, Lipp, 1999). In cases where a specific mech- serotonin), through either opioidergic or secondary anism underlying the neurotoxicity has been cytokine mechanisms (Schaefer et al., 2003). characterized, targeted treatment strategies have Though not well studied, IFN influences other been explored. For example, treatment with nal- cytokines, specifically inducing the production of trexone (i.e., a µ-opioid receptor antagonist) was interleukins IL-1, IL-2, IL-6 and TNF (Taylor & effective in relieving neurotoxic side-effects in Grossberg, 1998). These particular cytokines are seven of nine patients undergoing IFN treatment known to play a central role in stimulation of the for hematological malignancies (Valentine et al., HPA and suppression of the HPT (Zaloga et al., 1995). Musselman et al. (2001) demonstrated the 2001). As previously noted, HPA and HPT alterations benefit of pre-treatment with paroxetine (selec- may explain mood and cognitive changes during tive serotonin reuptake inhibitor) in minimizing treatment. Moreover, animal models suggest that depression in melanoma patients receiving IFN cytokines (e.g., IL-2) modulate serotonin levels in treatment. the prefrontal cortex (Lacosta et al., 2000). However, empirically supported therapies for persistent cognitive and neurobehavioral dysfunction are limited. Stimulant therapies have proven effective in treating the cognitive dysfunction that is Intervention strategies to prevent or common in cancer patients (Meyers et al., 1998). manage cognitive and neurobehavioral Other pharmacologic interventions commonly used dysfunction to treat other diseases affecting cognitive function are currently being explored (Barton & Loprinzi, Chemotherapy and immunotherapy are necessary 2002; see Chapter 22). Cognitive and behavioral components of the management and eradication intervention strategies that have been studied in of many types of cancer. Although not all patients the traditional rehabilitation literature with stroke will experience treatment-related neurotoxicity, for and traumatic brain injury survivors may also be a subpopulation of patients cognitive and/or behav- employed. These interventions often focus on com- ioral symptoms are distressing and disruptive. pensatory strategy training, stress management, Clearly, there is an opportunity and a very real need energy conservation and psycho-education (see to explore therapies that may prevent negative side- Chapters 20, 21). effects or minimize the impact and extent of symptoms that are already present. Ideally, these interventions should be tailored to the symptom (e.g., REFERENCES anemia, durable fatigue, memory-retrieval deficit) and be based on the hypothesized mechanism. Ahles TA, Saykin AJ, Noll WW et al. (2003). The rela- While determining the nature of the symptom is tionship of APOE genotype to neuropsychological per- often feasible, our understanding of the etiologic formance in long-term cancer survivors treated with mechanisms underlying these symptoms is limited standard dose chemotherapy. Psychooncology 12: 612– in most cases. 619. 110 Section 2. Effects of cancer and cancer treatment on cognition

Alcaraz A, Rey C, Concha A et al. (2002). Intrathecal adjuvant chemotherapy and tamoxifen. J Clin Exp Neu- vincristine: fatal myeloencephalopathy despite cere- ropsychol 26: 955–969. brospinal fluid perfusion. J Toxicol Clin Toxicol 40: 557– Chen Y, Lomnitski L, Michaelson DM et al. (1997). Motor 561. and cognitive deficits in apolipoprotein E-deficient Anderson-Hanley C, Sherman ML, Riggs R et al. (2003). mice after closed head injury. Neuroscience 80: 1255– Neuropsychological effects of treatments for adults with 1262. cancer: a meta-analysis and review of the literature. JInt Cirelly R, Tyring SK (1995). Major therapeutic uses of inter- Neuropsychol Soc 9: 967–982. ferons. Clin Immunother 3: 27–87. Barton D, Loprinzi C (2002). Novel approaches to pre- Clark JW (1996). Biological response modifiers. Cancer venting chemotherapy-induced cognitive dysfunction in Chemother Biol Response Modif 16: 239–273. breast cancer: the art of the possible. Clin Breast Cancer Crandall J, Sakai Y, Zhang J et al. (2004). 13-cis-Retinoic acid 3 [Suppl. 3]: S121–S127. suppresses hippocampal cell division and hippocampal- Bender CM, Yasko JM, Kirkwood JM et al. (2000). Cogni- dependent learning in mice. Proc Natl Acad Sci USA 101: tive function and quality of life in interferon therapy for 5111–5116. melanoma. Clin Nurs Res 9: 352–363. Cunningham RS (2003). Anemia in the oncology patient: Bender CM, Sereika SM, Berga SL et al. (2006). Cognitive cognitive function and cancer. Cancer Nurs 26: 38S–42S. impairment associated with adjuvant therapy in breast Delattre JY, Posner JB (1995). Neurological complications cancer. Psychooncology 15: 422–430. of chemotherapy and radiation therapy. In Aminoff Birgegard˚ G, Aapro MS, Bokemeyer C et al. (2005). Cancer- MJ (ed.). Neurology and General Medicine (2nd edn.) related anemia: pathogenesis, prevalence, and treat- (pp. 421–445). New York: Churchill Livingstone. ment. Oncology 68 [Suppl. 1], 3–11. Dieperink E, Willenbring M, Ho SB (2000). Neuropsy- Blalock JE, Smith EM (1980). Human leukocyte inter- chiatric symptoms associated with hepatitis C and feron: structural and biological relatedness to adrenocor- interferon alpha: a review. Am J Psychiatry 157: 867– ticotropic hormones and endorphins. Proc Natl Acad Sci 876. USA 77: 5972–5974. Doolittle ND, Peereboom DM, Christoforidis GA et al. Blalock JE, Stanton JD (1980). Common pathways of inter- (2006). Delivery of chemotherapy and antibodies across feron and hormonal action. Nature 283: 406–408. the blood-brain barrier and the role of chemoprotection, Brown MS, Stemmer SM, Simon JH et al. (1998). White mat- in primary and metastatic brain tumors: report of the ter disease induced by high-dose chemotherapy: longi- Eleventh Annual Blood-Brain Barrier Consortium meet- tudinal study with MR imaging and proton spectroscopy. ing. J Neurooncol 81: 81–91. Am J Neuroradiol 19: 217–221. Eberling JL, Wu C, Tong-Turnbeaugh R et al. (2004). Brown WS, Marsh JT, Wolcott D et al. (1991). Cognitive Estrogen- and tamoxifen-associated effects on brain function, mood and p3 latency: effects of the ameliora- structure and function. Neuroimage 21: 364–371. tion of anemia in dialysis patients. Neuropsychologia 29: EganMF,GoldbergTE,KolachanaBSet al. (2001). Effect 35–45. of COMT Val108/158 met genotype on frontal lobe func- Buntzel¨ J, Kuttner¨ K (1998). Chemoprevention with inter- tion and risk for schizophrenia. Proc Natl Acad Sci USA feron alfa and 13-cis retinoic acid in the adjunctive treat- 98: 6917–6922. ment of head and neck cancer. Auris Nasus Larynx 25: Egan MF, Kojima M, Callicott JH et al. (2003). The BDNF 413–418. val66met polymorphism affects activity-dependent Capuron L, Ravaud A, Miller AH et al. (2004). Baseline secretion of BDNF and human memory and hippocam- mood and psychosocial characteristics of patients devel- pal function. Cell 112: 257–269. oping depressive symptoms during interleukin-2 and/or Farkkila M, Iivanainen M, Roine R et al. (1984). Neurotoxic interferon-alpha cancer therapy. Brain Behav Immun 18: and other side effects of high dose interferon in amyo- 205–213. trophic lateral sclerosis. Acta Neurol Scand 70: 42–46. Caraceni A, Gangeri L, Martini C et al. (1998). Neurotoxic- Fliessbach K, Urbach H, Helmstaedter C et al. (2003). Cog- ity of interferon-α in melanoma therapy: results from a nitive performance and magnetic resonance imaging randomized clinical trial. Cancer 83: 482–489. findings after high-dose systemic and intraventricular Castellon SA, Ganz PA, Bower JE et al. (2004). Neurocogni- chemotherapy for primary central nervous system lym- tive performance in breast cancer survivors exposed to phoma. Arch Neurol 60: 563–568. Chapter 8. Cognitive dysfunction related to chemotherapy and biological response modifiers 111

Garcia-Tena J, Lopez-Andreu JA, Ferris J et al. (1995). Joshi G, Sultana R, Tangpong J et al. (2005). Free radical Intrathecal chemotherapy related myeloencephalopathy mediated oxidative stress and toxic side effects in brain in a young child with acute lymphoblastic leukemia. induced by the anti cancer drug adriamycin: insight into Pediatr Hematol Oncol 12: 377–385. chemobrain. Free Radic Res 39: 1147–1154. Gilbert MR, Armstrong TS (1996). Neurotoxicities. In Kirk- Juengling FD, Ebert D, Gut O et al. (2000). Prefrontal cor- wood J, Lotze M, Yasko J (eds.) Current Cancer Therapeu- tical hypometabolism during low-dose interferon alpha tics (2nd edn.) (pp. 364–371). Philadelphia, PA: Current treatment. Psychopharmacology 152: 383–389. Medicine. Kayl AE, Wefel JS, Meyers CA (2006). Chemotherapy and Goldman LS (1994). Successful treatment of interferon alfa- cognition: effects, potential mechanisms and manage- induced mood disorder with nortriptyline. Psychosomat- ment. Am J Ther 13: 362–369. ics 35: 412–413. Keime-Guibert F, Napolitano M, Delattre JY (1998). Neu- Green HJ, Pakenham KI, Headley BC et al. (2002). Altered rological complications of radiotherapy and chemother- cognitive function in men treated for prostate cancer apy. J Neurol 245: 695–708. with luteinizing hormone-releasing hormone analogues Kim YA, Chung HC, Choi HJ et al. (2006). Intermediate dose and cyproterone acetate: a randomized controlled trial. 5-fluorouracil-induced encephalopathy. Jpn J Clin Oncol BJU Int 90: 427–432. 36: 55–59. Groopman JE, Itri LM (1999). Chemotherapy-induced ane- Knobf MT (2006). Reproductive and hormonal sequelae of mia in adults: incidence and treatment. J Natl Cancer Inst chemotherapy in women. Am J Nurs 106 [Suppl. 3]: 60– 91: 1616–1634. 65. Haykin ME, Gorman M, van Hoff J et al. (2006). Diffusion- Krajinovic M, Robaey P, Chiasson S et al. (2005). Polymor- weighted MRI correlates of subacute methotrexate- phisms of genes controlling homocysteine levels and IQ related neurotoxicity. J Neurooncol 76: 153–157. score following treatment for childhood ALL. Pharma- Heflin LH, Meyerowitz BE, Hall P et al. (2005). Cancer as a cogenomics 6: 293–302. risk factor for long-term cognitive deficits and dementia. Krause I, Valesini G, Scrivo R et al. (2003). Autoimmune J Natl Cancer Inst 97: 854–856. aspects of cytokine and anticytokine therapies. Am J Med Honigsberger L, Fielding JW, Priestman TJ (1983). Neuro- 115: 390–397. logical effects of recombinant human interferon. Br J Kreukels BPC, Schagen SB, Ridderinkhof KR et al. (2006). Med 286: 719. Effects of high-dose and conventional-dose adjuvant Jansen C, Miakowski C, Dodd M et al. (2005). Potential chemotherapy in patients with breast cancer: an electro- mechanisms for chemotherapy-induced impairments in physiologic study. Clin Breast Cancer 7: 67–78. cognitive function. Oncol Nurs Forum 32: 1151–1163. Lacosta S, Merali Z, Anisman H (2000). Central monoamine Jenkins VA, Shilling V, Fallowfield L et al. (2004). Does hor- activity following acute and repeated systemic mone therapy for the treatment of breast cancer have interleukin-2 administration. Neuroimmunomodu- a detrimental effect on memory and cognition? A pilot lation 8: 83–90. study. Psychooncology 13: 61–66. Largillier R, Etienne-Grimaldi MC, Formento JL et al. Jenkins VA, Bloomfield DJ, Shilling VM et al. (2005). Does (2006). Pharmacogenetics of capecitabine in advanced neoadjuvant hormone therapy for early prostate cancer breast cancer patients. Clin Cancer Res 12: 5496– affect cognition? Results from a pilot study. BJU Int 96: 5502. 48–53. Lee BN, Dantzer R, Langley KE et al. (2004). A cytokine- Jenkins V, Shilling V, Deutsch G et al. (2006). A 3-year based neuroimmunologic mechanism of cancer-related prospective study of the effects of adjuvant treatments symptoms. Neuroimmunomodulation 11: 279–292. on cognition in women with early stage breast cancer. Linnebank M, Pels H, Klecza N et al. (2005). MTX-induced Br J Cancer 94: 828–834. white matter changes are associated with polymor- Jensen BV (2006). Cardiotoxic consequences of phisms of methionine metabolism. Neurology 64: 912– anthracycline-containing therapy in patients with 913. breast cancer. Semin Oncol 33 [Suppl. 8]: S15–S21. Lipp HP (1999). Neurotoxicity (including sensory toxicity) Johnson SA (2006). Anthracycline-induced cardiotoxicity in induced by cytostatics. In Lipp HP (ed.). Anticancer Drug adult hematologic malignancies. Semin Oncol 33 [Suppl. Toxicity: Prevention, Management, and Clinical Pharma- 8]: S22–S27. cokinetics (pp. 431–453). New York: Marcel Dekker, Inc. 112 Section 2. Effects of cancer and cancer treatment on cognition

Lipshultz SE (2006). Exposure to anthracyclines during Moleski M (2000). Neuropsychological, neuroanatom- childhood causes cardiac injury. Semin Oncol 33 [Suppl. ical, and neurophysiological consequences of CNS 8]: 8–14. chemotherapy for acute lymphoblastic leukemia. Arch Madhyastha S, Somayaji SN, Rao MS et al. (2002). Hip- Clin Neuropsychol 15: 603–630. pocampal brain amines in methotrexate-induced learn- Molina JR, Barton DL, Loprinzi CL (2005). Chemotherapy- ing and memory deficit. Can J Physiol Pharmacol 80: induced ovarian failure. Drug Safety 28: 401–416. 1076–1084. Moor BD (2005). Neurocognitive outcomes in survivors of Malek-Ahmadi P, Ghandour E (2004). Bupropion for treat- childhood cancer. J Pediatr Psychol 30: 51–63. ment of interferon-induced depression. Ann Pharma- Mottet N, Prayer-Galetti T, Hammerer P et al. (2006). Opti- cother 38: 1202–1205. mizing outcomes and quality of life in the hormonal Malik UR, Makower DF, Wadler S (2001). Interferon- treatment of prostate cancer. BJU Int 98: 20–27. mediated fatigue. Cancer 92 [Suppl]: 1664–1668. Musselman DL, Lawson DH, Gumnick JF et al. (2001). Massa E, Madeddu C, Lusso MR et al. (2006). Evaluation Paroxetine for the prevention of depression induced by of the effectiveness of treatment with erythropoietin on high-dose interferon alpha. NEnglJMed344: 961–966. anemia, cognitive functioning and functions studied by National Institutes of Health (2003). Biological Therapy. comprehensive geriatric assessment in elderly cancer Treatments that use your Immune System to Fight Cancer. patients with anemia related to cancer chemotherapy. [Brochure]. NIH Publication No. 03–5406. Crit Rev Oncol Hematol 57: 175–182. Nemeroff CB, Krishnan KR, Reed D et al. (1992). Adrenal Mayr N, Zeitlhofer J, Deecke L et al. (1999). Neurologi- gland enlargement in major depression: a computed cal function during long-term therapy with recombinant tomographic study. Arch Gen Psychiatry 49: 384–387. interferon alpha. J Clin Neurosci 11: 343–348. Okcu MF, Selvan M, Wang L et al. (2004). Glutathione McAllister TW, Ahles TA, Saykin AJ et al. (2004). Cognitive S-transferase polymorphisms and survival in primary effects of cytotoxic cancer chemotherapy: predisposing malignant glioma. Clin Cancer Res 10: 2618–2625. risk factors and potential treatments. Curr Psychiatr Rep O’Shaughnessy JA, Svetislava JV, Holmes FA et al. (2005). 6: 364–371. Feasibility of quantifying the effects of epoetin alpha McEwen BS, Alves SE (1999). Estrogen actions in the central therapy on cognitive function in women with breast can- nervous system. Endocr Rev 20: 279–307. cer undergoing adjuvant or neoadjuvant chemotherapy. Menzies R, Phelps C, Wiranowska M et al. (1996). The effect Clin Breast Cancer 5: 439–446. of interferon-alpha on the pituitary-adrenal axis. J Inter- Pavol MA, Meyers CA, Rexer JL et al. (1995). Pattern of neu- feron Cytokine Res 16: 619–629. robehavioral deficits associated with interferon alpha Meyers CA, Valentine AD (1995). Neurological and psychi- therapy for leukemia. Neurology 45: 947–950. atric effects of immunological therapy. CNS Drugs 3: 56– Penson RT, Kronish K, Duan Z et al. (2000). Cytokines IL- 68. 1beta, IL-2, IL-6, IL-8, MCP-1, GM-CSF and TNF-alpha in Meyers CA, Scheibel RS, Forman AD (1991). Persistent neu- patients with epithelial ovarian cancer and their relation- rotoxicity of systemically administered interferon-alpha. ship to treatment with paclitaxel. Int J Gynecol Cancer 1: Neurology 41: 672–676. 33–41. Meyers CA, Kudelka AP, Conrad CA et al. (1997). Neurotox- Pusztai L, Mendoza TR, Reuben JM et al. (2004). Changes icity of CI-980, a novel mitotic inhibitor. Clin Cancer Res in plasma level of inflammatory cytokines in response to 3: 419–422. paclitaxel chemotherapy. Cytokine 25: 94–102. Meyers CA, Weitzner MA, Valentine AD et al. (1998). Quinn CT, Kamen BA (1996). A biochemical perspective Methylphenidate therapy improves cognition, mood, of methotrexate neurotoxicity with insight on nonfolate and function of brain tumor patients. J Clin Oncol 16: rescue modalities. J Invest Med 44: 522–530. 2522–2527. Roe CM, Behrens MI, Xiong C et al. (2005). Alzheimer dis- Meyers JN, Whiteside TL (1996). Immunotherapy of squa- ease and cancer. Neurology 64: 895–898. mous cell carcinoma of the head and neck. In Meyers Rottenberg DA (ed.) (1991). Neurological Complica- EN, Suen JY (eds.). Cancer of the Head and Neck (pp. 805– tions of Cancer Treatment. Boston, MA: Butterworth- 817). Philadelphia, PA: WB Saunders Company. Heinemann. Mihich E (2000). Historical overview of biologic response Saykin AJ, Ahles TA, McDonald BC (2003). Mechanisms of modifiers. Cancer Invest 18: 456–466. chemotherapy-induced cognitive disorders: neuropsy- Chapter 8. Cognitive dysfunction related to chemotherapy and biological response modifiers 113

chological, pathophysiological, and neuroimaging per- SmedleyH,KatrakM,SikoraKet al. (1983). Neurological spectives. Semin Clin Neuropsychiatry 8: 201–216. effects of recombinant interferon alpha. Br Med J 286: Schaefer M, Engelbrecht MA, Gut O et al. (2002). Interferon 262–264. alpha (IFN-α) and psychiatric syndromes: a review. Prog Smith A, Tyrrell D, Coyle K et al. (1988). Effects of interferon Neuropsychopharmacol Biol Psychiatry 26: 731–746. alpha on performance in man: a preliminary report. Psy- Schaefer M, Schwiger M, Pich M et al. (2003). Neuro- chopharmacology 96: 414–416. transmitter changes by interferon-alpha and therapeu- Spiegel RJ (1989). The alpha interferons: clinical overview. tic implications. Pharmacopsychiatry 36 [Suppl. 3]: S203– Urology 34: 75–79. S206. Staat K, Segatore M (2005). The phenomenon of chemo Schagen SB, Muller MJ, Boogerd W et al. (2002). Cogni- brain. Clin J Oncol Nurs 9: 713–721. tive dysfunction and chemotherapy: neuropsychologi- Steinherz LJ, Yahalom J (1993). Cardiac complications of cal findings in perspective. Clin Breast Cancer 3 [Suppl]: cancer therapy. In DeVita VT, Jr., Hellman S, Rosenberg S100–S108. SA (eds.) Cancer Principles and Practice of Oncology (4th Scheibel RS, Valentine AD, O’Brien S et al. (2004). Cognitive edn.) (pp. 2370–2385). Philadephia, PA: J.B. Lippincott dysfunction and depression during treatment with inter- Company. feron alpha and chemotherapy. J Neuropsychiatry Clin Strite D, Valentine AD, Meyers CA (1997). Manic episodes Neurosci 16:1–7. in two patients treated with interferon alpha. J Neuropsy- Schneiderman B (2004). Hippocampal volumes smaller in chiatry 9: 273–276. chemotherapy patients. Lancet Oncol 5: 202. Sul JK, DeAngelis LM (2006). Neurologic complications of Shah RR (2005). Mechanistic basis of adverse drug reac- cancer chemotherapy. Semin Oncol 33: 324–332. tions: the perils of inappropriate dose schedules. Expert Taphoorn MJB, Klein M (2004). Cognitive deficits in adult Opinion Drug Safety 4: 103–128. patients with brain tumors. Lancet Neurol 3: 159–168. Shapiro CL, Recht A (2001). Drug therapy – side effects of Taylor JL, Grossberg SE (1998). The effects of interferon-α adjuvant treatment of breast cancer. New Engl J Med 344: on the production and action of other cytokines. Semin 1997–2008. Oncol 25 [Suppl. 1]: S3–S29. Sheline GE, Wara WM, Smith V (1980). Therapeutic irra- Tompkins WA (1999). Immunomodulation and therapeu- diation and brain injury. Int J Radiat Oncol Biol Phys 6: tic effects of the oral use of interferon-α: mechanism of 1215–1228. action. J Interferon Cytokine Res 19: 817–828. Shilling V, Jenkins V, Morris R et al. (2005). The effects Tsavaris N, Kosmas C, Vadiaka M et al. (2002). Immune of adjuvant chemotherapy on cognition in women with changes in patients with advanced breast cancer under- breast cancer – preliminary results of an observational going chemotherapy with taxanes. Br J Cancer 87: 21–27. longitudinal study. Breast 14: 142–150. Valentine AD, Meyers CA (2005). Neurobehavioral effects of Shin DM, Khuri FR, Murphy B et al. (2001). Combined interferon therapy. Curr Psychiatry Rep 7: 391–395. interferon-alpha, 13-cis-retinoic acid, and alpha toco- Valentine AD, Meyers CA, Talpaz M (1995). Treatment of pherol in locally advanced head and neck squamous cell neurotoxic side effects of interferon-alpha with naltrex- carcinoma: novel bioadjuvant phase II trial. J Clin Oncol one. Cancer Invest 13: 561–566. 19: 3010–3017. Valentine AD, Meyers CA, Kling MA et al. (1998). Mood Shin DM, Glisson BS, Khuri FR et al. (2002). Phase II and and cognitive side effects of interferon-α therapy. Semin biological study of interferon alpha, retinoic acid, and Oncol 25: 39–47. cisplatin in advanced squamous cell skin cancer. JClin Wagner LI, Sweet JJ, Desa J et al. (2005). Prechemother- Oncol 20: 364–370. apy hemoglobin (Hgb) and cognitive impairment among Shuper A, Stark B, Kornreich L et al. (2000). Methotrexate breast cancer patients. J Clin Oncol 23: 760S. treatment protocols and the central nervous system: sig- Wefel JS, Meyers CA (2005). Cancer as a risk factor for nificant cure with significant neurotoxicity. J Child Neu- dementia: a house built on shifting sand. J Natl Cancer rol 15: 573–580. Inst 97: 788–789. Silverman DH, Dy CJ, Castellon SA et al. (2007). Altered Wefel JS, Lenzi R, Theriault R et al. (2004a). “Chemobrain” frontocortical, cerebellar, and basal ganglia activity in in breast cancer? A prologue. Cancer 101: 466–475. adjuvant-treated breast cancer survivors 5–10 years after Wefel JS, Lenzi R, Theriault R et al. (2004b). The cog- chemotherapy. Breast Cancer Res Treat 103(3): 303–311. nitive sequelae of standard dose adjuvant chemother- 114 Section 2. Effects of cancer and cancer treatment on cognition

apy in women with breast cancer: results of a prospec- Yaffe K, Sawaya G, Lieberburg I et al. (1998). Estrogen ther- tive, randomized, longitudinal trial. Cancer 100: 2292– apy in postmenopausal women effects of cognitive func- 2299. tion and dementia. J Am Med Assoc 279: 688–695. Weiss RB, Vogelzang NJ (1993). Miscellaneous toxicities. In Yoshikawa E, Matsuoka Y, Inagaki M et al. (2005). No DeVita VT, Jr., Hellman S, Rosenberg SA (eds.) Cancer adverse effects of adjuvant chemotherapy on hippocam- Principles and Practice of Oncology (4th edn.) (pp. 2349– pal volume in Japanese breast cancer survivors. Breast 2358). Philadelphia, PA: JB Lippincott Company. Cancer Res Treat 92: 81–84. Wood LJ, Nail LM, Gilster A et al. (2006). Cancer Zaloga GP, Bhatt B, Marik P (2001). Critical illness and sys- chemotherapy-related symptoms: evidence to suggest a temic inﬂammation. In Becker KL (ed.) Endocrinology role for proinﬂammatory cytokines. Oncol Nurs Forum and Metabolism (3rd edn.) (pp. 2068–2076). Philadelphia, 33: 535–542. PA: Lippincott Williams & Wilkins. 9

Effect of hormones and hormonal treatment on cognition

Christien Schilder, Sanne Schagen, and Frits van Dam

Introduction Hormonal therapy is increasingly used for breast and prostate cancer, often for long periods and by This chapter will address the possible influences of elderly patients. Because intact cognitive function- hormonal therapy on cognitive functioning. Hor- ing is essential for independent living and activities monal therapy is an important treatment option of daily life, elucidation of the possible effects of the for, among others, breast and prostate cancer. various hormonal agents and treatment regimens While there is increasing evidence that chemother- is becoming increasingly important. The first part apy induces cognitive dysfunction in a subgroup of of this chapter will provide an introduction to the patients (Tannock et al., 2004), the effects of hor- influence of reproductive hormones on cognitive monal therapy on cognitive functioning have not functions from a neuropsychological point of view. been investigated thoroughly. Actually, it is con- In the second part, the mechanisms of action of the ceivable that hormonal therapy also influences cog- different hormonal agents that are used in cancer nitive performance. After all, hormonal therapies treatment are described and the neuropsychologi- interfere with serum levels of reproductive hor- cal literature on the impact of hormonal therapy on mones (particularly estrogens and androgens) or cognitive functioning is critically reviewed. Finally, with hormonal actions. There are indications that some methodological aspects of the investigation of reproductive hormones are important in cognitive the complex relations between hormone levels and functioning (Bender et al., 2001). The mechanisms neuropsychological test scores are addressed. of action of reproductive hormones on brain structures are not entirely understood. One of the possibilities is that these hormones act through estrogen The role of reproductive hormones and androgen receptors that are present in those in cognitive function brain structures important for cognitive function, for example the hippocampi and the cerebral cortex The influence of hormones begins in the prenatal (Norbury et al., 2003). Furthermore, it has been sug- period of life. Prenatal reproductive hormones exert gested that reproductive hormones have a benefi- long-lasting organizational influences on brain and cial effect on neurotransmitters that are involved in behavior. It is suggested that the early presence of cognitive processes (Cholerton et al., 2002; Norbury androgens may organize the male brain to enhance et al., 2003). certain spatial functions (Sanders et al., 2002). In

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

115 116 Section 2. Effects of cancer and cancer treatment on cognition

adulthood, fluctuations in reproductive hormone clear acute change in cognitive performance, most levels (ranging from daily to monthly to seasonally) prominently on aspects of short- and long-term may cause small fluctuations in cognitive function- verbal memory (Phillips & Sherwin, 1992; Sherwin, ing. Later in life, levels of most reproductive hor- 1988; Verghese et al., 2000). However, there are mones decline with age, and it is hypothesized that few data on the long-term effects of physiologi- this decline is associated with age-related decline cal or surgical menopause on cognitive functioning in cognitive functioning. Androgens and estrogens (Barrett-Connor & Kritz-Silverstein, 1993). are present at different levels in men and women. After the menopausal transition, estrogen levels In men, androgens, in particular testosterone, pre- are low, but in that low range individual hormone dominate. In women, estrogens are the predominat- levels vary between women. Some authors have ing reproductive hormones (Green et al., 2000). investigated the relationship between estrogen levels and cognitive performance in postmenopausal women. The most consistent finding is that higher The influence of reproductive hormones endogenous estrogen levels are related to better on cognition in healthy women scores on verbal memory tasks (Drake et al., 2000; In premenopausal women, the levels of estrogen Hogervorst et al., 2004; Wolf & Kirschbaum, 2002), fluctuate during the menstrual cycle. Neuropsycho- indicating that even small variations in the low logical studies suggest that these fluctuations have range of estrogen levels are related to cognitive per- an influence on cognitive performance. In general, formance. it is found that high levels of estrogens are beneficial In the last few decades, numerous studies have to performance on tasks at which women as a group evaluated the possible influences of hormone excel (particularly tasks on verbal memory, verbal replacement therapy (HRT, mostly prescribed for fluency, and some fine motor skills), but detrimen- relieving menopausal symptoms) on cognitive func- tal to tasks in which men as a group excel (partic- tioning and the risk of developing dementia. ularly tasks on mental rotation and spatial percep- Initially, observational studies showed a signifi- tion) (Hampson, 1990; Hausmann et al., 2000; Maki cantly reduced risk for developing dementia and et al., 2002). Studies on the influence of testosterone Alzheimer’s disease in HRT users (Yaffe et al., 1998). on cognitive functioning in women suggest that However, there were concerns about attributing daily fluctuations are associated with spatial perfor- the reduced risk to the use of HRT, because HRT mance: early in the morning, when testosterone lev- users tend to be healthier and better educated than els are relatively high, spatial performance is better non-users, factors that are in themselves protec- than later in the morning, when testosterone levels tive against cognitive decline. To avoid this “healthy are relatively lower (Sanders et al., 2002). user bias,” the effects of HRT were investigated in The declining levels of reproductive hormones a large randomized, placebo-controlled study, the during the menopausal transition give another Women’s Health Initiative Memory Study (WHIMS). opportunity to relate changes in hormone lev- In this study, the effects of estrogen with or with- els to cognitive functioning in women. During out a progestin on probable dementia and, sec- menopause, levels of estrogen decrease substan- ondarily, mild cognitive impairment were exam- tially but gradually. This gradual decline, how- ined. Contrary to the expectations, a significantly ever, does not result in clearly measurable declines increased risk of “probable dementia” was found in the performance on cognitive tests during the in the estrogen-plus-progestin group; the “estrogen years of the menopausal transition (Henderson alone” group showed a non-significant increase in et al., 2003; Meyer et al., 2003). However, surgi- probable dementia. Remarkably, the risk of mild cally induced menopause, in which estrogen lev- cognitive impairment, thought to be a precursor els drop abruptly, has been reported to induce a to Alzheimer’s disease, was not increased in either Chapter 9. Effect of hormones and hormonal treatment on cognition 117

group (Espeland et al., 2004; Rapp et al., 2003; when testosterone levels are higher (Sanders et al., Shumaker et al., 2003, 2004). After these surpris- 2002). The correlations are in the opposite direc- ing results were published, several authors tried to tion from the correlations found in women, indicat- explain them. One possible explanation is that the ing that there might be an optimum level of testos- women in the WHIMS were already too old at the terone for the performance of spatial tasks that is time HRT was initiated for it to have any protec- situated in the “lower male range.” tive effect. It is hypothesized that there is a limited In men, testosterone levels decline modestly but period of time after cessation of ovarian function consistently with age (Juul & Skakkebaek, 2002). over which HRT is likely to protect cognitive func- Because circulating estrogens arise through aroma- tioning. One possible explanation for such a “lim- tization of testosterone, estrogens decrease as well ited time window” is that the rather rapid deple- with increasing age. Ironically, because the testes tion of estrogen at the time of menopause may have never stop the secretion of testosterone entirely, a particularly pronounced effect on neurons. Hor- elderly men have higher levels of both testosterone mone replacement therapy could conceivably pre- and estrogen than elderly women (Sherwin, 2003). vent that detrimental effect. It is also possible that, It is hypothesized that because of this difference, after a long period of depletion of estrogen, neu- men are less susceptible to Alzheimer’s disease than rons become less sensitive to estrogen, or that older women (Bowen et al., 2005). neurons have reduced responsivity to the hormone This age-related decline of reproductive hor- (Sherwin, 2005). Another possible explanation for mones is used to investigate the impact of hor- the findings in WHIMS is the choice of HRT prepara- mone levels on cognitive function in men. Unfor- tion. In that study, conjugated equine estrogen (with tunately, studies that tried to find correlations or without a progestin) was used, an agent that ele- between endogenous estrogen and testosterone lev- vates steady-state levels of estrone (a less potent els and cognitive performance in aging men mainly estrogen) instead of replicating premenopausal show inconsistent results. In several studies, age variations in hormone levels. It is possible that these seems to play a modifying role in the relationship particular HRT preparations fail to induce a pos- between hormone levels and cognitive functions. itive effect on cognitive functions (Gleason et al., For example, one study only found a positive cor- 2005). Consequently, important questions about the relation between estrogen levels and spatial span impact of HRT on cognitive functioning and the performance, and between testosterone levels and risk on dementia remain. These questions cen- speed of information processing in the age range of ter around issues such as the timing of initiation, 61–72 years (Hogervorst et al., 2004). Another study duration of treatment and type of HRT regimen found a positive relationship between testosterone (Maki, 2004). levels and global cognitive performance as assessed by the Mini-Mental State Examination (MMSE) in their oldest age category, 70–80 years of age (Muller The influence of reproductive hormones et al., 2005). Other studies failed to find any signifi- on cognition in healthy men cant relationship between circulating hormone lev- In men, the influence of reproductive hormones els and cognitive performance (Fonda et al., 2005; on cognitive function has also been the subject Wolf & Kirschbaum, 2002), while one study only of investigation. One way to study this is to use found a correlation between bioavailable testos- the natural fluctuations in testosterone levels. As in terone (but not total testosterone) and cognitive women, testosterone levels in men show daily fluc- performance (Yaffe et al., 2002). tuations. It is found that young men score better on Apart from the possibility that estrogen may pro- spatial tasks late in the morning, when testosterone tect against cognitive decline and Alzheimer’s dis- levels are relatively low, than early in the morning, ease, there are questions concerning the role of 118 Section 2. Effects of cancer and cancer treatment on cognition

declining testosterone levels in the development of while another study suggests a detrimental effect cognitive decline and Alzheimer’s disease in aging (O’Connor et al., 2001). men. This becomes even more important because of the finding that men with Alzheimer’s disease show lower testosterone levels than men of the same age Neuroimaging studies without Alzheimer’s disease (Hogervorst et al., 2001; To unravel the complex associations between Rosario et al., 2004). These findings are ambiguous, levels of reproductive hormones and cognitive however, because of the possibility that depleted functioning, some authors used the assistance of testosterone levels in men with Alzheimer’s dis- neuroimaging techniques. These techniques are ease may actually be a consequence of the dis- predominantly used in studies on supplementa- ease rather than a cause. Research to resolve this tion of reproductive hormones. The results suggest ambiguity is ongoing. There are some preliminary that HRT in women is capable of altering brain results that indicate that a low free testosterone level structure and brain activation patterns, especially occurs before the diagnosis of Alzheimer’s disease in brain regions that are important for memory and is possibly a risk factor (Moffat et al., 2004), but (Cook et al., 2002; Eberling et al., 2000; Luoto et al., more prospective longitudinal studies are needed to 2000; Maki & Resnick, 2000; Resnick et al., 1998; assess the impact of long-term testosterone levels Schmidt et al., 1996; Shaywitz et al., 1999). Further- on cognitive decline and Alzheimer’sdisease. In fact, more, testosterone treatment in hypogonadal men if testosterone levels turn out to be a risk factor for increases cerebral perfusion in various brain regions cognitive decline and Alzheimer’s disease, this will (Azad et al., 2003). However, the clinical significance have important consequences for the role of andro- of these alterations is not clear and needs further gen substitution therapy in prevention and treat- investigation. ment. While the potential role of androgen substitution therapy for the prevention and treatment of Conclusions Alzheimer’s disease is unclear, the impact of short- term use of this therapy on cognitive functioning Although the relationships between reproductive has in fact been investigated in several studies. hormones and cognitive functioning have been Androgen substitution therapy is used by men of all addressed in many different studies, they appear to ages who experience symptoms of hypogonadism, be complex and, in many cases, unclear. There is such as decreased muscle mass and bone den- increasing evidence that, in general, reproductive sity, fatigue, decreased energy levels and decreased hormones play a modest role in cognitive function- libido (Juul & Skakkebaek, 2002). But studies of ing. For estrogen, the most consistent effects are theimpactofandrogensubstitutiononcognitive found on verbal memory performance in women functioning have also produced conflicting results. (Sherwin, 2000). The relationship between andro- Some authors found a beneficial effect on measures gens and cognitive functioning seems to be a com- of working memory (Janowsky et al., 2000), spa- plex one. Androgens particularly appear to affect tial memory (Cherrier et al., 2001), verbal memory spatial abilities but in a complex way: it has been (Cherrier et al., 2001), and verbal fluency (O’Connor suggested that there might be an optimal testos- et al., 2001), whereas other authors did not find any terone level, situated in the “lower male range.” significant correlation with scores on cognitive tests Both lower and higher levels of testosterone may (Haren et al., 2005; Kenny et al., 2002; Wolf et al., have a detrimental effect on spatial performance. 1997). Even within the cognitive domain of spatial Until now, many questions regarding the impact abilities, the findings are contradictory: one study of substitution of hormones on cognitive func- suggests a beneficial effect (Cherrier et al., 2001), tioning, and the relationship between reproductive Chapter 9. Effect of hormones and hormonal treatment on cognition 119

hormone levels and the risk of dementia are still releasing hormone (LHRH) analogs were intro- unanswered. duced. These drugs are used to produce “reversible” chemical castration in both sexes. In women, these drugs suppress ovarian production of estrogen, Hormonal therapy in breast and prostate while in men the production of testosterone by the cancer: is there an influence on cognitive testes is reduced. The most used LHRH agonists are function? leuprolide and goserelin (Hellerstedt & Pienta, 2002; Miller, 1996). The section above shows that neuropsychological In breast cancer treatment, tamoxifen is widely and neuroimaging studies provide evidence for a used in both pre- and postmenopausal patients. In relationship between reproductive hormones and the 1960s, when the drug was synthesized, it was cognitive functions. In addition, reproductive hor- demonstrated to have antiproliferating effects in the mones play a major role in the etiology and treat- breast. The drug appeared to be capable of binding ment in breast and prostate cancer. Hormonal ther- to the estrogen receptors in breast tissue, thereby apy interferes with reproductive hormone levels or preventing estrogen from initiating the estrogenic with the activity of reproductive hormones. There- effects. Tamoxifen thus became widely known as fore, it has been suggested that hormonal therapies an anti-estrogen. Since then, it has been discovered may have an effect on cognitive function as well. that it, paradoxically, has many estrogenic quali- This section will point out the possible effects of ties, including agonist effects on bone, blood lipids, hormonal therapy on cognitive function. After an and the endometrium (Osborne et al., 2000). This introduction to the mechanisms of action of vari- finding led to the development of new drugs with ous hormonal therapies in the treatment of breast specific and selective effects on the estrogen recep- and prostate cancer, an overview will be given of the tor function. Currently, tamoxifen and related drugs studies on the effects of hormonal therapies on cog- are collectively known as selective estrogen receptor nitive function. modulators (SERMs) (Goss & Strasser, 2001). SERMs such as tamoxifen are tolerated relatively well. How- ever, their estrogenic as well as anti-estrogenic qual- The role of hormonal therapies in cancer ities can lead to a variety of side-effects, including treatment thromboembolic events, hot flashes, and the risk of Hormonal therapy plays a prominent and increas- endometrial cancer with prolonged use. ing role in the treatment of breast and prostate The introduction of aromatase inhibitors as a cancer. The healthy breasts and prostate are new class of agents has extended the treatment dependent on reproductive hormones, as are a high options for breast cancer patients. The enzyme aro- percentage of the malignant tumors that originate matase is required for the peripheral conversion from these organs. Interference with the hormonal of testosterone and androstenedione to estrogen, milieu will slow down or stop the growth of the the final step in the estrogen biosynthesis path- tumor in many cases. There are several methods way (Visvanathan & Davidson, 2003). Aromatase to achieve this interference in the hormonal milieu inhibitors almost completely inhibit the action of (Tripathy & Benz, 2001). this enzyme. Consequently, aromatase inhibitors Originally, deprivation of reproductive hormones lower the level of circulating estrogen by almost was achieved by ovariectomy or irradiation of the 100% (Simpson & Dowsett, 2002). ovaries in women and orchiectomy in men. Despite In prostate cancer, hormonal therapy plays an the efficacy of these methods, patient acceptance is important role, especially in metastatic disease, usually poor because of the irreversibility of the although in the adjuvant and neoadjuvant settings intervention. In the 1980s, luteinizing-hormone- its role is increasing. The most frequently used 120 Section 2. Effects of cancer and cancer treatment on cognition

option is androgen deprivation by means of LHRH study in young women (mean age 27 years) did not agonists, such as leuprolide and goserelin. Treat- find any effect on cognition after similar treatment ment with LHRH agonists initially results in a rise in (Owens et al., 2002). For breast cancer patients, serum testosterone levels, potentially causing stim- data on the impact of LHRH agonists on cogni- ulation of tumor growth and accompanying side- tive function are scarce. In one study on adjuvant effects, such as increased bone pain. To prevent this tamoxifen and/or goserelin therapy, premenopausal “tumor flare,” LHRH treatment is often preceded by breast cancer patients (mean age 45 years) reported treatment with an “anti-androgen.” Anti-androgens on a questionnaire increased memory and con- block the effect of the testosterone surge on andro- centration problems during the period of treat- gen receptors, preventing testosterone from exert- ment. However, an increase was also found in the ing its growth-promoting effect on the tumor. The tamoxifen group and in an untreated control group most commonly used anti-androgens are flutamide (Nystedt et al., 2000). and bicalutamide. The use of anti-androgens can Data on the impact of tamoxifen on cognitive be continued for 2–4 weeks. Another possibil- function are also preliminary. The finding that ity is to combine LHRH agonist treatment with tamoxifen treatment often induced hot flashes led anti-androgen treatment for longer periods. Anti- to the hypothesis that tamoxifen acts as an estro- androgens can also be used as monotherapy (Sharifi gen antagonist within the central nervous system et al., 2005). In order to optimize the treatment and may in the long term lead to cognitive deficits outcome and reduce long-term toxicities related (Benson, 2002). Experimental evidence for a detri- to testosterone deficiency, “intermittent androgen mental effect on memory was found in two experi- suppression” (IAS) is introduced as a treatment ments with mice. The results suggest that tamoxifen option. In IAS, androgen suppression (6–9 months) impairs memory function (especially the retrieval is alternated with an off-treatment period in which of spatial information) in mice (Chen et al., 2002a, testosterone levels return to physiologic levels. As 2002b). the “prostate-specific androgen” (PSA) reaches a In four studies with patients that made use certain threshold, treatment is reinstated (Cherrier of neuropsychological tests, tamoxifen users were et al., 2003). included (see Table 9.1 for more detailed information). Paganini-Hill and Clark (2000) were the first to Hormonal therapies in breast cancer: investigate the impact of tamoxifen on cognitive the influence on cognitive functioning functioning. They mailed a questionnaire, includ- To date, there is only limited information about ing three neuropsychological tests (clock drawing, the impact of the different hormonal therapies for copying a box drawing, narrative writing), to breast breast cancer on cognitive function. Much of the cancer patients. They analyzed data from 1163 evidence for the impact of LHRH agonists on cog- women: 710 women had taken tamoxifen and 453 nitive functioning comes from studies with young had never used it. The tamoxifen group was divided women who received those agents for benign gyne- in two subgroups, past users (n = 428) and current cologic conditions. In one study, young female users (n = 241). Using a cross-sectional design, they LHRH-agonist users (mean age 32.4 years) reported found few differences between test scores of women a decrease in memory function (Newton et al., who had used tamoxifen for the standard 5 years 1996). Studies that used neuropsychological tests and never-users. However, more women who had show mixed results: one study showed a decrease used tamoxifen for 5 years or longer reported see- in verbal memory function in young women (mean ing their physician for memory problems than non- age 34.2 years) after 3 months of leuprolide treat- users. Current users also had significantly lower ment (Sherwin & Tulandi, 1996), while another mean complexity scores on the narrative writing Table 9.1. Overview of the studies on the impact of hormonal therapy on cognitive functioning in breast cancer. (ATAC Anastrozole, tamoxifen and combined trial, GHQ General Health Questionnaire, HRT hormone replacement therapy, MMSE Mini-Mental State Examination, MRI magnetic resonance imaging, NART National Adult Reading Test, PET positron emission tomography, WMS Wechsler Memory Scale)

Design and Study Goal Patients and comparisons Cognitive measures Biological measures statistical analysis

Paganini-Hill & Clark To assess the effect of Breast cancer patients Clock drawing task Cross-sectional design. (2000). Preliminary tamoxifen on cognitive 1. Past tamoxifen users Copying a box drawing Patients recruited assessment of function (short-term: <4years; Narrative writing to describe from a population- cognitive function in standard-term: 4–5 years; a pictured scene based case–control breast cancer long-term: 6+years) Tests were mailed to the study of women with patients treated with (n = 428) patients and completed at primary breast tamoxifen. Breast 2. Current tamoxifen users home cancer Cancer Res Treat 64: (short-term: <4years; Group comparison, 165–176. standard-term: 4–5 years; mean scores per test long-term: 6+years) Group comparison of (n = 241) percentage patients 3. Never users (n = 453) with errors

Affective measures Cognitive complaints Conclusion Remarks Geriatric Depression Scale Questions in questionnaire Little differences between Limitations of the study: r test scores of women who Current users had had used tamoxifen for the signiﬁcantly lower mean standard 5 years and complexity score on the never users narrative writing task r More women who had used The battery of tests was small tamoxifen for the standard and didn’t contain a test of term or longer reported verbal memory r seeing their physician for Patients were not randomly memory problems than assigned to receive non-users tamoxifen r It was impossible to classify over 20% of the patients by their duration of use r Tests were mailed to the patients, it is not sure that they did them without guidance r Non-responders were more likely to be less educated, single and not Caucasian (cont.) Table 9.1. (cont.)

Design and Study Goal Patients and comparisons Cognitive measures Biological measures statistical analysis

Ernst et al. (2002). The Assessment of the effects Three groups of age-matched Modiﬁed MMSE Four metabolites in the brain Cross-sectional design; effects of tamoxifen of tamoxifen and women from local Digit Symbol substitution test (myo-inositol; creatine; Mean scores per and estrogen on estrogen on the brain community Trailmaking A N-acetyl-containing test/mean brain metabolism in chemistry of elderly 1. Breast cancer patients compounds; concentration per elderly women. J women treated with tamoxifen choline-containing metabolite Natl Cancer Inst 20 mg/day for at least compounds) measured by 94(8): 592–597. 2 years (mean ± sd proton magnetic resonance = 4.4 ± 1.7 years) spectroscopy (n = 16) 2. Healthy women using HRT for at least 2 years (mean ± sd = 20.8 ± 10.5 years) (n = 27) 3. Healthy controls (n = 33)

Affective measures Cognitive complaints Conclusion Remarks No differences in scores on This article is criticized by other cognitive tests authors. It is stated that there Lower myo-inositol are alternative explanations concentration in HRT and for the lower myo-inositol tamoxifen group compared concentrations in the brain with controls, suggesting a similar, maybe neuroprotective effect of both agents Eberling et al. (2004). To evaluate the effects of 40 postmenopausal women, MMSE Each subject underwent an Cross-sectional design Estrogen- and estrogen and tamoxifen recruited by advertisement Verbal episodic memory MRI scan and a PET scan For the cognitive scores: tamoxifen-associated on cognitive testing, on Three groups (VEM) At the MRI scan, hippocampal group means were effects on brain PET measures on brain 1. Tamoxifen group: breast Semantic memory (object volumes were rated calculated structure and metabolism and MRI cancer patients using naming) PET data: group function. measures of tamoxifen (n = 10) Verbal attention span (VAS) comparisons of Neuroimage 21: hippocampus atrophy 2. Estrogen group: women Pattern recognition (PR) glucose metabolism 364–371. currently taking MRI data: ratings of unopposed estrogen hippocampal volume (n = 15) were compared 3. Women not taking between groups tamoxifen or estrogen (n = 15)

Affective measures Cognitive complaints Conclusion Remarks Center for Epidemiological The tamoxifen group showed The authors state that the Studies – Depression significantly poorer findings provide support for Scale (CES-D) performance on the an antagonistic role of semantic memory test tamoxifen on certain brain than the other two groups areas that are related to PET measures: the tamoxifen cognitive function group showed widespread Limitations: areas of hypometabolism • Relatively small sample sizes in some areas of the frontal • Differences in estrogen or lobe (inferior and dorsal tamoxifen use with respect to lateral areas) relative to the dosage and duration could other two groups affect the results MRI measures: the tamoxifen • Cognitive function of the group had smaller right tamoxifen group could be hippocampal volumes influenced by general than the estrogen group anesthesia, stress or (borderline significance) depression (cont.) Table 9.1. (cont.)

Design and Study Goal Patients and comparisons Cognitive measures Biological measures statistical analysis

Shilling V. et al. (2003). To establish whether Breast cancer patients from NART Cross-sectional design The effects of significant cognitive the ATAC trial: WMS III: logical memory Group comparisons hormone therapy on deficit exists in women 1. Patient group I + II Raw scores for each of cognition in breast receiving hormone (n = 94) consisting of three WMS III: faces I + II the cognitive cancer J Steroid therapy for breast cancer subgroups: WMS III: spatial span measures were Biochem Mol Biol 86: To develop a statistical • Tamoxifen + anastrozole WMS III: digit span converted to z-scores 405–412. package that is sensitive placebo WMS III: letter-number using the mean score to the potential effects of • Anastrozole + tamoxifen sequencing of the control group estrogen deficiency on placebo Kendrick digit copying task as a reference cognition • Anastrozole + placebo Correlational analyses 2. Healthy control group between cognitive (n = 35) measures and depression measures, and between length of treatment and cognitive performance

Affective measures Cognitive complaints Conclusion Remarks Beck Depression Inventory Broadbent Cognitive Failures The patient group did not The study was not designed or GHQ-12 Questionnaire (25 items) differ from the control powered to investigate group on measures of differences between working memory, attention treatment arms and visual memory, but was signiﬁcantly impaired on measures of verbal memory and processing speed. Cognitive performance was not signiﬁcantly related to length of time on trial or psychological morbidity Chapter 9. Effect of hormones and hormonal treatment on cognition 125

task. This study suggests that current use of tamox- imaging (MRI) measures on hippocampal atrophy. ifen may adversely affect cognition. However, the In addition, five neuropsychological tests were used interpretation of the results is problematic because [MMSE, object naming, attention span (Digit Span), the cognitive assessments were sent by mail and, verbal memory, and pattern recognition]. Subjects consequently, the data may be unreliable. Further- were 40 postmenopausal women (10 tamoxifen more, the selected tests were insensitive and not users, 15 estrogen users and 15 controls). The adequate with regard to the expected cognitive tamoxifen group showed significantly poorer per- functions affected by tamoxifen. For example, the formance on the naming test than the other two assessment did not include a test of verbal memory. groups. The other neuropsychological tests and the Ernst et al. (2002) found more or less opposite depression questionnaire showed no significant dif- results. They used proton magnetic resonance spec- ferences between groups. On the PET measures, troscopy, a neuroimaging technique that measures the tamoxifen group showed widespread parts of concentrations of biochemical markers associated hypometabolism in some parts of the frontal lobe with brain injury, to study the impact of tamox- (inferior and dorsal lateral areas) relative to the ifen and HRT on brain function. They compared other groups. In comparison with women not taking brain metabolism of women with breast cancer who estrogen or tamoxifen, the estrogen group showed received tamoxifen (n = 16) with healthy women higher rates of metabolism in the inferior frontal who had received HRT (n = 27) and with healthy cortex and temporal cortex. On the MRI measures, controls (n = 33). They looked at four biochemical the tamoxifen group had smaller right hippocam- markers. In addition, they used three neuropsycho- pal volumes than the estrogen group, but this effect logical tests (one screening instrument and two tests was of borderline significance. The authors con- for psychomotor speed). In the neuropsychological cluded that their findings provide support for an part of the study, they found no differences in group anti-estrogenic role of tamoxifen in certain brain means on the tests. In the spectroscopy part, they areas that are related to cognitive function. found reduced concentrations of myo-inositol in Although data on tamoxifen are sparse, the cog- the brains of women treated with tamoxifen and in nitive effects of another SERM (raloxifene) are the women who used HRT, compared with control well documented. Tamoxifen and raloxifene differ women. They suggested that patients might receive in their profiles of estrogenic and anti-estrogenic neuroprotective benefits both from HRT and from qualities. Raloxifene is primarily used in treatment tamoxifen. However, this study has been highly crit- and prevention of osteoporosis. Two large random- iqued, for both its design and its interpretation of ized placebo-controlled studies of postmenopausal the results, by Ganz et al. (2002), who state that alter- women with osteoporosis reported no significant native explanations for the lower myo-inositol con- detrimental effects of raloxifene on cognitive per- centrations in the brain in estrogen and tamoxifen formance (Nickelsen et al., 1999; Yaffe et al., 2001). users should be considered. The lower myo-inositol Moreover, raloxifene (120 mg/day) resulted in a concentrations could, for example, reflect the life- reduced risk of cognitive impairment (Yaffe et al., long exposure to endogenous and exogenous estro- 2005). Even though raloxifene does not play an gen. After all, older women that develop breast can- important role in the treatment of breast cancer often have higher circulating levels of estrogen cer, these results may be important for tamoxifen after menopause compared to women who do not users because of similarities between the agents develop breast cancer. (both have estrogenic and anti-estrogenic proper- In a neuroimaging study, Eberling et al. (2004) ties). However, it is possible that differences in the evaluated the effects of tamoxifen and estrogen on profiles of estrogenic and anti-estrogenic proper- positron emission tomography (PET) measures of ties result in distinctly different effects on cognitive brain glucose metabolism and magnetic resonance functioning. 126 Section 2. Effects of cancer and cancer treatment on cognition

The impact of aromatase inhibitors on cognitive cally investigate the impact of hormonal therapy functioning remains virtually unknown. Because in prostate cancer were Green et al. (2002). They these agents induce a substantial drop in circulating tested 65 prostate cancer patients who were ran- estrogens, it is supposed that alterations in cogni- domized between four treatment modalities: two tive functions are associated with these treatments. types of LHRH agonists (leuprorelin and gosere- After all, as stated earlier, variations in the low lin), anti-androgen monotherapy with cyproterone range of postmenopausal hormone levels are found acetate and close clinical monitoring. They com- to be associated with variations in cognitive func- pared baseline scores with scores after 6 months tions. Beside tamoxifen users, Shilling et al. (2003) of treatment. They found that goserelin users (one included patients treated with anastrozole (an aro- of the LHRH agonists) had improved their perfor- matase inhibitor) in their study on the effects of hor- mance on a task for visual memory. In the domain of mone therapy on cognition. They tested, in a cross- verbal memory, goserelin users had improved their sectional design, 94 breast cancer patients from score on one test of verbal memory (recalling prose theanastrozole,tamoxifenandcombined(ATAC) passages) but worsened their score on another ver- trial and 35 non-cancer controls. Patients were bal memory task (list learning). Patients random- randomized to receive tamoxifen, anastrozole or ized to “close clinical monitoring” also improved tamoxifen/anastrozole in combination. Cognitive on the prose verbal memory task. Besides compar- assessments consisted of a range of memory and ing mean scores, the authors examined the differ- attention functions. The patient group (consisting ences between baseline and follow-up scores for of tamoxifen users, anastrozole users and users of each patient. They found that none of the “close the combination) did not differ from the control monitoring” patients showed a significant change group on measures of working memory, attention, in cognitive variables over 6 months. For patients or visual memory, but had significantly impaired receiving active treatment, 24 out of 50 showed a verbal memory and processing speed compared to decrease on at least one cognitive task and 7 out of the control group. Cognitive performance was not 50 on two or more tasks. significantly related to the length of time in the Salminen et al. (2003, 2004, 2005) published three trial or measures of psychological morbidity. The studies on the influence of androgen deprivation authors state that the study was not designed or therapy on cognitive functioning. In the first study able to investigate differences in effects between (Salminen et al., 2003), they tested 25 prostate can- tamoxifen and anastrozole on cognitive function, cer patients at baseline (before starting therapy) and speculate that differences in test scores may and after 6 and 12 months of therapy. The therapy reflect the different activities of tamoxifen and anas- started with 250 mg flutamide (an anti-androgen) trozole. The overall conclusion from the study was three times a day for 4 weeks. An LHRH agonist that the adverse effect of hormone therapy on cog- was added after 2 weeks and lasted 12 months. nition seems to be specific (i.e., on verbal memory Patients acted as their own controls in the follow- and processing speed) rather than widespread. up. The authors did not find any impairment in cognitive functioning during androgen deprivation therapy; on the contrary, they found an improve- Hormonal therapies in prostate cancer: ment in the scores of an object recall task and the influence on cognitive functioning a semantic memory task. In two additional stud- Just as in hormonal therapy in breast cancer, infor- ies on an identical patient population, associations mation about the impact of hormonal therapy for between estrogen and testosterone levels and scores prostate cancer on cognitive functioning is lim- on cognitive tasks were investigated (Salminen et al., ited (see Table 9.2 for a description of the con- 2004, 2005). Serum samples of testosterone, free ducted studies). The first authors to systemati- testosterone, and estradiol were taken at baseline Table 9.2. Overview of the studies on the impact of hormonal therapy on cognitive functioning in prostate cancer. (AVLT Auditory Verbal Learning Test, CAMCOG cognitive and self-contained part of the Cambridge Examination for Mental Disorders of the Elderly, CPA cyproterone acetate, PSA prostate-specific antigen, SHBG sex hormone binding globulin, SOP subjected-ordered pointing, WAIS Wechsler Adult Intelligence Scale, WMS-R Wechsler Memory Scale, Revised)

Design and Study Goal Patients and comparisons Cognitive measures Biological measures statistical analysis

Green et al. (2002). To report the first 65 prostate cancer patients, WMS-Rvisualmemory Periodic serum assays of Longitudinal design. Altered cognitive systematic investigation randomly assigned to four WMS-Rverbalmemory testosterone and PSA were Cognitive function in men of the cognitive effects of conditions: AVLT collected to check patient assessments at treated for prostate LHRH analogs on 1. Treatment with leuprorelin Rey complex figure; compliance baseline and at cancer with prostate cancer patients (LHRH analog); n = 19 immediate and delayed 6 months. Group × luteinizing 2. Treatment with goserelin recall time analyses were hormone-releasing (LHRH analog); n = 20 WMS-R Attention and conducted for serum, hormone (LHRH) 3. Treatment with (steroidal concentration index cognitive and analogues and anti-androgen); n = 11 WAIS Digit Symbol emotional measures. cyproterone acetate: 4. Close monitoring; n = 15 Trail Making Test A and B The “reliable change arandomized Controlled Oral Word index” was used to controlled trial. Association Test more closely examine BJU Int 90: 427–432. Stroop test Victoria version individual results and WAIS-R 4 subtests to identify clinically significant cognitive changes

Affective measures Cognitive complaints Conclusion Conclusion (continued) Remarks Depression Anxiety Stress The results suggest that A signiﬁcant decrease on ≥1 Limitations of the study: Scales (DASS-21) patients treated with task was observed for 9 out • Small sample size goserelin improved of 19 patients on • Control group had higher performance on WMS-R leuprorelin, 9 out of 20 average intelligence and visual and verbal memory patients on goserelin, 6 out education than active measures but decreased of 11 patients on CPA and treatment groups performance on the AVLT none out of 15 patients on measure of verbal memory close monitoring. The close monitoring group increased their performance on the WMS verbal memory test (cont.) Table 9.2. (cont.)

Design and Study Goal Patients and comparisons Cognitive measures Biological measures statistical analysis

Cherrier et al. (2003). To evaluate the effects of Two groups: Spatial memory: Serum total testosterone levels Longitudinal design: The effects of androgen deprivation on 1. 19 patients with prostate • Route test and free testerone levels were cognitive testing combined androgen cognitive function. cancer (52–76 years of age) Spatial ability: obtained from the patient occurred twice before blockade on receiving intermittent • Block design test group the start of treatment, cognitive function androgen suppression • Mental rotation test after 9 months of during the ﬁrst cycle therapy Verbal memory: treatment, and after 3 of intermittent 2. 15 healthy community • Proactive interference test months of treatment androgen dwelling control • Story recall test suppression in participants Verbal ability: patients with Letter ﬂuency test prostate cancer Executive function: JUrol170: Stroop color word test 1808–1811. • Visual working memory task

Affective measures Cognitive complaints Conclusion Conclusion (continued) Remarks 9 months of combined Preliminary results androgen blockade because of: resulted in: • Small sample size • A beneﬁcial effect on verbal memory • A detrimental effect on one measure of spatial ability (mental rotation).

Salminen et al. (2003). To assess whether Two groups: WAIS similarities Longitudinal design: r Androgen cognitive functioning is • 25 prostate cancer patients WAIS Digit span Patients were tested at deprivation and impaired during 12 receiving neoadjuvant WAIS Digit-symbol baseline (before start of cognition in prostate months of androgen androgen depression Substitution subtest therapy); and at 6 and 12 cancer. Br J deprivation therapy therapy for 12 months in WAIS block design months of therapy r Cancer89: 971–976. connection with radical Verbal ﬂuency (animals) Patients acted as their own radiotherapy Picture naming task controls in the follow-up r Naming time (immediate and Healthy controls were tested delayed) once Word list recall • 52 healthy control males Benton Visual Recognition (baseline measurement (form C) only) Wechsler Visual Memory Span MMSE CogniSpeed Reaction times (4 subtests) Affective measures Cognitive complaints Conclusion Conclusion (continued) Remarks Beck Depression No impairment in cognitive Improvement was observed Limitations of the study: Inventory performances was found in object recall (naming • Small sample size Quality of Life forms during androgen time, immediate and • The authors didn’t control for (EORTC QLQ-C30) with deprivation therapy delayed); and semantic practice effects additional items on memory (similarities) • The healthy controls were sexual functioning, In QOL, impairment in only used for a comparison physical and mental physical functioning was with the cognitive discomfort, worry, role observed. performances of the patient functioning, limitations group at baseline, limiting in daily activities and the possibilities of bother due to prostate interpretation cancer or treatments

Salminen et al. (2004). To investigate the 23 prostate cancer patients WAIS similarities Serum samples for hormone Longitudinal design. Associations associations between receiving androgen WAIS Digit Span analysis were collected at Changes in hormone between serum serum testosterone deprivation therapy for 12 WAIS Digit-Symbol baseline and at 6 and 12 levels and in testosterone fall and decline and specific months: Substitution subtest months cognitive cognitive function in cognitive functions in • 250 mg of flutamide 3 times WAIS block design Serum testosterone, free performance from prostate cancer newly diagnosed a day for 4 weeks Verbal fluency (animals) testosterone and sexual baseline to 6 and 12 patients prostate cancer patients • From week 2: 11.25 mg Picture naming task hormone binding globulin months were Clin Cancer Res 10: treated with androgen leuprolide every Naming time (immediate and were determined calculated and 7575–7582. deprivation therapy 3 months delayed) compared Word list recall Regression analysis and Benton Visual correlations were Recognition (C) used to study the (cont.) Table 9.2. (cont.)

Design and Study Goal Patients and comparisons Cognitive measures Biological measures statistical analysis

Wechsler Visual associations between Memory Span changes in hormone MMSE levels and cognitive CogniSpeed Reaction times performances (4 subtests)

Affective measures Cognitive complaints Conclusion Conclusion (continued) Remarks Beck Depression Testosterone decline was Limitations of the study: Inventory associated with visuo • Small sample size motor slowing, slowed reaction times in some attentional domains including working, memory and impaired hit-rate in a vigilance test, impaired delayed recall and recognition speed of letters. Improvement in object recall

Salminen et al. (2005). To investigate the 23 prostate cancer patients WAIS similarities Serum samples for hormone Longitudinal design. Estradiol and association of androgen- receiving androgen WAIS Digit Span analysis were collected at Changes in hormone cognition during deprivation-induced deprivation therapy for 12 WAIS Digit-Symbol baseline and at 6 and 12 levels and in androgen estradiol decline with months: Substitution subtest months cognitive deprivation in men cognition in prostate • 250 mg of flutamide 3 times WAIS block design Serum testosterone and performance from with prostate carcinoma a day for 4 weeks Verbal fluency (animals) estradiol were determined baseline to 6 and 12 carcinoma. • from week 2: 11.25 mg Picture naming task months were Cancer 103 (7): leuprolide every 3 months Naming time (immediate and calculated and 1381–1387. delayed) compared Word list recall Regression analysis and Benton Visual correlations were Recognition (C) used to study the Wechsler Visual associations between Memory Span changes in hormone MMSE levels and cognitive CogniSpeed Reaction times performances (4 subtests) Affective measures Cognitive complaints Conclusion Conclusion (continued) Remarks Beck Depression Estrogen decline during Limitations of the Inventory androgen deprivation study: therapy was associated • Small sample size with a decline in visual memory and recognition speed at 6 months and with an improvement of verbal fluency at 12 months

Almeida et al. (2004). To clarify whether Prostate cancer patients who CAMCOG Blood was drawn every visit to Longitudinal design One year follow-up testosterone depletion were prescribed WMS-III word list monitor plasma levels of PSA Naturalistic study study of the and receptor blockade intermittent androgen WMS-III verbal paired estradiol, testosterone and Patients were tested 1 association between are associated with deprivation treatment with associates beta-amyloid week prior to chemical castration, changes in mood and leuprolide (an LHRH WMS-III visual reproduction baseline, at baseline sex hormones, cognitive functions in analog) every 3 months for WAIS-III block design and after 4, 12, 24, 36 beta-amyloid, humans 36 weeks) and ﬂutamide weeks of treatment. memory and 250 mg (an anti-androgen) At 36 weeks, depression in men. for 36 weeks. n = 40; mean treatment was Psychoneuroen- age = 72.4) discontinued and docrinology 29: patients were 1071–1081. re-assessed at 42, 48 and 54 weeks after baseline Multivariate analysis of variance for repeated measures; correlations (cont.) Table 9.2. (cont.)

Design and Study Goal Patients and comparisons Cognitive measures Biological measures statistical analysis

Affective measures Cognitive complaints Conclusion Conclusion (continued) Remarks Beck Depression Discontinuation of treatment No effect on the block design Limitations of the study: Inventory was linked to an test • Small sample size Beck Anxiety improvement on the • The authors didn’t control for Inventory CAMCOG, and on recall practice effects due to total scores on the word list repetitive testing test and the verbal paired associates test

Jenkins et al. (2005). To examine, in a 32 patients with localized Intelligence: At all three time points serum Longitudinal design Does neoadjuvant prospective study, the prostate cancer, treated in • National Adult Reading Test free and bound testosterone, Naturalistic study hormone therapy for inﬂuence that a neoadjuvant setting Verbal ability: beta-estradiol and Cognitive assessments early prostate cancer temporary reversible 18 healthy control subjects • Phonemic verbal ﬂuency sex-hormone binding at baseline, at 3 affect cognition? medical castration for task globulin (SHBG) levels were months of therapy Results from a pilot localized prostate cancer Verbal memory: determined and 9 months later study. BJU Int has on cognition, by • Rey Auditory Verbal Group comparisons of 96: 48–53. assessing whether Learning Test test scores were made temporary (3–5 months) Visual memory: using one-way and treatment with a LHRH • Complex Figure Task repeated measures agonist before radical Working memory: ANOVA or radiotherapy had a • WMS III Digit Span chi-squared tests short- or long-term • WMS III Spatial Span effect on cognitive Processing Speed: function • Kendrick Assessment of Cognitive Ageing battery

Affective/QOL measures Cognitive complaints Conclusion Conclusion (continued) Remarks General Health Semi-structured interview at Short-term LHRH therapy for Limitations of the study: Questionnaire the ﬁnal assessment early-stage prostate cancer • Small sample size Functional Assessment of has modest consequences • The authors didn’t control for Cancer Therapy – on cognitive functioning in practice effects due to Prostate (FACT – P) some men. The most repetitive testing Trial Outcome Index affected cognitive domain is spatial ability Beer et al. (2006). To examine the impact of 18 patients with Verbal memory: At the two time points serum Longitudinal design Testosterone loss androgen deprivation androgen-independent • Paragraph recall immediate total and free testosterone, Naturalistic study and estradiol and subsequent prostate cancer, starting recall estradiol and SHBG levels Cognitive assessments administration estradiol therapy on the second-line hormonal • Paragraph recall delayed were determined at baseline and 4 modify memory in long-term and working treatment with recall weeks later men. JUrol175: memory of patients with transdermal estradiol Working memory: Repeated measures 130–135. prostate cancer 18 prostate cancer patients • SOP ANOVA were used to on continuous androgen • Trail Making Test compare deprivation • MMSE performance across 17 community dwelling • WAIS-R Vocabulary the two study visits. healthy control subjects One-way ANOVA with Tukey-HSD post hoc corrections or paired t-tests with Bonferroni corrections were used to examine main effects Affective/QOL measures Cognitive complaints Conclusion Conclusion (continued) Remarks Geriatric Depression Patients on hormonal Patients that received Limitations of the study: Scale therapy performed worse estradiol therapy improved • Small sample size POMS on verbal memory and their verbal memory • Short study duration information-processing performance compared to speed. Performance on baseline working memory measures do not differ between groups (cont.) Table 9.2. (cont.)

Design and Study Goal Patients and comparisons Cognitive measures Biological measures statistical analysis

Taxel et al. (2004). The To determine the effect of 27 patients receiving Verbal memory: At the two time points serum Longitudinal design effect of short-term estrogen alone (without treatment with LHRH • Rey AVLT testosterone, estrogen and Cognitive assessments estradiol therapy on the inﬂuence of agonists for localized Visual memory: SHBG levels were at baseline and 9 cognitive function in testosterone) on prostate cancer after • Benton visual retention test determined weeks later older men receiving cognitive function in primary treatment, or Executive function: Repeated measures hormonal older prostate cancer planning to receive • Stroop color word test ANOVA were used to suppression therapy patients neoadjuvant treatment • Trail Making Test examine changes in for prostate cancer. with LHRH agonists for • Controlled oral word hormone levels over J Am Geriatr Soc 52: stage B or C prostate association test time. Cognitive 269–273. cancer MMSE function data were Patients were randomized analyzed with between 1 mg/day 17-β repeated measures micronized estradiol MANOVA versus placebo

Affective/ QOL measures Cognitive complaints Conclusion (Conclusion continued) Remarks Beck Depression Cognitive failures Short-term treatment with No differences between the Limitations of the Inventory questionnaire low-dose estrogen resulted groups in any other measure. study: in a improvement of 2 out No difference in • Small sample size of 17 test scores self-reported cognitive • Short study duration (information-processing deﬁcits or depressive speed, executive symptoms functioning) Chapter 9. Effect of hormones and hormonal treatment on cognition 135

and at 6 and 12 months. Significant associations ting. Patients with localized prostate cancer (n = were found between testosterone decline and slow- 32, mean age 67.5 years) had cognitive assessments ing on some cognitive tasks, mostly in the domain of before the start of treatment with LHRH agonists, attention. Testosterone decline was also associated after 3 months of therapy and 9 months later. Eight- with less careful performance on a vigilance task. een healthy control men (mean age 65.4 years) com- Furthermore, an association was found between pleted the cognitive assessments at the same times. testosterone decline and improvement in object The tests covered a broad range of cognitive func- recall. Estradiol decline during androgen depriva- tions. The results did not show an overall group tion therapy was associated with a decline in visual effect of the treatment, but revealed that 47% of the memory and recognition speed at 6 months of ther- patients, versus 17% of the controls, showed cog- apy, and an improvement in verbal fluency after nitive decline in at least one task after 3 months 12 months of therapy. of treatment. The decline was most often in the Two studies evaluated the influence of combined domains of spatial memory and spatial ability. Nine androgen deprivation therapy on cognitive func- months later the proportions of subjects with cogni- tioning. Cherrier et al. (2003) investigated patients tive decline did not differ between both groups (34% (n = 19) who received as initial therapy flutamide versus 28%). (an anti-androgen; 250 mg three times daily). After Two studies evaluated the influence of estro- 2 weeks, monthly injections of 7.5 mg leuprolide gen on cognitive functioning in prostate cancer (an LHRH agonist) were administered for 9 months patients. Beer et al. (2006) analyzed working mem- in addition to flutamide. They used a healthy ory and long-term memory in patients with andro- control (n = 15, mean age of patients and controls gen independent prostate cancer (n = 18); assess- together: 65 years). After 9 months, the therapy was, ments took place before the start of second-line depending on PSA levels, continued or discontin- hormonal therapy with transdermal estradiol and ued and re-initiated later. Patients were tested twice 4 weeks later. The same assessments were per- before the start of treatment, after 9 months of treat- formed in an age-matched patient group undergo- ment and 3 months after discontinuation of treating continued androgen deprivation therapy (n = ment. They concluded that 9 months of androgen 18) and a control group consisting of age-matched deprivation therapy resulted in a detrimental effect healthy community dwelling men (n = 17). The on one measure of spatial ability (mental rotation authors concluded that patients on hormonal ther- test). Discontinuation of the therapy resulted in a apy performed worse on measures of verbal mem- beneficial effect on verbal memory. Almeida et al. ory and information processing speed compared (2004) assessed 40 patients (mean age 72.4 years) to healthy controls, but not on measures of work- twice before the start of treatment and four times ing memory. Patients who received estradiol ther- during combined androgen deprivation treatment. apy improved their verbal memory performance After discontinuation, of treatment, patients were compared to baseline. Taxel et al. (2004) examined re-assessed another three times. Their results sug- the influence of estrogen on cognitive function in gest that not initiation, but discontinuation, of treat- prostate cancer patients receiving LHRH agonists. ment was associated with significant improvements Patients treated with LHRH agonists were random- in cognitive performance, particularly on a cogni- ized between additional estradiol treatment (n = 13) tive screening test and a verbal memory test. Visuo- and placebo (n = 10). Cognitive assessments took spatial abilities were not influenced by introduction place before the start of the estradiol therapy and 9 or discontinuation of treatment. weeks later. In this study, no clear effect of the estro- Jenkins et al. (2005) studied the relationship gen therapy on cognitive functioning was found: on between temporary hormonal therapy (3–5 months) only 2 out of 17 cognitive measures (on information- and cognitive performance in a neoadjuvant set- processing speed and executive function) did the 136 Section 2. Effects of cancer and cancer treatment on cognition

estradiol-treated patients show a statistically signifi- ual changes over time associated with a particu- cant increase in performance compared to placebo- lar treatment. The studies on prostate cancer do treated patients. not have this shortcoming: all include a baseline measurement. However, prospective studies con- tend with other methodological problems, such as Methodological aspects (possibly selective) loss of subjects to follow-up. Furthermore, repeated administrations of neu- Although the conducted studies give some evi- ropsychological tests can yield what appears to be dence that hormonal therapy in the treatment of improvement in cognitive function due to practice cancer impacts cognitive functioning, many ques- that is not a true change in cognitive status. In gen- tions about the character, extent, and reversibility eral, an adequate method to correct for practice of the effects remain. Furthermore, the mechanisms effects was not used in the reviewed prospective of action of hormones and hormonal therapy on studies. In fact, stable test scores in a patient group brain structures are complex and have only been when improvement is to be expected could actually partially elucidated. In addition, the neuropsycho- reflect a decrease in cognitive functioning. Please logical studies on this topic show methodological refer to Chapter 23 for a more extensive description weaknesses that probably cause under- or overes- of test characteristics that are important to consider timation of the reported effects. The following are in longitudinal trials. Finally, the value of the base- several methodological issues. line measurement should not be overestimated: at the time of baseline measurement, patients often have to cope with the diagnosis of cancer and may Sample size undergo other treatments as well, such as surgery or Most of the studies had small sample sizes. Of 11 radiotherapy. These factors themselves could influ- studies, 9 had sample sizes of only 10–37 patients ence cognitive performance. (Almeida et al., 2004; Cherrier et al., 2003; Eberling et al., 2004; Ernst et al., 2002; Green et al., 2002; Selection of cognitive measures Jenkins et al., 2005; Salminen et al., 2003; Salminen et al., 2004; Salminen et al., 2005). If a healthy con- The neuropsychological tests and the number of trol group was used, sample sizes were also small tests used vary highly between the studies. One (15–35 subjects), and sometimes the control sub- study only used three tests (Paganini-Hill & Clark, jects differed from the patient groups in terms of age 2000), while others used an extensive battery of and intelligence. It is likely that small study popu- tests. Often no rationale for the test selection was lations make it difficult to detect significant differ- given. Since little is known about the influence of ences in test scores. However, several small but well- hormonal treatments on cognitive functioning, it is designed studies (for example Eberling et al., 2004) important to use a battery of tests that cover a broad found significant differences even in small groups range of cognitive functions so as not to miss pos- of subjects, supporting the hypothesis that effects of sible cognitive effects through shortcomings in test hormonal therapy on cognitive function go beyond selection. From the literature, there is evidence that normal variations in test scores. reproductive hormones have an impact on distinct cognitive domains; estrogens particularly impact verbal memory, while androgens possibly influence Cross-sectional versus prospective studies visuospatial functions in particular. Therefore, it is Studies of breast cancer are all cross-sectional and remarkable that in the studies with breast cancer lack a baseline measurement. As a consequence, patients only one out of four studies included a the results of these studies do not reflect individ- verbal memory test (Shilling et al., 2003). Another Chapter 9. Effect of hormones and hormonal treatment on cognition 137

problem arises when tests are categorized under vidual variations in test scores. A subgroup of different cognitive domains. A specific cognitive patients with deviant test scores, or a subgroup domain can be represented by a variety of tests that whose scores improve or decline more than those can vary highly in degree of complexity and that may of others may not be detected by comparing means. not be assessing exactly the same cognitive ability. Only two studies used a method of determining This leads to disparity in the findings and, as a con- individual decline (Green et al., 2002; Jenkins et al., sequence, confusion about the role of the hormonal 2005), and showed clearly that more patients in the therapy (Jenkins et al., 2005). active treatment groups suffered from significant decreases on one or more test scores compared to a non-treatment patient group or a non-cancer con- Self-reported cognitive complaints trol group. Neuropsychological test scores frequently lack a clear association with self-reported cognitive problems in daily life. Most of the studies described Conclusions above lack data on self-reported cognitive problems. Therefore, the question arises as to the extent to Although many questions remain unanswered which the neuropsychological test scores actually regarding the influence of reproductive hormones reflect cognitive problems that patients have to deal on cognitive functioning, there is increasing evi- with in daily life. dence that reproductive hormones, and therapies that act on these hormones, can have a rather modest effect on cognitive functioning. In healthy men Anxiety, depression, fatigue, and and women, the most consistent results come from psychosocial distress studies on natural fluctuations and, in addition, on It is known that mood problems, anxiety, fatigue, surgically induced menopause in women. These and psychosocial distress can have detrimental studies suggest that higher estrogen levels are asso- effects on cognitive test performance. In addition, ciated with better performance on verbal memory cancer patients are more likely to suffer from these tasks. Testosterone levels are mostly found to influ- symptoms than healthy controls. Moreover, hor- ence visuospatial abilities, in a way that suggests monal therapy itself possibly influences mood, psy- an optimal testosterone level, situated in the “lower chosocial distress, and fatigue (Almeida et al., 2004; male range.” In any case, the relationships between Herr & O’Sullivan, 2000). For a good understand- reproductive hormones and cognitive functioning ing of associations between psychosocial factors, are complex and, to date, only partially elucidated. mood factors, and cognitive performance, a thor- Although many neuropsychological studies ough assessment and description of these different have been conducted on hormone substitution in factors and their mutual dependencies is required healthy persons, the results of these studies are far (Schagen, 2002). from conclusive. The results of these supplementation studies vary from a beneficial effect or no effect at all, to a detrimental effect on cognitive func- Comparisons of group means versus analysis tioning. Neuroimaging studies show that estrogens of test scores of individuals and testosterone are capable of changing brain- In most studies, mean test scores of groups are com- activation patterns, but the clinical significance of pared (Almeida et al., 2004; Cherrier et al., 2003; these results is unknown. Eberling et al., 2004; Ernst et al., 2002; Paganini-Hill Many questions remain unanswered about the & Clark, 2000; Salminen et al., 2003; Shilling et al., possible influence of hormonal therapies in the 2003). Comparing group means can obscure indi- treatment of cancer on cognitive functioning. In 138 Section 2. Effects of cancer and cancer treatment on cognition

breast cancer studies, three out of four studies REFERENCES suggest a slightly detrimental effect on cognitive functioning. On the basis of these studies, it is Almeida OP, Waterreus A, Spry N et al. (2004). One year impossible to draw conclusions about the cogni- follow-up study of the association between chemical tive domains that are most vulnerable to the effects castration, sex hormones, beta-amyloid, memory and depression in men. Psychoneuroendocrinology 29: 1071– of the hormonal agents. In prostate cancer studies, 1081. mixed results were found without a clear and under- Azad N, Pitale S, Barnes WE et al. (2003). Testosterone treat- standable pattern. Detrimental as well as beneficial ment enhances regional brain perfusion in hypogonadal effects were found. men. J Clin Endocrinol Metab 88: 3064–3068. Most of the studies showed methodological short- Barrett-Connor E, Kritz-Silverstein D (1993). Estrogen comings, among others small sample sizes, a lack of replacement therapy and cognitive function in older baseline measurements, a lack of control for prac- women. J Am Med Assoc 269: 2637–2641. tice effects, inadequate selection of cognitive tests, Beer TM, Bland LB, Bussiere JR et al. (2006). Testosterone inadequate determination of cognitive decline, and loss and estradiol administration modify memory in a lack of data on mood, anxiety, fatigue, and psy- men. JUrol175: 130–135. chosocial distress. Bender CM, Paraska KK, Sereika SM et al. (2001). Cognitive function and reproductive hormones in adjuvant ther- The role of hormonal therapy in cancer treat- apy for breast cancer: a critical review. J Pain Symptom ment is increasing, and the medical grounds for pre- Manag 21: 407–424. scribing this kind of treatment are expanding. As a Benson JR (2002). Re: The effects of tamoxifen and estrogen consequence, increasing numbers of, often elderly, on brain metabolism in elderly women. J Natl Cancer Inst patients use hormonal therapy. Because intact cog- 94: 1336–1337. nitive functioning is essential for independent liv- Bowen RL, Martins RN, Gregory CW et al. (2005). Dementia ing and activities in daily life, it is important that the and testosterone levels in men. J Am Med Assoc 293: 551. effects on cognition of the various hormonal agents Chen D, Wu CF, Shi B et al. (2002a). Tamoxifen and and treatment regimens are included in long-term toremifene cause impairment of learning and mem- safety and quality-of-life studies. Large-scale studies ory function in mice. Pharmacol Biochem Behav 71: that use appropriate controls and that include mea- 269–276. Chen D, Wu CF, Shi B et al. (2002b). Tamoxifen and sures of symptoms of depression, anxiety, psychoso- toremifene impair retrieval, but not acquisition, of spa- cial distress, and fatigue are needed. Such stud- tial information processing in mice. Pharmacol Biochem ies should, among others, include verbal memory Behav 72: 417–421. and spatial ability tasks, because these cognitive Cherrier MM, Asthana S, Plymate S et al. (2001). Testos- domains are probably the most vulnerable to hor- terone supplementation improves spatial and verbal monal effects. Furthermore, research that addresses memory in healthy older men. Neurology 57: 80–88. mechanisms that might explain the results from Cherrier MM, Rose AL, Higano C (2003). The effects of neuropsychological studies is needed. Finally, prob- combined androgen blockade on cognitive function dur- ably the most important issue that needs attention ing the first cycle of intermittent androgen suppres- is the experience of patients who use the various sion in patients with prostate cancer. JUrol170: 1808– hormonal agents. After all, little is known about their 1811. Cholerton B, Gleason CE, Baker LD et al. (2002). Estrogen cognitive complaints during therapy or the relation- and Alzheimer’s disease: the story so far. Drugs Aging 19: ship between their complaints and test scores, psy- 405–427. chosocial features, and health characteristics. Infor- Cook IA, Morgan ML, Dunkin JJ et al. (2002). Estrogen mation from these studies can be used to make replacement therapy is associated with less progression patients and clinicians aware of any potentially of subclinical structural brain disease in normal elderly harmful cognitive side-effects that have to be bal- women: a pilot study. Int J Geriatr Psychiatry 17: 610– anced against benefits of the hormonal treatments. 618. Chapter 9. Effect of hormones and hormonal treatment on cognition 139

Drake EB, Henderson VW, Stanczyk FZ et al. (2000). Asso- Hellerstedt BA, Pienta KJ (2002). The current state of hor- ciations between circulating sex steroid hormones and monal therapy for prostate cancer. CA Cancer J Clin 52: cognition in normal elderly women. Neurology 54: 599– 154–179. 603. Henderson VW, Guthrie JR, Dudley EC et al. (2003). Estro- Eberling JL, Reed BR, Coleman JE et al. (2000). Effect gen exposures and memory at midlife: a population- of estrogen on cerebral glucose metabolism in post- based study of women. Neurology 60: 1369–1371. menopausal women. Neurology 55: 875–877. Herr HW, O’Sullivan M (2000). Quality of life of asymp- Eberling JL, Wu C, Tong-Turnbeaugh R et al. (2004). tomatic men with nonmetastatic prostate cancer on Estrogen- and tamoxifen-associated effects on brain androgen deprivation therapy. JUrol163: 1743–1746. structure and function. Neuroimage 21: 364–371. Hogervorst E, Williams J, Budge M et al. (2001). Serum total Ernst T, Chang L, Cooray D et al. (2002). The effects of testosterone is lower in men with Alzheimer’s disease. tamoxifen and estrogen on brain metabolism in elderly Neuroendocrinol Lett 22: 163–168. women. J Natl Cancer Inst 94: 592–597. Hogervorst E, De Jager C, Budge M et al. (2004). Serum Espeland MA, Rapp SR, Shumaker SA et al. (2004). Con- levels of estradiol and testosterone and performance in jugated equine estrogens and global cognitive function different cognitive domains in healthy elderly men and in postmenopausal women: Women’s Health Initiative women. Psychoneuroendocrinology 29: 405–421. Memory Study. J Am Med Assoc 291: 2959–2968. Janowsky JS, Chavez B, Orwoll E (2000). Sex steroids modify Fonda SJ, Bertrand R, O’Donnell A et al. (2005). Age, working memory. J Cogn Neurosci 12: 407–414. hormones, and cognitive functioning among middle- Jenkins VA, Bloomﬁeld DJ, Shilling VM et al. (2005). Does aged and elderly men: cross-sectional evidence from the neoadjuvant hormone therapy for early prostate cancer Massachusetts Male Aging Study. J Gerontol A Biol Sci affect cognition? Results from a pilot study. BJU Int 96: Med Sci 60: 385–390. 48–53. Ganz PA, Castellon SA, Silverman DH (2002). Estrogen, Juul A, Skakkebaek NE (2002). Androgens and the ageing tamoxifen, and the brain. J Natl Cancer Inst 94: 547–549. male. Hum Reprod Update 8: 423–433. Gleason CE, Cholerton B, Carlsson CM et al. (2005). Neuro- Kenny AM, Bellantonio S, Gruman CA et al. (2002). Effects protective effects of female sex steroids in humans: cur- of transdermal testosterone on cognitive function and rent controversies and future directions. Cell Mol Life Sci health perception in older men with low bioavailable 62: 299–312. testosterone levels. J Gerontol A Biol Sci Med Sci 57: Goss PE, Strasser K (2001). Aromatase inhibitors in the M321–M325. treatment and prevention of breast cancer. J Clin Oncol Luoto R, Manolio T, Meilahn E et al. (2000). Estrogen 19: 881–894. replacement therapy and MRI-demonstrated cerebral Green HJ, Pakenham KI, Gardiner RA (2000). Effects of infarcts, white matter changes, and brain atrophy in luteinizing hormone releasing hormone analogs on cog- older women: the Cardiovascular Health Study. JAm nition in women and men: a review. Psychol Health Med Geriatr Soc 48: 467–472. 5: 407–418. Maki PM (2004). Hormone therapy and risk for dementia: Green HJ, Pakenham KI, Headley BC et al. (2002). Altered where do we go from here? Gynecol Endocrinol 19: 354– cognitive function in men treated for prostate cancer 359. with luteinizing hormone-releasing hormone analogues Maki PM, Resnick SM (2000). Longitudinal effects of estro- and cyproterone acetate: a randomized controlled trial. gen replacement therapy on PET cerebral blood ﬂow and BJU Int 90: 427–432. cognition. Neurobiol Aging 21: 373–383. Hampson E (1990). Variations in sex-related cognitive abil- Maki PM, Rich JB, Rosenbaum RS (2002). Implicit mem- ities across the menstrual cycle. Brain Cogn 14: 26–43. ory varies across the menstrual cycle: estrogen effects in Haren MT, Wittert GA, Chapman IM et al. (2005). Effect of young women. Neuropsychologia 40: 518–529. oral testosterone undecanoate on visuospatial cognition, Meyer PM, Powell LH, Wilson RS et al. (2003). A population- mood and quality of life in elderly men with low-normal based longitudinal study of cognitive functioning in the gonadal status. Maturitas 50: 124–133. menopausal transition. Neurology 61: 801–806. Hausmann M, Slabbekoorn D, Van Goozen SH et al. (2000). Miller WR (1996). Estrogens and Endocrine Therapy for Sex hormones affect spatial abilities during the men- Breast Cancer. Estrogen and Breast Cancer. Heidelberg: strual cycle. Behav Neurosci 114: 1245–1250. Springer Verlag. 140 Section 2. Effects of cancer and cancer treatment on cognition

Moffat SD, Zonderman AB, Metter EJ et al. (2004). Free Salminen E, Portin R, Korpela J et al. (2003). Androgen testosterone and risk for Alzheimer disease in older men. deprivation and cognition in prostate cancer. Br J Can- Neurology 62: 188–193. cer 89: 971–976. Muller M, Aleman A, Grobbee DE et al. (2005). Endogenous Salminen EK, Portin RI, Koskinen A et al. (2004). Asso- sex hormone levels and cognitive function in aging men: ciations between serum testosterone fall and cognitive is there an optimal level? Neurology 64: 866–871. function in prostate cancer patients. Clin Cancer Res 10: Newton C, Slota D, Yuzpe AA et al. (1996). Memory 7575–7582. complaints associated with the use of gonadotropin- Salminen EK, Portin RI, Koskinen AI et al. (2005). Estradiol releasing hormone agonists: a preliminary study. Fertil and cognition during androgen deprivation in men with Steril 65: 1253–1255. prostate carcinoma. Cancer 103: 1381–1387. Nickelsen T, Lufkin EG, Riggs BL et al. (1999). Raloxifene Sanders G, Sjodin M, de Chastelaine M (2002). On the elu- hydrochloride, a selective estrogen receptor modula- sive nature of sex differences in cognition: hormonal tor: safety assessment of effects on cognitive function influences contributing to within-sex variation. Arch Sex and mood in postmenopausal women. Psychoneuroen- Behav 31: 145–152. docrinology 24: 115–128. Schagen SB (2002). Cognitive function following Norbury R, Cutter WJ, Compton J et al. (2003). The neu- chemotherapy: a study in breast cancer patients. roprotective effects of estrogen on the aging brain. Exp Ph.D Thesis. University of Amsterdam. Gerontol 38: 109–117. Schmidt R, Fazekas F, Reinhart B et al. (1996). Estrogen Nystedt M, Berglund G, Bolund C et al. (2000). Random- replacement therapy in older women: a neuropsycholog- ized trial of adjuvant tamoxifen and/or goserelin in ical and brain MRI study. J Am Geriatr Soc 44: 1307–1313. premenopausal breast cancer – self-rated physiological Sharifi N, Gulley JL, Dahut WL (2005). Androgen depriva- effects and symptoms. Acta Oncol 39: 959–968. tion therapy for prostate cancer. J Am Med Assoc 294: O’Connor DB, Archer J, Hair WM et al. (2001). Activational 238–244. effects of testosterone on cognitive function in men. Shaywitz SE, Shaywitz BA, Pugh KR et al. (1999). Effect of Neuropsychologia 39: 1385–1394. estrogen on brain activation patterns in postmenopausal Osborne CK, Zhao H, Fuqua SA (2000). Selective estro- women during working memory tasks. J Am Med Assoc gen receptor modulators: structure, function, and clini- 281: 1197–1202. cal use. J Clin Oncol 18: 3172–3186. Sherwin BB (1988). Estrogen and/or androgen replace- Owens JF, Matthews KA, Everson SA (2002). Cognitive func- ment therapy and cognitive functioning in surgi- tion effects of suppressing ovarian hormones in young cally menopausal women. Psychoneuroendocrinology women. Menopause 9: 227–235. 13: 345–357. Paganini-Hill A, Clark LJ (2000). Preliminary assessment of Sherwin BB (2000). Oestrogen and cognitive function cognitive function in breast cancer patients treated with throughout the female lifespan. Novartis Found Symp tamoxifen. Breast Cancer Res Treat 64: 165–176. 230: 188–196. Phillips SM, Sherwin BB (1992). Effects of estrogen on Sherwin BB (2003). Steroid hormones and cognitive func- memory function in surgically menopausal women. Psy- tioning in aging men: a mini-review. J Mol Neurosci 20: choneuroendocrinology 17: 485–495. 385–393. Rapp SR, Espeland MA, Shumaker SA et al. (2003). Effect of Sherwin BB (2005). Estrogen and memory in women: how estrogen plus progestin on global cognitive function in can we reconcile the findings? Horm Behav 47: 371–375. postmenopausal women: the Women’s Health Initiative Sherwin BB, Tulandi T (1996). “Add-back” estrogen reverses Memory Study: a randomized controlled trial. JAmMed cognitive deficits induced by a gonadotropin-releasing Assoc 289: 2663–2672. hormone agonist in women with leiomyomata uteri. Resnick SM, Maki PM, Golski S et al. (1998). Effects of estro- J Clin Endocrinol Metab 81: 2545–2549. gen replacement therapy on PET cerebral blood flow and Shilling V, Jenkins V, Fallowfield L et al. (2003). The effects neuropsychological performance. Horm Behav 34: 171– of hormone therapy on cognition in breast cancer. 182. J Steroid Biochem Mol Biol 86: 405–412. RosarioER,ChangL,StanczykFZet al. (2004). Age- Shumaker SA, Legault C, Rapp SR et al. (2003). Estrogen related testosterone depletion and the development of plus progestin and the incidence of dementia and mild Alzheimer disease. J Am Med Assoc 292: 1431–1432. cognitive impairment in postmenopausal women: the Chapter 9. Effect of hormones and hormonal treatment on cognition 141

Women’s Health Initiative Memory Study: a randomized Visvanathan K, Davidson NE (2003). Aromatase inhibitors controlled trial. J Am Med Assoc 289: 2651–2662. as adjuvant therapy in breast cancer. Oncology (Hunt- Shumaker SA, Legault C, Kuller L et al. (2004). Conjugated ingt) 17: 335–342, 347. equine estrogens and incidence of probable demen- Wolf OT, Kirschbaum C (2002). Endogenous estradiol and tia and mild cognitive impairment in postmenopausal testosterone levels are associated with cognitive perfor- women: Women’s Health Initiative Memory Study. JAm mance in older women and men. Horm Behav 41: 259– Med Assoc 291: 2947–2958. 266. Simpson ER, Dowsett M (2002). Aromatase and its Wolf OT, Neumann O, Hellhammer DH et al. (1997). Effects inhibitors: signiﬁcance for breast cancer therapy. Recent of a two-week physiological dehydroepiandrosterone Prog Horm Res 57: 317–338. substitution on cognitive performance and well-being in Tannock IF, Ahles TA, Ganz PA et al. (2004). Cogni- healthy elderly women and men. J Clin Endocrinol Metab tive impairment associated with chemotherapy for can- 82: 2363–2367. cer: report of a workshop. J Clin Oncol 22: 2233– Yaffe K, Sawaya G, Lieberburg I et al. (1998). Estrogen 2239. therapy in postmenopausal women: effects on cognitive Taxel P, Stevens MC, Trahiotis M et al. (2004). The effect function and dementia. J Am Med Assoc 279: 688–695. of short-term estradiol therapy on cognitive function Yaffe K, Krueger K, Sarkar S et al. (2001). Cognitive func- in older men receiving hormonal suppression ther- tion in postmenopausal women treated with raloxifene. apy for prostate cancer. J Am Geriatr Soc 52: 269– NEnglJMed344: 1207–1213. 273. Yaffe K, Lui LY, Zmuda J et al. (2002). Sex hormones and Tripathy D, Benz CC (2001). Hormones and cancer. cognitive function in older men. J Am Geriatr Soc 50: 707– In Greenspan FS, Gardner DG (eds.) Basic and 712. Clinical Endocrinology. New York: Lange Medical Yaffe K, Krueger K, Cummings SR et al. (2005). Effect Books. of raloxifene on prevention of dementia and cognitive Verghese J, Kuslansky G, Katz MJ et al. (2000). Cognitive impairment in older women: the Multiple Outcomes of performance in surgically menopausal women on estro- Raloxifene Evaluation (MORE) randomized trial. Am J gen. Neurology 55: 872–874. Psychiatry 162: 683–690. 10

Low-grade gliomas

Martin J. B. Taphoorn and Charles G. Niel¨

Introduction Epidemiology and biology, pathology and genetics, clinical and imaging features, prognostic factors in LGG Low-grade gliomas (LGG) are diffusely infiltrating primary tumors of the cerebral hemispheres, and Epidemiology and biology originate from glial tissue (Kleihues & Cavanee, 2000). Patients with these tumors, like any patient The percentage of low-grade tumors amongst with a brain disease, may experience cognitive com- gliomas, the most common primary brain tumor, plaints and have cognitive deficits on examination. ranges between 15% and 20% (Kleihues & Cavanee, In LGG patients, who usually have a paucity of neu- 2000). The incidence of gliomas in adults is 5 to rological deficits, these cognitive complaints and 7 per 100 000 (Bondy & Wrensch, 1996). This fig- deficits may be particularly prominent, in contrast ure has remained stable for many years, unlike that to patients with high-grade gliomas (HGG). In HGG of other brain tumors such as primary central ner- patients, the rapidly growing tumor typically gives vous system (CNS) lymphoma, which is increasing rise to hemiparesis or increased intracranial pres- in incidence. Low-grade glioma typically develops sure, which may overshadow more subtle cognitive in young adults. There is a slight preponderance in deficits (Ashby & Shapiro, 2004; Rees, 2002). More- men (male:female ratio 1.3:1). Only the diffusely over, LGG patients have a relatively good prognosis growing LGG of the cerebral hemispheres in adults with median survival rates ranging from 5 to more will be discussed here. The pilocytic astrocytoma than 15 years. Long-term-surviving LGG patients (mainly occurring in children in the posterior fossa), run the risk of late toxicity of treatment. Tumor and and the optic nerve glioma, both regarded as treatment effects may impair cognitive functioning LGG, should be treated differently from the hemi- in these patients during the course of their disease spheric LGG in adults, and they have a differ- and have a deleterious impact on the quality of life ent prognosis (Kleihues & Cavanee, 2000). Also, of the patient and their family. relatively rare tumors such as the gangliocytoma

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

142 Chapter 10. Low-grade gliomas 143

and the dysembryoplastic neuro-ectodermal tumor 80% of all gliomas. Histological variants of astro- (DNET), considered as belonging to the group of cytomas are the fibrillary astrocytoma, the gemis- LGG, will not be discussed here due to their dif- tocytic astrocytoma, and the protoplasmatic astro- ferent behavior and prognosis (Rees, 2002). The cytoma. The gemistocytic variant is particularly same holds true for brainstem astrocytomas and prone to dedifferentiation into a HGG. Subse- gliomatosis cerebri, which both may show low- quently, a distinction is made between low-grade grade features on histopathological examination and high-grade tumors. The absence of high-grade (Ashby & Shapiro, 2004). features such as mitoses, necrosis, nuclear atypia, In contrast to HGG, which are rapidly growing and microvascular proliferation implies a low-grade tumors leading to neurological deficit and increased tumor. In the current World Health Organization intracranial pressure, LGG in the adult may remain (WHO) classification the formerly applied grading silent for a long time (Ashby & Shapiro, 2004; Rees, system (grades I–IV for astrocytic tumors and grades 2002). Still, LGG are not benign tumors; by far the A–D for oligodendroglial tumors) has been aban- majority of LGG patients will, in due time, die from doned. A LGG is thus denominated as “astrocy- progressive tumor growth after dedifferentiation to toma,” “oligodendroglioma” or “oligo-astrocytoma” a high-grade tumor. Although there is increasing (Kleihues & Cavanee, 2000). knowledge on molecular tumor biology in general Although the histopathology of gliomas has prog- and on genetic abnormalities in all kinds of tumors, nostic significance, this analysis is by definition sub- including gliomas, the etiology of gliomas remains jective because it is based on visual criteria. More- still unknown (Rasheed et al., 1999). Environmen- over, due to the heterogeneous nature of gliomas, tal factors, such as food substances, alcohol, coffee, high-grade features may be missed on histopatho- smoking, and the use of cellular phones, have been logical examination, especially in small (stereotac- studied, with no or only a minor increased risk for tic) biopsy specimens (“sampling error”) (Jackson glioma revealed (Christensen et al., 2005; Efird et al., et al., 2001). Therefore, the definite diagnosis should 2004; Hardell et al., 2007; Huncharek et al., 2003; always be based on the combination of histopatho- Rasheed et al., 1999; Takebayashi et al., 2008). The logical, clinical, and radiological features. only known risk factor for gliomas is radiation ther- It is believed that genetic analysis of gliomas, next apy in the past (Salvati et al., 1991). Many years fol- to histopathological examination, will play a major lowing therapeutic irradiation of the head there is a role in future neuro-oncology (Fuller et al., 2002; small chance that a glioma may occur. Godard et al., 2003; Idbaih et al., 2007; Iwadate et al., Gliomas are more likely to arise in patients with 2004). As we have not in the past been able to define genetic tumor syndromes such as neurofibromato- subsets of (low-grade) gliomas that are responsive to sis, and the rarely occurring Turcot syndrome and specific treatments, the value of these therapies may Li-Fraumeni syndrome (Kleihues & Cavanee, 2000). have been underestimated. Also, siblings of glioma patients without any known In LGG, frequent mutations of the tumor sup- genetic syndrome have an increased risk for gliomas pression gene p53 have been known to exist for a (Hemminki & Li, 2003). long time and are found in approximately 75% of cases. Other genetic changes include a gain of chromosome 7q, amplification of 8q, loss of heterozy- Pathology and genetics gosity (LOH) on 10p and 22q, and overexpression Based on histopathological examination, which is of platelet-derived growth factor receptor (PDGFR) the gold standard for the diagnosis of LGG (Kleihues (Kleihues & Cavanee, 2000). As these tumors dedif- & Cavanee, 2000), the diffusely growing gliomas ferentiate to high-grade tumors many other muta- are classified as astrocytomas, oligodendrogliomas tions will occur, reflecting the aggressive behavior or mixed gliomas. Astrocytomas account for about of these tumors (Rees, 2002). This genetic pathway 144 Section 2. Effects of cancer and cancer treatment on cognition

is quite different from that of tumors that are high ing have their limitations in determining whether grade from the start, such as the de novo glioblas- a lesion is an LGG. An arachnoid cyst or an infarc- toma multiforme, in which p53 mutations are very tion with atypical clinical presentation may be mis- rare, but where epidermal growth factor receptor taken for an LGG. Also, a ganglioglioma and a DNET, (EGFR) overexpression and PTEN (phosphatase and which are rare tumors in adults and are most fre- tensin homolog deleted on chromosome 10) muta- quently located in the (medial) temporal lobe, may tion is frequently observed (Kleihues & Cavanee, have features similar to LGG. Still, these tumors are 2000). Only recently has a link been made between more sharply demarcated from surrounding tissue specific genetic abnormalities in anaplastic oligo- than LGG and a ganglioglioma may enhance with dendroglioma (allelic loss of chromosomes 1p and contrast agent. More importantly, the distinction 19q), the response to chemotherapy, and survival between a high-grade and a low-grade tumor may (Cairncross et al., 1998). Also in LGG, both astro- be difficult: histopathological examination of about cytoma and oligodendroglioma types, genetic aber- 35% of “typical” LGG on imaging reveals high-grade rations of chromosomes 1p and 19q have been features, with the highest probability in patients described (Hirose et al., 2003; Kaloshi et al., 2007; overage40(Ginsberget al., 1998; Scott et al., 2002). Smith et al., 2000; Watanabe et al., 2002). More recently developed imaging modalities such as diffusion-weighted imaging (DWI), MR spectroscopy (MRS), and functional positron emis- Clinical and imaging features sion tomography (PET) have been unable to clearly The majority of LGG patients present with one or improve the specificity of standard MRI (Herholz more (focal) seizures but have no other symptoms et al., 1998; Knopp et al., 1999; Kono et al., 2001; (Ashby & Shapiro, 2004; Rees, 2002). Typically, no Vuori et al., 2004), although perfusion-weighted MR abnormalities are found on neurological examina- imaging (PWI) might be of help in differentiating tion; the features of a rapidly growing tumor such as low-grade from high-grade tumors (Maia et al., a HGG leading to neurological deficit and increased 2005). intracranial pressure are absent. Even so, many patients with LGG appear to have problems with Prognostic factors in LGG cognitive functioning at presentation (Pahlson et al., 2003; Taphoorn & Klein, 2004). In some patients the The natural history of an LGG is unpredictable. tumor may be detected by chance, as imaging was Some patients may remain free from clinical and performed for other reasons (e.g., head trauma, ten- radiological signs of tumor progression for several sion headache, dizziness). decades, whereas in others progression and de- On imaging, LGG are hypodense lesions on differentiation to a high-grade tumor occur within computed tomography (CT) with relatively little weeks to months from the initial presentation mass effect and no contrast enhancement. On (Ashby & Shapiro, 2004; Rees, 2002). Tumor pro- magnetic resonance imaging (MRI), T2-weighted gression results in an increased incidence of epilep- sequences or fluid-attenuated inversion recovery tic seizures, neurological deficit, and/or increased (FLAIR) images demonstrate a diffusely infiltrat- intracranial pressure. The majority of LGG patients ing high-signal lesion with slight mass effect (Fig- will die from their disease sooner or later, unless ure 10.1). There is no contrast enhancement of the intercurrent diseases occur. The median survival low-signal lesion on T1-weighted sequences (Recht of LGG patients ranges from 5 to 10 years accord- et al., 1992). Compared with CT, MRI is more sen- ing to a large number of studies (Ashby & Shapiro, sitive in detecting LGG. The presence of calcifica- 2004; Johannesen et al., 2003; Rees, 2002). This wide tions in the lesion suggests an oligodendroglioma range may be explained by the earlier diagnosis rather than an astrocytoma. Both CT and MR imag- of LGG with MRI currently compared to CT. Chapter 10. Low-grade gliomas 145

(a) (b)

Figure 10.1. A 37-year-old male, presenting with focal seizures. (a) On MRI (FLAIR image): left temporal lesion, no surrounding edema, no mass effect; (b) T1-weighted image after gadolinium: no contrast enhancement

One study of oligodendroglioma even indicated a loss of chromosome 1p/19q (Brown et al., 2004; median survival of more than 15 years (Olson et al., Kaloshi et al., 2007; Komine et al., 2003; Levin et al., 2000). 2006). Prognostic factors in LGG are age (patients over the age of 35–40 years have a worse prognosis than younger patients) and tumor histology (oligo- Therapeutic management dendroglioma has a better prognosis compared to Surgery astrocytoma) (Leighton et al., 1997; Vecht, 1993). Also, LGG patients who present solely with epilepsy Surgery in LGG may have two goals: a tissue biopsy have a better outlook than those who have neuro- is necessary to make the histopathological diagnosis logical deficit on presentation. Next to these fac- (and to perform molecular genetics); and reduction tors, tumor size and location (i.e., tumor extend- of tumor mass may be intended to relieve neurologi- ing into the other hemisphere) may have prognostic cal symptoms and signs and/or to enhance survival. implications as well, according to the analysis of a Surgery to reduce tumor mass is controversial large sample of LGG patients (Pignatti et al., 2002). in LGG (Ashby & Shapiro, 2004; Dropcho, 2004; Sorting the patients from this sample according Rees, 2002). This especially holds true for surgery to five negative prognostic factors (age >40 years, to improve survival in cases without increased neurological deficit, astrocytic histology, tumor size intracranial pressure. The majority of LGG patients >6 cm, tumor crossing the midline) into a low-risk do not have neurological symptoms and signs that group (0–2 negative factors) and a high-risk group may improve following reduction of tumor mass, (3–5 factors) resulted in a median survival of 7.7 but some do. Examples are patients with cystic years in the favorable group versus 3.2 years in the tumors causing neurological deficit or patients with unfavorable group. medication-resistant epilepsy that may be relieved More recently observed prognostic factors are by tumor resection (“epilepsy-surgery”). cognitive function, the activity of the DNA repair Despite technical developments in surgery, enzyme O-6-methylguanine methyltransferase and such as intraoperative-guided imaging and 146 Section 2. Effects of cancer and cancer treatment on cognition

functional mapping, the infiltrative growth pattern Radiotherapy of LGG prevents surgical cure. A review of the extent of resection as a factor influencing outcome in Based on several retrospective analyses in LGG LGG suggested that there is no proof that surgery patients, (focal) external radiotherapy resulted improves survival (Keles et al., 2001) but a more in improved survival (Cairncross, 2000; Shaw recent review contested this (Sanai and Berger et al., 1989). As other retrospective studies did not 2008). However, a recent retrospective study found demonstrate a survival benefit of radiotherapy, that extent of resection was associated with signif- selection bias was presumed to be a confounding icantly longer overall survival and minimal mor- factor (Cairncross, 2000). Adversaries of radiation bidity, suggesting improved patient outcome with therapy in LGG also point to irreversible long- maximal resection of hemispheric LGG (Smith et al., term side-effects of radiation that may result in 2008). severe cognitive deterioration (Choucair et al., Using intraoperative-guided imaging and func- 1997). tional mapping in patients with LGG, surgery even At present, the controversy on (early) radiation in eloquent brain locations is feasible with very lim- treatment in LGG has been ended to some extent by ited lasting neurological deficits due to the opera- the results of prospective randomized trials in both tion (Duffau, 2003a, 2003b, 2004). Europe and the USA. Two trials comparing high- Since neither radiation therapy nor chemother- dose with low-dose radiation (59.4 versus 45 Gy; apy following surgery has been demonstrated to 64.8 versus 50.4 Gy) did not demonstrate a survival improve survival in the majority of LGG patients, benefit for high-dose radiation (Karim et al., 1996; a so-called wait and see policy is often adopted Shaw et al., 2002). Moreover, based on the results of in LGG patients with favorable prognostic fac- the European trial (EORTC 22844) high-dose radia- tors (Ashby & Shapiro, 2004; Dropcho, 2004; Recht tion resulted in a worse health-related quality of life et al., 1992; Rees, 2002). This conservative pol- (Kiebert et al., 1998). Research on the dose depen- icy is in contrast to the earlier statement that dency of neurotoxicity and of treatment is, however, LGGarenotbenigntumors,andthatsomeare hampered by the constant technical evolution in known to be responsive to treatment due to cer- radiation oncology. The earlier-mentioned prospec- tain genetic features. A “wait and see” policy should tive randomized trials, one starting in 1984 (Karim only be advocated in patients under 40 years et al., 1996) and the other in 1986 (Shaw et al., old who present with epilepsy and have no abnor- 2002), failed to demonstrate better results from a malities on neurogical examination. Also, the typ- higher radiation dose. This might, in part, be caused ical features of an LGG should be found on imag- by suboptimal dose definition compared to current ing. In these cases, even a biopsy may be deferred standards: at the time these studies were performed, (Reijneveld et al., 2001). A second MRI should be dose definition was merely an obligation to describe performed after a 3-month interval to rule out the radiation dose in a uniform way, resulting in good development of a high-grade tumor. In the case of a inter-institutional dose comparability in multisite stable clinical situation and no signs of tumor pro- trials (ICRU report 29, 1979) rather than the clear gression on imaging, the “wait and see” approach and uniform individual dose description that is nec- may be continued with clinical and MRI follow- essary to study dose-related clinical outcome. It was up once or twice a year, until progression occurs. not until 1993 that the International Convention on A slow and slight increase in tumor size is nor- Radiation Units (ICRU) defined radiation dose in an mally observed over time in these patients on crit- unequivocal way. ical measurements of tumor volume (Mandonnet Of even more importance is the EORTC 22845 ran- et al., 2003). This, however, does not necessarily domized trial of over 300 LGG patients, compar- imply dedifferentiation of the tumor. ing early radiotherapy to observation only following Chapter 10. Low-grade gliomas 147

surgery or biopsy. Progression-free survival was sig- potential long-term adverse effects of radiation and nificantly longer in the radiotherapy group than the observation that LGG responds to chemother- in the observation group; however, overall survival apy, randomized studies comparing radiation and was not different as reported in two interim anal- chemotherapy in (progressive) LGG are currently yses (Karim et al., 2002; Van den Bent et al., 2005). running in the USA as well as in Europe. As quality of life was not studied, it is unknown whether clinical deterioration was postponed by early radiotherapy. Based on the results of these Neurocognitive disturbances in LGG trials, early radiotherapy in LGG patients with favor- Neurocognitive deficits in LGG patients can be able prognostic factors has been largely abandoned. caused by the tumor, by tumor-related epilepsy Patients with histologically proven LGG who are (Klein et al., 2003), and treatment (neurosurgery, older than 40 and/or have (progressive) neurolog- radiotherapy, anti-epileptics, chemotherapy, or ical deficit are likely to benefit from early radia- corticosteroids), as well as by psychological distress. tion (Pignatti et al., 2002). Focal radiation is cur- More likely, a combination of these factors will rently applied with total doses ranging from 45 to contribute to neurocognitive dysfunction. Also 60 Gy in fractions of up to 2 Gy. Also, patients with tumor regrowth (either locally or diffuse), lep- medication-refractory epilepsy due to the tumor tomeningeal metastasis, or metabolic disturbances may benefit from radiation of their LGG (Rogers may negatively affect neurocognitive function. et al., 1993).

Chemotherapy Primary tumor as a cause of neurocognitive deficits Systemic chemotherapy is not a primary treatment for LGG patients currently. In line with The tumor itself is an important contributor to the effective chemotherapeutic treatment of high- neurocognitive deficits, which holds true for both grade oligodendroglioma and high-grade oligoas- high-grade and low-grade tumors. Neurocognitive trocytoma with PCV chemotherapy (procarbazine, deficits are a prominent clinical feature in slowly lomustine, also known as CCNUR , and vincristine) growing tumors, such as LGG, or in diffusely infil- or temozolomide, these agents have also been trating tumors, such as primary CNS lymphoma or applied to low-grade oligodendroglial tumors in gliomatosis cerebri. In rapidly growing, high-grade several phase II studies (Buckner et al., 2003; Hoang- tumors focal neurological deficits and high intracra- Xuan et al., 2004; Kaloshi et al., 2007; Levin et al., nial pressure may overshadow more subtle cognitive 2006; Pace et al., 2003; Quinn et al., 2003; Stege deficits. et al., 2005; Van den Bent et al., 1998, 2003). Objec- In many studies of neurocognitive function in tive responses on imaging have been reported but patients with brain tumor, conclusions about the may be difficult to appreciate due to the absence role of the tumor cannot be easily made, because of contrast enhancement of LGG. MR spectroscopy data are only gathered after treatment. In a series of may add to the evaluation of treatment response 139 patients with different brain tumors, neurocog- (Murphy et al., 2004). In a recent study of low-grade nitive disturbances were observed in 91% before oligodendroglial tumors, an objective response was treatment was initiated (Tucha et al., 2000). denoted in 29%–52% of patients. In contrast to the Neurocognitive testing in these studies is often relation between response to chemotherapy and directed more toward the functions of the dom- loss of chromosomes 1p and 19q in high-grade inant hemisphere, therefore it is not surprising oligodendroglioma, this has not been observed in that patients with tumors in the dominant hemi- all studies on LGG (Buckner et al., 2003). Due to the sphere reportedly have more cognitive deficits 148 Section 2. Effects of cancer and cancer treatment on cognition

than those with non-dominant hemisphere lesions nitive function and health-related quality of life in a (Hahn et al., 2003; Taphoorn et al., 1992). case-matched control study (Reijneveld et al., 2001). Unlike stroke patients, who tend to have site- specific deficits, glioma patients have may have more diffuse, milder and variable deficits, which Radiation as the cause of cognitive deficit may be explained by the diffuse growth of tumor cells infiltrating normal brain tissue (Anderson Cranial irradiation can cause several adverse et al., 1990). Additionally, acute neurotransmitter changes in normal brain tissue (Sheline et al., 1980; changes and chronic degeneration of fiber tracts Taphoorn & Klein, 2004). Radiation-induced white caused by damage to certain brain areas may impair matter disease due to demyelination and/or small neuronal responses in remote undamaged cortical vessel damage may result in (severe) cognitive dete- regions (i.e., diaschisis). rioration several months to years following cranial Neurocognitive deficits may also be the first man- radiotherapy (see Chapter 7 for a more extensive ifestation of tumor recurrence, even before struc- discussion of radiation injury to the brain). As tural changes are observed on imaging (Armstrong LGG patients have a relatively good prognosis, this et al., 2003; Meyers & Hess, 2003). long-term complication is much feared (Surma-Aho et al., 2001). The risk of long-term adverse effects of radiotherapy in LGG patients has always been Surgery as the cause of neurocognitive deficit an important issue in the discussion of early versus delayed radiation treatment of these patients. Neurosurgery and peri-operative injuries may cause From a clinical and morphological point of view (transient) neurological deficits due to damage of four types of damage caused by irradiation can be normal surrounding tissue. Many neurosurgeons discerned: acute reaction (transient edema), early are therefore hesitant to operate on patients with delayed reaction, late delayed reaction, and focal brain tumors in eloquent brain areas. According radiation necrosis. The type of damage and its to Scheibel and co-workers (1996) neurosurgery in clinical impact are time dependent. Mainly early patients with glioma leads to focal neurocognitive delayed damage, occurring between 1 and 6 months deficits, in contrast to more diffuse neurocognitive after irradiation, and late delayed damage (after disturbances caused by radiation and chemother- 6 months to several years post-irradiation) have a apy. Recent studies that use intraoperative-guided negative impact on neurocognitive function. The imaging and functional mapping in patients with first is a reversible process of de- and remyelination, LGG in eloquent brain locations show that a high whereas the late delayed damage results in a dif- percentage of them have post-operative neurocog- fuse encephalopathy, of which neurocognitive dis- nitive deficits (Duffau et al., 2003a, 2003b). However, turbances are the hallmark (Armstrong et al., 2002; most of these deficits resolved within 3 months, Behin´ & Delattre, 2003; Vigliani et al., 1996). presumably owing to the plasticity of the normal The pathophysiology of the diffuse encephalopa- brain and recovery from the acute effects of surgery thy following radiotherapy is not completely under- (Duffau et al., 2003a, 2003b; Duffau, 2006). stood. Both vascular structures and glial cells are In line with these data, a study on a large group thought to be the target of radiation damage, as his- of patients with LGG who had biopsy or neurosur- tological examination can reveal both demyelina- gical tumor resection at least 1 year before indicated tion and vascular damage. In the glial hypothesis, that neurosurgery did not contribute to neurocogni- oligodendrocytes are the primary target of radiation tive disability (Klein et al., 2002). By contrast, neuro- damage resulting in demyelination, whereas the surgery in patients with histologically proven versus vascular hypothesis is based on histological proof suspected LGG had a negative effect on neurocog- of blood-vessel dilatation and wall thickening with Chapter 10. Low-grade gliomas 149

hyalinization, endothelial cell loss, and a decrease In contrast, 45 of the 59 non-irradiated patients in vessel density leading to white-matter necrosis. were still alive, indicating that the irradiated group The pathogenesis is probably far more complex, was highly selected. Of the 44 irradiated patients and may also involve effects on neural stem cell alive at the time of assessment, 28 patients had production with resulting hippocampal dysfunction neurocognitive deficits, of whom 19 had received (Monje & Palmer, 2003; Monje et al., 2007). whole-brain irradiation (40 Gy), followed by focal On imaging modalities late delayed encephalop- irradiation up to a total dose of 64–75 Gy. Whole- athy may result in cerebral atrophy and/or diffuse brain irradiation is now obsolete in treating LGG white matter disease. To some extent there is a rela- since when it is applied in 2-Gy fractions followed tion between the abnormalities on MRI and cogni- by a focal boost of irradiation to a higher dose, tive status (Postma et al., 2002). Diffusion tensor MR it results in “neurocognitive disturbances.” In con- imaging may be of additional value in this respect trast to this, Klein et al.(Kleinet al., 2002) ret- (Nagesh et al., 2008). rospectively evaluated the impact of focal radia- The severity of neurocognitive deficits due to late tion on neurocognitive function in LGG patients. delayed radiation encephalopathy ranges from mild In this study, 195 LGG patients were compared to or moderate neurocognitive deficits all the way to 100 patients with a hematological malignancy with dementia. Mild to moderate neurocognitive deficits comparable overall survival, and with 195 healthy result in attention or short-term memory distur- controls. Focal radiation treatment itself was not bances as the main features. Because studies on this found to be a significant factor in causing neurocog- subject vary greatly in the neuropsychological test nitive disturbances as long as the radiation fraction procedures, the populations studied, and the dura- size did not exceed 2 Gy. In contrast, the presence tion of follow-up (Behin´ & Delattre, 2003; Vigliani of the tumor itself increased the risk for neurocog- et al., 1999) the clinical picture and incidence of this nitive impairment compared to the hematologi- complication are hard to define exactly. cal malignancy group. One might expect that with Following cranial irradiation, diffuse white mat- current focal irradiation techniques, such as three- ter changes (leukoencephalopathy on CT and MRI) dimensional conformal radiotherapy and intensity- may occur in as many as 40%–50% of patients modified radiotherapy, the amount of normal brain (Constine et al., 1988). The greater the amount of tissue irradiated will be even less compared to the healthy brain that is exposed to irradiation, the patients from the Klein et al. (2002) study (Niel¨ greater the likelihood of these imaging changes and et al., 2005). Studies looking at neurocognitive out- resulting neurocognitive deficits developing. Focal comes with these newer techniques are not yet irradiation causes fewer imaging and neurocogni- available. tive changes compared to whole-brain irradiation The total radiation dose given is, after volume (Swennen et al., 2004). Therefore, caution should be of treatment and fraction size, the third factor that used in interpreting the results of studies into the influences the amount of neurotoxicity from irradi- impact of cranial irradiation if the irradiated vol- ation. In a study by Corn et al. (1994) the amount of ume is not well defined: focal irradiation is current white matter change (but not neurocognitive func- practice in the treatment of LGG, whereas in the tioning) was analysed in dose escalation studies per- 1980s and 1990s, some of these patients received formed by the Radiation Therapy Oncology Group whole-brain irradiation. These latter patients have (RTOG). Mild white matter changes did not seem to been included in retrospective analyses, such as a occur more frequently with higher radiation doses study of the neurocognitive function of 160 patients than with lower doses, whereas the most severe treated from 1980 to 1992 (Surma-Aho et al., 2001). white matter changes (necrosis) occurred signifi- Of these patients, 101 had received irradiation and cantly more frequently in the highest dose groups. 57 of them had died by the time of the analysis. A focal irradiation dose of 54 Gy resulted in less 150 Section 2. Effects of cancer and cancer treatment on cognition

(a) (b)

Figure 10.2. (a) A 42-year old female, who had been operated on for a left parietal low-grade oligodendroglioma 5 years before. She had focal radiation treatment (54 Gy, fraction size 1.8 Gy) when she had a regrowth of the tumor (T1-weighted image after gadolinium); (b) 2 years following radiation she experienced mental slowing, had no recurrence of tumor on imaging but had extensive white matter disease (T2-weighted image)

than 8% of moderate to severe white matter changes memory was demonstrated in 50% of patients. (Corn et al., 1994). White matter abnormalities on T2-weighted MRI Apart from radiation factors determining the risk were observed at 6 months after treatment and of diffuse encephalopathy, such as total dose, frac- beyond, but did not progress after 3 years from base- tion dose, volume of brain irradiated, and total line (Armstrong et al., 2002). Torres and co-work- duration of treatment, patient factors may also ers (2003) reported subtle neurocognitive deﬁcits contribute to this adverse effect of treatment. These in patients with LGG that were already present at factors are age greater than 60, pre-existing white the start of radiotherapy, without further decline matter disease (e.g., vascular white matter disease, during a 2-year follow-up (Figure 10.2). Analysis of multiple sclerosis), vascular risk factors (smoking, cognitive functions in patients with LGG treated diabetes, hypertension), and possibly a genetic pre- with lower dose radiation or higher dose radiation disposition (Peterson et al., 1993; Swennen et al., (randomly assigned to 50.4 Gy or 64.8 Gy) revealed 2004). that Mini-Mental State Examination scores in 101 In addition to the study by Klein et al. (2002) patients with a median follow-up of 7.4 years after demonstrating that radiotherapy was not the main radiotherapy showed little cognitive deterioration reason for cognitive disturbances in LGG survivors, (Brown et al., 2003). Moreover, extensive psychome- several other studies have shown similar results. tric testing in a subgroup of 20 patients revealed sta- Armstrong and colleagues (2002) reported on a ble cognitive functioning during 3 years of follow-up prospective study regarding neurocognitive func- (Laack et al., 2005). tion and MRI ﬁndings in 26 patients with LGG. Despite these reassuring data regarding adverse After treatment a highly selective decline in visual effects of radiotherapy on cognitive functioning in Chapter 10. Low-grade gliomas 151

LGG patients, neurocognitive decline may still result proven to be effective treatments (Buckner et al., from therapeutic irradiation, especially in the long 2003; Pace et al., 2003; Quinn et al., 2003; Stege et al., run (Klein et al., 2006). Therefore, randomized stud- 2005; Van den Bent et al., 1998, 2003). ies are underway both in the USA and Europe, com- Cognitive side-effects of temozolomide have so paring radiotherapy with chemotherapy in the treat- far not been described, and encephalopathy with ment of LGG. cognitive deficits caused by PCV chemotherapy has been reported only with high-dose regimens (Postma et al., 1998). Medical therapy as a cause of The late neurotoxic effects of chemotherapy neurocognitive disturbances may be difficult to discern from those of radio- Anti-epileptic drugs therapy, because many LGG patients treated with chemotherapy have already been treated with radio- Epileptic seizures are the first symptom of an therapy. In contrast to several other brain tumors, intracranial tumor in 30%–90% of patients, and this LGG is not treated with intra-arterial drug regimens, holds especially true for LGG. Among LGG patients, local chemotherapy or intrathecal chemotherapy. about 70% take one or more anti-epileptic drugs These applications may increase the likelihood of (Klein et al., 2003). The older anti-epileptic drugs, neurotoxicity, compared to systemic chemotherapy such as phenytoin, carbamazepine, and valproic (oral or intravenous) (Wen, 2003). acid, decrease neurocognitive functioning. Impair- ments of attention and cognitive slowing may result in memory deficits by reducing the efficiency of Mood disorder as a cause of encoding and retrieval (Drane & Meador, 2002; neurocognitive disturbances Meador, 2002). Of the newer drugs, such as lam- Like any cancer patient, brain tumor patients have otrigine, levetiracetam, and topiramate, data on feelings of anxiety, depression, and future uncer- cognitive side-effects are still scarce. Apart from tainty as psychological reactions to the disease anti-epileptic drugs, cognitive function may be neg- (Anderson et al., 1999). As patients with extracra- atively affected by the seizures themselves (Jokeit & nial tumors do not have structural brain lesions that Ebner, 1999). In a study of 156 long-term LGG sur- cause cognitive deficits, neuropsychological distur- vivors without signs of tumor recurrence, deficits in bances in these patients are more likely related information-processing speed, psychomotor func- to mood disorders than to central nervous sys- tioning, executive functioning, and working mem- tem lesions. LGG patients report lower levels of ory capacity were significantly related to the use panic, depression, anxiety, and fear of death than of anti-epileptic drugs or the severity of epilepsy. do patients with high-grade tumors. These mood As patients in this study who took anti-epileptic disturbances may lead to deficits in attention, vig- drugs had cognitive disturbances even in the ilance, and motivation that subsequently affect sev- absence of seizures, the use of drugs primarily eral cognitive domains (Anderson et al., 1999). seems to affect cognitive function (Klein et al., 2003). Conclusion

Chemotherapy Compared to their high-grade counterparts, LGG In glioma patients, combination therapy with PCV may behave as rather indolent tumors with a rela- for the subset of oligodendroglial tumors proved tively long median survival time. Due to the absence to be a breakthrough in treatment. In LGG, both of focal neurological deﬁcits and increased intracra- PCV chemotherapy and temozolomide have been nial pressure in the majority of LGG patients, 152 Section 2. Effects of cancer and cancer treatment on cognition

cognitive complaints and deficits may be very Buckner JC, Gesme D, O’Fallon JR et al. (2003). Phase II prominent. Neurotoxicity, mainly due to radiother- trial of procarbazine, lomustine, and vincristine as initial apy, has for a long time been thought to be the therapy for patients with low-grade oligodendroglioma main cause of cognitive deficits in LGG patients. and oligoastrocytoma: efficacy and associations with An important finding from several studies on LGG chromosomal abnormalities. J Clin Oncol 21: 251–255. Cairncross JG (2000). Understanding low-grade glioma: a is that, although radiation has to some extent an decade of progress. Neurology 54: 1402–1403. adverse effect on cognitive function, other tumor Cairncross JG, Ueki K, Zlatescu MC et al. (1998). Spe- and treatment factors deserve more attention. Also, cific chromosomal losses predict chemotherapeutic cognitive function is being recognized as an inde- response and survival in patients with anaplastic oligo- pendent factor in LGG patients, and may even be dendrogliomas. J Natl Cancer Inst 90: 1473–1479. a first indication of tumor regrowth following treat- Choucair AK, Scott C, Urtasun R et al. (1997). Quality of ment. life and neuropsychological evaluation for patients with malignant astrocytomas: RTOG 91–14. Int J Radiat Oncol Biol Phys 38: 9–20. Christensen HC, Schuz¨ J, Kosteljanetz M et al. (2005). Cel- REFERENCES lular telephones and risk for brain tumors: a population- based, incident case-control study. Neurology 64: 1189– Anderson SI, Taylor R, Whittle IR (1999). Mood disor- 1195. ders in patients after treatment for primary intracranial Constine LS, Konski A, Ekholm S et al. (1988). Adverse tumours. Br J Neurosurg 13: 480–485. effects of brain irradiation correlated with MR and CT Anderson SW, Damasio H, Tranel D (1990). Neuropsycho- imaging. Int J Radiat Oncol Biol Phys 15(2): 319–330. logical impairments associated with lesions caused by Corn BW, Yousem DM, Scott CB et al. (1994). White matter tumororstroke.Arch Neurol 47: 397–405. changes are correlated significantly with radiation dose. Armstrong CL, Hunter JV, Ledakis GE et al. (2002). Late cog- Obervations from a randomized dose-escalation trial for nitive and radiographic changes related to radiotherapy: malignant glioma (Radiation Therapy Oncology Group initial prospective findings. Neurology 59: 40–48. 83-02). Cancer 74 (10): 2828–2835. Armstrong CL, Goldstein B, Shera D et al. (2003). The pre- Drane LD, Meador KJ (2002). Cognitive and behavioural dictive value of longitudinal neuropsychologic assess- effects of antiepileptic drugs. Epilepsy Behav 3: 49–53. ment in the early detection of brain tumor recurrence. Dropcho EJ (2004). Low-grade gliomas in adults. Curr Treat Cancer 97: 649–656. Options Neurol 6: 265–271. Ashby LS, Shapiro WR (2004). Low-grade glioma: supraten- Duffau H (2006). New concepts in surgery of WHO grade torial astrocytoma, oligodendroglioma, and oligoastro- II glioma: functional brain mapping, connectionism and cytoma in adults. Curr Neurol Neurosci Rep 4: 211–217. plasticity – a review. J Neurooncol 79: 77–115. Behin´ A, Delattre JY (2003). Neurologic sequelae of radio- Duffau H, Capelle L, Denvil D et al. (2003a). Functional therapy of the nervous system. In Schiff D, Wen PY recovery after surgical resection of low grade gliomas (eds.) Cancer Neurology in Clinical Practice (pp. 173– in eloquent brain: hypothesis of brain compensation. J 192). Totowa: Humana Press. Neurol Neurosurg Psychiatry 74(7): 901–907. Bondy ML, Wrensch MR (1996). Epidemiology of primary Duffau H, Capelle L, Denvil D et al. (2003b). Usefulness malignant brain tumours. In Yung WKA (ed.) Cerebral of intraoperative electrical subcortical mapping during Gliomas (pp. 251–270). London: Balliere` Tindall. surgery for low-grade gliomas located within eloquent Brown PD, Buckner JC, O’Fallon JR et al. (2003). Effects of brain regions: functional results in a consecutive series radiotherapy on cognitive function in patients with low- of 103 patients. J Neurosurg 98(4): 764–778. grade glioma measured by the Folstein mini-mental state Duffau H, Khalil I, Gatignol P et al. (2004). Surgical removal examination. J Clin Oncol 21: 2519–2524. of corpus callosum infiltrated by low-grade glioma: func- Brown PD, Buckner JC, O’Fallon JR et al. (2004). Impor- tional outcome and oncological considerations. J Neuro- tance of baseline mini-mental state examination as a surg 100: 431–437. prognostic factor for patients with low-grade glioma. Int Efird JT, Friedman GD, Sidney S et al. (2004). The risk J Radiat Oncol Biol Phys 59: 117–125. for malignant primary adult-onset glioma in a large Chapter 10. Low-grade gliomas 153

multiethnic, managed-care cohort: cigarette smoking Jackson RJ, Fuller GN, Abi-Said D et al. (2001). Limita- and other lifestyle behaviors. J Neurooncol 68: 57–69. tions of stereotactic biopsy in the initial management of Fuller GN, Hess KR, Rhee CH et al. (2002). Molecular gliomas. Neurooncology 3: 193–200. classification of human diffuse gliomas by multidimen- Johannesen TB, Langmark F, Lote K (2003). Progress in sional scaling analysis of gene expression profiles paral- long-term survival in adult patients with supratento- lels morphology-based classification, correlates with sur- rial low-grade gliomas: a population-based study of 993 vival and reveals clinically relevant novel glioma subsets. patients in whom tumors were diagnosed between 1970 Brain Pathol 12: 108–116. and 1993. J Neurosurg 99: 854–862. Ginsberg LE, Fuller GN, Hashni M et al. (1998). The signif- Jokeit H, Ebner A (1999). Long term effects of refractory icance of lack of MR contrast enhancement of supraten- temporal lobe epilepsy on cognitive abilities: a cross sec- torial brain tumors in adults: histopathological evalua- tional study. J Neurol Neurosurg Psychiatry 67: 44–50. tionofaseries.Surg Neurol 49: 436–440. Kaloshi G, Benouaich-Amiel A, Diakite F et al. (2007) Temo- Godard S, Getz G, Delorenzi M et al. (2003). Classification zolomide for low-grade gliomas: predictive impact of of human astrocytic gliomas on the basis of gene expres- 1p/19q loss on response and outcome. Neurology 68: sion: a correlated group of genes with angiogenic activity 1831–1836. emerges as a strong predictor of subtypes. Cancer Res 63: Karim ABMF, Maat B, Hatlevoll R et al. (1996). A ran- 6613–6625. domised trial on dose response in radiation therapy of Hahn CA, Dunn RH, Logue PE et al. (2003). Prospec- low grade cerebral glioma: EORTC study 22844. Int J tive study of neuropsychologic testing and quality-of- Radiat Oncol Biol Phys 36: 549–556. life assessment of adults with primary malignant brain Karim ABMF, Afra D, Cornu P et al. (2002). Randomized tumors. Int J Radiat Oncol Biol Phys 55: 992–999. trial on efficacy of radiotherapy for cerebral low-grade Hardell L, Carlberg M, Soderqvist¨ F, Mild KH, Morgan glioma in the adult. EORTC study 22845 with the MRC (2007). Long-term use of cellular phones and brain study BR04: an interim analysis. Int J Radiat Oncol Biol tumours: increased risk associated with use for ≥10 Phys 52: 316–324. years. Occup Environ Med 64: 626–632. Keles GE, Lamborn KR, Berger MS (2001). Low-grade hemi- Hemminki K, Li X (2003). Familial risks in nervous sys- spheric gliomas in adults: a critical review of extent of tem tumors. Cancer Epidemiol Biomarkers Prev 12: 1137– resection as a factor influencing outcome. J Neurosurg 1142. 95: 735–745. Herholz K, Holzer¨ T, Bauer B et al. (1998). 11C-Methionine Kiebert GM, Curran D, Aaronson NK et al. (1998). Qual- PET for differential diagnosis of low-grade gliomas. Neu- ity of life after radiation therapy of low-grade glioma of rology 50: 1316–1322. the adult: results of a randomised phase III trial on dose Hirose Y, Aldape KD, Chang S et al. (2003). Grade II astro- response (EORTC 22844). Eur J Cancer 34: 1902–1909. cytomas are subgrouped by chromosome aberrations. Kleihues P, Cavenee WK (eds.) (2000). World Health Orga- Cancer Genet Cytogenet 142:1–7. nization Classification of Tumours of the Nervous System. Hoang-Xuan K, Capelle L, Kujas M et al. (2004). Temo- Lyon: WHO/ARC Press. zolomide as initial treatment for adults with low-grade Klein M Heimans JJ, Aaronson NK et al. (2002). Effect oligodendrogliomas or oligoastrocytomas and correla- of radiotherapy and other treatment-related factors on tion with chromosome 1p deletions. J Clin Oncol 22: mid-term to long-term cognitive sequelae in low-grade 3133–3138. gliomas: a comparative study. Lancet 360: 1361–1368. Huncharek M, Kupelnick B, Wheeler L (2003). Dietary Klein M, Engelberts NH, Van Der Ploeg HM et al. (2003). cured meat and the risk of adult glioma: a meta-analysis Epilepsy in low-grade gliomas: the impact on cognitive of nine observational studies. J Environ Pathol Toxicol function and quality of life. Ann Neurol 54: 514–520. Oncol 22: 129–137. Klein M, Fagel S, Taphoorn MJB, Postma TJ, Heimans JJ Idbaih A, Omuro A, Ducray F, Hoang-Xuan K (2007). (2006). Neurocognitive function in long-term low-grade Molecular genetic markers as predictors of response to glioma. Survivors: a six-year follow-up study. Neuroon- chemotherapy in gliomas. Curr Opin Oncol 19: 606–611. cology 8 (4): 321 (abstract). Iwadate Y, Sakaida T, Hiwasa T et al. (2004). Molecular Knopp EA, Cha S, Johnson G et al. (1999). Glial neoplasms: classification and survival prediction in human gliomas dynamic contrast-enhanced T2-weighted imaging. Radi- based on proteome analysis. Cancer Res 64: 2496–2501. ology 211: 791–798. 154 Section 2. Effects of cancer and cancer treatment on cognition

Komine C, Watanabe T, Katayama Y et al. (2003). Pro- Niel¨ CGJH, Struikmans H, van Santvoort J (2005). Radio- motor hypermethylation of the DNA repair gene O6- therapy for primary brain tumors: clinical potential of methylguanine-DNA methyltransferase is an indepen- novel treatment techniques. Neurooncology 7(3): 391 dent predictor of shortened progression free survival (abstract). in patients with low-grade diffuse astrocytomas. Brain Olson JD, Riedel E, DeAngelis LM (2000). Long-term Pathol 13: 176–184. outcome of low-grade oligodendroglioma and mixed Kono K, Inoue Y, Nakayama K et al. (2001). The role glioma. Neurology 54: 1442–1448. of diffusion-weighted imaging in patients with brain Pace A, Vidiri A, Galie E et al. (2003). Temozolomide tumors. Am J Neuroradiol 22: 1081–1088. chemotherapy for progressive low-grade glioma: clinical Laack NN, Brown PD, Ivnik RJ et al. (2005). Cogni- beneﬁts and radiological response. Ann Oncol 14: 1722– tive function after radiotherapy for supratentorial low- 1726. grade glioma: a North Central Cancer Treatment Group Pahlson A, Ek L, Ahlstrom G et al. (2003). Pitfalls in the prospective study. Int J Radiat Oncol Biol Phys 63: 1175– assessment of disability in individuals with low-grade 1183. gliomas. J Neurooncol 65: 149–158. Leighton C, Fisher B, Bauman G et al. (1997). Supratentorial Peterson K, Rosenblum MK, Powers JM et al. (1993). Effect low-grade glioma in adults: an analysis of prognostic fac- of brain irradiation on demyelinating lesions. Neurology tors and timing of radiation. J Clin Oncol 15: 1294–1301. 43: 2105–2112. Levin N, Lavon I, Zelikovitsch B et al. (2006). Progressive Pignatti F, van den Bent MJ, Curran D et al. (2002). Prog- low-grade oligobendrogliomas: response to temozolo- nostic factors for survival in adult patients with cerebral mide and correlation between genetic proﬁle and O-6- low-grade glioma. J Clin Oncol 20: 2076–2084. methyguanine DNA methyltransferase protein expres- Postma TJ, van Groeningen CJ, Witjes RJ et al. (1998). sion. Cancer 106: 1759–1765. Neurotoxicity of combination chemotherapy with pro- Maia AC Jr., Malheiros SM, da Rocha AJ et al. (2005). MR carbazine, CCNU and vincristine (PCV) for recurrent cerebral blood volume maps correlated with vascular glioma. J Neurooncol 38: 69–75. endothelial growth factor expression and tumor grade in Postma TJ, Klein M, Verstappen CC et al. (2002). nonenhancing gliomas. AJNR Am J Neuroradiol 26: 777– Radiotherapy-induced cerebral abnormalities in 783. patients with low-grade glioma. Neurology 59: 121– Mandonnet E, Delattre JY, Tnaguy ML et al. (2003). Contin- 123. uous growth of mean tumor diameter in a subset of grade Quinn JA, Reardon DA, Friedman AH et al. (2003). Phase II gliomas. Ann Neurol 53: 524–528. II trial of temozolomide in patients with progressive low Meador KJ (2002). Cognitive outcomes and predictive fac- grade glioma. J Clin Oncol 21: 646–651. tors in epilepsy. Neurology 58: 21–26. Rasheed BK, Wiltshire RN, Bigner SH et al. (1999). Molecu- Meyers CA, Hess KR (2003). Multifaceted end points in lar pathogenesis of malignant gliomas. Curr Opin Oncol brain tumor clinical trials: cognitive deterioration pre- 11: 162–167. cedes MRI progression. Neurooncology 5: 89–95. Recht LD, Lew R, Smith TW (1992). Suspected low-grade Monje ML, Palmer T (2003). Radiation injury and neuroge- glioma: is deferring treatment safe? Ann Neurol 31: 431– nesis. Curr Opin Neurol 16: 129–134. 436. Monje ML, Vogel H, Masek M, Ligon KL, Fisher PG, Palmer Rees JH (2002). Low-grade gliomas in adults. Curr Opin TD (2007). Impaired human hippocampal neurogenesis Neurol 15: 657–661. after treatment for central nervous system malignancies. Reijneveld JC, Sitskoorn MM, Klein M et al. (2001). Cogni- Ann Neurol 62: 515–520. tive status and quality of life in patients with suspected Murphy PS, Viviers L, Abson C et al. (2004). Monitoring versus proven low-grade glioma. Neurology 56: 618–623. temozolomide treatment of low-grade glioma with pro- Rogers LR, Morris HH, Lupica K (1993). Effect of cranial ton magnetic resonance spectroscopy. Br J Cancer 90: irradiation on seizure frequency in adults with low-grade 781–786. astrocytoma and medically intractable epilepsy. Neurol- Nagesh Y, Tsien CI, Chevenert TL et al. Radiation-induced ogy 43: 1599–1601. changes in normal-appearing white matter in patients Sanai N, Berger MS (2008). Glioma extent of resection and with cerebral tumors: a diffusion tensor imaging study. its impact on patient outcome. Neurosurgery 62(4): 753– Int J Radiat Oncol Biol Phys 70(4): 1002–1010. 764. Chapter 10. Low-grade gliomas 155

SalvatiM,ArticoM,CarusoRet al. (1991). A report on glioma: a preliminary report. J Neurol Neurosurg Psychi- radiation-induced gliomas. Cancer 67: 392–397. atry 55: 372–376. Scheibel RS, Meyers CA, Levin VA (1996). Cognitive dys- Torres JJ, Mundt AJ, Sweeney PJ et al. (2003). A longitudi- function following surgery for intracerebral glioma: nal neuropsychological study of partial brain radiation in influence of histopathology, lesion location, and treat- adults with brain tumors. Neurology 60: 1113–1118. ment. J Neurooncol 30(1): 61–69. Tucha O, Smely C, Preier M, Lange KW (2000). Cogni- Scott JN, Brasher PM, Sevick RJ et al. (2002). How often are tive deficits before treatment among patients with brain nonenhancing supratentorial gliomas malignant? A pop- tumors. Neurosurgery 47: 324–333. ulation study. Neurology 59: 947–949. Van den Bent MJ, Kros JM, Heimans JJ et al. (1998). Shaw EG, Daumas-Duport C, Scheithauer BW et al. (1989). Response rate and prognostic factors of recurrent Radiation therapy in the management of low-grade oligodendroglioma treated with procarbazine, CCNU, supratentorial astrocytomas. J Neurosurg 70: 853–861. and vincristine chemotherapy. Neurology 51: 1140–1145. Shaw E, Arusell R, Scheithauer B et al. (2002). Prospective Van den Bent MJ, Taphoorn MJB, Brandes AA et al. (2003). randomized trial of low- versus high-dose radiation ther- Phase II study of first-line chemotherapy with temo- apy in adults with supratentorial low-grade glioma: ini- zolomide in recurrent oligodendroglial tumors: EORTC tial report of a NCCTG/RTOG/ECOG study. J Clin Oncol Brain Tumor Group Study 26971. J Clin Oncol 21: 2525– 20: 2267–2276. 2528. Sheline GE, Wara WM, Smith V (1980). Therapeutic irradia- Van den Bent MJ, Afra D, de Witte O et al. (2005). Long- tionandbraininjury.Int J Radiat Oncol Biol Phys 6: 1215– term efficacy of early versus delayed radiotherapy for 1228. low-grade astrocytoma and oligodendroglioma in adults: Smith JS, Perry A, Borell TJ et al. (2000). Alterations of chro- the EORTC 22845 randomised trial. Lancet 366: 985–990. mosome arms 1p and 19q as predictors of survival in Vecht CJ (1993). Effect of age on treatment decisions in low- oligodendrogliomas, astrocytomas and mixed oligoas- grade glioma. J Neurol Neurosurg Psychiatry 56: 1259– trocytoma. J Clin Oncol 18: 636–645. 1264. Smith JS, Chang EF, Lamborn KR et al. (2008). Role of Vigliani MC, Sichez N, Poisson M, Delattre JY (1996). extent of resection in the long-term outcome of low- A prospective study of cognitive functions following grade hemispheric gliomas. J Clin Oncol 26: 1338–1345. conventional radiotherapy for supratentorial gliomas in Stege EM, Kros JM, de Bruin HG et al. (2005). Successful young adults: 4-year results. Int J Radiat Oncol Biol Phys treatment of low-grade oligodendroglial tumors with a 35: 527–533. chemotherapy regimen of procarbazine, lomustine, and Vigliani MC, Duyckaerts C, Hauw JJ et al. (1999). Dementia vincristine. Cancer 103: 802–809. following treatment of brain tumors with radiotherapy Surma-Aho O, Niemela M, Vilkki J et al. (2001). Adverse administered alone or in combination with nitrosourea- long-term effects of brain radiotherapy in adult low- based chemotherapy: a clinical and pathological study. J grade glioma patients. Neurology 56: 1285–1290. Neuro Oncol 41: 137–149. Swennen MHJ, Bromberg JEC, Witkamp TD et al. (2004). Vuori K, Kankaanranta L, Hakkinen AM et al. (2004). Low- Delayed radiation toxicity after focal or whole brain grade gliomas and focal cortical developmental malfor- radiotherapy for low-grade glioma. J Neurooncol 66: 333– mations: differentation with proton MR spectroscopy. 339. Radiology 230: 703–708. Takebayashi T, Varsier N, Kikuchi Y et al. (2008). Mobile Watanabe T, Nakamura M, Kros JM et al. (2002). Phenotype phone use, exposure to radiofrequency electromagnetic versus genotype correlation in oligodendroglioma and field, and brain tumour: a case–control study. Br J Cancer low-grade diffuse astrocytoma. Acta Neuropathol 103: 98: 652–659. 267–275. Taphoorn MJB, Klein M (2004). Cognitive deficits in adult Wen PY (2003). Central nervous system complications of brain tumour patients. Lancet Neurol 3: 159–168. cancer therapy. In Schiff D, Wen PY (eds.) Cancer Neu- Taphoorn MJB, Heimans JJ, Snoek FJ et al. (1992). Assess- rology in Clinical Practice (pp. 215–231. Totowa: Humana ment of quality of life in patients treated for low-grade Press. 11

High-grade gliomas

Michael J. Glantz and John W. Conlee

Introduction the bone marrow as the most important dose- limiting end organ for cancer therapy in general, After more than 30 years of intensive clinical and for therapy directed at central nervous sys- and laboratory research, high-grade gliomas (HGG) tem tumors in particular. For the large number of [glioblastoma multiforme (GBM), anaplastic astro- children (Bhat et al., 2005; Lannering et al., 1990; cytoma (AAs), anaplastic oligodendroglioma, and Packer et al., 2003), and for the still small but grow- anaplastic mixed glioma] retain a well-deserved ing number of adult long-term survivors of malig- reputation for poor response to therapy, rapid nant primary brain tumors, the nervous system tumor recurrence, and short overall survival. Cur- rather than the hematopoietic system more fre- rently available treatments, including surgery, cran- quently affects the quality of survival and the eco- ial irradiation, and chemotherapy, extend survival nomic productivity of survivors. Thus the combina- measurably, but are almost always non-curative, tion of disease-related and treatment-related ner- and are associated with substantial toxicity. In vous system insults presents a daunting challenge this context, maintaining good quality of life has to patients and health care providers. The spectrum assumed an increasingly prominent role in selecting of problems includes focal neurologic deﬁcits, cog- treatments and in designing clinical trials (Report nitive impairment, affective disorders, and associ- of the Brain Tumor Progress Review Group, 2005; ated medical complications (Table 11.1). This chap- Taphoorn et al., 2005), and the paradigm of com- ter will focus on the frequency, assessment, causes, pressing morbidity and “rectangularizing” the sur- and potential therapies of cognitive impairment in vival curve (Fries, 1980) is increasingly seen as the adults with high-grade gliomas. central goal of cancer therapy (Figure 11.1). Traditionally, myelosuppression and its attendant problems have been the dose-limiting and most Scope of the problem important toxicities of radiation and chemotherapy. In the late 1990s and early 2000s, however, An estimated 21 810 patients will be diagnosed with the availability of colony stimulating factors, and primary brain tumors in the United States in 2007 dramatic improvements in transfusion medicine, (Central Brain Tumor Registry of the United States, antibiotic therapy, and supportive care made the 2005; Jemal et al., 2008). Of these, approximately bone marrow more robust. Today, a strong case 60% will be HGG (with glioblastomas compris- can be made that the nervous system has replaced ing 50% of the total). While median survival has

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

156 Chapter 11. High-grade gliomas 157

(a)

100 Score (%) Karnofsky Performance

0 1 3 5 7 9 11 13 15 Survival (months)

(b)

100 Score (%) Karnofsky Performance

0 0 246 8 101214 Survival (months)

(c)

100 Score (%) Karnofsky Performance

0 0 2 4 6 8 10 12 14 Survival (months)

Figure 11.1. Quality of life versus survival (a) the ideal outcome, (b) the worst case, and (c) the realistic goal 158 Section 2. Effects of cancer and cancer treatment on cognition

Table 11.1. Spectrum of nervous system problems in patients with high-grade gliomas

Category of problem Examples Potential causes

Focal deficits Hemiparesis, aphasia, Tumor, surgery, radiation necrosis hemianopsia Cognitive impairment Diminished short-term memory, Tumor, surgery, radiation, executive function, fine motor chemotherapy function, information processing Affective disorders Depression, anxiety Tumor, radiation, chemotherapy, emotional response to the disease Medical complications Proximal myopathy, fatigue, Tumor, corticosteroids, endocrine deficiencies chemotherapy, radiation

changed very little in the last decade (10–12 months 2000). Similarly, 90.5% of patients with brain meta- for patients with GBM, 2–3 years for AA), a surpris- stases scored in the abnormal range on one ing variability in individual survival occurs (Carson or more tests of cognitive function prior to et al., 2007; Curran et al., 1993). Almost two-thirds of receiving cranial irradiation (Meyers et al., patients with AAs survive at least 18 months, 25% of 2004). Much research has focused on the role patients with AAs aged 64 or younger survive longer of tumor-directed therapy, particularly cranial than 10 years, and 6% of patients with GBMs survive irradiation, as a cause of cognitive deficits. As more than 4 years (Central Brain Tumor Registry of demonstrated by the almost universal presence the United States, 2005). Similarly, the recent large, of cognitive deficits in newly diagnosed and in randomized EORTC/NCCI trial reported a 26.5% 2- post-operative patients, however, HGG themselves year survival for patients with GBM (Stupp et al., contribute substantially to the burden of cognitive 2005). While increased survival is a welcome out- impairment. come of new therapeutic strategies, long-term sur- Neurologic and cognitive deficits related to the vivors are potentially at increasing risk for devel- tumor itself may manifest as a global decline in level oping neurocognitive deficits related to the delayed of alertness when the tumor is causing increased effects of therapy. intracranial pressure or obstruction of spinal fluid pathways, or as focal deficits (for example, aphasia, verbal learning and memory difficulties, visual Causes of cognitive impairment in patients field defects, hemiparesis, or neglect) when the with high-grade gliomas cause is local pressure effects and vascular changes related to the location of the lesion. The conse- Neurocognitive dysfunction in patients with HGG quences of brain tumor surgery are also typically is multifactorial (Table 11.2). Patients with primary thought of as focal and related to the site of tumor brain tumors frequently report a variety of cognitive resection (Archibald et al., 1994; Imperato et al., complaints, and recent studies have documented a 1990; Levin et al., 2002; Scheibel et al., 1996). In spectrum of deficits ranging from subtle to blatant contrast, treatment-related neurocognitive deficits in 34% of patients when assessed by a relatively (both radiation- and chemotherapy-induced) are insensitive measurement tool (Folstein Mini-Mental primarily related to subcortical white matter dys- State Examination – MMSE) (Brown et al., 2006) function, and include impairment in short-term and in 59%–100% of patients when more compre- memory, executive function, sustained attention, hensive cognitive assessments are performed (Klein speed of information processing, and bilateral fine et al., 2001, 2003b; Levin et al., 2002; Meyers et al., motor control (Archibald et al., 1994; Crossen et al., Chapter 11. High-grade gliomas 159

Table 11.2. Instruments available for assessing neurocognitive function in patients with high-grade gliomas

Time to administer Test Characteristic measured (min)

Cognitive function Hopkins Verbal Learning Test (Benedict et al., 1998) Verbal memory 5 Trail Making Test Part A (Lezak et al., 1994) Visual-motor speed 2 Trail Making Test Part B (Lezak et al., 1994) Executive function 5 Controlled Oral Word Association (Benton & Hamsher, 1989) Verbal ﬂuency 5 Digit Span Test Tota (Wechster, 1981) Attention and concentration 10 Folstein MMSE (Folstein et al., 1975) Global cognitive function 10 Grooved Pegboard (dominant and non-dominant hands) Motor dexterity and speed 2–3 (Lezak, 1995) Categoric Word Fluency (Benton, 1968) Executive function 5

Overall performance Barthel Index (Wade & Collin, 1988) Activities of daily living 5 Karnofsky Performance Score (Karnofsky et al., 1948) Functional Status 2 ECOG Performance Score (Oken et al., 1982) Toxicity 2

Quality of life Functional Assessment of Cancer Therapy (FACT) (Br) Quality of life 5 (Weitzner et al., 1995) SF-36 (Cella et al., 1993; Ware & Sherbourne, 1992) Health-related quality of life 10 Pediatric Quality of Life Inventory 4.0 (Varni et al., 2007) Health-related quality of life <20 Edmonton Symptom Assessment Scale (Bruera et al., 1991) Multiple health-related symptoms 3 Women’s Health Initiative Insomnia Rating Scale Insomnia 3 (Levine et al., 2003) EORTC Quality of Life Questionnaire Core-30 Quality of life 7 (Mauer et al., 2007) EORTC Brain Cancer Module-20 (Mauer et al., 2007) Health-related quality of life 5 Cancer Fatigue Scale (Okuyama et al., 2000) Fatigue 5 Brief Fatigue Inventory (Mendoza et al., 1999) Fatigue 3

Emotional status Proﬁle of Mood States (McNair et al., 1992) Multiple subscalesa 10 Beck Depression Inventory (Beck & Steer, 1993) Mood (depression) 5 Hamilton Depression Rating Scale (Hamilton, 1960) Mood (depression) 5 State-Trait Anxiety Inventory (Spielberger et al., 1970) Mood (anxiety) 5 Hamilton Anxiety Rating Scale (Hamilton, 1959) Mood (anxiety) 5 Beck Anxiety Inventory (Beck et al., 1990) Mood (anxiety) 5 a Depression, anxiety, fatigue, vigor, confusion, anger, overall mood

1994; Hochberg & Slotnick, 1980; Imperato et al., lia, deﬁcits in attention and executive function, 1990; Salander et al., 1995), as well as apathy, depres- and other characteristics of a frontal lobe dysfunc- sion, and other alterations in personality and mood tion – are also common in patients with brain (Marin 1991, 1996). In practice, symptoms of more tumors, even when those tumors are not located widespread cortical dysfunction – in particular, abu- in the frontal lobes (Lilja et al., 1992), and even 160 Section 2. Effects of cancer and cancer treatment on cognition

before radiation or chemotherapy has been admin- convulsants). Despite these shortcomings, when istered. Recently, studies employing magnetoen- taken as a group, these studies suggest that at least cephalography, diffusion tensor imaging, and func- some of the neurocognitive deficits seen in patients tional MRI have started to elucidate the physiologic with HGG are attributable to tumor progression and substrate for these clinical observations by reveal- are not, as previously asserted, due entirely to the ing widespread loss of functional connectivity in the effects of therapy. In addition, those studies in which brains of patients with malignant gliomas, partic- baseline assessments of neurocognitive status were ularly when the tumor is located in the dominant performed prior to the start of cranial irradia- hemisphere (Bartolomei et al., 2006; Rozental et al., tion and chemotherapy have shown unequivocally 1990; Wei et al., 2007; Young, 2007). that neurocognitive deficits are present, in vary- Several recent studies in patients with HGG have ing degrees, in essentially all patients, even when suggested that tumor progression contributes sig- more traditional and less sensitive measures of per- nificantly to neurocognitive impairment, and that formance (Karnofsky Performance Score, Eastern in the absence of tumor progression, the level Cooperative Oncology Group Performance Score, of neurocognitive functioning remains relatively Barthel Index) or global cognitive function (MMSE) stable (Brown et al., 2006; Steinbach et al., 2006; are normal. These findings are supported by analo- Torres et al., 2003). Similar findings have been gous findings in patients with brain metastases and reported in some studies of patients with brain with neoplastic meningitis. metastases (Meyers et al., 2004; Patchell et al., The challenge of distinguishing between tumor- 1998). These studies suggest that cranial irra- and treatment-related causes of cognitive impair- diation, as applied with modern-day treatment- ment in patients with HGG is further complicated planning techniques, partial-brain treatment plans, by a long list of common and potentially contrib- and modest daily fraction and total dose tar- utory etiologies (Taphoorn & Klein, 2004) (Table gets, is not as neurotoxic as earlier studies sug- 11.3). One of the most common of these is con- gested (Archibald et al., 1994; Crossen et al., 1994; current medications. As with any medical patient, DeAngelis et al., 1989; Hochberg & Slotnick, 1980). anxiolytic and sedative drugs, analgesics (including Some of these studies have used insensitive mea- non-steroidal anti-inflammatory agents), antide- sures of cognitive dysfunction. Many are retro- pressants, antihistamines, and anticholinergics spective; lack control groups; suffer from sub- are common offenders (Wen et al., 2006). Anticon- stantial patient drop-out, short patient follow-up, vulsants, particularly the first-generation agents and small numbers of patients at later endpoints; (phenytoin, phenobarbital, and carbamazepine) are include highly selected patients (for example, pre- commonly implicated (Mattson, 2004), although dominantly young patients); and have not required the newer agents (in particular, topiramate) are masked assessment. As a result, the risks of con- not immune (Bosma et al., 2007). Seizures them- founding (e.g., there may be shared risk factors for selves, sometimes subtle enough to be clinically survival, tumor progression, and cognitive decline); inapparent, can produce sustained cognitive and selection bias (e.g., if cognitive decline is itself a risk behavioral deficits (Klein et al., 2003a). Cortico- factor for shortened survival, longer survivors are steroids have also been demonstrated, both less likely to show impaired cognition); investigator clinically and in laboratory models, to produce bias; and lack of generalizability are considerable. sustained deficits in memory and learning (Brunner Most studies have also failed to assess the contri- et al., 2006; Lupien et al., 1999; Young et al., butions of patient variables (e.g., level of education; 1999). A wide array of chemotherapeutic agents, pre-existing impairment; vascular risk factors) and including many that are frequently used in brain additional competing causes of neurocognitive dys- tumor patients, have also been associated with function (e.g., chemotherapy, corticosteroids, anti- both acute and chronic neurocognitive deficits Chapter 11. High-grade gliomas 161

Table 11.3. Differential diagnosis of behavioral in patients with cancer. CMV, Cytomegalovirus; HSV,herpes simplex virus; JC, John Cunningham virus; MRI, Magnetic resonance imaging; SIADH, Syndrome of inappropriate antidiuretic hormone

Etiology Comment

Endocrinopathy Hypothyroidism, hyperprolactinemia, cortisol, and testosterone deﬁciencies are all common Infection Meningitis Listeria is common; ventricular reservoirs or shunts predispose Brain abscess Sepsis Usually in the setting of high fever, hypoxia, or hypotension Encephalitis Particularly HSV, CMV, and progressive multifocal leukoencephalopathy (JC virus) Metabolic Hypomagnesemia Common with cisplatin use, often 3–8 days after therapy; also after bisphosphonate administration Hyponatremia Common following neurosurgery; also vincristine, carbamazepine, oxcarbazepine, SIADH, cyclophosphamide, or chemotherapy requiring co-administration of large ﬂuid volumes Hypocalcemia Following cisplatinum or bisphosphonate administration Drug-induced Cisplatin Possibly related to electrolyte disturbances; may be delayed up to 2 weeks after completion of treatment Vincristine, Etoposide Uncommon Busulfan, Methotrexate Interleukin-2, Interferon-α Particularly during high-dose therapy Ifosfamide Especially in patients with renal failure or hypoalbuminemia Thienamycin antibiotics Up to 6% of all patients (imipenem and others) Radiation-related Radiation necrosis MRI may be indistinguishable from recurrent tumor Intracranial Thrombocytopenia Related to disease, chemotherapy, or drugs (e.g., heparin) Hemorrhage Coagulopathy Disseminated intravascular coagulation Tumor-related New or progressive disease Including neoplastic meningitis and brain metastases Paraneoplastic Encephalitis Increasingly recognized; most common with small-cell lung cancer and ovarian teratoma Other Posterior reversible Seen in multiple settings, including bone marrow transplant, Leukoencephalopathy immunosuppressive therapy with ciclosporin, tacrolimus, syndrome interferon-α, and others. Abrupt increases in blood pressure may predispose

(Cavaliere & Schiff, 2006; Hildebrand, 2006). son et al., 1999; Raida et al., 2001; Shehata et al., While some of these effects are indirect (medi- 1999), methotrexate (Ulrich et al., 2001, 2003), ated through chemotherapy-associated anemia, and the selective serotonin reuptake inhibitors seizures, or vasculopathy for example) many agents (SSRIs) (Murphy et al., 2003). Interactions between appear to directly produce behavioral and neu- radiation therapy and a variety of patient risk rocognitive changes (Table 11.3). Increasingly, factors (including hypercholesterolemia, diabetes, individual pharmacogenetic variations under- pre-existing Alzheimer’s disease, hyperhomocys- lying these adverse cognitive effects are being teinemia, and the presence of the apoE4 allele) have identiﬁed. Examples include 5-ﬂuorouracil (John- also been suggested. An interesting paradigm of the 162 Section 2. Effects of cancer and cancer treatment on cognition

interaction between a chemotherapy agent and an to correlate with shortened survival in patients with underlying inherited condition conspiring to pro- primary and metastatic brain tumors (Brown et al., duce dramatic cognitive impairment can be found 2003, 2006; Murray et al., 2000; Shaw et al., 2002; in patients with adult-onset fragile X syndrome who Taylor et al., 1998). No single, simple, brief screening are treated with cis-platinum – a convergence of test currently available is capable of reliably identi- factors that may result in fulminate encephalopa- fying impairment; distinguishing among and appor- thy (O’Dwyer et al., 2005). Other indirect effects tioning responsibility between the multiple poten- of tumor therapy, such as infection (including tial causes of impairment; allowing the evolution of chronic meningitis), radiation-related secondary impairment to be tracked; and predicting the ulti- and tertiary endocrinopathies, and (very rarely in mate course of that impairment. Relatively concise patients with primary brain tumors) paraneoplastic batteries of tests have, however, been composed encephalopathy, also enter the differential diagno- which do fulfill these criteria, and which can be sis. Of particular note, because of their frequency of taught to and performed by health care providers occurrence and lack of recognition, is the contribu- without advanced training in neuropsychology tion of fatigue, depression, and anxiety to apparent (Meyers et al., 2004). Neurocognitive testing (com- neurocognitive deficits (Cull et al., 1996; Klein et al., bined with a careful patient history and examina- 2001; Wellisch et al., 2002). tion) is also reliably able to identify deficits related to alterations in mood and state such as depression, anxiety, and fatigue. Table 11.2 lists some commonly Assessment of cognitive function in patients used tests of cognitive function, mood, and qual- with high-grade gliomas ity of life. These tests can be administered by any- one who has been properly trained, but can only be Just as the differential diagnosis of cognitive impair- interpreted by an appropriate professional. ment is complex, so too is the measurement of cognitive function in patients with HGG. Accurate assessment is complicated by the presence of com- Importance of neurocognitive assessment peting causes of impairment (sometimes producing similar or overlapping patterns of disability); by Neurocognitive assessment contributes in at least neurologic deficits that limit the ability of patients three essential areas to the management of patients to participate fully in testing; by the complexity of with HGG: by providing a clinically relevant and some test batteries; by limitations in time, training, quantifiable outcome measure for patients enrolled and inclination on the part of health care profes- in clinical trials or receiving therapy outside of the sionals; and by deficiencies in the test instruments investigational setting; by providing a measure of themselves (Brown et al., 2003; Klein & Heimans, the neurotoxicity of standard or investigational ther- 2004; Meyers & Brown, 2006; Meyers & Wefel, 2003). apies (Meyers et al., 1997); and by refining our abil- Simple and widely used dementia screening tools ity to predict patient outcomes. In diseases where such as the MMSE are insensitive to clinically mean- therapies often differ only modestly in efficacy, neu- ingful improvements and deteriorations in cog- rocognitive function has the potential to become a nitive function (particularly the subcortical white primary endpoint in clinical trials, and is gaining matter deficits most commonly produced by radi- acceptance from regulatory agencies for this reason ation), may be susceptible to practice effects, and (Report of the Brain Tumor Progress Review Group, are influenced significantly by the level of education 2005). In addition, in a number of recent studies, and the presence of language impairment in test performance on neurocognitive testing in patients subjects. Nevertheless, the MMSE has been shown with low-grade (Brown et al., 2003; Shaw et al., in some studies to presage tumor recurrence, and 2002), and high-grade (Brown et al., 2006; Meyers Chapter 11. High-grade gliomas 163

et al., 2000; Taylor et al., 1998) primary brain tumors, In part this may be related to an altered perception brain metastases (Meyers et al., 2004), and neo- of functioning produced by cognitive impairments plastic meningitis (Sherman et al., 2002) has been (Taphoorn et al., 1992). More importantly, however, shown to help predict both tumor recurrence and a patient’s appreciation of what constitutes worth- overall survival. However, caution is advised when while “quality” evolves over the course of his or her interpreting these results as some of the studies illness, so that even in the face of mounting physical, hadalargeamountofmissingdata,usedinsen- neurological, and cognitive deficits, patients may sitive measures, lacked a control group, had only feel that their quality of life remains good enough short-term follow-up assessments, or did not con- to justify continued tumor-directed therapy. Thus trol for other factors that might affect cognitive evaluation of neurocognitive function is critical in function, such as the use of adjuvant medications. identifying treatments with less neurocognitive tox- Similarly, a recent analysis of patients with glioblas- icity, and those which improve neurocognitive, eco- tomas participating in the EORTC phase III trial nomic, and social functioning, but it cannot sub- comparing cranial irradiation alone with irradiation stitute for quality of life assessments, which more plus daily temozolomide has raised doubts about directly register the patient’s own appreciation of whether cognitive functioning, social functioning, the disease process and the therapies used to fight it. and global health status truly add predictive power to prognostic models including the more traditional variables of age, performance status, and histology Treatment of cognitive deficits in patients (Mauer et al., 2007). While additional prospective with high-grade gliomas study is necessary, these findings do suggest that the results of neurocognitive function testing may help The underlying biochemistry and neurophysiology to provide a rationale for clinicians to continue or of neurocognitive deficits in patients with HGG abandon therapies in the face of radiographically is still poorly understood and undoubtedly multi- stable disease, and may also assist in the important factorial. Radiation, chemotherapy, and the tumor role of prognostication. itself probably conspire to produce brain injury Prolonging survival remains a valid long-range through damage to the neurovascular endothelium goal in neuro-oncology. However, when aggressive and multiple cell populations (oligodendroglio- therapies are administered to patients with no or cytes, neural stem cells, neurons, astrocytes, and very limited curative potential, attention to alterna- microglia), disruption of cellular DNA, and disruptive outcome measures becomes critical. One such tion of the neuronal microenvironments, includ- alternative concept, first proposed in another con- ing alterations in cytokine and neurotransmitter text by Fries (1980), suggests that the “compression expression (Abdallah et al., 2007; Armstrong et al., of morbidity” within a given lifespan may be equally 1995; Belka et al., 2001; O’Connor & Mayberrg, 2000; as important as overall survival given the limitations Monje et al., 2002, 2007; Moulder et al., 1998; Tofilon of existing therapies. This idea of “rectangularizing” & Fike 2000). Host factors and the response of dam- the survival curve by prolonging a high level of func- aged tissues to injury are also involved. Neverthe- tioning for as much of the lifespan as possible res- less, the phenotypic similarity of some symptoms onates loudly with cancer patients in general, and to certain neurological and psychiatric disorders brain tumor patients and their families in particular in non-tumor populations, including Alzheimer’s (Davies & Clarke, 2005; Steinbach et al., 2006). Inter- disease, has encouraged the investigative use of estingly, measures of neurocognitive, social, and therapies useful in those disorders in patients work-related function on the one hand and quality with brain tumors. Thus conventional antidepres- of life on the other have been found, in at least some sant therapy (particularly employing SSRIs and studies, to correlate poorly (Steinbach et al., 2006). combined serotonin and norepinephrine reuptake 164 Section 2. Effects of cancer and cancer treatment on cognition

inhibitors) has been used with success equal to A small, open-label phase II trial of donepezil that in non-brain tumor patients (Gill & Hatcher, (5 mg daily for 6 weeks, then 10 mg daily for 18 2000; Wen et al., 2006). Improvements in energy, weeks) in 35 cognitively impaired patients with cognitive function, fine motor performance, depres- brain tumors (most low grade, all but one primary) sion and global measures of performance and daily who had survived at least 6 months from the com- functioning have all been documented in small pletion of radiation therapy has recently demon- studies using methylphenidate and modafinil. In a strated significant improvement in cognitive func- trial involving 30 predominantly young (mean age tion (including measures of attention and concen- 40.3 years) adult patients with malignant primary tration, verbal memory, and figural memory), mood, brain tumors, there was a suggestion of a dose– and health-related quality of life (Shaw et al., 2006). response relationship in patients receiving 10, 20, or Trends toward improvement were also seen in ver- 30 mg of methylphenidate twice daily (Meyers et al., bal fluency, fatigue, and anger, although no change 1998). Remarkably, improvements occurred despite in MMSE was identified. Toxicity was modest. The radiographic evidence of tumor progression in 7 young age (mean 45 years), high dropout rate (11 (23%) patients and treatment-related white matter of 35 patients), absence of a control group, and injury in 8 (27%) patients. Toxicity was minimal. In unmasked assessment represent significant prob- particular, no worsening of seizure control was seen. lems with this study. In addition, a second study of A larger, randomized, placebo-controlled, double- donepezil failed to show any benefit over placebo in blind crossover study in 80 pediatric long-term sur- improving fatigue in a cohort of 142 patients with vivors of brain tumors (43 patients) or acute lym- various malignancies (Bruera et al., 2007). Neverthe- phoblastic leukemia also demonstrated improve- less, the findings are encouraging enough to merit ments in attention, social skills, and school perfor- further study. mance with the use of methylphenidate over the The possibility of preventing treatment-induced course of 3 weeks, although no difference between brain injury and the resulting neurocognitive low (0.3 mg/kg) and moderate (0.6 mg/kg) dose deficits is also being actively explored. An intriguing therapy was identified (Mulhern et al., 2004). In recent series of studies in rats (Monje et al., 2002) contrast, a short-term (7-day) randomized trial of and in humans (Monje et al., 2007) with medul- methylphenidate in 112 patients with non-nervous- loblastoma or acute myelogenous leukemia treated system cancer showed substantial improvement in with cranial irradiation has shown that radiation fatigue in both the placebo and methylphenidate inhibits hippocampal neurogenesis, diverts neural treatment arms, with no significant difference stem cell differentiation away from neuron pro- between treatment arms (Bruera et al., 2006). Most duction, and increases microglia production and recently, a 30-patient randomized trial of modafinil inflammation – changes that can be reversed by in 30 patients with a variety of high-grade and low- inflammatory blockade with indomethacin (Monje grade primary brain tumors showed clinically and et al., 2003). These findings have stimulated a statistically significant improvements from base- prophylaxis trial which is now underway. line in a comprehensive battery of neurocogni- These encouraging findings must still be treated tive tests and fatigue scales at doses of modafinil as preliminary, but certainly underscore the urgent ranging from 200–400 mg/day for the first 3 weeks need for larger, well-designed, randomized trials of the trial to (after a 1-week washout period) with masked outcome assessment in a wider spec- 500–600 mg/week for an additional 8-week open- trum of patients, including pediatric and elderly label extension period. Headache (43% of patients), individuals. Incorporation of biochemical response insomnia (27%), dizziness (23%), and dry mouth measures, functional imaging, and PET scanning (23%) were common but generally mild in severity offer the potential for more effective selection of (Kaleita et al., 2006). patients and agents, and a more sophisticated Chapter 11. High-grade gliomas 165

understanding of the mechanisms underlying neu- is the result of interactions between the tumor rocognitive dysfunction in patients with brain itself, direct and indirect effects of surgery, radi- tumors. Examination of other acetylcholinesterase ation, and chemotherapy, ancillary drugs such as inhibitors, N-methyl-d-aspartate (NMDA) antago- anticonvulsants and corticosteroids, and alterations nists such as memantine, dopamine agonists, and in mood and state. Recent investigations employ- combinations of agents, as well as prophylaxis ing carefully selected neuropsychological test pan- trials are all important avenues of investigation els have given rise to several paradigm-altering con- for the future. With the increasing availability of cepts. First, neurocognitive impairment exists in intensity-modulated radiation therapy and other most patients with HGG even prior to the start highly conformal radiation techniques, the possibil- of brain tumor therapy. Second, tumor progres- ity of designing radiation fields that spare partic- sion often contributes substantially to accumulat- ularly sensitive and functionally critical structures ing neurocognitive deficits during the course of such as the hippocampus should also be explored therapy. Third, changes in neurocognitive perfor- in the clinical trial setting. mance may precede and even presage radiographic Finally, research on effective cognitive rehabili- and clinical exam evidence of tumor recurrence. tation strategies for brain tumor patients remains Finally, neurocognitive testing plays an important scant. The purely descriptive reports that largely role in evaluating the success of conventional and populate the current literature must be replaced investigational therapies, particularly in this set- by studies that link neuropsychological diagno- ting where cures are rarely achieved, and avoid- sis with specific cognitive rehabilitation interven- ance of accumulating deficits is an important goal of tions (Report of the Brain Tumor Progress Review treatment. Practical, validated testing batteries are Group, 2005). The cognitive deficits resulting from currently available to clinicians and should be a brain tumor are frequently subtle and are often incorporated into both routine care and clinical brought to light only by careful neuropsycholog- trials. Several small but encouraging investigations ical assessment (Meyers & Boake, 1993). At the suggest that both pharmacological and neurocogni- same time, these psychometrically subtle deficits tive interventions are available which improve neu- may produce significant functional impairments in rocognitive function and overall quality of life. the less structured settings of home and workplace (Mesulam, 2000). The milder and more variable cognitive impairments present in some brain tumor REFERENCES patients (Jagaroo et al., 2000; Kayl & Meyers, 2003; Maldjian et al., 2001) may make them ideal candi- Abdallah NM-BB, Slomianka L, Lipp H-P (2007). Reversible dates for cognitive rehabilitation (Meyers & Boake, effect of X-irradiation on proliferation, neurogenesis, 1993; Sherer et al., 1997). Cognitive rehabilitation and cell death in the dentate gyrus of adult mice. Hip- interventions used successfully in treating patients pocampus 17(12): 1230–1240. with brain injuries and strokes may provide tar- Archibald YM, Lulnn D, Ruttan LA et al. (1994). Cognitive geted strategies for the remediation of specific cog- functioning in long-term survivors of high-grade glioma. nitive impairments in patients with brain tumors J Neurosurg 80: 247–253. Armstrong C, Ruffer J, Corn B, DeVires K, Mollman J (Sohlberg & Maateer, 1989). (1995). Biphasic patterns of memory deficits following moderate-dose partial-brain irradiation: neuropsychologic outcome and proposed mechanisms. J Clin Oncol Summary 13: 2263–2271. Bartolomei F, Bosma I, Klein M et al. (2006). How do Neurocognitive impairment in patients with brain brain tumors alter functional connectivity? A magne- tumors, in particular HGG, is almost universal, and toencephalography study. Ann Neurol 59: 128–138. 166 Section 2. Effects of cancer and cancer treatment on cognition

Beck AT, Steer RA (1993). Beck Depression Inventory.San tumor therapy CNS consortium phase I and II clinical tri- Antonio, TX: Psychological Corp. als. J Clin Oncol 25: 2601–2606. Beck AT, Brown G, Epstein N, Steer RA (1990). An inventory Cavaliere R, Schiff D (2006). Neurologic toxicities of cancer for measuring clinical anxiety: psychometric properties. therapies. Curr Neurol Neurosci Rep 6: 218–226. J Consult Clin Psychol 56: 893–897. Cella DF, Tulsky DS, Gray G et al. (1993). The functional BelkaC,BudachW,KortmannRDet al. (2001). Radiation assessment of cancer therapy scale: development and induced CNS toxicity – molecular and cellular mecha- validation of the general measure. J Clin Oncol 11: 570– nisms. Br J Cancer 85: 1233–1239. 577. Benedict RHB, Schretlen D, Groninger L et al. (1998). Hop- Central Brain Tumor Registry of the United States (2005). kins Verbal Learning Test-Revised: normative data and Primary Brain Tumors in the United States Statistical analysis of inter-form and test-retest reliability. Clin Neu- Report 1998–2002 (pp. 8–50). Chicago, IL: Central Brain ropsychol 12: 43–55. Tumor Registry of the United States. Benton AD (1968). Differential behavioral effects in frontal Crossen JR, Garwood D, Glatstein E et al. (1994). Neurobe- lobe disease. Neuropsychologia 6: 53–60. havioral sequelae of cranial irradiation in adults: a review Benton AL, Hamsher K de S (1989). Multilingual Aphasia of radiation-induced encephalopathy. J Clin Oncol 12: Examination. Iowa City, IA: AJA Associates. 627–642. Bhat SR, Goodwin TL, Burwinkle TM et al. (2005). Proﬁle Cull A, Hay C, Love SB et al. (1996). What do cancer patients of daily life in children with brain tumors: an assess- mean when they complain of concentration and mem- ment of health-related quality of life. J Clin Oncol 23: ory problems? Br J Cancer 74: 1674–1679. 5493–5500. Curran WJ Jr., Scott CB, Horton J et al. (1993). Recursive Bosma I, Vos MJ, Heimans JJ et al. (2007). The course partitioning analysis of prognostic factors in three Radi- of neurocognitive functioning in high-grade glioma ation Therapy Oncology Group malignant gliomas trials. patients. Neurooncology 9: 53–62. J Natl Cancer Inst 85: 704–710. Brown PD, Buckner JC, O’Fallon JR et al. (2003). Effects of Davies E, Clarke C (2005). Views of bereaved relatives radiotherapy on cognitive function in patients with low- about quality of survival after radiotherapy for malignant grade glioma measured by the Folstein mini-mental state cerebral glioma. J Neurol Neurosurg Psychiatry 76: 555– examination. J Clin Oncol 21: 2519–2524. 561. Brown PD, Jensen AW, Felten SJ et al. (2006). Detrimen- DeAngelis LM, Delattre J-Y, Posner JB (1989). Radiation- tal effects of tumor progression on cognitive function of induced dementia in patients cured of brain metastases. patients with high-grade glioma. J Clin Oncol 24: 5427– Neurology 39: 789–796. 5433. Folstein MF, Folstein SE, McHugh PR (1975). “Mini-mental Bruera E, Kuehn N, Miller M, Selmser P,Macmillan K (1991). state”: a practical method for grading the cognitive state The Edmonton Symptom Assessment System (ESAS): of patients for the clinician. J Psychiatr Res 12: 189–198. a simple method for the assessment of palliative care Fries JF (1980). Aging, natural death, and the compression patients. J Palliative Care 7:6–9. of morbidity. NEnglJMed303: 130–135. Bruera E, Valero V, Driver L et al. (2006). Patient-controlled Gill D, Hatcher S (2000). Antidepressants for depression in methylphenidate for cancer fatigue: a double-blind, ran- medical illness. Cochrane Database Syst Rev 4: CD001312. domized, placebo-controlled trial. J Clin Oncol 24: 2073– Hamilton M (1959). The assessment of anxiety state by rat- 2078. ing. Br J Med Psychol 32: 50–55. Bruera E, Osta BE, Valero V et al. (2007). Donepezil for Hamilton M (1960). A rating scale for depression. J Neurol cancer fatigue: a double-blind, randomized, placebo- Neurosurg Psychiatry 23: 56–62. controlled trial. J Clin Oncol 25: 3475–3481. Hildebrand J (2006). Neurological complications of cancer Brunner R, Schaefer D, Hess K, Parzer P, Resch F, Schwab S chemotherapy. Curr Opin Oncol 18: 321–324. (2006). Effect of high-dose cortisol on memory functions. Hochberg FH, Slotnick B (1980). Neuropsychologic impair- Ann N Y Acad Sci 1071: 434–437. ment in astrocytoma survivors. Neurology 30: 172–177. Carson KA, Grossman SA, Fisher JD, Shaw EG (2007). Prog- Imperato JP, Palelogos NA, Vick NA (1990). Effects of treat- nostic factors for survival in adult patients with recur- ment on long-term survivors with malignant astrocy- rent gliomas enrolled onto the new approaches to brain tomas. Ann Neurol 28: 818–822. Chapter 11. High-grade gliomas 167

Jagaroo V, Rogers MP, Black PM (2000). Allocentric visuo- Lilja A, Brun A, Salford LG et al. (1992). Neuropsychologi- spatial processing in patients with cerebral gliomas: a cal indexes of a partial frontal syndrome in patients with neurocognitive assessment. J Neurooncol 49: 235–248. nonfrontal gliomas. Neuropsychology 6: 315–326. Jemal A, Siegel R, Ward E et al. (2007). Cancer statistics, Lupien SJ, Gillin CJ, Hauger RL (1999). Working mem- 2008. CA Cancer J Clin 58: 71–96. ory is more sensitive than declarative memory to the Johnson MR, Hageboutros A, Wand K, High L, Smith JB, acute effects of corticosteroids: a dose-response study in Diasio RB (1999). Life-threatening toxicity in a dihy- humans. Behav Neurosci 113: 420–430. dropyrimidine dehydrogenase-deficient patient after Maldjian JA, Detre JA, Killgore WES et al. (2001). Neuropsy- treatment with topical 5-fluorouracil. Clin Cancer Res 5: chologic performance after resection of an activation 2006–2011. cluster involved in cognitive memory function. AJR Am Kaleita TA, Wellich DK, Graham CA et al. (2006). Pilot study J Roentgenol 176: 541–544. of modafinil for treatment of neurobehavioral dysfunc- Marin RS (1991). Apathy: a neuropsychiatric syndrome. J tion and fatigue in adult patients with brain tumors. Proc Neuropsychiatry Clin Neurosci 3: 243–254. ASCO 24(18S): 58S, [abstract 1503]. Marin RS (1996). Apathy: concept, syndrome, neural mech- Karnofsky DA, Abelmann WH, Craver LF (1948). The use anisms, and treatment. Semin Clin Neuropsychiatry 1: of nitrogen mustards in the palliative treatment of car- 304–314. cinoma. Cancer 1: 634–656. Mattson RH (2004). Cognitive, affective, and behavioral Kayl AE, Meyers CA (2003). Does brain tumor histology side events in adults secondary to antiepileptic drug use. influence cognitive function? Neurooncology 5: 255–260. Rev Neurol Dis 1 [Suppl 1]: S10–S17. Klein M, Heimans JJ (2004). The measurement of cognitive MauerM,StuppR,TaphoornMJet al. (2007). The prog- functioning in low-grade glioma patients after radiother- nostic value of health-related quality-of-life data in pre- apy [Letter]. J Clin Oncol 22: 966–967. dicting survival in glioblastoma cancer patients: results KleinM,TaphoornMJB,HeimansJJet al. (2001). Neurobe- from an international randomised phase III EORTC Brain havioral status and health-related quality of life in newly Tumour and Radiation Oncology Groups, and NCIC diagnosed high-grade glioma patients. J Clin Oncol 19: Clinical Trials Group study. Br J Cancer 97: 302–307. 4037–4047. McNair DM, Lorr M, Droppleman LF (1992). Profile of Mood Klein M, Engelberts NHJ, van der Ploeg HM et al. (2003a). States Manual. San Diego, CA: Educational and Industrial Epilepsy in low-grade gliomas: the impact on cognitive Testing Service. function and quality of life. Ann Neurol 54: 514–520. Mendoza TR, Wang XS, Cleeland CS et al. (1999). The rapid Klein M, Postma TJ, Taphoorn MJB et al. (2003b). The prog- assessment of fatigue severity in cancer patients: use of nostic value of cognitive functioning in the survival of the Brief Fatigue Inventory. Cancer 85: 1186–1196. patients with high-grade glioma. Neurology 61: 1796– Mesulam M (2000). Attentional networks, confusional 1798. states, and neglect syndromes. In Mesulam M (ed.) Prin- Lannering B, Marky I, Lundberg A et al. (1990). Long-term ciples of Behavioral and Cognitive Neurology (pp. 174– sequelae after pediatric brain tumors: their effect on dis- 256). New York: Oxford University Press. ability and quality of life. Med Pediatr Oncol 18: 304–310. Meyers CA, Boake C (1993). Neurobehavioral disorders in Lezak MD. (1995). Neuropsychological Assessment (3rd brain tumor patients: rehabilitation strategies. Cancer edn.). New York: Academic Press. Bull 45: 362–364. Lezak MD, Howieson DB, Loring DW (1994). Neuropsycho- Meyers CA, Brown PD (2006). Role and relevance of neu- logical Assessment. New York: Oxford University Press. rocognitive assessment in clinical trials of patients with Levin VA, Yung WK, Bruner J et al. (2002). Phase II study of CNS tumors. JClinOncol24: 1305–1309. accelerated fractionation radiation therapy with carbo- Meyers CA, Wefel JS (2003). The use of the mini-mental platin followed by PCV chemotherapy for the treatment state examination to assess cognitive functioning in can- of anaplastic gliomas. Int J Radiat Oncol Biol Phys 53: 58– cer trials: no ifs, ands, buts, or sensitivity. J Clin Oncol 21: 66. 3357–3358. Levine DW, Kripke DF, Kaplan RM et al. (2003). Reliability Meyers CA, Kudelka AP, Conrad CA et al. (1997). Neurotox- and validity of the Women’s Health Initiative Insomnia icity of CI-980, a novel mitotic inhibitor. Clin Cancer Res Rating Scale. Psychol Assess 15: 137–148. 3: 419–422. 168 Section 2. Effects of cancer and cancer treatment on cognition

Meyers CA, Weitzner MA, Valentine AD, Levin VA (1998). dimensional, self-rating scale for assessment of fatigue Methylphenidate therapy improves cognition, mood, in cancer patients. J Pain Symptom Manage 19: 5–14. and function of brain tumor patients. J Clin Oncol 16: Packer RJ, Gurney JG, Punyko JA et al. (2003). Long-term 2522–2527. neurologic and neurosensory sequelae in adult survivors Meyers CA, Hess KR, Yung WKA, Levin VA (2000). Cognitive of a childhood brain tumor: childhood cancer survivor function as a predictor of survival in patients with recur- study. J Clin Oncol 21: 3255–3261. rent malignant glioma. J Clin Oncol 18: 646–650. Patchell RA, Tibbs PA, Regine WF et al. (1998). Postopera- Meyers CA, Smith JA, Bezjak A et al. (2004). Neurocognitive tive radiotherapy in the treatment of single metastases to function and progression in patients with brain metas- the brain: a randomized trial. J Am Med Assoc 280: 1485– tases treated with whole-brain radiation and motexafin 1489. gadolinium: results of a randomized phase III trial. JClin Raida M, Schwabe W, Hausler P et al. (2001). Prevalence of a Oncol 22: 157–165. common point mutation in the dihydropyrimidine dehy- Monje ML, Mizumatsu S, Fike JR, Palmer TD (2002). Irra- drogenase (DPD) gene within the 5-splice donor site of diation induces neural precursor-cell dysfunction. Nat intron 14 in patients with severe 5-fluorouracil (5-FU)- Med 8: 955–962. related toxicity compared with controls. Clin Cancer Res Monje ML, Toda H, Palmer TD (2003). Inflammatory block- 7: 2832–2839. ade restores adult hippocampal neurogenesis. Science Report of the Brain Tumor Progress Review Group (2005). 302: 1760–1765. National Institute of Neurologic Disorders and Stroke. Monje ML, Vogel H, Masek M, Ligon KL, Fisher PG, Palmer Accessed 27 March, 2008 from http://www.ninds.nih. TD (2007). Impaired human hippocampal neurogenesis gov/find people/groups/brain tumor prg/btprgreport. after treatment for central nervous system malignancies. htm. Ann Neurol 62(5): 515–520. Rozental JM Levine RL, Nickles RJ, Dobkin JA, Hanson JM Moulder JE, Robbins ME, Cohen EP et al. (1998). Pharma- (1990). Cerebral diaschisis in patients with malignant cologic modification of radiation-induced late normal glioma. J Neurooncol 8: 153–161. tissue injury. Cancer Treat Res 93: 129–151. Salander P, Karlsson T, Bergenheim T et al. (1995). Mulhern RK, Khan RB, Kaplan S et al. (2004). Short-term Long-term memory deficits in patients with malignant efficacy of methylphenidate: a randomized, double- gliomas. J Neurooncol 25: 227–238. blind, placebo-controlled trial among survivors of child- Scheibel RS, Meyers CA, Levin VA (1996). Cognitive dys- hood cancer. J Clin Oncol 22: 4795–4803. function following surgery for intracerebral glioma: MurphyGM,KremerC,RodriguesHE,SchatzbergAF influence of histopathology, lesion location, and treat- (2003). Pharmacogenetics of antidepressant medication ment. J Neurooncol 30: 61–69. intolerance. Am J Psychiatry 160: 1830–1835. Shaw E, Arusell R, Scheithauer B et al. (2002). Prospective Murray KJ, Scott C, Zachariah B et al. (2000). Importance randomized trial of low- versus high-dose radiation ther- of the mini-mental status examination in the treatment apy in adults with supratentorial low-grade glioma: ini- of patients with brain metastases: a report from the tial report of a North Central Cancer Treatment Group/ Radiation Therapy Oncology Group protocol 91–04. Int Radiation Therapy Oncology Group/Eastern Coopera- J Radiat Oncol Biol Phys 48: 59–64. tive Oncology Group study. J Clin Oncol 20: 2267– O’Connor NM, Mayberrg MR (2000). Effects of radia- 2276. tion on cerebral vasculature: a review. Neurosurgery 46: Shaw EG, Rosdhal R, D’Agostino RB Jr et al. (2006). Phase 138–151. II study of donepezil in irradiated brain tumor patients: O’Dwyer JP, Clabby C, Crown J, Barton DE, Hutchinson effect on cognitive function, mood, and quality of life. J M (2005). Fragile X-associated tremor/ataxia syndrome Clin Oncol 24: 1415–1420. presenting in a woman after chemotherapy. Neurology Shehata N, Pater A, Tang SC (1999). Prolonged severe 65: 331–332. 5-fluorouracil-associated neurotoxicity in a patient with Oken MM, Creech RH, Tormey DC et al. (1982). Toxicity dihydropyrimidine dehydrogenase deficiency. Cancer and response criteria of the Eastern Cooperative Oncol- Invest 17: 201–205. ogy Group. Am J Clin Oncol 5: 649–655. Sherer M, Meyers CA, Bergloff P (1997). Efficacy of post- Okuyama T, Akechi T, Kugaya A et al. (2000). Development acute brain injury rehabilitation for patients with pri- and validation of the cancer fatigue scale: a brief, three- mary malignant brain tumors. Cancer 80: 250–257. Chapter 11. High-grade gliomas 169

Sherman AM, Jaeckle K, Meyers CA (2002). Pre-treatment Ulrich CM, Robien K, Sparks R (2003). Pharmacogenetics cognitive performance predicts survival in patients with and folate metabolism – a promising direction. Pharma- leptomeningeal disease. Cancer 95: 1311–1316. cogenomics 3: 299–313. Sohlberg MM, Maateer CA (1989). Introduction to Cognitive Varni JW, Limbers CA, Burnwinkle TM (2007). How young Rehabilitation.NewYork:Guilford. can children reliably and validly self-report their health- Spielberger C, Gorsuch R, Lushene R (1970). State-Trait related quality of life? An analysis of 8,591 children across Anxiety Inventory. Palo Alto, CA: Consulting Psycholo- age subgroups with the PedsQL 4.0 Generic Core Scales. gists. Health Qual Life Outcomes 5: 1–13. Steinbach JP, Blaicher H-P, Herrlinger U et al. (2006). Sur- Wade DT, Collin C (1988). The Barthel ADL Index: a stan- viving glioblastoma for more than 5 years: the patient’s dard measure of physical disability. Int Disabil Stud 10: perspective. Neurology 66: 239–242. 64–67. Stupp R, Mason WP, van den Bent MJ et al. (2005). Radio- Ware JE, Sherbourne CD (1992). The MOS 36-item short- therapy plus concomitant and adjuvant temozolomide form health survey (SF-36): I. Conceptual framework and for glioblastoma. New Engl J Med 352: 987–996. item selection. Med Care 30: 473–483. Taphoorn MJ, Klein M (2004). Cognitive deﬁcits in adult Wechsler D (1981). Wechsler Adult Intelligence Scale- patients with brain tumours. Lancet Neurol 3: 159–168. Revised. San Antonio, TX: The Psychological Corpora- Taphoorn MJ, Heimans JJ, Snoek FL et al. (1992). Assess- tion. ment of quality of life in patients treated for low-grade Wei CW, Gui G, Mikulis DJ (2007). Tumor effects on cerebral glioma: a preliminary report. J Neurol Neurosurg Psychi- white matter as characterized by diffusion tensor trac- atry 55: 372–376. tography. Can J Neurol Sci 34: 62–68. Taphoorn MJ, Stupp R, Coens C et al. (2005). Health-related Weitzner MA, Meyers CA, Gelke CK et al. (1995). The quality of life in patients with glioblastoma: a random- Functional Assessment of Cancer Therapy (FACT) scale: ized controlled trial. Lancet Oncol 6: 937–944. development of a brain subscale and revalidation of the Taylor BV, Buckner JC, Cascino TL et al. (1998). Effects general version (FACT-G) in patients with primary brain of radiation and chemotherapy on cognitive function in tumors. Cancer 75: 1151–1161. patients with high-grade glioma. J Clin Oncol 16: 2195– Wellisch DK, Kaleita TA, Freeman D, Cloughesy T, Gold- 2201. man J (2002). Predicting major depression in brain tumor Toﬁlon PJ, Fike JR (2000). The radioresponse of the central patients. Psychooncology 11: 230–238. nervous system: a dynamic process. Radiat Res 153: 357– Wen PY, Schiff D, Kesari S, Drappatz J, Gigas DC, Doherty 370. L (2006). Medical management of patients with brain Torres IJ, Mundt AJ, Sweeney PJ et al. (2003). A longitudi- tumors. J Neurooncol 80: 313–332. nal neuropsychological study of partial brain radiation in Young AH, Sahakian BJ, Robbns TW et al. (1999). The effects adults with brain tumors. Neurology 60: 1113–1118. of chronic administration of hydrocortisone on cognitive Ulrich CM, Yasui Y, Storb R et al. (2001). Pharmacogenet- functioning normal male volunteers. Psychopharmacol- ics of methotrexate: toxicity among marrow transplanta- ogy 145: 260–266. tion patients varies with the methylenetetrahydrofolate Young GS (2007). Advanced MRI of adult brain tumors. reductase C677T polymorphism. Blood 98: 231–234. Neurol Clin N Am 25: 947–973. 12

Brain metastases

Deepak Khuntia, Beela S. Mathew, Christina A. Meyers, Sterling Johnson, and Minesh P. Mehta

Introduction preventive and therapeutic strategies based on current understanding. Brain metastasis is the commonest intracranial tumor in adults. In the United States, approximately 170 000 patients are diagnosed with brain metas- Overview of brain metastases tases every year (Greenberg et al., 1999; Mehta & Tremont-Lukas, 2004). The rise in incidence Epidemiology is attributed to a number of factors including increased life expectancy, improved control of sys- Brain metastasis is a major debilitating complica- temic disease, and better imaging capabilities that tion affecting cancer patients. Autopsy data indicate facilitate diagnosis of smaller lesions (Wen et al., that approximately 24% of adult cancer patients 2001). While it is recognized that the overall progno- develop metastatic brain disease during the course sis of these patients remains poor, newer treatment of their cancer (Posner, 1995). The peak age group is methods have led to improved survival in subsets 55–65 years, reflecting the incidence of primary can- of patients. Patients with brain metastases often cers that occur mainly in the fifth and sixth decades suffer from a variety of neurological, cognitive, and of life (Lim et al., 2004). Among the primary tumor emotional difficulties. It is known that even subtle types, lung cancer accounts for about 50% of all impairments of cognitive function can adversely brain metastases (Ellis & Gregor, 1998; Kelly & Bunn, affect the quality of life, an issue that was largely 1998; Postmus et al., 2000). Other tumors caus- ignored earlier due to the dismal outcome. However, ing brain metastases are breast cancer (15%–20%), in the changing scenario of improved survival, rec- melanoma (10%), unknown primary (10%–15%), ognizing the effects of the disease and its therapies colorectal cancer (2%–12%), kidney (1%–8%) and on neurocognitive outcomes is important in for- thyroid (1%–10%) (Cappuzo et al., 2000; Lassman & mulating treatment modifications and strategies for DeAngelis, 2003; Wen et al., 2001). In children the rehabilitation that will enable patients to maximize incidence of brain metastases is approximately 6%– their functional ability. This chapter briefly reviews 10% and the most common associated solid primary the incidence and management of brain metastases tumors are sarcoma, Wilms’ tumor, neuroblastoma, as relevant to neurocognitive problems in cancer and germ cell tumor (Graus et al., 1983; Vannucci & patients, discusses the etiology and pathogenesis Baten, 1974). The majority of patients have involve- of cognitive deficits in these patients, and suggests ment of the cerebral hemispheres (80%) while the

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

170 Chapter 12. Brain metastases 171

Table 12.1. Classiﬁcation of brain metastases patients Table 12.2. Response to whole-brain radiation by prognosis (RPA analysis) (Gaspar et al., 1997). KPS, therapy in patients with brain metastases from various Karnofsky Performance Scale; RPA, recursive tumor types (Nieder et al., 1997) (n = 108 patients) partitioning analysis Percentage of patients Median survival showing response by Prognostic group (months) Source of primary tumor CT (%)

Class I Small-cell carcinoma 37 (Age<65, KPS ≥70, controlled primary, 7.1 Breast cancer 35 no extracranial metastases) Squamous cell carcinoma 25 Adenocarcinoma (non-breast) 14 Class II Renal cell carcinoma 0 (KPS ≥70 with age ≥65 OR 4.2 Melanoma 0 uncontrolled primary OR extracranial metastases)

Class III Factors associated with longer survival in brain (KPS <70) 2.3 metastases patients are younger age, absence of extracranial disease, better performance status, and cerebellum is involved in 15% and the brainstem in single lesions (Gaspar et al., 1997). Improvement <5% of patients (Sawaya & Bindal, 2001). in cognitive function following treatment has also been reported to be a good prognostic factor Clinical presentation (Curran et al., 2002). Response to treatment may be influenced by the Brain metastases should be considered in the dif- tumor type being treated. Nieder and colleagues ferential diagnosis of any cancer patient develop- reported CT responses in 108 patients based on ing new neurological symptoms or signs. Com- tumor type following whole-brain radiation therapy mon presentations include headache (24%–53%), (WBRT) alone (Nieder et al., 1997) (see Table 12.2). focal weakness (16%–40%), altered mental status Complete responses were noted in 24% of all (24%–31%), seizures (15%), and ataxia (9%–20%) patients after WBRT with partial responses in 35%. (Nussbaum et al., 1996; Schellinger et al., 1999). Small cell lung cancers had the best response rate with 37% being responders, while renal cell and Prognosis malignant melanoma were the worst with 0%. The prognosis of patients with brain metastases Despite higher response rates with small cell lung is generally poor. The overall median survival for cancer that has metastasized to the brain, patients an untreated patient and patients receiving radi- do similarly poorly, with median survival ranging ation is approximately 1 month and 4–6 months 6.5–8.5 months after treatment (Quan et al., 2004). respectively (Nussbaum et al., 1996; Posner, 1995; For patients undergoing surgery, however, histol- Sundstrom et al., 1998; Zimm et al., 1981). Ret- ogy has not been found to be an important factor rospective recursive partitioning analysis (RPA) of (Agboola et al., 1998). prognostic factors performed on more than 1100 patients enrolled in Radiation Therapy Oncol- Treatment ogy Group (RTOG) trials identified three well- defined prognostic classes with significantly dif- The aims of treatment of brain metastases are to ferent median survivals (Gaspar et al., 1997) (see maintain quality of life and functional status while Table 12.1). maximizing duration of survival (Renschler et al., 172 Section 2. Effects of cancer and cancer treatment on cognition

2003). Traditionally the mainstay of treatment has and several retrospective and prospective trials consisted of fractionated external WBRT. Phase III have proven that radiosurgery along with WBRT trials have reported that WBRT results in a median improves local control and survival for patients with survival of 4–6 months and improves neurologic unresectable solitary or multiple brain metastases function in about half of patients (Borgelt et al., if they have favorable characteristics (Alexander & 1980; Gaspar et al., 1997). Common fractionation Loeffler, 1999; Aoyama et al., 2006; Breneman et al., schemes include 30 Gy in 10 fractions, 37.5 Gy in 1997; Chougule et al., 2000; Kondzoilka et al., 1999; 15 fractions, and 40 Gy in 20 fractions. Randomized Mehta & Tremont-Lukas, 2004; Seung et al., 1998; trials by Kondzoilka et al. (1999) and Patchell et al. Young, 1998). Sanghavi et al. (2001) analyzed radio- (1990) indicate that actuarial local control at 1 year surgical data from 10 institutions and stratified 502 after WBRT alone ranges from 0% to 14%. These data patients into three RPA classes as defined by the suggest that long-term control of gross brain metas- RTOG. Patients treated with radiosurgery boost in tases after WBRT alone is unlikely. Owing to the poor addition to WBRT showed improved median sur- survival outcomes with WBRT alone, newer surgi- vival (p < 0.05) in all three classes compared to cal, radiation, and chemotherapeutic approaches patients who received WBRT alone. Three prospec- have been tried with promising results in subsets of tive randomized trials comparing stereotactic radio- patients. surgeryplusWBRTtoWBRThavebeencon- Patchell et al. (1990) demonstrated superior local ducted (Andrews et al., 2004; Chougule et al., 2000; control (20% vs. 52%, p < 0.02) and median survival Kondzoilka et al., 1999). Two studies with 27 and (40 vs. 15 weeks, p < 0.01) as well as prolonged time 104 patients respectively demonstrated improved to brain metastases recurrence (59 vs. 21 weeks, local control rates favoring the radiosurgery arm but p < 0.005) with the addition of surgery for patients showed no significant difference in median survival with solitary brain metastases. Likewise another (Chougule et al., 2000; Kondzoilka et al., 1999). How- study of 63 patients also reported improvement in ever, results from a phase III multi-institutional trial overall survival and functional independence for (RTOG 9508) of 333 patients reported statistically combined treatment with surgery and radiation significant improvement in median survival with compared to radiation alone (Noordjik et al., 1994). the addition of radiosurgery to WBRT for patients Patchell and colleagues (1998) have also compared with solitary brain metastases (6.5 vs. 4.9 months, surgery alone versus surgery and WBRT. This study p = 0.039), higher local control at 1 year (82% vs. noted significant differences in favor of combined 71%, p = 0.01) and an increase in the likelihood of treatment for patients with tumor recurrence (18% stable or improved KPS and decreased steroid use vs. 70%, p < 0.001), median time to recurrence (220 at 6 months (43% vs. 27%, p = 0.03) for all patients. weeks vs. 26 weeks, p < 0.001) and fewer deaths due No differences were seen when assessing mental to neurological causes (14% vs. 44%, p = 0.003). Cur- status, largely because only the Mini-Mental Status rently surgical resection followed by WBRT is rec- Examination was used for evaluation. On multivari- ommended for patients with a single brain metas- ate analysis RPA class I patients (p < 0.0001) and tasis who have none or minimal systemic disease those with favorable histology (p = 0.0121) did sig- and a Karnofsky Performance Score (KPS) higher nificantly better (Andrews et al., 2004). Thus radio- than 70. surgery boost is recommended as a standard treat- Radiosurgery, a technique that delivers accurately ment for patients with a single unresectable brain targeted highly conformal radiation to a defined metastasis or for RPA class I patients. A phase III lesion, has been evaluated as a non-invasive alter- trial comparing radiosurgery alone to radiosurgery native to surgery in solitary brain metastases and and WBRT has recently been completed in Japan more recently in multiple lesions too. High doses, (Aoyama et al., 2006). In this trial, Aoyama and generally in a single treatment, are prescribed colleagues have demonstrated that the addition of Chapter 12. Brain metastases 173

WBRT to radiosurgery decreased local recurrence and course of cognitive dysfunction is limited from 46.8% in the WBRT + stereotactic radiosurgery owing to a lack of formal and systematic evaluation (SRS) group compared to 76.4% for the SRS-alone of neuropsychometric morbidity in a population of group (p < 0.001). Furthermore, salvage treatments patients expected to have short longevity. Newer were less likely in patients receiving the whole- treatment approaches have clearly demonstr- brain radiation. Despite the improvement in local ated improved survival in subpopulations of brain control, no difference in overall survival was rea- metastases patients with favorable prognostic lized. Another study comparing SRS with or with- factors. In this context neurocognitive and quality out WBRT is currently accruing patients through of life issues are a growing concern for survivors. the American College of Surgeons Oncology Group Further, neurocognitive function has been demon- (ACOSOG-Z0300). strated to be a predictor of survival for patients It is generally believed that most chemotherapeu- with brain metastases and primary brain tumor tic agents are not useful in brain tumors due to patients, and is considered to be a sensitive, viable their inability to penetrate the blood–brain barrier and important endpoint that measures clinical (BBB). Recent understanding that metastatic tumor benefit on patient functioning (Meyers & Hess, growth causes the upregulation of angiogenic fac- 2003). tors and neovascularization with a disrupted BBB in the new vessels has led to a renewed inter- Incidence est in using chemotherapy for brain metastases. Temozolomide, a novel alkylating agent with 100% Although comprehensive data on the magnitude of bioavailability after oral administration and high cognitive dysfunction in patients with brain metas- cerebrospinal fluid (CSF) penetrability, has been tases are limited, it has been demonstrated that tested with promising results (Abrey et al., 2001; the majority have significant neurocognitive defects Antonadou et al., 2002a, 2002b; Christodoulou et al., compromising their quality of life. Neurocognitive 2001; Verger et al., 2003). impairment affecting functional independence may Radiosensitizers are agents that enhance the even be more common than physical disability effects of radiation. Two agents – motexafin gadolin- (Meyers & Boake, 1993). Deficits range from sub- ium (MGd), a redox modulator that induces apopto- tle problems with concentration, memory, affect, sis, and RSR13, an allosteric modifier of hemoglobin and personality to severe dementia. One of the early that augments oxygenation of hypoxic tissues – reports of cognitive decline in patients with brain have been tested in brain metastases patients metastases documented dementia in 11% patients (Bradley & Mehta, 2004; Mehta et al., 2002, 2003, who survived 1 year after WBRT (DeAngelis et al., 2006). Although neither agent could demonstrate 1989). This was considered to be a function of the an overall survival advantage over WBRT alone, large fraction sizes employed for WBRT at that time; cohorts of patients who may benefit have been in fact none of the patients treated with conven- identified, and prospective trials evaluating the tional schedules and doses developed serious long- efficacy of radiosensitizers in these subsets are term dementia. Prospective studies conducted in underway. small cell lung carcinoma (SCLC) patients receiving prophylactic cranial irradiation (PCI) demonstrated cognitive dysfunction prior to radiotherapy. Neurocognitive impairment in brain Komaki et al. (1995) reported impaired cognitive metastases tests in 97% patients at baseline. Two large random- The majority of brain metastases patients suffer ized trials of PCI that incorporated neuropsycho- from some degree of neuropsychological impair- metric testing into the protocols found that 40%– ment. Our knowledge about the nature, severity, 60% patients showed significant abnormalities at 174 Section 2. Effects of cancer and cancer treatment on cognition

the time of randomization (Arriagada et al., 1995; of cognitive deficits in a significant number of Gregor et al., 1997). Cognitive deficits were unre- patients prior to treatment and correlation of cog- lated to age, gender or previous therapy (Gregor nitive decline with tumor progression in patients et al., 1997). That being said, recent data from who have received WBRT both clearly indicate that Li et al. (2006) have shown that with treatment and the cancer itself is an important cause of cognitive subsequent reduction in tumor burden in the brain, dysfunction in brain metastases patients. In some neurocognitive function can be improved with instances the cognitive deficits pre-date the devel- time. opment of metastatic disease to the brain, possi- Recognizing the importance of neurobehavioral bly suggesting the presence of unidentified parane- outcome measures in brain tumor trials, the RTOG oplastic processes (Arriagada et al., 1995; Komaki conducted a phase II feasibility study (BR0018) et al., 1995; Regine et al., 2001). The type and degree of systematic neurocognitive assessments in 55 of deficits depend on the location of the lesion, brain metastases patients with excellent compli- but they also correlate with the total volume of ance rates prior to (95%), upon completion of (84%), brain metastatic disease (as opposed to the num- and 1 month after (70%) WBRT (Regine et al., ber of brain metastases alone) (Turkheimer et al., 2004). Another single institution study in 30 patients 1990). Patients with left hemispheric tumors gen- reported 100% compliance, proving the feasibility erally have language dysfunction, verbal, learning, of conducting such assessments in a routine clin- memory and right-sided motor dexterity impair- ical context (Herman et al., 2003). Subsequently, ment. Right hemispheric tumors cause impairment prospective cognitive testing was incorporated into of visual-perception skills and left-sided motor dex- a large randomized pharmaceutical trial of 401 terity. Changes in mentation are common with brain metastases patients in which Meyers et al. tumors of the frontal lobe (Scheibel et al., 1996). (2004) demonstrated baseline cognitive impairment Cerebellar cancers are also implicated in cognitive in 90.5% of patients, with multiple abnormalities in malfunction (Gottwald et al., 2004; Karatekin et al., most patients. This study establishes that neurocog- 2000). nitive dysfunction is a significant problem for brain Tumor-related events such as edema and hem- metastases patients. orrhage cause disruption of sensitive afferent and efferent connections between the frontal region and other parts of the brain (Herman et al., 2003). This Etiology and pathogenesis accounts for the impairment in executive frontal The etiology of neurocognitive impairment in brain lobe functions manifested as apathy, lack of motiva- metastases patients is diverse. Etiologies include tion and spontaneity, impaired attention and mem- direct damage due to the cancer, indirect effects ory, which are noted even in the absence of a frontal of cancer (paraneoplastic), and effects of can- tumor (Lilja et al., 1992). cer treatments on the brain. In addition, these Apart from the direct insult exerted by the tumor, patients often have co-existing neurological and neuropsychiatric difficulties could also be a result of psychiatric disorders that affect their cognition and other processes such as paraneoplastic effects par- mood. ticularly in patients with lung cancer (Komaki et al., 1995). Central nervous system (CNS) paraneoplastic disorders are grouped into a clinico-pathological Neurocognitive deficits due to cancer entity of paraneoplastic encephalomyelitis (PEM) Cognitive problems in brain tumor patients pre- characterized by multifocal autoimmune injury ceding cancer treatments such as surgery, radia- to the brain, spinal cord, dorsal root ganglia, tion or chemotherapy are typically related to the and autonomic ganglia. Patients with brain damage inflicted by the tumor itself. The evidence lesions often present with memory loss, affective Chapter 12. Brain metastases 175

disorders, and cognitive deterioration (Dropcho, lar interactions between vasculature and parenchy- 2005). mal cells particularly oligodendrocytes, which are important for myelination. Oligodendrocyte death occurs due to either a direct p53-dependent radi- Cognitive sequelae of cancer therapy ation apoptosis, or exposure to radiation-induced tumor necrosis factor α (TNF-α) (Cammer, 2000; Brain damage consequent to cancer therapy is Chow et al., 2000). Post-radiation injury to the vas- a well-recognized cause of cognitive decline in culature involves damage to the endothelium lead- cancer patients. Contributory factors include spe- ing to platelet aggregation and thrombus formation cific antineoplastic treatments such as radiation, initially, followed by abnormal endothelial prolifer- chemotherapy, hormonal therapy, immunotherapy, ation and intraluminal collagen deposition (Burger and surgery as well as supportive medications such et al., 1979; Crossen et al., 1994). Hippocampal- as corticosteroids and anticonvulsants. dependent functions of learning, memory, and spatial information processing seem to be preferen- tially affected by radiation (Monje & Palmer, 2003). Animal studies have shown that doses as low as Radiotherapy 2 Gy produce apoptosis in the proliferating cells in Most patients with brain metastases receive whole- the hippocampus leading to decreased repopula- brain radiotherapy in current practice. The delete- tive capacity (Peissner et al., 1999). The reader is rious effects of radiation on the CNS are well docu- directed to Chapter 7 for a more detailed review mented. The response to irradiation of the brain has of the biological bases of radiation injury to the been classically divided into three categories based brain. on the timing of the onset of symptoms (Sheline Studies of neurocognitive functioning in patients et al., 1980). Acute effects occur during the first few surviving 1 year after radiation yield conflicting weeks of treatment and are characterized by drowsi- results. DeAngelis and colleagues reported demen- ness, headache, nausea, vomiting, and worsening tia in 11% patients who received WBRT using daily focal deficits. These symptoms are believed to be fractions of 3–6 Gy (DeAngelis et al., 1989). No- due to cerebral edema and are reversed by treat- patients receiving <3 Gy per fraction experienced ment with corticosteroids. Subacute encephalopa- this. Likewise no significant cognitive decline was thy (early delayed reaction) occurring at 1–6 months reported for treatments with 40 Gy in 20 fractions after completion of radiation is thought to be sec- (Penitzka et al., 2002), 30 Gy in 10 fractions or ondary to diffuse demyelination (Boldrey & She- 54.4 Gy in an accelerated hyperfractionated regimen line, 1967; Van der Kogel, 1986). Typical symptoms (Regine et al., 2001). Data from prospective trials of include headache, somnolence, fatigability, and PCI in SCLC patients showed no significant cogni- deterioration of pre-existing deficits that resolve tive deterioration following radiotherapy with frac- within several months. Late delayed effects appear tion sizes of less than 3 Gy (Arriagada et al., 1995; more than 6 months after radiation and are irre- Gregor et al., 1997; Komaki et al., 1995; Van Ooster- versible and progressive (Kramer, 1968). This is hout et al., 1995). Fraction size, advanced age (>60 thought to be a result of white matter damage due to years), higher total dose, volume of brain irradiated, vascular injury, demyelination, and necrosis. Symp- chemotherapy, and co-morbid vascular risk factors toms range from mild lassitude to significant mem- such as diabetes mellitus influence the incidence ory loss, and severe dementia (Schultheiss et al., of radiation-induced injury to the brain and may 1995). The pathophysiology of radiation-induced account for the differences in reported incidences neurocognitive impairment thus involves dynamic, of cognitive deficits (Crossen et al., 1994; Lee et al., complex processes including inter- and intracellu- 2002). 176 Section 2. Effects of cancer and cancer treatment on cognition

Systemic anti-cancer therapy Positron emission tomography (PET) imaging has demonstrated greater prefrontal hypometabolism The majority of brain metastases patients receive in women treated with chemotherapy and tamox- systemic anticancer therapies for control of primary ifen compared to chemotherapy alone (Silverman or extracranial metastatic disease, either before or et al., 2003). The hippocampus contains testos- after the diagnosis of brain metastases. Many agents terone receptors as well as estradiol receptors. Hor- administered are known to have effects on brain monal challenges through luteinizing-hormone- function. releasing (LHRH) agonists thus have the potential Chemotherapy-related cognitive impairment has to affect hippocampal function. However, there are been reported in 17%–75% patients (Ahles et al., inconsistent findings regarding neurocognitive dys- 2002; Brezden et al., 2000; Schagen et al., 1999; function in the patients treated with LHRH agonists Wieneke & Dienst, 1995): rather than dementia, (Green et al., 2002; Salminen et al., 2003). subtle neurocognitive variations are more com- Other systemic antineoplastic agents that have mon. A prospective trial evaluating neuropsycho- the potential to produce neurotoxic effects include logical function in breast cancer patients under- cytokines such as interferons and interleukins. going chemotherapy reported cognitive decline Cytokines are reported to have direct and in- between baseline and short-term (6 months post direct effects on CNS function through alteration chemotherapy) assessments in 61% of patients of, for example, neurotransmitters and neuroen- (Wefel et al., 2004b). The most commonly affected docrine function (Kelley et al., 2003). More than 50% domains included attention, learning, and process- of patients receiving cytokine therapy are reported ing speed, consistent with disruption of frontal net- to have neurocognitive impairment (Meyers & work systems (Wefel et al., 2004b). Etiologic mecha- Abbruzzese, 1992). Functional neuroimaging stud- nisms underlying chemotherapy-induced cognitive ies have demonstrated frontal region abnormalities dysfunction may differ according to the agents used. consistent with the neurocognitive deficits in such Drugs such as methotrexate and 5-fluorouracil are patients (Juengling et al., 2000; Meyers et al., 1994). particularly neurotoxic. Cisplatin, etoposide and The interested reader is directed to Chapter 8 for a vincristine may cause white matter injury (Komaki more detailed review of chemotherapy-related cog- et al., 1995). Potential reasons for brain damage nitive impairment. include direct injury to the gray and white matter, microvascular injury, and secondary insults due to Surgery immune-mediated inflammatory responses (Wefel et al., 2004a). Surgery has the potential to induce cognitive dys- Alterations of an individual’s hormonal milieu function secondary to brain injury and is most likely are associated with neurocognitive impairments related to the tumor’s location. There are conflict- (Wefel et al., 2004a). Hormonal manipulation for ing reports as to the contribution of surgery to cog- control of systemic cancer is common in can- nitive decline in brain tumor patients. While there cers of the breast and prostate. Estrogen recep- are reports of cognitive decline after neurosurgery tors have been detected in areas of the brain in children (Fontanella et al., 2003; Grill et al., 2004; important for cognitive functioning including the Peace et al., 1997), other investigators could not hypothalamus, anterior pituitary, amygdala, and demonstrate any impairment of cognitive function hippocampus (McEwen & Alves, 1999). Assess- following surgery compared to the pre-operative ments of neurocognitive function in women receiv- status (Friedman et al., 2003; Hutter et al., 1997; ing anti-estrogens have demonstrated impairment Tucha et al., 2003). There are no data regarding in memory, executive function, and motor co- the effect of surgery on cognitive function in brain ordination (Rich & Maki, 1999; Varney et al., 1998). metastases patients. Chapter 12. Brain metastases 177

Adjuvant medications agent include insomnia, cardiac symptoms, and anxiety or nervousness. It inhibits the metabolism of Apart from the neurotoxic effects of specific anti- certain drugs such as warfarin and tricyclic antide- cancer therapy, other medications such as steroids, pressants, and therefore plasma levels of these anticonvulsants, and pain medications used as agents must be monitored. The drug also holds a adjuncts in brain metastases patients may cause relative contraindication in patients receiving mon- neurocognitive and behavioral symptoms. Gluco- amine oxidase inhibitors. corticoids are associated with a 5%–50% inci- In a series by Meyers et al. (1998), 30 patients with dence of psychiatric syndromes including eupho- primary brain tumors with evidence of neurocogni- ria, mania, insomnia, restlessness, and increased tive deficits were administered methylphenidate at motor activity and memory dysfunction (Kershner doses between 10 mg and 30 mg twice daily (bid). & Wang-Cheng, 1989; Wefel et al., 2004a; Wolkowitz In this study, subjective improvement was seen et al., 1990). Anticonvulsants are associated with a in all patients receiving 30 mg bid. This included sixfold increased risk of deficits in perception, psy- improvement in energy, concentration, mood, and chomotor speed, attention and executive function- ambulation. Toxicity from the drug in this trial was ing (Taphoorn, 2003). Klein et al. (2003) demon- quite minimal. The presumed mechanism for this strated that both seizure frequency and the use of improvement in neurocognitive function relates to anticonvulsants have an adverse impact on neu- the dopaminergic innervation of the mesolimbic rocognitive function. Pain medications may cause system, which mediates subcortical function, and sedation and associated difficulties in neurocogni- methylphenidate could help with improvement in tive function. motivation and drive by stimulating this system. Therefore, methylphenidate could be considered in the management of neurocognitive deficits in brain Management of neurocognitive deficits metastasis patients. Certainly, larger trials will be necessary to evaluate the use of methylphenidate in Knowledge of management strategies for therapy- the adjuvant setting. induced neurocognitive deficits is limited. Since standard therapy has traditionally resulted in poor survival, few patients live long enough to develop Donepezil late side-effects of therapy. Though no standard Donepezil, a cholinesterase inhibitor, has shown exists for the management of neurocognitive efficacy in Alzheimer’s disease (Feldman et al., deficits, medical management, oxygen therapy, 2005). This drug inhibits acetylcholinesterase, and astrocyte cell transplantation have been allowing for increased availability of the neuro- investigated. transmitter acetylcholine. It is relatively safe, with headaches, fatigue, dizziness, and gastrointestinal upset being the most common toxicities. Because Methylphenidate of its metabolism through the liver, caution must Methylphenidate, a dopamine agonist used in be used when prescribed with drugs that block patients with narcolepsy and attention deficit dis- liver enzymes responsible for its metabolism (e.g., order, has been studied in patients with primary ketoconazole, quinidine). Also, the activity of this brain tumors experiencing neurocognitive dysfunc- drug can be decreased when combined with agents tion. The drug acts as a stimulant, which is useful in such as carbamazepine, phenytoin, and rifampin, addressing fatigue, concentration, and depression which increase its metabolism in the body. in patients with side-effects from brain irradiation As a result of its efficacy in Alzheimer’s dis- (Weitzner et al., 1995). The main toxicities of this ease, efforts have been made to test the drug in 178 Section 2. Effects of cancer and cancer treatment on cognition

cancer patients. Recently, a prospective randomized All patients underwent a comprehensive battery of double-blind placebo control study was attempted baseline neuropsychological testing and were then to evaluate the efficacy of this drug in patients with randomized to either immediate hyperbaric treat- SCLC (Jatoi et al., 2005). However, because of poor ment with post-therapy neurocognitive testing ver- accrual (only 9 of 104 patients) the study was closed sus delayed treatment (by 3 months) (Hulshof et al., prematurely. 2002). Six of the seven patients ultimately realized Currently, the National Cancer Institute (NCI) is a benefit from the therapy, however there was no sponsoring a phase II trial looking at donepezil and statistically significant benefit to earlier versus later EGb761 as agents that may improve neurocogni- intervention. As response to this therapy may take tive function in patients who have undergone radia- months, this therapy should only be considered in tion for either primary brain tumors or brain metas- patients who have reasonable performance status tases. EGb761 is an extract from Ginkgo biloba that and no active extracranial disease. has been found to have efficacy in the management of dementia and improvement in general cognition Transplantation of purified oligodendrocytes (Hoerr, 2003; LeBars, 2003; Mix & Crews, 2002). A more detailed discussion is presented in Chapter In addition to vascular changes induced by radi- 22. ation, there may also be direct damage to neural tissue. Classical teachings suggest that neural cells are incapable of repair. However, more recent data Hyperbaric oxygen suggest that the introduction of transplanted oligo- Radiation can decrease perfusion of tissues because dendrocytes may allow for remyelination of dam- of narrowing of the vasculature. As a result, tis- aged neural cells (Groves et al., 1993). Groves and sues may be deficient of oxygen and other nutrients colleagues described a process by which they were that are necessary to help with the tissue’s recov- able to expand purified populations of oligodendro- ery process (Kohshi et al., 2003; Roman, 2000). Oxy- cyte type-2 astrocytes (O-2A) in vitro. These cells genation increases tissue oxygen levels and results were injected into demyelinated lesions of rat spinal in increased fibroblast proliferation, which pro- cords. They observed remyelination of the damaged motes angiogenesis allowing for neovascularization cords. Others have described similar experiences (Knighton et al., 1983; Marx et al., 1985). Generally, in animal models where transplanted oligodendro- patients are given 100% oxygen at above-sea-level cytes can either promote remyelination or even pre- pressures in an enclosed chamber. These sessions vent white matter damage (Bambakidis & Miller, are referred to as dives (as they simulate below- 2004; Blakemore et al., 2003; Magy et al., 2003). How- sea-level pressure). Each dive takes about 2 h and ever, few data currently exist with transplantation in often 20–30 treatments are necessary. Because of the humans. expense and labor necessary to do this, it is only offered in select institutions in the United States. NMDA receptor Data showing efficacy in improving radiation- related toxicities following soft-tissue and bone The NMDA (N-methyl-d-aspartate) receptor assists injury after head and neck radiotherapy are rela- in a variety of functions in the mammalian CNS. tively well established, but few data are available Of particular interest is the long-term potentiation showing the efficacy of hyperbaric oxygen for neu- of synapses within the hippocampus (Kauer et al., rotoxicity following cranial irradiation. In a phase I- 1988). This process plays a critical role in mem- II trial conducted in the Netherlands, seven patients ory and learning (Izquierdo et al., 1992; Levin et al., experiencing post-radiation neurocognitive deficits 2003; Riedel & Reymann, 1996). Memantine is a drug were treated with 30 sessions of hyperbaric oxygen. approved by the FDA for use in the management Chapter 12. Brain metastases 179

Table 12.3. Randomized trials of whole-brain radiation therapy alone for brain metastases (adapted from Shaw et al., 2003)

Randomization Median survival time Study Year No. of patients Gy/fractions (months)

Harwood & Simson, 1977 1977 101 30/10 vs. 10/1 4.0–4.3 Kurtz et al., 1981 1981 255 30/10 vs. 50/20 3.9–4.2 Borgelt et al., 1980 1980 138 10/1 vs. 30/10 vs. 40/20 4.2–4.8 Borgelt et al., 1981 1981 64 12/2 vs. 20/5 2.8–3.0 Chatani et al., 1986 1986 70 30/10 vs. 50/20 3.0–4.0 Haie-Meder et al., 1993 1993 216 18/3 vs. 36/6 or 43/13 4.2–5.3 Chatani et al., 1994 1994 72 30/10 vs. 50/20 or 20/5 2.4–4.3 Murray et al., 1997 1997 445 54.4/34 vs. 30/10 4.5 of moderate to severe Alzheimer’s disease (Danysz an accelerated hyperfractionated regimen (Penitzka et al., 2000). The drug acts as an NMDA-receptor et al., 2002; Regine et al., 2001). antagonist, which may assist in the excitation of the NMDA receptor, resulting in improvement of memory and learning. Currently, the use of this drug in Radiosurgery alone the management of neurocognitive deficits following treatment of brain metastases remains investi- The use of radiosurgery alone is controversial, but gational. may be reasonable in the appropriately selected A more detailed discussion with regards to phar- patient. For patients with solitary lesions, in whom macologic interventions is presented in Chapter 22. there is a high degree of certainty that there are no other intracranial metastases, radiosurgery alone may be adequate treatment. Also, patients with Prevention of neurocognitive deficits significant small vessel disease with few metastases may also be candidates for radiosurgery Radiation dose and fractionation alone to reduce the neurocognitive toxicities of A variety of WBRT treatment schedules, from WBRT. Pirzkall and colleagues, in a retrospective hypofractionation to hyperfractionation, from low study of 236 patients, found no survival difference dose (10 Gy) to high dose (54 Gy), and from small between patients with radiosurgery plus WBRT ver- fields to large fields have been evaluated. Table 12.3 sus radiosurgery alone (Pirzkall et al., 1998). How- shows the variation in radiation treatment sched- ever, there was a trend for improved survival in those ules studied by the RTOG. The finding of these patients without extracranial disease who received studies suggests that differences in dose, timing, both WBRT and radiosurgery, suggesting that such and fractionation have not significantly altered the an approach should involve careful selection of median survival time following WBRT treatment of patients. brain metastases. As described earlier, patients with Recently, a report from the Japanese Radiation fraction sizes greater than 3 Gy are at higher risk of Oncology Study Group randomized patients with developing neurotoxicity from whole-brain therapy brain metastases to either radiosurgery alone or (DeAngelis et al., 1989); therefore, hypofractionated radiosurgery plus whole-brain radiation (Aoyama regimens should be avoided. No significant cogni- et al., 2006). As mentioned earlier, patients receiv- tive decline was reported for treatments with 40 Gy ing radiosurgery alone were at a higher risk of devel- in 20 fractions, 30 Gy in 10 fractions, or 54.4 Gy in oping recurrence not only in the untreated brain, 180 Section 2. Effects of cancer and cancer treatment on cognition

Table 12.4. Up-front WBRT decreases risk of brain and local relapse. NS, Not signiﬁcant; RS, radiosurgery, WBRT, whole-brain radiation therapy

Treatment RS RS + WBRT P value Risk ratio

n 67 65 – – Median 7.6 7.9 NS – survival (months) Local control 70 86 0.0001 1.22 (%) Brain failure 52 18 0.0001 1.7 (%)

but also at the site of radiosurgery. There was no difference in overall survival, neurological function, or cause of death (see Table 12.4). Therefore, it is Figure 12.1. Identiﬁcation of the hippocampus. Area in not unreasonable to consider radiosurgery alone in white denotes the mapped out hippocampus on MRI the carefully selected patient with a solitary brain metastasis, as long as they are followed closely.

Avoidance Region Erythropoietin 30 Gy Recently, erythropoietin (EPO) has been described 3 Gy as a possible radiation protectant in the brain 30 GY (Senzer, 2002). Reactive oxygen and nitrogen inter- IMRT with tomotherapy mediates resulting from radiation treatment play a achieves significant dose reduction (hippocampus), role in the development of neurotoxicity. In a mouse while delivering 30 Gy to model, EPO has been found to improve perfor- the rest of the brain mance and prevent cognitive impairment. Erythro- poietin is an ideal drug for neurotoxicity preven- Figure 12.2. Hippocampus avoidance with intensity tion, as it crosses the BBB and also because recep- modulated radiotherapy (provided by Drs. Hazim Jaradat tors for EPO are present on astrocytes, neurons, and and Wolfgang Tome, University of Wisconsin). See color endothelial cells in the brain. Recently, however, the version in color plate section use of EPO in patients with metastatic breast cancer receiving first-line chemotherapy was found to radiation (2 Gy or less) can damage the hippocam- decrease survival (Leyland-Jones et al., 2005). There- pus (Peissner et al., 1999). Current investigations fore, routine use of this drug in patients is not rec- are looking at means of using image-guided tech- ommended, as further investigations are necessary nology to avoid the hippocampus in the delivery of to validate its efficacy. WBRT. Because the hippocampus can be difficult to identify (see Figure 12.1), autocontouring algo- Conformal avoidance of the hippocampus rithms are being developed to make this type of As mentioned earlier, the hippocampus plays a vital delivery practical on a large scale. With the use of role in memory and learning. Very low doses of intensity-modulated radiotherapy, it may be Chapter 12. Brain metastases 181

possible to treat the entire brain to standard erative stem cells as a possible source of amelio- doses of whole-brain radiation, while keeping the rating this condition, the possibility of oligoden- radiation dose to the hippocampus very low (see drocytic/astrocytic cell transplantation, and other Figure 12.2). Clinical trials implementing this new approaches promise possible future interventional technology are currently under development. avenues. Further details on neuroprotection can be found in Chapter 22. REFERENCES Conclusion Abrey LE, Olson JD, Raizer JJ et al. (2001). A phase II trial of Cognitive dysfunction is an important compo- temozolomide for patients with recurrent or progressive nent of many malignancies, and occurs in the brain metastases. J Neurooncol 53: 259–265. vast majority of patients with brain metastases. Agboola O, Benoit B, Cross P et al. (1998). Prognostic fac- This results in major quality of life disruption, tors derived from recursive partitioning analysis (RPA) of and recent data suggest that the level of neu- Radiation Therapy Oncology Group (RTOG) brain metas- rocognitive dysfunction may be predictive of over- tases trials applied to surgically resected and irradiated brain metastatic cases. Int J Radiat Oncol Biol Phys 42: all outcome. Decline in neurocognition is com- 155–159. monly encountered in these patients and the Ahles TA, Saykin AJ, Furstenberg CT et al. (2002). Neuropsy- etiology of this is multifactorial, including the dis- chologic impact of standard dose systemic chemother- ease process itself, treatment effects, the conse- apy in long-term survivors of breast cancer and lym- quences of supportive treatment measures, and phoma. J Clin Oncol 20: 485–493. underlying physiological processes such as anemia, Alexander E, Loeffler JS (1999). The case for radiosurgery. etc. The incidence and severity of neurocognitive Clin Neurosurg 45: 32–40. deficits are generally underestimated, as few studies Andrews DW, Scott CB, Sperduto PW et al. (2004). Whole have prospectively rigorously evaluated cognitive brain radiation therapy with or without stereotac- function. With the development of “user-friendly” tic radiosurgery boost for patients with one to three test batteries, it is now relatively easy to measure brain metastases: phase III results of the RTOG 9508 randomized trial. Lancet 363(9422): 1665– and follow deficits and changes in neurocogni- 1672. tive function. As management for brain metasta- Antonadou D, Coliarakis N, Paraskevaidis M et al. (2002a). sis continues to evolve, whole-brain radiotherapy Whole brain radiotherapy alone or in combination remains a mainstay of treatment. Though efficacy with temozolomide for brain metastases. A phase III is realized with WBRT, survival is short, and toxic- study [abstract]. Int J Radiat Oncol Biol Phys 54: ities are real. As multimodal approaches including 93–94. WBRT, radiosurgery, systemic, and local chemother- Antonadou D, Paraskevaidis M, Sarris G et al. (2002b). apy evolve and outcomes improve, long-term tox- Phase II randomized trial of temozolomide and con- icities will become even more of an issue. Knowl- current radiotherapy in patients with brain metastases. edge of mechanisms of this toxicity is expanding. J Clin Oncol 20: 3644–3650. This understanding, along with advances in radia- AoyamaH,ShiratoH,TagoMet al. (2006). Stereotac- tic radiosurgery plus whole-brain radiation therapy vs tion technology, will likely improve our outcomes stereotactic radiosurgery alone for treatment of brain with this difficult disease. In particular, newer phar- metastases: a randomized controlled trial. JAmMed macologic interventions, based on seminal research Assoc 295(25): 2483–2491. in Alzheimer’s disease, a better appreciation of Arriagada R, LeChevalier T, Borie F et al. (1995). Prophy- the exquisite radiosensitivity of the hypothalamus lactic cranial irradiation for patients with small cell lung and the development of “hippocampus-avoidance” cancer in complete remission. J Natl Cancer Inst 87: 183– radiotherapy techniques, the recognition of regen- 190. 182 Section 2. Effects of cancer and cancer treatment on cognition

Bambakidis NC, Miller RH (2004). Transplantation of oligo- knife radiosurgery, whole brain radiotherapy or both dendrocyte precursors and sonic hedgehog results in [abstract]. Int J Radiat Oncol Biol Phys 48: 114 improved function and white matter sparing in the Chow BM, Li YQ, Wong CS. (2000). Radiation-induced spinal cords of adult rats after contusion. Spine J 4(1):16– apoptosis in the adult central nervous system is p53- 26. dependent. Cell Death Differ 7(8): 712–720. Blakemore WF,Gilson JM, Crang AJ (2003). The presence of Christodoulou C, Bafaloukos D, Kosmidis P et al. (2001). astrocytes in areas of demyelination influences remyeli- Phase II study of temozolomide in heavily pretreated nation following transplantation of oligodendrocyte pro- cancer patients with brain metastases. Ann Oncol 12: genitors. Exp Neurol 184(2): 955–963. 249–254. Boldrey E, Sheline G (1967). Delayed transitory clinical Crossen JR, Garwood D, Glatstein E et al. (1994). Neurobe- manifestation after radiation treatment of intracranial havioral sequelae of cranial irradiation in adults: a review tumors. Acta Radiol 5: 5–10. of radiation induced encephalopathy. J Clin Oncol 12: Borgelt B, Gelber R, Kramer S et al. (1980). The palliation 627–642. of brain metastases: final results of the first two studies Curran W, Mehta MP, Terhaard C et al. (2002). Predictors by the Radiation Therapy Oncology Group. Int J Radiat of Survival for Patients with Brain Metastases: Results of OncolBiolPhys6:1–9. a Randomized Phase III Trial [abstract 155]. Int J Radiat Borgelt B, Gelber R, Larson M et al. (1981). Ultra-rapid Oncol Biol Phys 54: 93. high dose irradiation schedules for the palliation of brain Danysz W, Parsons CG, Mobius JH et al. (2000). Neuro- metastases: final results of the first two studies by the protective and symptomatological action of memantine Radiation Therapy Oncology Group. Int J Radiat Oncol relevant for Alzheimer’s disease: a unified glutamatergic Biol Phys 7: 1633–1638. hypothesis on the mechanism of action. Neurotoxicity Bradley KA, Mehta MP (2004). Management of brain metas- Res 2: 85–97. tases. Semin Oncol 31: 693–701. DeAngelis LM, Delattre JY, Posner JB (1989). Radiation Breneman JC, Warnick RE, Albright REJ et al. (1997). Stereo- induced dementia in patients cured of brain metastases. tactic radiosurgery for the treatment of brain metastases. Neurology 39: 789–796. Results of a single institution series. Cancer 79: 551–557. Dropcho EJ (2005). Paraneoplastic disorders of the nervous Brezden CB, Phillips KA, Abdolell M (2000). Cognitive system. In Black PM, Loeffler JS (eds.) Cancer of the Ner- function in breast cancer patients receiving adjuvant vous System, (pp. 691–714). Philadelphia, PA: Lippincott chemotherapy. J Clin Oncol 18: 2695–2701. Williams & Wilkins. Burger PC, Mahaley MS Jr., Dudka L et al. (1979). The mor- Ellis R, Gregor A (1998). The treatment of brain metastases phologic effects of radiation administered therapeuti- from lung cancer. Lung Cancer 20: 81–84. cally for intracranial gliomas: a postmortem study of 25 Feldman H, Gauthier S, Hecker J et al. (2005). Efficacy cases. Cancer 44: 1256–1279. and safety of donepezil in patients with more severe Cammer W (2000). Effects of TNFalpha on immature and Alzheimer’s disease: a subgroup analysis from a ran- mature oligodendrocytes and their progenitors invitro. domized placebo-controlled trial. Int J Geriatr Psychiatry Brain Res 864(2): 213–219. 20(6): 559–569. Cappuzo F, Mazzoni F, Maestri A et al. (2000). Medical Fontanella M, Perozzo P, Ursone R et al. (2003). Neu- treatment of brain metastases from solid tumors. Forum ropsychological assessment after microsurgical clip- (Genova) 10: 137–148. ping or endovascular treatment for anterior commu- Chatani M, Teshima T, Hata K et al. (1986). Prognostic fac- nicating artery aneurysm. Acta Neurochir 145: 867– tors in patients with brain metastases from lung carci- 872. noma. Strahlenther Onkol 162: 157–161. Freidman MA, Meyers CA, Sawaya R (2003). Neuropsycho- Chatani M, Matayoshi Y, Masaki N et al. (1994). Radi- logical effects of third ventricular tumor surgery. Neuro- ation therapy for brain metastases from lung carci- surgery 52: 791–798. noma. Prospective randomized trial according to the Gaspar L, Scott C, Rotman M et al. (1997). Recursive par- level of lactate dehydrogenase. Strahlenther Onkol 170: titioning analysis (RPA) of prognostic factors in three 155–161. Radiation Therapy Oncology Group (RTOG) brain metas- Chougule PB, Burton-Williams M, Saris S et al. (2000). tases trials. Int J Radiat Oncol Biol Phys 37(4): 745– Randomized treatment of brain metastases with gamma 751. Chapter 12. Brain metastases 183

Gottwald B, Wilde B, Mihajlovic Z et al. (2004). Evidence for Izquierdo I, da Cunha C, Rosat R et al. (1992). Neurotrans- distinct cognitive deﬁcits after focal cerebellar lesions. mitter receptors involved in post-training memory pro- J Neurol Neurosurg Psychiatry 75: 1524–1531. cessing by the amygdala, medial septum, and hippocam- Graus F, Walker RW, Allen JC (1983). Brain metastases in pus of the rat. Behav Neural Biol 58(1): 16–26. children. J Pediatr 103(4): 558–561. Jatoi A, Kahanic SP, Frytag S et al. (2005). Donepezil and Green HJ, Pakenham KI, Headley BC et al. (2002). Altered vitamin E for preventing cognitive dysfunction in small cognitive function in men treated for prostate cancer cell lung cancer patients: preliminary results and sugges- with luteinising hormone-releasing hormone analogues tions for future study designs. Support Care Cancer 13(1): and cyproterone acetate: a randomized controlled trial. 66–69. Br J Urol 90: 427–432. Juengling FD, Ebert D, Gut O et al. (2000). Prefrontal cor- Greenberg H, Chandler WF, Sandler HM (1999). Brain tical hypometabolism during low-dose interferon alpha metastases. In: Greenberg H, Chandler WF, Sandler HM treatment. Psychopharmacology (Berl) 152: 383–389. (eds.) Brain Tumors, (pp. 299–317, Vol. 54). New York: Karatekin C, Lazareff JA, Asarnow RF (2000). Relevance of Oxford University Press. cerebellar hemispheres for executive functions. Pediatr Gregor A, Cull A, Stephens RJ et al. (1997). Prophylactic cra- Neurol 22: 106–112. nial irradiation is indicated following complete response Kauer JA, Manlenka RC, Nicoll RA (1988). NMDA applica- to induction therapy in small cell lung cancer: results of tion potentiates synaptic transmission in the hippocam- a multicenter randomized trial. United Kingdom Coor- pus. Nature 334(6179): 250–252. dinating Committee for Cancer Research (UKCCCR) and Kelly K, Bunn PAJ (1998). Is it time to reevaluate our the European Organisation for Research and Treatment approach to the treatment of brain metastases in of Cancer (EORTC). Eur J Cancer 33: 1752–1758. patients with non-small cell lung cancer? Lung Cancer Grill J, Viguier D, Kieffer V et al. (2004). Critical risk fac- 20: 85–91. tors for intellectual impairment in children with poste- Kelley KW, BlutheRM,DantzerR´ et al. (2003). Cytokine- rior fossa tumors: the role of cerebellar damage. J Neuro- induced sickness behavior. Brain Behav Immun 17: surg 101: 152–158. S112–S118. Groves AK, Barnett SC, Franklin RJ et al. (1993). Repair of Kershner P, Wang-Cheng R (1989). Psychiatric side-effects demyelinated lesions by transplantation of puriﬁed O-2A of steroid therapy. Psychosomatics 30: 135–139. progenitor cells. Nature 362(6419): 453–455. Klein M, Engelberts NHJ, van der Ploeg HM et al. (2003). Haie-Meder C, Pellae-Cosset B, Laplanche A et al. (1993). Epilepsy in low grade gliomas: the impact on cognitive Results of a randomized clinical trial comparing two function and quality of life. Ann Neurol 54: 514–520. radiation schedules in the palliative treatment of brain Knighton DR, Hunt TK, Schenestuhl H et al. (1983). Oxy- metastases. Radiother Oncol 26: 111–116. gen tension regulates the expression of angiogenesis fac- Harwood AR, Simson WJ (1977). Radiation therapy of cere- tor by macrophages. Science 221: 1283–1289. bral metastases: a randomized prospective clinical trial. Kohshi K, Imada H, Nomoto S et al. (2003). Successful treat- Int J Radiat Oncol Biol Phys 2: 1091–1094. ment of radiation-induced brain necrosis by hyperbaric Herman MA, Tremont-Lukats I, Meyers CA et al. (2003). oxygen therapy. J Neurol Sci 209(1–2): 115–117. Neurocognitive and functional assessment of patients Komaki R, Meyers CA, Shin DM et al. (1995). Evaluation of with brain metastases. Am J Clin Oncol 26(3): 273–279. cognitive function in patients with limited small cell lung Hoerr R (2003). Behavioral and psychological symptoms of cancer prior to and shortly following prophylactic cranial dementia (BPSD): effects of EGb 761. Pharmacopsychia- irradiation. Int J Radiat Oncol Biol Phys 33(1): 179–182. try 36 [Suppl. 1]: S56–S61. Kondzoilka D, Patel A, Lunsford LD et al. (1999). Stereo- Hulshof MC, Stark NM, van der Kleij A et al. (2002). Hyper- tactic radiosurgery plus whole brain radiotherapy ver- baric oxygen therapy for cognitive disorders after irradi- sus radiotherapy alone for patients with multiple brain ation of the brain. Strahlenther Onkol 178(4): 192–198. metastases. Int J Radiat Oncol Biol Phys 45(2): 427–434. Hutter BO, Spetzger U, Bertalanffy H et al. (1997). Cog- Kramer S (1968). The hazards of therapeutic irradiation of nition and quality of life in patients after transcallosal the central nervous system. Clin Neurosurg 15: 301–318. microsurgery for midline tumors. J Neurosurg Sci 41: Kurtz JM, Gelber R, Brady LW et al. (1981). The palliation 123–129. of brain metastases in a favorable patient population: 184 Section 2. Effects of cancer and cancer treatment on cognition

a randomized clinical trial by the Radiation Therapy Mehta MP, Rodrigus P, Terhaard CHJ et al: (2003). Survival Oncology Group. Int J Radiat Oncol Biol Phys 7: 891–895. and neurologic outcomes in a randomized trial of Motex- Lassman AB, DeAngelis LM (2003). Brain metastases. Neu- afin Gadolinium and whole brain radiation therapy in rol Clin 21(I): I,vii. brain metastases. J Clin Oncol 21: 2529–2536. LeBars PL (2003). Response patterns of EGb 761 in Mehta MP, Gervais R, Chabot P et al. (2006). Motex- Alzheimer’s disease: influence of neuropsychological afin gadolinium (MGd) combined with prompt whole profiles. Pharmacopsychiatry Jun; 36 [Suppl 1]: S44–S49. brain radiation therapy (RT) prolongs time to neuro- Lee AW, Kwong DLW, Leung SF et al. (2002). Factors affect- logic progression in non-small cell lung cancer (NSCLC) ing risk of symptomatic temporal lobe necrosis: signifi- patients with brain metastases: Results of a phase III trial cance of fractional dose and treatment time. Int J Radiat [abstract]. Proc Am Soc Clin Oncol 24 [Suppl. 18S]: 7014. OncolBiolPhys53: 75–85. Meyers CA, Abbruzzese JL (1992). Cognitive functioning in Levin ED, Sledge D, Baruah A et al. (2003). Ventral hip- cancer patients: effect of previous treatment. Neurology pocampal NMDA blockade and nicotinic effects on 42: 434–436. memory function. Brain Res Bull 61(5): 489–495. Meyers CA, Boake C (1993). Neurobehavioral disorders Leyland-Jones B, Semiglazov V, Pawlicki M et al. (2005). experienced by brain tumor patients: rehabilitation Maintaining normal hemoglobin levels with epoetin alfa strategies. Cancer Bull 45: 362–364. in mainly nonanemic patients with metastatic breast Meyers CA, Hess KR (2003). Multifaceted endpoints in cancer receiving first-line chemotherapy: a survival brain tumor clinical trials: cognitive deterioration pre- study. J Clin Oncol 23: 5960–5972. cedes MRI progression. Neurooncology 5: 89–95. Li J, Bentzen SM, Renschler M, Mehta MP (2006). Improve- Meyers CA, Valentine AD, Wong FCL et al. (1994). ment in neurocognitive function (NCF) correlates with Reversible neurotoxicity of interleukin-2 and tumor tumor regression after whole brain radiation therapy necrosis factor: correlation of SPECT with neuropsycho- (WBRT) for brain metastases (BM) [abstract]. Proc Am logical testing. J Neuropsychiatry Clin Neurosci 6: 285– Soc Clin Oncol 24 [Suppl. 18S]: 1504. 288. Lilja A, Smith GJ, Salford LG (1992). Microprocesses in per- Meyers CA, Weitzner MA, Valentine et al. (1998). ception and personality. J Nerv Ment Dis 180: 82–88. Methylphenidate therapy improves cognition, mood, Lim LC, Rosenthal MA, Maartens N et al. (2004). Manage- and function of brain tumor patients. J Clin Oncol 16(7): ment of brain metastases. Intern Med J 34: 270–278. 2522–2527. Magy L, Mertens C, Avellana-Adalid V et al. (2003). MeyersCA,SmithJA,BezjakAet al. (2004). Neurocognitive Inducible expression of FGF2 by a rat oligodendrocyte function and progression in patients with brain metas- precursor cell line promotes CNS myelination in vitro. tases treated with whole brain radiation and Motexafin Exp Neurol 184(2): 912–922. Gadolinium: results of a randomized phase III trial. JClin Marx RE, Johnson RP, Kline SN (1985). Prevention of osteo- Oncol 22(1): 157–165. radionecrosis: a randomized prospective clinical trial of Mix JA, Crews WD (2002). A double-blind, placebo- hyperbaric oxygen versus penicillin. J Am Dent Assoc 111; controlled, randomized trial of Ginkgo biloba extract 49–54. EGb 761 in a sample of cognitively intact older adults: McEwen BS, Alves SE (1999). Estrogen actions in the central neuropsychological findings. Hum Psychopharmacol nervous system. Endocr Rev 20: 279–307. 17(6): 267–277. Mehta MP, Tremont-Lukas I (2004). Radiosurgery for single Monje ML, Palmer T (2003). Radiation injury and neuro- and multiple metastases. In Sawaya R (ed.) Intracranial genesis. Curr Opin Neurol 16: 129–134. Metastases: Current Management Strategies, (pp. 139– Murray KJ, Scott C, Greenberg HM et al. (1997). A random- 164). Malden MA: Blackwell. ized phase III study of accelerated hyperfractionation Mehta MP, Shapiro WR, Glantz M et al. (2002). Lead-in versus standard in patients with unresected brain metas- phase to randomized trial of Motexafin Gadolinium and tases: a report of the Radiation Therapy Oncology Group whole brain radiation for patients with brain metastases. (RTOG) 9104. Int J Radiat Oncol Biol Phys 39: 571–574. Centralized assessment of magnetic resonance imaging, Nieder C, Berberich W, Schnabel K (1997). Tumor-related neurocognitive and neurologic end points. J Clin Oncol prognostic factors for remission of brain metastases after 20: 3445–3453. radiotherapy. Int J Radiat Oncol Biol Phys 39: 25–30. Chapter 12. Brain metastases 185

Noordjik EM, Vecht CJ, Haaxma-Reiche H et al. (1994). The Renschler M, Mehta MP, Donald DM et al. (2003). Treat- choice of treatment of single brain metastases should be ment intent for brain metastases: surveys of medical based on extracranial tumor activity and age. Int J Radiat and radiation oncologists indicate that maintaining neu- Oncol Biol Phys 29: 711–717. rologic and neurocognitive function is more important Nussbaum ES, Djalilian HR, Cho K et al. (1996). Brain than prolonging survival. Proc Am Soc Clin Oncol 22: 552 metastases: histology, multiplicity, surgery and survival. (no. 2222). Cancer 78(8): 1781–1788. Rich JB, Maki P (1999). Estrogen, testosterone and the Patchell RA, Tibbs PA, Walsh JW et al. (1990). A randomized brain: a review of neuropsychological, aging and neu- trial of surgery in the treatment of single metastases to roimaging studies. J Int Neuropsychol Soc 6: iii. the brain. N Eng J Med 322(8): 494–500. Riedel G, Reymann KG (1996). Metabotropic glutamate Patchell RA, Tibbs PA, Regine WF et al. (1998). Postopera- receptors in hippocampal long-term potentiation and tive radiotherapy in the treatment of single metastases to learning and memory. Acta Physiol Scand 157(1): 1–19. the brain: a randomized trial. J Am Med Assoc 280: 1485– Roman G (2000). Perspectives in the treatment of vascular 1489. dementia. Drugs Today (Barc) 36(9): 641–653. Peace KA, Orme SM, Thompson AR (1997). Cognitive dys- Salminen E, Porten R, Korpela J et al. (2003). Androgen function in patients treated for pituitary tumors. JClin deprivation and cognition in prostate cancer. Br J Can- Exp Neuropsychol 19:1–6. cer 89: 971–976. Peissner W, Kocher M, Treuer H (1999). Ionizing radiation- Sanghavi SN, Miranpuri SS, Chappell R et al. (2001). Radio- induced apoptosis of proliferating stem cells in the den- surgery for patients with brain metastases: a multi- tate gyrus of the adult rat hippocampus. Brain Res Mol institutional analysis stratified by the RTOG recursive Brain Res 71(1): 61–68. partitioning analysis method. Int J Radiat Oncol Biol Phys Penitzka S, Steinvorth S, Sehlleier S et al. (2002). Assess- 51: 426–434. ment of cognitive function after preventive and thera- Sawaya R, Bindal R (2001). Metastatic brain tumors. In peutic whole brain irradiation using neuropsychological Laws E, Kaye AH (eds.) Brain Tumors: an Encyclope- testing. Strahlenther Onkol 178: 252–258. dic Approach (pp. 923–946). Edinburgh: Churchill Living- Pirzkall A, Debus J, Lohr F et al. (1998). Radiosurgery alone stone. or in combination with whole-brain radiotherapy for Schagen SB, vanDam FS, Muller MJ et al. (1999). Cognitive brain metastases. J Clin Oncol 16: 3563–3569. deficits after postoperative chemotherapy for breast can- Posner JB (1995). Neurologic Complications of Cancer. cer. Cancer 85: 640–650. Philadelphia, PA: FA Davis. Scheibel RS, Meyers CA, Levin VA (1996). Cognitive dys- Postmus PE, Haaxma-Reiche H, Smit EF et al. (2000). function following surgery for intracerebral glioma: Treatment of brain metastases of small-cell lung cancer: influence of histopathology, lesion location and treat- comparing teniposide and teniposide with whole-brain ment. J Neurooncol 30: 61–69. radiotherapy. A phase III study of the European Organi- Schellinger PD, Meinck HM, Thron A (1999). Diagnostic zation for the Research and Treatment of Cancer Lung accuracy of MRI compared to CCT in patients with brain Cancer Cooperative Group. J Clin Oncol 18: 3400–3408. metastases. J Neurooncol 44(3): 275–281. Quan AL, Videtic GM, Suh JH (2004). Brain metastases in Schultheiss TE, Kun LE, Ang KK et al. (1995). Radiation small cell lung cancer. Oncology 18: 961–72. response of the central nervous system. Int J Radiat Regine WF, Scott C, Murray K et al. (2001). Neurocogni- OncolBiolPhys31: 1093–1112. tive outcome in brain metastases patients treated with Senzer N (2002). Rationale for a phase III study of erythro- accelerated-fractionation vs accelerated hyperfraction- poietin as a neurocognitive protectant in patients with ated radiotherapy: an analysis of RTOG study 91–04. Int J lung cancer receiving prophylactic cranial irradiation. Radiat Oncol Biol Phys 51: 711–717. Semin Oncol 6 [Suppl. 19]: 47–52. Regine WF, Schmitt FA, Scott CB et al. (2004). Feasibil- Seung SK, Sneed PK, McDermott MW et al. (1998). ity of neurocognitive outcome evaluations in patients Gamma Knife radiosurgery for malignant melanoma with brain metastases in a multi-institutional coopera- brain metastases. Cancer J Sci Am 4: 103–109. tive group setting: results of Radiation Therapy Oncology Shaw E, Scott C, Suh J et al. (2003). RSR13 plus cra- Group trial BR-0018. Int J Radiat Oncol Biol Phys 58(5): nial radiation therapy in patients with brain metas- 1346–1352. tases: comparison with the Radiation Therapy Oncology 186 Section 2. Effects of cancer and cancer treatment on cognition

Group recursive partitioning analysis brain metastases Verger E, Gil M, Yaya R et al. (2003). Concomitant temozolo- database. J Clin Oncol 21: 2364–2371. mide (TMZ) and whole brain radiotherapy (WBRT) in Sheline GE, Wara WM, Smith V (1980). Therapeutic irra- patients with brain metastases (BM): randomized multi- diation and brain injury. Int J Radiat Oncol Biol Phys 6: centric phase II study [abstract]. Proc Am Soc Clin Oncol 1215–1228. 22: 101. Silverman DH, Castellon SA, Abraham L et al. (2003). Wefel JS, Kayl AE, Meyers CA (2004a). Neuropsychologi- Abnormal regional brain metabolism in breast cancer cal dysfunction associated with cancer and cancer thera- survivors after adjuvant chemotherapy is associated with pies: a conceptual review of an emerging target. Br J Can- cognitive changes. Proc Am Soc Clin Oncol 22: 12. cer 90: 1691–1696. Sundstrom JT, Minn H, Lertola KK et al. (1998). Prognosis of Wefel JS, Lenzi R, Theriault RL et al. (2004b). The cogni- patients treated for intracranial metastases with whole- tive sequelae of standard-dose adjuvant chemotherapy brain irradiation. Ann Med 30: 296–299. in women with breast carcinoma: results of a prospec- Taphoorn MJB (2003). Neurocognitive sequelae in the tive randomized longitudinal trial. Cancer 100(11): 2292– treatment of low grade gliomas. Semin Oncol 30 [6 Suppl. 2299. 19] 45–48. Weitzner MA, Meyers CA, Valentine AD (1995). Tucha O, Smely C, Preier M (2003). Preoperative and post- Methylphenidate in the treatment of neurobehavioral operative cognitive functioning in patients with frontal slowing associated with cancer and cancer treatment. meningioma. J Neurosurg 98: 21–31. J Neuropsychiatry Clin Neurosci 7: 347–350. Turkheimer E, Yeo RA, Jones CL et al. (1990). Quantitative Wen PY, Black PM, Loefﬂer JS (2001). Metastatic brain can- assessment of covariation between neuropsychological cer. In Devita VT Jr., Hellman S, Rosenberg SA (eds.) function and location of naturally occurring lesions in Cancer: Principles and Practice of Oncology (6th edn.) human. J Clin Exp Neuropsychol 12: 549–565. (pp. 2655–2670). Philadelphia PA: Lippincott Williams & Van Der Kogel AJ (1986). Radiation-induced damage in the Wilkins. central nervous system: an interpretation of target cell Wieneke MH, Dienst ER (1995). Neuropsychologi- responses. Br J Cancer 7: 207–217. cal assessment of cognitive functioning following Van Oosterhout AG, Boon PJ, Houx PJ et al. (1995). Fol- chemotherapy for breast cancer. Psychooncology 4: low up of cognitive functioning in patients with small 61–66. cell lung cancer. IntJRadiatOncolBiolPhys31: 911– Wolkowitz OM, Reus VI, Weingartner H et al. (1990). Cogni- 914. tive aspects of corticosteroids. Am J Psychiatry 147: 1297– Vannucci RC, Baten M (1974). Cerebral metastatic disease 1303. in childhood. Neurology 24(10): 981–985. Young RF (1998). Radiosurgery for the treatment of brain Varney NR, Syrop C, Kubu CS et al. (1998). Neuropsycho- metastases. Semin Surg Oncol 14: 70–78. logic dysfunction in women following leuprolide acetate Zimm S, Wampler GL, Stablein D et al. (1981). Intracerebral induction of hypoestrogenism. J Assist Reprod Genet 10: metastases in solid tumor patients: natural history and 53–57. results of treatment. Cancer 48: 384–394. 13

Primary central nervous system lymphoma

Denise D. Correa

Introduction of increased intracranial pressure, headache, or progressive cognitive decline may also occur at Primary central nervous system lymphoma (PCNSL) presentation. The diagnosis of PCNSL involves is a relatively rare non-Hodgkin’s lymphoma that stereotactic brain biopsy or demonstration of arises within the CNS. Until recently, it accounted malignant lymphocytes in the cerebrospinal fluid foronly1%ofallprimarybraintumors,butits (CSF) or vitreous in cases of ocular involvement. incidence increased threefold in immunocompe- Surgical resection is often not beneficial, and the tent populations from 1988 to the time of writ- deep-seated nature of most lesions increases the ing (Eby et al., 1988). With a median age at diag- risk of possible surgical complications (DeAngelis nosis of 60 years (Peterson & DeAngelis, 1997), et al., 1990). Glucocorticoids produce only tem- PCNSL is a disease of middle and late adult life, porary clinical and radiographic response. and it is slightly more common in males (O’Neill Conventional treatment of PCNSL has consisted & Illig, 1989). It is an infiltrative tumor most often of whole-brain radiation therapy (WBRT) due to located in the periventricular region and subcortical the multifocal growth and spread of the disease. A gray matter (Grant & Isaacson, 1992) (Figure 13.1). dose of 40–50 Gy is used in most centers, yield- Leptomeningeal involvement is present in approxi- ing a median survival of 12–24 months (DeAngelis mately one-third of patients at diagnosis (Peterson et al., 1992; Nelson, 1999). Most patients have a com- & DeAngelis, 1997), and the eye is another site plete or major partial response to this treatment, of multifocal CNS involvement in about 25% of but the tumor typically relapses within the first patients (Peterson et al., 1993). On neuroimaging year post-WBRT (DeAngelis et al., 1990). Unlike pri- studies, PCNSL is identified as contrast-enhancing mary glial tumors, PCNSL often relapses at a loca- in 90% of cases, and multifocal lesions occur in tion distant from the original site of disease and approximately 40% of patients (DeAngelis, 1995). may show widespread infiltration. In recent years, chemotherapy in conjunction with radiotherapy has been more frequently used to treat this disease. Diagnosis and treatment The use of agents that cross the blood–brain barrier (BBB) has been found to be necessary in order Patients may present with focal neurological signs, to gain access to disease residing behind an intact such as weakness, gait disturbance, language BBB. Regimens that include high-dose methotrex- dysfunction, or seizures. Generalized symptoms ate (HD-MTX), high-dose cytarabine (HD-ARA-C),

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

187 188 Section 2. Effects of cancer and cancer treatment on cognition

was reported in 74% of patients and median survival was 40 months (Dahlborg et al., 1996; Doolittle et al., 2000; McAllister et al., 2000). However, the long-term efﬁcacy of HD-MTX-based regimens with or without BBB disruption remains to be conﬁrmed (Ferreri et al., 2002; Herrlinger et al., 2005), as many patients relapse and require additional treatment with radiotherapy or chemotherapy (Abrey et al., 2000; Pels et al., 2003; Sandor et al., 1998). (a) (b)

Figure 13.1. (a, b) T1-weighted axial MRIs revealing Treatment-related delayed neurotoxicity gadolinium-enhancing lesions in two patients with primary central nervous system lymphoma (PCNSL) Radiotherapy often produces irreversible and progressive damage to the CNS through vascular as well as intrathecal MTX and WBRT have shown injury causing ischemia of surrounding tissue, and favorable results, with a median survival of 3–4 years demyelination of white matter and necrosis (Sheline and a 30% 5-year survival (Abrey et al., 1998; Boiardi et al., 1980). These are delayed effects of radia- et al., 1993; DeAngelis et al., 1992). Most prospective tion that become apparent a few months to many trials have used chemotherapy followed by WBRT, years after treatment (Sheline, 1980). Pathologi- as there is some evidence that HD-MTX admin- cal changes include multifocal areas of coagula- istered subsequent to WBRT increases the risk of tive necrosis in the deep white matter with loss of late neurotoxicity (Blay et al., 1998). Although com- myelin, axonal swelling, fragmentation, and gliosis. bined modality regimens prolong survival, there is Suggested mechanisms include depletion of glial an increased risk for long-term neurotoxicity that progenitor cells and perpetuation of oxidative stress increases with advanced age and in patients with induced by radiation (Toﬁlon & Fike, 2000). Radi- prolonged disease-free survival (Abrey et al., 1998; ation may diminish the reproductive capacity of Pels & Schlegel, 2006). the O-2A progenitors of oligodendrocytes, disrupt- Chemotherapy based on HD-MTX without ing the normal turnover of myelin (Van Der Maazen WBRT is efﬁcacious in the treatment of PCNSL and et al., 1993). This progressive demyelination may reduces the risk of delayed neurotoxicity (Freilich take months to produce symptoms, contributing to et al., 1996; Kraemer et al., 2002); it has been used the latency in onset of neurotoxicity and its progres- more frequently in elderly PCNSL patients (Abrey sive nature. The prevalence of radiation-induced et al., 2000; Hoang-Xuan et al., 2003). In prospective brain injury appears to increase with volume of radi- trials, HD-MTX alone produced a 52%–100% ated tissue, dose of radiation, dose per fraction, con- response rate and a 2-year survival rate of about comitant administration of chemotherapy, and age 60% (Cher et al., 1996; Guha-Thakurta et al., 1999; (Constine et al., 1988). See Chapter 7 for a detailed Herrlinger et al., 2002); HD-MTX-based poly- discussion of radiation injury to the brain. chemotherapy regimens resulted in a 65%–100% The interactions between radiation and HD-MTX response rate and a 2-year survival rate of 65%–78% are the most clearly demonstrated (Keime-Guibert (Sandor et al., 1998; Schlegel et al., 2001). Consider- et al., 1998), as WBRT may have a synergistic ing that MTX does not readily cross an intact BBB, effect when combined with HD-MTX (Crossen Neuwelt and colleagues (1991) have used intra- et al., 1992). Other chemotherapy agents that when arterial MTX administration after transiently open- combined with radiation may produce CNS dam- ing the BBB. A complete response to this treatment age are nitrosoureas, cytosine arabinoside, and Chapter 13. Primary central nervous system lymphoma 189

1998; Peterson & DeAngelis, 1997; Poortmans et al., 2003). The specific contribution of the disease itself and of various treatment modalities to the development of neurotoxicity remains to be elucidated, as the neurotoxic potential of combined treatments is difficult to determine when each modality can produce CNS damage individually (DeAngelis & Shapiro, 1991). Neurological sequelae of treatment usually develop more than 1 year after therapy, and can only be established in the absence of tumor (a) (b) (DeAngelis et al., 1989). Therefore, its incidence Figure 13.2. (a, b) T1-weighted axial MRI showing diffuse is proportional to the percentage of patients with white matter abnormalities in a 61-year-old PCNSL disease-free survival (DeAngelis et al., 1992), and patient 5 years post-WBRT and HD-MTX-based is related to both long-term survival and advanced chemotherapy (HD-MTX, High-dose methotrexate; WBRT, age (DeAngelis & Hormigo, 2004). When a combina- whole-brain radiation therapy) tion of WBRT and chemotherapy is used, the incidence of delayed neurotoxicity ranges from 8% to vincristine (DeAngelis & Shapiro, 1991). In PCNSL 50% of patients in different PCNSL clinical trials patients treated with combined modality therapy, (Abrey et al., 1998; Glass et al., 1994; O’Brien et al., magnetic resonance imaging (MRI) studies most 2000; Sarazin et al., 1995). It has been recognized commonly show diffuse white matter abnormalities as a significant problem in long-term survivors over (Figure 13.2), but cerebral atrophy, communicating 60 years of age (Abrey et al., 1998; Batchelor & hydrocephalus, and radionecrotic lesions can also Loeffler, 2006; Besell et al., 2001; Blay et al., 1998; be seen. Several chemotherapeutic agents, partic- Ferreri et al., 2002; Peterson & DeAngelis, 1997; ularly HD-MTX and HD-ARA-C, have been shown Poortmans et al., 2003), and reported to range to produce periventricular white matter abnormal- between 50% and 80% in older patients treated ities, but often less extensive than seen after com- with WBRT and HD-MTX-based chemotherapy, and bined modality treatment. The pathophysiological between 5% and 8% in patients treated with MTX- mechanisms of chemotherapeutic agents are not based chemotherapy alone (Abrey et al., 2000; well understood, but several have been hypothe- Hoang-Xuan et al., 2003). The variability in the sized including demyelination, secondary inflam- reported incidence of neurotoxicity across studies matory response, and microvascular injury. In a may be in part related to differences in both the type recent series of five autopsied PCNSL patients and severity of symptoms documented. treated with WBRT and chemotherapy (i.e., HD- MTX, HD-ARA-C, doxorubicin, or etoposide) who died of leukoencephalopathy (Lai et al., 2004), there Cognitive functions in PCNSL was myelin and axonal loss, pallor, rarefaction, spongiosis, and gliosis of the cerebral hemispheric The development of cognitive dysfunction in white matter. In addition, all patients had fibrotic PCNSL patients is most likely related to multiple thickening of small blood vessels in the deep white factors including the effects of the tumor itself matter, and four patients had atherosclerosis of the given its infiltrative pattern, age (i.e., median age large cerebral blood vessels in the circle of Willis. at diagnosis is 60 years), and the delayed effects Delayed neurotoxicity has been recognized as of treatment with WBRT and HD-MTX-based a significant problem as effective treatment for chemotherapy either combined or alone (O’Neill, PCNSL has increased survival rates (Abrey et al., 2004). It is currently considered the most frequent 190 Section 2. Effects of cancer and cancer treatment on cognition

complication among long-term survivors (Behin & BBB disruption (n = 7 with complete response) Delattre, 2003), and may interfere with the patient’s displayed cognitive deterioration despite limited ability to function at pre-morbid levels profession- survival (i.e., 16 months). The investigators used ally and socially, despite adequate disease control an extensive battery of neuropsychological tests (Correa et al., 2004; Harder et al., 2004). Consistent and included pre- and post-treatment evaluations, with the current status of the research on patients but only a small number of patients were availa- with other brain tumor histologies, there is also a ble for long-term follow-up in these studies. In paucity of information regarding cognitive func- addition, the findings were discussed regarding tioning in PCNSL patients. The majority of studies the presence or absence of cognitive decline (i.e., reported performance status, patients’ complaints, overall summary index of change from baseline), clinical observations, and mental status examina- and little information about specific cognitive tions (Corry et al., 1998; Herrlinger et al., 1998), domains that may have been differentially affected but systematic cognitive evaluations have rarely by treatment was reported. In a more recent study, been included. These methods have low sensitivity McAllister and colleagues (2000) reported the to detect cognitive dysfunction in patients with results of pre- and post-therapy cognitive evalua- brain tumors (Meyers & Wefel, 2003; Weitzner & tions on 23 PCNSL patients who had a complete Meyers, 1997). Therefore, only severe neurotoxicity response at least 1 year after treatment with HD- resulting in significant disability was documented MTX chemotherapy with BBB disruption. The in most clinical trials, suggesting that the true inci- mean time interval between the two evaluations dence of cognitive dysfunction in PCNSL is most was 16.5 months (SD = 10.9 months). All patients probably higher than reported (Blay et al., 1998; showed improvement in overall cognitive func- Keime-Guibert et al., 1998; Laack & Brown, 2004). tions (summary z-scores) at follow-up. However, The few studies that described cognitive outcome an evaluation of individual test scores revealed in PCNSL involved a relatively small number of no significant changes from baseline in verbal patients at follow-up, mostly as a result of increased learning, cognitive flexibility, and motor skills; in drop-out rates due to disease relapse or death. seven patients there was a mild decline in motor performance (at least 1 SD below the mean) at follow-up. Prospective studies Fliessbach and colleagues (2003) performed Neuwelt and his colleagues have investigated longitudinal neuropsychological evaluations on neuropsychological abilities in PCNSL patients a group of PCNSL patients treated with a poly- who were treated with osmotic BBB disruption chemotherapy regimen that included HD-MTX. with or without radiotherapy, and performed Although 20 patients were available for pre- pre-treatment and long-term follow-up on several treatment evaluations, only 10 patients (median patients (Crossen et al., 1992; Dahlborg et al., 1996, age = 60 years, range = 27–67) achieved durable 1998; Neuwelt et al., 1991; Roman-Goldstein et al., remissions without relapse more than 1 year after 1995). Initial studies (Dahlborg et al., 1996; Neuwelt completion of treatment (median follow-up = 36 et al., 1991) showed that PCNSL patients who months). The authors assessed verbal learning obtained a complete response to treatment with and recall, non-verbal recognition memory, word HD-MTX-based chemotherapy with BBB disruption fluency, visuoconstruction, and attention. There (n = 15) had no significant overall cognitive decline was evidence of cognitive improvement (i.e., pri- over a period of several years (median survival = marily in attention and verbal memory) in 4 of 41 months) regardless of age. In contrast, patients the 5 patients who had impaired performance at whose disease has recurred after initial radiotherapy baseline, and no change at follow-up was noted and who subsequently received chemotherapy with in the 3 patients who had intact cognition prior to Chapter 13. Primary central nervous system lymphoma 191

therapy; 2 patients could not be assessed at base- uated at baseline, 4 months, 12 months, and 15–41 line. In the 4 patients who relapsed within 1 year months after completion of treatment (8 patients after treatment, significant cognitive impairment had a complete response); these patients showed was noted in 3 patients after additional treatment stable or improved summary cognitive test scores at (i.e., 2 had chemotherapy, 1 had WBRT). A subse- the last follow-up (median = 95, range = 89–107; 100 quent study including 5 additional patients treated ± 10 as reference value). Nine patients also had pre- with the same regimen (Pels et al., 2003) reported served cognitive functions during follow-up, but no no significant cognitive decline at long-term specific intervals or summary scores were reported. follow-up. One patient required additional chemotherapy due In a more recent study, Fliessbach and colleagues to disease relapse and developed severe cognitive (2005) reported the results of prospective cogni- impairment 21 months after therapy. No informative evaluations in 23 PCNSL patients (median tion was available regarding the patients’ perfor- age = 54 years, range = 28–68) who achieved a mance on each cognitive test domain. complete remission after treatment with HD-MTX- In a prospective study of patients with PCNSL based polychemotherapy. At baseline, 16 patients treated with high-dose chemotherapy and stem had moderate to severe cognitive impairments (i.e., cell transplantation without WBRT (Abrey et al., reduced verbal fluency, attention/executive, and 2003), 14 patients (mean age = 53.6 years, SD = memory abilities), 3 had mild impairment, and 4 6.8) received prospective neuropsychological evalu- had intact cognitive abilities. At long-term follow- ations. Patients performed in the impaired range (z- up (median follow-up = 44 months, range = 17– scores 2 SD below normative sample means) prior 96), 6 patients had moderate to severe cognitive to treatment on several cognitive domains includ- impairment, 5 had mild cognitive difficulties, and ing executive function, memory, and psychomotor 12 had intact cognitive functions; only 4 patients speed. Improvements across all cognitive domains reported diminished quality of life. The authors con- were documented after induction chemotherapy cluded that cognitive functions either improved or with HD-MTX and HD-ARA-C for the 7 patients remained stable at follow-up, suggesting that HD- who had no progressive disease and were available MTX-based chemotherapy is not associated with for initial follow-up (Correa et al., 2003); this was long-term neurotoxicity. However, not all patients consistent with neuroimaging evidence of reduc- received the same tests, and timed measures of psy- tion in lesion size in response to treatment. Three chomotor speed, cognitive flexibility, and working patients who had a complete response to treatment memory were given only to a subset of patients. remained cognitively stable up to 18 months post- It is possible that the limited assessment of exec- transplant (i.e., scores within 1 SD below the nor- utive abilities in these two studies underestimated mative mean). The small number of patients seen the degree of cognitive dysfunction in their patient for long-term cognitive follow-up in this study pre- cohort. cludes any conclusions regarding the possible neu- Schlegel and colleagues (2001) assessed cogni- rotoxicity of this treatment regimen. tive outcome in 20 PCNSL patients (median age = 64 years, range = 27–71) treated with HD-MTX Retrospective studies and HD-ARA-C. The neuropsychological test battery included measures of attention, verbal and Pels and colleagues (2000) described a series of non-verbal memory, verbal fluency, and visuocon- 27 PCNSL patients (age range = 27–74 years) struction; a global index score of cognitive function treated with various regimens including WBRT, was generated by transforming raw scores into stan- high-dose chemotherapy (i.e., HD-MTX, HD-ARA- dard values according to normative data and aver- C), or combined modality therapy. Patients’ survival aging them (mean = 100). Ten patients were eval- time ranged from 2 to 95 months. The cognitive 192 Section 2. Effects of cancer and cancer treatment on cognition

evaluations were conducted subsequent to treat- attention, executive function, and motor speed. Ten ment, but the follow-up intervals were not reported. patients were on disability, four worked at a lower There was evidence of cognitive deficits in 8 of the level, and two worked less than before diagnosis. 13 patients who received WBRT alone or in combi- Correa and colleagues (2004) investigated cogni- nation with high-dose chemotherapy; however, 4 of tive functioning in 28 survivors of PCNSL who were these patients had either disease progression or only treated with WBRT and HD-MTX-based chemother- a partial response to therapy, suggesting that both apy, or high-dose chemotherapy alone, and had the tumor and treatment may have contributed to no disease recurrence. In the study, 18 patients the cognitive impairment. Of the 14 patients treated received WBRT ± HD-MTX-based chemotherapy with high-dose chemotherapy alone, 10 had cog- (median age = 53 years, range = 36–73; mean nitive deficits; 8 of these patients had a complete post-treatment interval = 61 months, SD = 40), response to treatment. The authors included no and 10 patients had HD-MTX-based chemother- information regarding specific cognitive tests used apy alone (median age = 71 years, range = 59–82; or the cognitive domains affected by tumor and/or mean post-treatment interval = 18 months, SD = treatment. 16). Patients who received combined modality treat- Herrlinger and colleagues (2005) studied neu- ment showed mild to moderate cognitive impair- ropsychological functions and quality of life in ments (i.e., scores at least 1.5 SD below the nor- 6 PCNSL patients (age range = 56 to 63 years) mative sample) on tests of complex attention and who survived for at least 48 months and had no executive functions, memory, psychomotor speed, active disease (follow-up range = 55–69 months); all and naming. In comparison to patients treated patients were treated with HD-MTX and one patient with chemotherapy alone, their performance was also had WBRT. There was evidence of mild to mod- more impaired on tests of memory, and atten- erate cognitive impairment in all patients, partic- tion and executive functions, regardless of time ularly in attention and memory functions. Cogni- since completion of treatment. Patients treated with tive dysfunction was more pronounced in the two HD-MTX-based chemotherapy alone had moderate patients with marked leukoencephalopathy (one impairment in psychomotor speed, but performed had combined modality therapy). Three patients within 1 SD below the normative sample on other reported moderate restriction in quality of life, par- cognitive domains. The memory performance of ticularly in cognitive and social functioning; the patients treated with WBRT ± chemotherapy did patient treated with WBRT described significantly not differ according to time since treatment com- decreased quality of life. pletion (≤60 months, n = 9; ≥60 months, n = 9), but Harder and colleagues (2004) studied cognitive was significantly (p < 0.05) more impaired than the abilities in 19 PCNSL patients (median age = 44 performance of patients treated with chemother- years, range = 24–63) treated with HD-MTX-based apy alone (≤60 months), who were also older. Eval- chemotherapy followed by WBRT; patients were uation of quality of life showed that half of the evaluated at least 6 months after treatment com- patients were either not employed or were work- pletion and had no recurrent disease (mean = 23 ing at a lower capacity as a consequence of their months, SD = 14). It was reported that 63% of disease and treatment. The findings suggested that patients showed mild to moderate cognitive impair- combined modality therapy was associated with ments (i.e., four to six test indices 2 SD below the more severe cognitive impairment than chemother- normative mean). In comparison to matched con- apy alone, regardless of time since treatment com- trols with hematological malignancies treated with pletion. The assessment of the specific contribu- systemic chemotherapy or non-CNS radiotherapy, tions of the disease and treatment, and time of onset PCNSL patients obtained lower scores on cognitive and course of neurotoxicity is relatively limited in domains involving verbal and non-verbal memory, this study given its retrospective nature. Chapter 13. Primary central nervous system lymphoma 193

Overall, the findings suggest that combined ter abnormalities in eight patients after HD-MTX- modality treatment for PCNSL with WBRT and HD- based polychemotherapy, but these were not corre- MTX-based chemotherapy results in more severe lated with cognitive performance. Roman-Goldstein cognitive impairment than chemotherapy alone, and colleagues (1995) documented white matter particularly in the areas of attention, executive abnormalities in only one of nine PCNSL patients function, memory, and psychomotor speed. Stud- following treatment with chemotherapy with BBB ies involving patients treated with MTX-based disruption; there was no evidence of cognitive chemotherapy with or without BBB disruption impairment in any of the patients. Schlegel and col- reported variable findings ranging from no signifi- leagues (2001) reported that 6 of 20 PCNSL patients cant cognitive decline or improvement from base- treated with HD-MTX and ARA-C developed conflu- line to cognitive impairment in some patients. How- ent white matter lesions 1–6 months after initiation ever, these trials included a relatively small number of treatment; only 1 of these patients had cognitive of patients who completed long-term follow-up, impairment. Neuwelt and colleagues (2005) docu- some studies evaluated patients with recurrent dis- mented that peri-tumor-enhancing abnormalities ease, and in some studies patients treated with com- were associated with cognitive dysfunction at diag- bined modality therapy had a history of relapse or nosis in 15 PCNSL patients, but not after a complete partial response to therapy suggesting a possible response to MTX-based chemotherapy with BBB selection bias. In addition, not all studies assessed or disruption. Long-term follow-up (n = 9) showed described the specific cognitive domains that may that some patients developed post-treatment dif- have been differentially affected by treatment, such fuse or focal bilateral periventricular abnormali- as executive functions and motor speed. ties, but these were not associated with cognitive performance (overall summary score), which remained stable or improved more than 2 years after Cognitive outcome and treatment-related diagnosis. white matter abnormalities Pels and colleagues (2000) reported that 5 of 13 PCNSL patients treated with WBRT alone or in com- The association between diffuse treatment-related bination with high-dose chemotherapy developed white matter abnormalities and the presence or confluent white matter lesions and marked cog- severity of neuropsychological dysfunction in brain nitive deficits following treatment. Three patients tumor patients is unclear. It may vary in severity, treated with chemotherapy alone developed white ranging from no abnormal clinical findings to pro- matter disease but it was not associated with cog- gressive global cognitive decline (Dropcho, 1991). nitive dysfunction; 10 of the 14 patients treated A moderate association between treatment-related with this regimen had cognitive dysfunction with- white matter changes and cognitive impairment out evidence of significant white matter disease. was found in some but not all studies involving In a subsequent study (Pels et al., 2003), 20 of PCNSL patients (Correa et al., 2004; Fliessbach et al., 57 patients developed white matter abnormalities 2003, 2005; Harder et al., 2004; Pels et al., 2000). during treatment with HD-MTX-based chemother- Fliessbach and colleagues (2003) documented apy, and these remained stable at follow-up; only the development of white matter changes in four 2 of these patients had severe cognitive dysfunc- of ten patients who received a polychemother- tion. Harder and colleagues (2004) reported white apy regimen; in three patients there were con- matter abnormalities and cortical atrophy in 14 fluent subcortical white matter changes; cogni- PCNSL patients (78% of the patient sample) fol- tive functions were within the normal range. In lowing treatment with HD-MTX-based chemother- a subsequent study (Fliessbach et al., 2005), the apy and WBRT; cortical atrophy, but not white authors observed bilateral confluent white mat- matter disease, was significantly correlated with 194 Section 2. Effects of cancer and cancer treatment on cognition

cognitive impairment, age, and Karnofsky Perfor- impairment and diffuse white matter abnormalities. mance Score. Correa and colleagues (2004) found Cognitive dysfunction after treatment with high- that more extensive white matter abnormalities dose chemotherapy alone was reported in some on MRI were significantly correlated with greater but not all studies; it is also unclear if this regi- impairment in executive function, memory, and men results in a compromise of long-term disease language abilities in 28 PCNSL patients; white mat- control. Future collaborative, prospective longitu- ter changes were more extensive in the 18 patients dinal studies are required to determine the inci- treated with WBRT and HD-MTX-based chemother- dence of cognitive dysfunction associated with vari- apy than in the 10 patients who received HD-MTX- ous treatment modalities in patients with PCNSL. At based chemotherapy alone. present, there is no effective therapy for treatment- The variable findings in the literature may related cognitive dysfunction, but recent studies be attributed in part to methodological factors have reported some benefit from psychopharma- (Desmond, 2002), as different scales and MRI cological interventions (Chapter 22) and cognitive sequences were used to measure white matter rehabilitation (Chapter 20). abnormalities across studies. In addition, several studies reported cognitive function as a summary score, and no correlations between white REFERENCES matter abnormalities and specific cognitive functions (e.g., executive function, processing speed) Abrey LE, DeAngelis LM, Yahalom J (1998). Long-term sur- were reported. Nevertheless, the results suggest vival in primary CNS lymphoma. J Clin Oncol 16: 859– that in comparison to chemotherapy-alone regi- 863. mens, WBRT alone or in combination with HD- Abrey LE, Yahalom J, DeAngelis LM (2000). Treatment of MTX-based chemotherapy produces more exten- primary CNS lymphoma: the next step. J Clin Oncol 18: sive white matter abnormalities, which are associ- 3144–3150. ated with cognitive impairment. These observations Abrey L, Moskowitz CH, Mason WP et al. (2003). Inten- are consistent with evidence that more extensive sive methotrexate and cytarabine followed by high- white matter disease may be necessary to produce dose chemotherapy with autologous stem-cell rescue in patients with newly diagnosed primary CNS lymphoma: measurable cognitive deficits, and that only spe- an intent-to-treat analysis. J Clin Oncol 21: 4151–4156. cific cognitive domains, such as executive functions Batchelor T, Loeffler JS (2006). Primary CNS lymphoma. and processing speed, are disrupted by diffuse white J Clin Oncol 24: 1281–1288. matter disease (Tullberg et al., 2004). Behin A, Delattre J-Y (2003). Neurologic sequelae of radiotherapy on the nervous system. In Schiff D, Wen PY (eds.) Cancer Neurology in Clinical Practice (pp. 173– Conclusion 191). Totowa, NJ: Humana Press Inc. Besell EM, Graus F, Lopez-Guillermo A et al. (2001). Treatment-related neurotoxicity has been recog- CHOD/BVAM regimen plus radiotherapy in patients with nized as a significant problem as therapy regimens primary CNS non-Hodgkin’s lymphoma. Int J Radiat for PCNSL have prolonged survival. However, the Oncol Biol Phys 50: 457–464. Blay J-Y, Conroy T, Chevreau C et al. (1998). High-dose incidence of cognitive dysfunction in this popula- methotrexate for the treatment of primary cerebral lym- tion may have been underestimated as only a small phomas: analysis of survival and late neurologic toxicity number of clinical trials have included formal neu- in a retrospective series. JClinOncol16: 864–871. ropsychological assessment as an outcome mea- Boiardi A, Silvani A, Valentini S et al. (1993). Chemother- sure. The studies reviewed suggest that treatment apy as a first treatment for primary non-Hodgkin’s lym- involving a combination of WBRT and HD-MTX- phoma of the central nervous system: preliminary data. based chemotherapy is associated with cognitive J Neurol 241: 96–100. Chapter 13. Primary central nervous system lymphoma 195

Cher L, Glass J, Harsh GR et al. (1996). Therapy of primary bined treatment with chemotherapy and radiotherapy. CNS lymphoma with methotrexate-based chemotherapy Neurology 40(1): 80–86. and deferred chemotherapy: preliminary results. Neurol- DeAngelis LM, Yahalom J, Thaler HT et al. (1992). Com- ogy 46: 1757–1759. bined modality therapy for primary CNS lymphoma. Constine LS, Konski A, Ekholm S et al. (1988). Adverse J Clin Oncol 10: 635–643. effects of brain irradiation correlated with MR and CT Desmond DW (2002). Cognition and white matter lesions. imaging. Int J Radiat Oncol Biol Phys 15: 319–330. Cerebrovasc Dis 13: 53–57. Correa DD, Anderson ND, Glass A et al. (2003). Cog- Doolittle ND, Miner ME, Siegal T et al. (2000). Safety nitive functions in primary central nervous system and efficacy of a multicenter study using intraarterial lymphoma patients treated with chemotherapy and chemotherapy in conjunction with osmotic opening of stem cell transplantation: preliminary findings. Clin Adv the blood-brain barrier for the treatment of patients with Hematol Oncol 1: 490. malignant brain tumors. Cancer 88: 637–647. Correa DD, DeAngelis LM, Shi W et al. (2004). Cognitive Dropcho EJ (1991). Central nervous system injury by thera- functions in survivors of primary central nervous system peutic irradiation. Neurol Clin 9: 69–88. lymphoma. Neurology 62: 548–555. Eby NL, Grufferman S, Flannelly CM et al. (1988). Increas- Corry J, Smith JG, Wirth A et al. (1998). Primary central ing incidence of primary brain lymphoma in the U.S. nervous system lymphoma: age and performance status Cancer 62: 2461–2465. are more important than treatment modality. Int J Radiat Ferreri AJM, Reni M, Pasini F et al. (2002). A multicenter Oncol 41: 615–620. study of treatment of primary CNS lymphoma. Neurol- Crossen JR, Goldman DL, Dahlborg SA et al. (1992). Neu- ogy 58: 1513–1520. ropsychological asssessment outcomes of nonacquired Fliessbach K, Urbach H, Helmstaedter C et al. (2003). Cog- immunodeficiency syndrome patients with primary cen- nitive performance and magnetic resonance imaging tral nervous system lymphoma before and after blood- findings after high-dose systemic and intraventricular brain barrier disruption chemotherapy. Neurosurgery 30: chemotherapy for primary central nervous system lym- 23–29. phoma. Arch Neurol 60: 563–568. Dahlborg SA, Henner WD, Crossen JR et al. (1996). Non- Fliessbach K, Helmstaedter C, Urbach H et al. (2005). Neu- AIDS primary CNS lymphoma: first example of a durable ropsychological outcome after chemotherapy for pri- response in a primary brain tumor using enhanced mary CNS lymphoma: a prospective study. Neurology 64: chemotherapy delivery without cognitive loss and with- 1184–1188. out radiotherapy. Cancer J Sci Am 2: 166–174. Freilich RJ, Delattre J-Y, Monjour A, DeAngelis LM (1996). Dahlborg SA, Petrillo A, Crossen JR et al. (1998). The poten- Chemotherapy without radiation therapy as initial treat- tial for complete and durable response in nonglial pri- ment for primary CNS lymphoma in older patients. Neu- mary brain tumors in children and young adults with rology 46: 435–439. enhanced chemotherapy delivery. Cancer J Sci Am 4(2): Glass J, Gruber ML, Cher L, Hochberg F (1994). Preirradia- 110–124. tion methotrexate chemotherapy of primary central ner- DeAngelis LM (1995). Current management of primary vous system lymphoma: long-term outcome. J Neurosurg central nervous system lymphoma. Oncology 9: 63–71. 81: 88–195. DeAngelis LM, Hormigo A (2004). Treatment of primary Grant JW, Isaacson PG (1992). Primary central nervous sys- central nervous system lymphoma. Semin Oncol 31: 684– tem lymphoma. Brain Pathol 2: 97–109. 692. Guha-Thakurta N, Damek D, Pollack C et al. (1999). Intra- DeAngelis LM, Shapiro WR (1991). Drug/radiation inter- venous methotrexate as initial treatment for primary actions and central nervous system injury. In Gutin PH, central nervous system lymphoma: response to therapy Leibel SA, Sheline GE (eds.) Radiation Injury to the Ner- and quality of life of patients. J Neurooncol 43: 259– vous System (pp. 361–382). New York: Raven Press. 268. DeAngelis LM, Delattre J-Y, Posner JB (1989). Radiation- Harder H, Holtel H, Bromberg JEC et al. (2004). Cognitive induced dementia in patients cured of brain metastases. status and quality of life after treatment for primary CNS Neurology 39: 789–796. lymphoma. Neurology 62: 544–547. DeAngelis LM, Yahalom J, Heinemann MH, Cirrincione C, Herrlinger U, Schabet M, Clemens M et al. (1998). Clini- Thaler HT, Krol G (1990). Primary CNS lymphoma: com- cal presentation and therapeutic outcome in 26 patients 196 Section 2. Effects of cancer and cancer treatment on cognition

with primary CNS lymphoma. Acta Neurol Scand 97: irradiation in primary CNS lymphoma. J Clin Oncol 18: 257–264. 519–526. Herrlinger U, Schabet M, Brugger W et al. (2002). German O’Neill BP (2004). Neurocognitive outcomes in primary Cancer Society Neuro-Oncology Working Group NOA-03 CNS lymphoma (PCNSL). Neurology 62: 532–533. multicenter trial of single-agent high-dose methotrexate O’Neill BP, Illig JJ (1989). Primary central nervous system for primary central nervous system lymphoma. Ann Neu- lymphoma. Mayo Clinic Proc 64: 1005–1020. rol 51: 247–252. Pels H, Schlegel U (2006). Primary central nervous system Herrlinger U, Kuker W, Uhl M et al. (2005). NOA-03 trial of lymphoma. Curr Treat Options Neurol 8: 346–357. high-dose methotrexate in primary central nervous sys- Pels H, Deckert-Schulter M, Glasmacher A et al. (2000). Pri- tem lymphoma: ﬁnal report. Ann Neurol 57: 843–847. mary central nervous system lymphoma: a clinicopatho- Hoang-Xuan K, Taillandier L, Chinot O et al. (2003). logical study of 28 cases. Hematol Oncol 18: 21–32. Chemotherapy alone as initial treatment for CNS lym- Pels H, Schmidt-Wolf IGH, Glasmacher A et al. (2003). Pri- phoma in patients older than 60 years: a multicenter mary central nervous system lymphoma: results of a phase II study (26952) of the European Organization for pilot study and phase II study of systemic and intraven- Research and Treatment of Cancer Brain Control Group. tricular chemotherapy with deferred radiotherapy. JClin J Clin Oncol 21: 2726–2731. Oncol 21: 4489–4495. Keime-Guibert F, Napolitano M, Delattre JY (1998). Neu- Peterson K, DeAngelis LM (1997). Primary cerebral lym- rological complications of radiotherapy and chemother- phoma. In Vecht CJ (ed.) Handbook of Clinical Neurology apy. J Neurol 245: 695–708. (Vol. 24, pp. 257–268). Part II: Neuro-Oncology. Amster- Kraemer DF, Fortin D, Neuwelt EA (2002). Chemothera- dam: Elsevier Science. peutic dose intensiﬁcation for treatment of malignant Peterson K, Gordon KB, Heinemann M-H et al. (1993). The brain tumors: recent developments and future direc- clinical spectrum of ocular lymphoma. Cancer 72: 843– tions. Curr Neurol Neurosci Rep 2: 216–224. 849. Laack NN, Brown PD (2004). Cognitive sequelae of brain Poortmans PMP, Kluin-Nelemans HC, Haaxma-Reiche H radiationinadults.Semin Oncol 31: 702–713. et al. (2003). High-dose methotrexate-based chemother- Lai R, Abrey LE, Rosenblum MK, DeAngelis LD (2004). apy followed by consolidating radiotherapy in non-AIDS- Treatment-induced leukoencephalopathy in primary related primary central nervous system lymphoma: CNS lymphoma. Neurology 62: 451–456. European Organization for Research and Treatment of McAllister LD, Doolittle ND, Guastadisegni PE et al. (2000). Cancer Lymphoma Group phase II Trial 20962. J Clin Cognitive outcomes and long-term follow-up results Oncol 21: 4483–4488. after enhanced chemotherapy delivery for primary cen- Roman-Goldstein SM, Mitchell P, Crossen JR et al. tral nervous system lymphoma. Neurosurgery, 46: 51–61. (1995). MR and cognitive testing of patients undergoing Meyers CA, Wefel JS (2003). The use of the mini-mental osmotic blood-brain barrier disruption with intraarterial state examination to assess cognitive functioning in can- chemotherapy. Am J Neuroradiol 16: 543–553. cer trials: no ifs, ands, buts, or sensitivity. J Clin Oncol 21: Sandor V, Stark-Vancs V, Pearson D et al. (1998). Phase 3557–3558. II trial of chemotherapy alone for primary CNS Nelson DF (1999). Radiotherapy in the treatment of pri- and intraocular lymphoma. J Clin Oncol 16: 3000– mary central nervous system lymphoma (PCNSL). J Neu- 3006. rooncol 43: 241–247. Sarazin M, Ameri A, Monjour A et al. (1995). Primary central Neuwelt EA, Goldamn DA, Dahlborg SA et al. (1991). Pri- nervous system lymphoma: treatment with chemother- mary CNS lymphoma treated with osmotic blood-brain apy and radiotherapy. Eur J Cancer 31A: 2003–2007. barrier disruption: prolonged survival and preservation Schlegel U, Pels H, Glasmacher A et al. (2001). Combined of cognitive function. J Clin Oncol 9: 1580–1590. systemic and intraventricular chemotherapy in primary Neuwelt EA, Guastadisegni PE, Varallyay P et al. (2005). CNS lymphoma: a pilot study. J Neurol Neurosurg Psychi- Imaging changes and cognitive outcome in primary CNS atry 71: 118–122. lymphoma after enhanced chemotherapy delivery. Am J Sheline G (1980). Irradiation injury of the human brain: a Neuroradiol 26: 258–265. review of clinical experience. In Gilbert HA, Kagan AR O’Brien P, Roos D, Pratt G et al. (2000). Phase II multicen- (eds.) Radiation Damage to the Nervous System (pp. 39– ter study of a brief single agent methotrexate followed by 68). New York: Raven Press. Chapter 13. Primary central nervous system lymphoma 197

Sheline G, Wara WM, Smith V (1980). Therapeutic irradia- Van Der Maazen RW, Kleiboer BJ, Verhagen I et al. (1993). tion and brain injury. Int J Radiat Oncol Biol Physics 6: Repair capacity of adult rat glial progenitor cells deter- 1215–1228. mined by an in vitro clonogenic assay after in vitro or Toﬁlon PJ, Fike JR (2000). The radioresponse of the central in vivo fractionated irradiation. Int J Radiat Biol 63: 661– nervous system: a dynamic process. Radiat Res 153: 357– 666. 370. Weitzner MA, Meyers CA (1997). Cognitive function- Tullberg M, Fletcher E, DeCarli C et al. (2004). White mat- ing and quality of life in malignant glioma patients: ter lesions impair frontal lobe function regardless of their a review of the literature. Psychooncology 6: 169– location. Neurology 63: 246–253. 177. 14

Childhood brain tumors

H. Stacy Nicholson, Louise Penkman Fennell, and Robert W. Butler

In long-term survivors of childhood and adolescent CNS tumor survivors were among the most likely to central nervous system (CNS) tumors, neuropsy- have chronic health conditions and multiple other chological and psychosocial late effects of therapy chronic conditions; in addition, they often have occur in a milieu of numerous medical late compli- functional impairments (Ness et al., 2005). Long- cations (Anderson et al., 2001). With improvements term medical surveillance of survivors is critically in treatment and survival rates for these patients important so that late effects of therapy may be since the 1980s, most children and adolescents with detected while still possibly amenable to interven- CNS tumors currently diagnosed will become long- tion (Oefﬁnger et al., 2004). term survivors (Pollack, 1994). This is particularly true for children with medulloblastoma or low- grade astrocytoma. Therefore, concerns about late complications of therapy are increasingly important Medical late effects to survivors and their families. In addition to the late consequences of radiation therapy and chemother- Although 5-year disease-free survival is the outcome apy, which are similar for all survivors of child- measure most often used in clinical trials, it may not hood cancer, the singular susceptibility of the brain correlate with a normal life expectancy. In one large and spinal cord to injury causes several late con- cohort study of adult survivors of childhood can- sequences unique to long-term survivors of CNS cer (Nicholson et al., 1994), survivors of CNS tumors tumors. were much more likely to die during adulthood than Late complications may be due to the tumor, were survivors of all other childhood and adoles- surgery, radiation therapy, chemotherapy, or the cent malignancies, except for those with Hodgkin’s psychologicaltraumaofdealingwithamalignancy, disease. In this study CNS tumor survivors had a and the late effects following CNS tumors include 9.2-fold excess risk of death from causes other than medical, psychological, neuropsychological, and their primary cancer diagnosis during their thirties. psychosocial problems. Some late effects may be Non-tumor causes of death included trauma, pneu- life threatening. In fact, long-term survivors of CNS monia and other respiratory diseases, cerebrovas- tumors have an excess risk of mortality relative cular disease, and aspiration of emesis. In a more to survivors of other cancers (Mostow et al., 1991; recently treated cohort, CNS tumor survivors had an Nicholson et al., 1994). In a large cohort study of almost 16-fold increase in mortality, and survivors childhood cancer survivors (Oefﬁnger et al., 2006), were still experiencing an excess risk of mortality

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

198 Chapter 14. Childhood brain tumors 199

25 years after diagnosis (Mertens et al., 2001). These as secondary cancers, may impact future treat- mortality data underscore the serious nature of late ment decisions. In a study with excellent treatment complications in this population. outcomes for 63 children with medulloblastoma treated with chemotherapy [lomustine (CCNU), cisplatin, and vincristine] and craniospinal radiother- Secondary malignancies apy, three patients developed secondary malignan- Secondary malignancies are increasingly problem- cies [two CNS tumors and one acute myelogenous atic for cancer survivors (Goldstein et al., 1997; leukemia (AML)] (Packer et al., 1994). The risk of Peterson et al., 2005; Stavrou et al., 2001; Travis a secondary malignancy may also be due to fac- et al., 2006). In CNS tumor survivors, the most tors other than the anticancer therapy, such as an frequent secondary malignancy is another brain underlying genetic predisposition to malignancy tumor (Neglia et al., 2006). These are typically asso- (Goldstein et al., 1994). For such patients, the risk ciated with radiotherapy, and the most common of a secondary cancer will likely be greater than it radiation-induced tumors are high-grade gliomas is for other children with the same primary can- and meningiomas. In addition, radiation may also cer. Although most children with brain tumors do invoke thyroid cancer (Ronckers et al., 2006). Also, not have a known genetic predisposition to can- as chemotherapy has been increasingly used, sec- cer, there are rare genetic syndromes associated ondary leukemia has occurred (Packer et al., 1994). with pediatric brain tumors (Goldstein et al., 1994; Estimates of the risk of a subsequent cancer after Hamilton et al., 1995). These include Gorlin syn- CNS tumors are best based on epidemiologic stud- drome, in which children tend to be diagnosed with ies. In a population-based study of 1262 histolog- medulloblastoma at a particularly young age. These ically confirmed cases of medulloblastoma in the children are particularly susceptible to basal cell United States and Sweden (Goldstein et al., 1997), carcinomas in the radiation fields. Turcot syndrome 20 secondary malignancies occurred. This corre- includes brain tumors and multiple colonic polyps sponded to a 5.4-fold excess of secondary neo- (Hamilton et al., 1995). A family history can be par- plasms (95% confidence interval, CI: 3.3–8.4) rela- ticularly helpful in ascertaining which patients are tive to the number expected based on population at increased risk of secondary malignancies. The data, and the median latency between the medul- family history should be updated at each annual loblastoma and the secondary malignancy was 73 follow-up visit for long-term survivors. months (range 8 months to 36 years). These second malignancies included cancers of the salivary Cardiac complications glands, uterine cervix, CNS, thyroid gland, and acute lymphoblastic leukemia (ALL); 46% of secondary Although anthracycline-induced cardiomyopathy malignancies occurred in or near the radiation field. is of concern for most childhood cancer survivors In another study, secondary carcinomas were less (Lipshultz et al., 1991), these agents are not rou- problematic for survivors of CNS tumors when com- tinely used to treat brain tumors. Therefore, this pared to other childhood cancer survivors (Bassal serious toxicity usually does not occur in survivors et al., 2006). The cumulative risk of secondary of CNS tumors. However, other chemotherapeutic malignancies will likely increase with time as sur- agents, such as cyclophosphamide, have also been vivor cohorts age. Latency periods of more than associated with cardiac dysfunction. Radiation six decades for radiation-induced CNS tumors have is also known to damage the heart and lead to been reported (Kleinschmidt-DeMasters & Lillehei, premature atherosclerosis. This is particularly 1995). true for long-term survivors of Hodgkin’s disease Similarly, follow-up beyond the usual 5-year treated with high doses of radiation (Donaldson disease-free survival outcome for clinical trials will & Kaplan, 1982). Children with brain tumors who be important because severe late effects, such received spinal irradiation have had some cardiac 200 Section 2. Effects of cancer and cancer treatment on cognition

abnormalities documented, presumably due to the small, 1 of 7 patients (14%) who received CCNU exit beam. In one study that included 26 patients without spinal radiotherapy, compared with 13 of who had received spinal radiotherapy, cardiac 21 of those (62%) who received spinal radiother- evaluations included electrocardiography, 24-h apy (with or without lomustine), developed pul- ambulatory electrocardiography, echocardiogra- monary fibrosis (Jakacki et al., 1995). This included phy, and exercise testing. Of the 16 patients who 4 of 8 who did not receive CCNU. In this study, pul- were exercise-tested, 75% achieved a maximal monary fibrosis was most associated with a history cardiac index below the fifth percentile. In addition, of spinal irradiation and not with CCNU exposure. 31% had pathologic Q waves, and there was an Thus, survivors who received spinal radiotherapy excess of elevated posterior wall stress (Jakacki et for their CNS tumor have a risk of late pulmonary al., 1993). Although the long-term significance of complications. In a large study not limited to CNS these findings is not yet known, this study points tumor survivors (Mertens et al., 2002), those who out that cardiac function needs to be followed in had received chest radiotherapy had a 3.5% 20-year survivors who received spinal radiotherapy. In the cumulative incidence of pulmonary fibrosis. More Childhood Cancer Survival Study (CCSS) cohort, precise estimates of risk in CNS tumor survivors are 18% of survivors reported a cardiovascular compli- needed. cation, including stroke (relative risk, RR = 42.8), blood clots (RR = 5.7) and angina (RR = 2) (Gurney Endocrine complications et al., 2003a). These findings have been confirmed by others (Bowers et al., 2002). Whether CNS Hormonal deficiencies are among the most com- tumor survivors have an increased risk of obesity is mon late effects of therapy in long-term survivors debated (Gurney et al., 2003b; Heikens et al., 2000), of childhood brain tumors (Oberfield et al., 1996; but those who are overweight would be at increased Sklar, 1995, 1997; Sklar & Constine, 1995), with risk of cardiovascular disease. 43% of survivors reporting at least one endocrine abnormality in the CCSS cohort (Gurney et al., 2003a). Endocrine late effects can result from the Pulmonary complications tumor itself, surgery, or radiotherapy. Tumors of the Few data exist regarding late pulmonary com- pituitary or hypothalamic region often present with plications in long-term survivors of pediatric pituitary dysfunction. Furthermore, the surgical CNS tumors. However, many patients, including treatment of tumors in this region also poses a risk those with astrocytoma and medulloblastoma, are of pituitary damage. In these patients, either the exposed to the nitrosoureas, which are known to tumor itself or surgery may result in panhypopitu- be associated with pulmonary fibrosis. The risk itarism. However, by far the most common cause of apparently does not decline with time, and fatal late endocrinologic sequelae is radiotherapy (Sklar pulmonary fibrosis has been described as long as & Constine, 1995). Focal radiotherapy that includes 17 years after exposure (O’Driscoll et al., 1990). In the pituitary gland, whole-brain radiotherapy, addition to the nitrosoureas, other chemotherapeu- and/or craniospinal radiotherapy all involve the tic agents are known to have pulmonary fibrosis pituitary and carry a risk of hormonal deficiencies. as a potential complication, including cyclophos- The deficiencies, which are dose-dependent, most phamide. The risk from agents other than the commonly include growth hormone deficiency. In nitrosoureas is not likely to be high. addition, low-dose cranial irradiation (18 Gy) may Pulmonary fibrosis may also be associated with be associated with the premature onset of puberty, spinal radiotherapy. In a study of 28 survivors of and higher doses of irradiation (more than 40 Gy) childhood brain tumors, half had significant pul- may lead to deficiencies of gonadotropins and monary fibrosis. Although the sample sizes are thyroid-hormone-releasing hormone, as well as Chapter 14. Childhood brain tumors 201

hyperprolactinemia (Oberfield et al., 1996). Fur- However, severe cisplatin ototoxicity in the speech thermore, the risk of developing hormonal compli- frequencies can often be prevented with careful cations of irradiation does not appear to decrease monitoring and dose adjustment during therapy. In with time. Thus, careful follow-up of growth, puber- the CCSS, a retrospective cohort study comparing tal development, and thyroid function is critical survivors to a control group of siblings without a because hormonal deficiencies are universally history of cancer, childhood brain tumor survivors treatable. had an elevated risk of hearing impairment (RR = Thyroid function should be followed for life in 17.3), legal blindness in at least one eye (RR = 14.8), all children who received irradiation to the thy- cataracts (RR = 11.9), and double vision (RR = 8.8) roid gland, regardless of whether or not the brain (Packer et al., 2003). was irradiated, and regardless of the radiotherapy As part of routine follow-up, all children who dose to the pituitary gland (Sklar & Constine, 1995). received cisplatin should have their hearing tested As with other hormonal deficiencies, the risk of because hearing loss can occur or worsen even hypothyroidism does not appear to decrease over when audition was previously documented to be time (Hancock et al., 1991). In addition, radiation sufficient; furthermore, many survivors can bene- therapy may be associated with both benign and fit from amplification. Also, as hearing impairment malignant thyroid nodules. Thus, as part of the may affect school performance, the results of the annual checkup, these patients should have careful hearing evaluation should be shared with the neu- thyroid palpation. ropsychologist so that neurocognitive results can be In general, hormonal deficiencies can be treated effectively integrated with educational recommen- with replacement therapy (Hancock et al., 1991; dations. Oberfield et al., 1996; Sklar, 1995, 1997; Sklar & Constine, 1995; Vassilopoulou et al., 1995). In fact, Renal complications children with growth hormone deficiency due to Some chemotherapeutic agents used in the treat- radiotherapy respond to replacement therapy as ment of CNS tumors are nephrotoxic, such as well as do children with idiopathic growth hormone cisplatin and ifosfamide (Daugaard & Abildgaard, deficiency (Vassilopoulou et al., 1995), although 1991). Cisplatin can both decrease the glomerular those survivors who have received spinal radiother- filtration rate (GFR) and cause electrolyte abnor- apy will not respond as well as those who have malities. Also, wasting of magnesium and potas- not (Brownstein et al., 2004). The use of growth sium are common in children who have received hormone appears not to be associated with an cisplatin, and some of these survivors may require increased risk of relapse, although there may be a long-term supplementation. Similarly, patients who slight increase in risk of secondary malignancies receive ifosfamide may develop renal Fanconi syn- (Ergun-Longmire et al., 2006). Consultation with an drome, including renal tubular acidosis, phospha- endocrinologist is important if endocrine deficits turia, and glucosuria. The development of elec- are known or suspected. trolyte abnormalities long after chemotherapy has ended is not expected, and whether the GFR Disorders of hearing and sight decreases with time is not well understood.

Depending on tumor location, either hearing or Gastrointestinal and hepatic complications sight can be affected. In addition, chemotherapy with cisplatin can also cause sensorineural hearing There is generally little risk of late gastrointestinal loss in virtually all patients (Skinner et al., 1990). sequelae in long-term survivors of brain tumors. This hearing loss is irreversible and is likely to be As noted above, rarely patients may have their exacerbated by radiation therapy to the inner ear. brain tumor as the initial manifestation of Turcot 202 Section 2. Effects of cancer and cancer treatment on cognition

syndrome (Hamilton et al., 1995); therefore, any in survivors was 39 times greater than that in sib- patient with obstructive symptoms or bloody stool ling controls (Bowers et al., 2006), with a rate of should be considered to be at risk for colonic polyps. nearly 268 events per 100 000 person-years. The risk As with other cancer survivors who received blood increased with increasing doses of radiotherapy. products, there is a small risk of blood-borne infec- Neurological complications in survivors are quite tions, including hepatitis. Patients transfused prior common. In the CCSS cohort, 49% of survivors to 1992, when routine screening for hepatitis C was reported co-ordination problems, 26% had motor implemented, should be screened for hepatitis C problems, and 25% had seizures (Packer et al., 2003). (Luban, 1998). Finally, in a large retrospective cohort study, survivors of brain tumors were at increased risk of being hospitalized for psychiatric problems (Ross et al., Neurological complications 2003). Children with brain tumors have malignancies and therapy that both directly affect the brain, and neurological sequelae are common. These Neuropsychological late effects include seizures, paralysis, radiation necrosis, and migraine-like symptoms (Martins et al., 1977; The neuropsychological and psychosocial seque- Shuper et al., 1995). Some of the neurological lae of childhood brain tumors and their treatment complications may be severe and life-threatening remain one of the most significant challenges to (Mostow et al., 1991; Nicholson et al., 1994). managing late effects. Disease-free survival in this Although the tumor and surgery can also cause population is increasing (Ries et al., 2005). Unfor- neurological damage, most late neurological com- tunately, the incidence of childhood brain tumors plications can be traced to radiation therapy also appears to be increasing. Treatment for child- (Kramer & Lee, 1974). Whether injuries result from hood leukemias has evolved to a point where cra- direct radiation damage to neurons and/or glial niospinal irradiation is largely avoided as a CNS cells, or to the vasculature, or to a combination prophylactic treatment. This has, correspondingly, of both is not well understood. Radiation necro- resulted in fewer neuropsychological late effects in sis can be a particular problem for the survivor of this population (Mulhern & Butler, 2004). However, CNS tumors requiring steroids or surgery (Martins treatment for the more common childhood brain et al., 1977). These lesions often mimic a tumor, and tumors involves relatively high doses of whole and whether a new mass in the radiation bed represents focal brain irradiation. Correspondingly, neuropsy- recurrent tumor or radiation necrosis can be diffi- chological and psychosocial late effects continue cult to ascertain by CT or MRI. This complication to be prominent, and the nature of the deficits usually occurs between 9 months and 2 years after is reasonably well understood at this time. Cur- radiation, and symptoms vary, depending on the rent research efforts are increasingly being directed location in the brain. Radiation necrosis can cause towards the development and testing of interven- headache, behavioral changes, seizures, lethargy, tions designed to lessen cognitive and social impair- hemiparesis, ataxia, and/or increased intracranial ment (Penkman, 2004). pressure. In addition, vascular changes may occur following Neurocognitive late effects radiation and in the most severe cases may lead to astroke(Reinholdet al., 1990). Moyamoya disease, Earlier studies were summarized in a review paper in which the small blood vessels have the abnormal authored by Nicholson and Butler (2001). It is appearance of a “puff of smoke,” can also occur in important to keep in mind that cranial irradiation this setting. In the CCSS cohort, the risk of stroke is not the only CNS insult associated with brain Chapter 14. Childhood brain tumors 203

tumors as identified above. In terms of irradiation Age is a significant risk factor for neuropsycholog- (RT), a pattern of deficits that includes attentional ical dysfunction following treatment with RT. Mulh- dysfunction (particularly under conditions of vig- ern and colleagues (Mulhern et al., 2001) examined ilance), non-dominant hemisphere deficits (such the role of age and white matter loss in young chil- as visual-motor integration difficulties), declines dren treated for medulloblastoma. Children receiv- in performance intelligence, and spatial awareness ing RT for treatment of medulloblastoma prior to impairments frequently result in the presentation of 4 years of age were at the greatest risk for neu- a non-verbal learning disability (Butler et al., 1994; ronal and glial cell damage. In another study by this Packer et al., 1989; Radcliffe et al., 1992; Ris & Noll, group (Reddick et al., 2005), age at time of RT and 1994). Other related insults such as resection and the eventual non-development of normal appear- chemotherapy are poorly understood. ing white matter (NAWM) were shown to be sig- The majority of childhood brain tumors, as noted nificantly related. They suggest that the process of previously, are posterior fossa in nature. Neverthe- myelination is halted at an earlier stage for young less, supratentorial tumors that are located in the children, and likely results in more severe intellec- cortex do occur. Correspondingly, cognitive impair- tual impairment. ment and declines are associated with tumor loca- As the research base examining the neurocog- tion. Thus, memory impairment, language deficits, nitive outcomes of treatment for brain tumors and motor dysfunction can also be present. accumulates, an evolution in focus of study has More current literature has further advanced occurred. Initially, researchers primarily investi- our understanding of tumor- and treatment-related gated global intellectual functioning (IQ scores) as neurocognitive late effects in the pediatric malig- the sum measure of neuropsychological outcome nancy population. There is considerable research (Ellenberg et al., 1987; Jannoun & Bloom, 1990; evidencing the significant cognitive declines and Mulhern et al., 1992). This was followed by a more subsequent academic failures experienced by chil- detailed and comprehensive approach with tests dren who are treated with RT (Mulhern et al., assessing many areas of specific cognitive function 2001; Ris et al., 2001). In fact, radiation injury to (see Ris & Noll, 1994, for a comprehensive review the brain is regarded as one of the most serious of the earlier literature). As knowledge amasses, complications of this treatment, and is considered studies are now beginning to implement theory- the major limitation in delivering high-dose radia- driven approaches examining processes that would tion (Strother et al., 2002). Radiation-induced brain be expected to be compromised, given the underly- injury includes edema formation, damage to glial ing neuropathology of radiation-induced injury. cells that inhibits the development of myelin, and It is now well accepted that global IQ is sig- vascular damage leading to white matter necrosis. nificantly impacted by RT in children with brain Magnetic resonance imaging studies have shown tumors. The estimates vary, but 22-point (Walter less white matter in children treated with RT as et al., 1999) and 17.4-point drops in full-scale IQ compared to children treated with surgery alone have been reported in the literature (Ris et al., (Mulhern et al., 1999) or healthy controls (Reddick 2001). Ris and colleagues (2001) estimated the et al., 2005). Radiation-induced injury is believed rate of change per year to be a reduction of 4 to be a progressive process, as opposed to a static IQ points within verbal, performance, and full- injury, as evidenced by the observation that most scale IQ domains in a group of 43 children with children do not manifest measurable deficits until average-risk tumors of the posterior fossa treated 1–3 years after RT has been completed, and because with reduced-dose RT and adjuvant chemother- their measured deficit appears to increase over time apy. More recently, Reimers and colleagues (Reimers (Hoppe-Hirsch et al., 1995; Mulhern & Butler, 2004, et al., 2003) reported a mean full-scale IQ score 2005). nearly 1 SD below the population mean of 100 in a 204 Section 2. Effects of cancer and cancer treatment on cognition

large group of pediatric brain tumor survivors. This sustained attention abilities of survivors of malig- decrement in IQ scores has been shown to be due nant brain tumors treated with RT using the Con- to a decline in rate of learning rather than a loss of ners’ Continuous Performance Test. Their group previously acquired skills (Palmer et al., 2001). The demonstrated poor performance and reduced white decline in IQ has also been demonstrated to corre- matter volume in the prefrontal cortex and cingulate with degree of white matter loss (Mulhern et al., late gyrus, areas typically activated during attention 2001; Reddick et al., 2003). tasks. A model has been proposed whereby intel- Studies using neuropsychological test batter- lectual and academic deficits could be explained ies have identified deficits in a number of areas by core deficits in attention and memory. It was beyond global IQ. Impairment has been identified reported that the primary consequence of reduced in visuomotor and visual perceptual skills, atten- NAWM in a group of pediatric patients treated for tion, memory, language, and executive functions brain tumors with RT was decreased attentional (Anderson et al., 2001; Bordeaux et al., 1988; Butler abilities, and that this deficit led to reduced IQ and et al., 1994; Copeland et al., 1999; Ris et al., 2001; academic achievement (Reddick et al., 2003). Oth- Riva et al., 1989). These children also seem to ers have examined attention, working memory, and struggle with math at school more than they do with processing speed in a group of children treated for the acquisition and maintenance of reading abili- malignant brain tumors with RT and chemother- ties (Buono et al., 1998; Butler et al., 1994; Fletcher & apy, a group of children treated with surgery alone, Copeland, 1988; Jankovic et al., 1994). As described and a group of children treated for a non-CNS above, some have likened the deficits observed in cancer as a control group (Mabbott et al., 2005). this group of children to a non-verbal learning dis- Although there were trends for the RT group to ability (NVLD) because they evidence a high rate evidence lower scores in all areas, only processing of difficulties with visual-spatial problem solving speed resulted in a statistically significant differ- and arithmetic (Anderson et al., 2000; Buono et al., ence, with the children receiving RT evidencing the 1998). This implicates greater dysfunction of the poorest performance. Hierarchical regression anal- non-dominant hemisphere. ysis revealed that processing speed accounted for It is now evident that brain tumor survivors unique variance in intellectual functioning. treated with RT evidence deficits in core neuropsy- Very recent published research from Dr. Mulhern’s chological processes such as attention, process- group and others is further elucidating the relation- ing speed, and working memory skills that result ship between memory, attention, and new learn- in the secondarily observed deficits in knowledge ing following treatment for medulloblastoma, the acquisition and ultimately academic performance. most common brain tumor in children. It is becom- Attention, memory, and processing speed have been ing increasingly apparent that damage to normal implicated as areas of deficit that emerge several white matter following irradiation, and possibly years following RT for treatment of a brain tumor chemotherapy treatments are primarily respon- (Mulhern et al., 1998). Processes such as attention sible for neurocognitive deficits in children. These and speed of processing are thought to rely on dis- individuals clearly suffer deficits in sustained atten- tributed neural networks that are dependent on tion, reaction time, and processing speed (Reeves white matter tracts for efficient processing. There- et al., 2006). The neuropsychological pattern of fore, it follows that deficits in these areas would attentional difficulties with reduced processing be observed. Reeves and colleagues (Reeves et al., speed and slowed reaction time is undoubtedly 2006) reported impaired sustained attention in the due to white matter damage (Filley, 2001). Sus- context of no impairment in verbal memory in tained attention and processing speed, in addition a group of survivors of medulloblastoma. Mulh- to reaction time difficulties, are clearly present in ern and colleagues (Mulhern et al., 2004) assessed the majority of children who receive treatment for Chapter 14. Childhood brain tumors 205

the most common pediatric brain tumors. These child/adolescent patient from a systems perspec- neurocognitive deficits are being recognized as hav- tive. The system extends inside and outside of ing a clear impact on intellectual and adaptive func- the household. As Eiser comments, “not only are tioning (Beebe et al., 2005; Palmer et al., 2003). parents unprepared and disappointed, but also con- Taken together, this growing literature suggests fronted by lack of sympathy. Teachers who under- that intellectual functioning is impacted in chil- standably know little about the disease and treat- dren treated for malignant brain tumors with RT. ment may be at a loss as to why the child makes Mounting evidence is implicating impairment in so little progress. They also have to deal with many several core cognitive processes as underlying the other children with learning or behavioral difficul- decline in functioning. This appears to be primar- ties and may pay more attention to those showing ily related to disruption in the normal development disruptive tendencies rather than worry about the of white matter by RT. We do not, however, have child treated for cancer.” In support of Dr. Eiser’s evidence regarding the relative impact of resection impressions, there is an excellent study that iden- and chemotherapy independent of RT at this time. tified predictors of child behavior problems and Next, we will summarize some of the significant psy- adaptive functioning in the pediatric brain tumor chosocial deficits that are suffered by pediatric brain population (Carlson-Green et al., 1995). This study tumor survivors. demonstrated that not only illness-related issues, but also family variables were predictive of eventual intellectual functioning. More specifically, fam- Psychosocial late effects ily stress, the ability of the mother to cope with Preliminary evidence provided clues that pediatric the stress, the number of parents in the home, and brain tumor survivors were at extremely high risk socio-economic status were all related to psychoso- for difficulties in psychosocial adjustment (Mostow cial adjustment in the child/adolescent survivor. et al., 1991; Mulhern et al., 1989). A very frequently A recent review of behavioral, social, and psy- cited study (Hoppe-Hirsch et al., 1990) indicated chological adjustment in childhood brain tumor that nearly one-half of pediatric brain tumor sur- survivors has identified a total of 31 published vivors, specifically those who had been treated for manuscripts that addressed social, emotional, and medulloblastoma, displayed significant deficits in behavioral functioning in children who had been social competence, adjustment, and also continued diagnosed and treated for a CNS tumor (Fuemmeler to have behavioral and adaptive deficits as they et al., 2002). A rather wide discrepancy between transitioned into adulthood. Unfortunately, few significant adjustment difficulties was noted across studies have attempted to replicate these findings, the various studies. More specifically, the authors and many childhood brain tumor survivors are reported a range of 25%–93% significant distress lost to follow-up as they become adults. Little is among participants. While the authors described known about vocational status, which is particularly difficulties in summary interpretations, there did relevant given that many of these individuals are appear to be an increased risk for internalizing not able to enter college/university. The first and problems as opposed to externalizing problems in third authors of this chapter have begun a research brain tumor survivors. Internalizing psychopathol- project that is specifically designed to assess voca- ogy is typically characterized by disorders in the tional readiness in adolescents and young adult depression and anxiety groupings. Externalizing survivors of brain tumors, due to this gap in our problems refer to behavior control and conduct knowledge. disturbances. Overall, it was concluded that chil- The importance of familial integrity is becoming dren treated for a brain tumor are at increased increasingly relevant for survivors of brain tumors. risk for socialization difficulties, and problems with Eiser (2004) has emphasized the need to view the peer relationships. This appears to be present both 206 Section 2. Effects of cancer and cancer treatment on cognition

on parent- and teacher-rated measures. The over- and Copeland (2002). There is an increasing interest all impression is that these competence deficits in hybrid approaches that will, for example, com- are associated with degree of CNS impairment. It bine stimulant medication with holistic cognitive was further concluded that survivors continue to rehabilitation. Given that family issues are becom- exhibit deficits in adjustment following the transi- ing increasingly apparent, direct treatment towards tion to adulthood. Employment, marriage, parent- the parents and siblings is also gaining interest. hood, and post-secondary academic placements all Progress in the area of brain injury rehabilitation, are areas in which the survivors lag behind their both with children/adolescents and adults, is tra- peers. Diagnosis and treatment prior to the age of ditionally slow and laborious. Nevertheless, this is 3–5 years increases the risk for eventual significant an exciting field, and the authors are very pleased late effects. to be describing increased efforts towards treat- The review further identified a number of direc- ment of late effect disabilities, as opposed to con- tions for future research. More specifically, they rec- tinued description of severity of impairment, and ommend three areas that are in need of increased patterns of impairment. In sum, while treatment attention: first, assessment of more specific neu- progress regarding neurocognitive and psychosocial rological deficits as they relate to psychosocial late effects remains in its infancy, a greater number issues; second, greater attention to the environ- of researchers are directing their attention towards mental demands that children treated for a CNS rehabilitation. This is a much needed trend because malignancy experience; and, third, increased atten- it will further serve a growing group of courageous tion to parental and family adjustment. In sup- survivors, parents, and siblings. port of this latter point, there is an emerging literature within childhood traumatic brain injury that is Summary clearly implicating family adjustment as an important moderator of improvement and recovery from Survivors of childhood and adolescent CNS tumors injury in the child/adolescent (Yeates et al., 2004). have an increased risk of multiple medical con- These findings are likely to be extremely relevant sequences and neuropsychological problems com- to all childhood brain injury populations, including pared to other cancer survivors. These issues the CNS cancer patients. have serious consequences for survivors’ quality of life and contributions to society. Although many problems have multifactorial causes, radiotherapy Interventions underlies many of the most severe late effects, We would like to emphasize the need for treat- including the cognitive issues that these survivors ment and rehabilitation efforts with this popula- face. Future treatment strategies that decrease the tion. As identified in this chapter, these survivors use of radiotherapy would benefit this survivor are at an extremely high risk for neurocognitive and population. In addition, lifelong annual medical psychosocial difficulties. Researchers must begin to follow-up for all, and periodic neuropsychological devote more attention to treatment in addition to evaluation and rehabilitative treatment for those assessment. Butler and Mulhern (2005) have sum- experiencing educational or vocational difficulties marized intervention efforts directed towards child- are important. To facilitate such follow-up, the Chil- hood cancer survivors, and most participants in dren’s Oncology Group has developed evidence- the reviewed studies were brain tumor survivors. based screening guidelines that advise clinicians Therapeutic efforts have generally been directed on screening tests to include in the medical eval- towards traditional cognitive remediation methods, uation, based on therapy received (Landier et al., psychoactive medications (mainly the stimulant 2004); these guidelines can be downloaded from drugs), or a holistic approach as described by Butler www.survivorshipguidelines.org. Chapter 14. Childhood brain tumors 207

REFERENCES ate, and systemic methotrexate in childhood cancer. J Clin Oncol 12(12): 2621–2629. Anderson DM, Rennie KM, Ziegler RS et al. (2001). Medical Carlson-Green B, Morris RD, Krawiecki N (1995). Fam- and neurocognitive late effects among survivors of child- ily and illness predictors of outcome in pediatric brain hood central nervous system tumors. Cancer 92(10): tumors. J Pediatr Psychol 20(6): 769–784. 2709–2719. Copeland DR, deMoor C, Moore BD 3rd et al. (1999). Neu- Anderson VA, Godber T, Smibert E et al. (2000). Cognitive rocognitive development of children after a cerebellar and academic outcome following cranial irradiation and tumor in infancy: a longitudinal study. J Clin Oncol chemotherapy in children: a longitudinal study. Br J Can- 17(11): 3476–3486. cer 82(2): 255–262. Daugaard G, Abildgaard U (1991). Renal Morbidity of BassalM,MertensAC,TaylorLet al. (2006). Risk of selected Chemotherapy. Oxford: Butterworth-Heinemann. subsequent carcinomas in survivors of childhood can- Donaldson SS, Kaplan HS (1982). Complications of treat- cer: a report from the Childhood Cancer Survivor Study. ment of Hodgkin’s disease in children. Cancer Treat Rep J Clin Oncol 24(3): 476–483. 66(4): 977–989. Beebe DW, Ris MD, Armstrong FD et al. (2005). Cognitive Eiser C (2004). Neurocognitive sequelae of childhood can- and adaptive outcome in low-grade pediatric cerebel- cers and their treatment: a comment on Mulhern and lar astrocytomas: evidence of diminished cognitive and Butler. Pediatr Rehabil 7(1): 15–16. adaptive functioning in National Collaborative Research Ellenberg L, McComb JG, Siegel SE et al. (1987). Factors Studies (CCG 9891/POG 9130). J Clin Oncol 23(22): 5198– affecting intellectual outcome in pediatric brain tumor 5204. patients. Neurosurgery 21(5): 638–644. Bordeaux JD, Dowell RE Jr., Copeland DR et al. (1988). Ergun-Longmire B, Mertens AC, Mitby P et al. (2006). A prospective study of neuropsychological sequelae in Growth hormone treatment and risk of second children with brain tumors. J Child Neurol 3(1): 63–68. neoplasms in the childhood cancer survivor. J Clin Bowers DC, Mulne AF, Reisch JS et al. (2002). Nonperiop- Endocrinol Metab 91(9): 3494–3498. erative strokes in children with central nervous system Filley CM (2001). The Behavioral Neurology of White Mat- tumors. Cancer 94: 1094–1101. ter. New York: Oxford University Press. Bowers DC, Liu Y, Leisenring W et al. (2006). Late-occurring Fletcher JM, Copeland DR (1988). Neurobehavioral effects stroke among long-term survivors of childhood leukemia of central nervous system prophylactic treatment of can- and brain tumors: a report from the Childhood Cancer cer in children. J Clin Exp Neuropsychol 10(4): 495–537. Survivor Study. J Clin Oncol 24(33): 5277–5282. Fuemmeler BF, Elkin TD, Mullins LL (2002). Survivors Brownstein CM, Mertens AC, Mitby PA et al. (2004). Fac- of childhood brain tumors: behavioral, emotional, and tors that affect ﬁnal height and change in height standard social adjustment. Clin Psychol Rev 22(4): 547–585. deviation scores in survivors of childhood cancer treated Goldstein AM, Pastakia B, DiGiovanna JJ et al. (1994). Clin- with growth hormone: a report from the childhood can- ical ﬁndings in two African-American families with the cer survivor study. J Clin Endocrinol Metab 89(9): 4422– nevoid basal cell carcinoma syndrome (NBCC). Am J Med 4427. Genet 50(3): 272–281. BuonoL,MorrisM,MorrisRet al. (1998). Evidence for the Goldstein AM, Yuen J, Tucker MA (1997). Second cancers syndrome of nonverbal learning disabilities in children after medulloblastoma: population-based results from with brain tumors. Child Neuropsychol 4(2): 144–157. the United States and Sweden. Cancer Causes Control Butler RW, Copeland DR (2002). Attentional processes and 8(6): 865–871. their remediation in children treated for cancer: a lit- Gurney J, Kadan-Lottick N, Packer R et al. (2003a). erature review and the development of a therapeutic Endocrine and cardiovascular late effects among adult approach. J Int Neuropsychol Soc 8(1): 115–124. survivors of childhood brain tumors: Childhood Cancer Butler RW, Mulhern RK (2005). Neurocognitive interven- Survivor Study. Cancer 97: 663–673. tions for children and adolescents surviving cancer. J Gurney JG, Ness KK, Stovall M et al. (2003b). Final height Pediatr Psychol 30(1): 65–78. and body mass index among adult survivors of child- Butler RW, Hill JM, Steinherz PG et al. (1994). Neuropsycho- hood brain cancer: childhood cancer survivor study. J logic effects of cranial irradiation, intrathecal methotrex- Clin Endocrinol Metab 88(10): 4731–4739. 208 Section 2. Effects of cancer and cancer treatment on cognition

Hamilton SR, Liu B, Parsons RE et al. (1995). The molecular Luban NL (1998). An update on transfusion-transmitted basis of Turcot’s syndrome. NEnglJMed332(13): 839– viruses. Curr Opin Pediatr 10(1): 53–59. 847. Mabbott D, Penkman L, Witol A et al. (2005). Atten- Hancock SL, Cox RS, McDougall IR (1991). Thyroid diseases tion, working memory and processing speed in children after treatment of Hodgkin’s disease. NEnglJMed325(9): with posterior fossa tumours. Poster presented at the 599–605. 2005 International Neuropsychological Society Meeting, Heikens J, Ubbink M, van der Pal H et al. (2000). Long term Dublin, Ireland, 6–9 July, 2005. survivors of childhood brain cancer have an increased Martins AN, Johnston JS, Henry JM et al. (1977). Delayed risk for cardiovascular disease. Cancer 88: 2116–2121. radiation necrosis of the brain. J Neurosurg 47(3): 336– Hoppe-Hirsch E, Renier D, Lellouch-Tubiana A et al. (1990). 345. Medulloblastoma in childhood: progressive intellectual Mertens AC, Yasui Y, Neglia JP et al. (2001). Late mortality deterioration. Childs Nerv Syst 6(2): 60–65. experience in five-year survivors of childhood and ado- Hoppe-Hirsch E, Brunet L, Laroussinie F et al. (1995). Intel- lescent cancer: the Childhood Cancer Survivor Study. J lectual outcome in children with malignant tumors of Clin Oncol 19(13): 3163–3172. the posterior fossa: influence of the field of irradiation Mertens AC, Yasui Y, Liu Y et al. (2002). Pulmonary compli- and quality of surgery. Childs Nerv Syst 11(6): 340–345; cations in survivors of childhood and adolescent cancer. discussion 345–346. A report from the Childhood Cancer Survivor Study. Can- Jakacki RI, Goldwein JW, Larsen RL et al. (1993). Cardiac cer 95(11): 2431–2441. dysfunction following spinal irradiation during child- Mostow EN, Byrne J, Connelly RR et al. (1991). Quality of hood. J Clin Oncol 11(6): 1033–1038. life in long-term survivors of CNS tumors of childhood Jakacki RI, Schramm CM, Donahue BR et al. (1995). Restric- and adolescence. J Clin Oncol 9(4): 592–599. tive lung disease following treatment for malignant brain Mulhern RK, Butler RW (2004). Neurocognitive sequelae of tumors: a potential late effect of craniospinal irradiation. childhood cancers and their treatment. Pediatr Rehabil J Clin Oncol 13(6): 1478–1485. 7(1): 1–14; discussion 15–16. Jankovic M, Brouwers P, Valsecchi MG et al. (1994). Mulhern R, Butler RW (2005). Neuropsychological late Association of 1800 cGy cranial irradiation with intel- effects. In Brown RT (ed.) Pediatric Hematology/ lectual function in children with acute lymphoblas- Oncology: A Biopsychosocial Approach.NewYork: tic leukaemia. ISPACC. International Study Group on Oxford University Press. Psychosocial Aspects of Childhood Cancer. Lancet Mulhern RK, Wasserman AL, Friedman AG et al. (1989). 344(8917): 224–227. Social competence and behavioral adjustment of chil- Jannoun L, Bloom HJ (1990). Long-term psychological dren who are long-term survivors of cancer. Pediatrics effects in children treated for intracranial tumors. Int J 83(1): 18–25. Radiat Oncol Biol Phys 18(4): 747–753. Mulhern RK, Hancock J, Fairclough D et al. (1992). Neu- Kleinschmidt-DeMasters BK, Lillehei KO (1995). Radiation- ropsychological status of children treated for brain induced meningioma with a 63-year latency period. Case tumors: a critical review and integrative analysis. Med report. J Neurosurg 82(3): 487–488. Pediatr Oncol 20(3): 181–191. Kramer S, Lee KF (1974). Complications of radiation ther- Mulhern RK, Kepner JL, Thomas PR et al. (1998). Neu- apy: the central nervous system. Semin Roentgenol 9(1): ropsychologic functioning of survivors of childhood 75–83. medulloblastoma randomized to receive conventional Landier W, Bhatia S, Eshelman DA et al. (2004). Devel- or reduced-dose craniospinal irradiation: a Pediatric opment of risk-based guidelines for pediatric cancer Oncology Group study. J Clin Oncol 16(5): 1723–1728. survivors: the Children’s Oncology Group Long-Term Mulhern RK, Reddick WE, Palmer SL (1999). Neurocog- Follow-Up Guidelines from the Children’s Oncology nitive deficits in medulloblastoma survivors and white Group Late Effects Committe and Nursing Discipline. J matter loss. Ann Neurol 46(6): 834–841. Clin Oncol 22(24): 4979–4990. Mulhern R, Palmer S, Reddick W et al. (2001). Risks of young Lipshultz SE, Colan SD, Gelber RD et al. (1991). Late cardiac age for selected neurocognitive deficits in medulloblas- effects of doxorubicin therapy for acute lymphoblastic toma are associated with white matter loss. J Clin Oncol leukemia in childhood. NEnglJMed324(12): 808–815. 19(2): 472–479. Chapter 14. Childhood brain tumors 209

Mulhern RK, White HA, Glass JO et al. (2004). Attentional medulloblastoma: a longitudinal analysis. J Clin Oncol functioning and white matter integrity among survivors 19(8): 2302–2308. of malignant brain tumors of childhood. J Int Neuropsy- Palmer SL, Gajjar A, Reddick WE et al. (2003). Predicting chol Soc 10(2): 180–189. intellectual outcome among children treated with 35–40 Neglia JP, Robison LL, Stovall M et al. (2006). New pri- Gy craniospinal irradiation for medulloblastoma. Neu- mary neoplasms of the central nervous system in sur- ropsychology 17(4): 548–555. vivors of childhood cancer: a report from the Childhood Penkman L (2004). Remediation of attention deficits in Cancer Survivor Study. J Natl Cancer Inst 98(21): 1528– children: a focus on childhood cancer, traumatic brain 1537. injury and attention deficit disorder. Pediatr Rehabil Ness KK, Mertens AC, Hudson MM et al. (2005). Limitations 7(2): 111–123. on physical performance and daily activities among Peterson K, Shao C, McCarter R et al. (2005). An analysis of long-term survivors of childhood cancer. Ann Intern Med SEER data of increasing risk of secondary malignant neo- 143(9): 639–647. plasms among long-term survivors of childhood brain Nicholson HS, Butler R (2001). Late effects of therapy in tumors. Pediatric Blood Cancer 47: 83–88. long-term survivors. In Keating R, Goodrich J, Packer R Pollack IF (1994). Brain tumors in children. NEnglJMed (eds.) Tumors of the Pediatric Central Nervous System. 331(22): 1500–1507. New York: Thieme. Radcliffe J, Packer RJ, Atkins TE et al. (1992). Three- and Nicholson HS, Fears TR, Byrne J (1994). Death during adult- four-year cognitive outcome in children with noncorti- hood in survivors of childhood and adolescent cancer. cal brain tumors treated with whole-brain radiotherapy. Cancer 73(12): 3094–3102. Ann Neurol 32(4): 551–554. Oberfield SE, Soranno D, Nirenberg A et al. (1996). Age Reddick WE, White HA, Glass JO et al. (2003). Developmen- at onset of puberty following high-dose central ner- tal model relating white matter volume to neurocogni- vous system radiation therapy. Arch Pediatr Adolesc Med tive deficits in pediatric brain tumor survivors. Cancer 150(6): 589–592. 97(10): 2512–2519. O’Driscoll BR, Hasleton PS, Taylor PM et al. (1990). Active Reddick WE, Glass JO, Palmer SL et al. (2005). Atypical lung fibrosis up to 17 years after chemotherapy with car- white matter volume development in children following mustine (BCNU) in childhood. NEnglJMed323(6): 378– craniospinal irradiation. Neurooncology 7(1): 12–19. 382. Reeves CB, Palmer SL, Reddick WE et al. (2006). Attention Oeffinger KC, Mertens AC, Hudson MM et al. (2004). Health and memory functioning among pediatric patients with care of young adult survivors of childhood cancer: a medulloblastoma. J Pediatr Psychol 31: 272–280. report from the Childhood Cancer Survivor Study. Ann Reimers TS, Ehrenfels S, Mortensen EL et al. (2003). Cog- Fam Med 2(1): 61–70. nitive deficits in long-term survivors of childhood brain Oeffinger KC, Mertens AC, Sklar CA et al. (2006). Chronic tumors: identification of predictive factors. Med Pediatr health conditions in adult survivors of childhood cancer. Oncol 40(1): 26–34. NEnglJMed355(15): 1572–1582. Reinhold HS, Calvo W, Hopewell JW et al. (1990). Devel- Packer RJ, Sutton LN, Atkins TE et al. (1989). A prospective opment of blood vessel-related radiation damage in the study of cognitive function in children receiving whole- fimbria of the central nervous system. Int J Radiat Oncol brain radiotherapy and chemotherapy: 2-year results. J Biol Phys 18(1): 37–42. Neurosurg 70(5): 707–713. Ries LAG, Eisner M, Kosary C et al. (2005). SEER Can- Packer RJ, Sutton LN, Elterman R et al. (1994). Outcome cer Statistics Review, 1975–2002. Bethesda, MD: for children with medulloblastoma treated with radia- National Cancer Institute. http://seer.cancer.gov/ tion and cisplatin, CCNU, and vincristine chemotherapy. CSR/1975 2002/, based on November 2004 SEER data J Neurosurg 81(5): 690–698. submission, posted to the SEER website, 2005. Packer R, Gurney J, Punyko J et al. (2003). Long-term neu- Ris MD, Noll RB (1994). Long-term neurobehavioral out- rologic and neurosensory sequelae in adult survivors of come in pediatric brain-tumor patients: review and a childhood brain tumor: Childhood Cancer Survivor methodological critique. J Clin Exp Neuropsychol 16(1): Study. J Clin Oncol 21: 3255–3261. 21–42. Palmer SL, Goloubeva O, Reddick WE et al. (2001). Patterns Ris MD, Packer R, Goldwein J et al. (2001). Intellectual out- of intellectual development among survivors of pediatric come after reduced-dose radiation therapy plus adjuvant 210 Section 2. Effects of cancer and cancer treatment on cognition

chemotherapy for medulloblastoma: a Children’s Cancer Sklar CA, Constine LS (1995). Chronic neuroendocrinolog- Group study. J Clin Oncol 19(15): 3470–3476. ical sequelae of radiation therapy. Int J Radiat Oncol Biol Riva D, Pantaleoni C, Milani N et al. (1989). Impairment Phys 31(5): 1113–1121. of neuropsychological functions in children with medul- Stavrou T, Bromley C, Nicholson HS et al. (2001). Prog- loblastomas and astrocytomas in the posterior fossa. nostic factors and secondary malignancies in childhood Childs Nerv Syst 5(2): 107–110. medulloblastoma. J Pediatr Hematol Oncol 23: 431–436. Ronckers CM, Sigurdson AJ, Stovall M et al. (2006). Thyroid Strother D, Pollack I, Fisher PG et al. (2002). Tumors of the cancer in childhood cancer survivors: a detailed evalua- central nervous system. In Poplack PA (ed.) Principles tion of radiation dose response and its modifiers. Radiat and Practice of Pediatric Oncology (4th edn.). Philadel- Res 166(4): 618–628. phia, PA: Lippincott Williams & Wilkins. Ross L, Johansen C, Dalton S et al. (2003). Psychiatric hos- Travis LB, Rabkin CS, Brown LM et al. (2006). Cancer sur- pitalizations among survivors of cancer in childhood or vivorship – genetic susceptibility and second primary adolescence. New Engl J Med 349: 650–657. cancers: research strategies and recommendations. J Shuper A, Packer RJ, Vezina LG et al. (1995). “Compli- Natl Cancer Inst 98(1): 15–25. cated migraine-like episodes” in children following cra- Vassilopoulou S, Klein MJ, Moore BD et al. (1995). Efficacy nial irradiation and chemotherapy. Neurology 45(10): of growth hormone replacement therapy in children with 1837–1840. organic growth hormone deficiency after cranial irradia- Skinner R, Pearson AD, Amineddine HA et al. (1990). Oto- tion. Horm Res 43(5): 188–193. toxicity of cisplatinum in children and adolescents. Br J Walter AW, Mulhern RK, Gajjar A et al. (1999). Survival and Cancer 61(6): 927–931. neurodevelopmental outcome of young children with Sklar CA (1995). Growth following therapy for childhood medulloblastoma at St Jude Children’s Research Hospi- cancer. Cancer Invest 13(5): 511–516. tal. J Clin Oncol 17(12): 3720–3728. Sklar CA (1997). Growth and neuroendocrine dysfunc- Yeates KO, Swift E, Taylor HG et al. (2004). Short- and long- tion following therapy for childhood cancer. Pediatr Clin term social outcomes following pediatric traumatic brain North Am 44(2): 489–503. injury. J Int Neuropsychol Soc 10(3): 412–426. 15

Neuroﬁbromatosis

Bartlett D. Moore, III and John M. Slopis

Introduction Viskochil et al., 1990). The molecular genetic basis of distinguishing clinical features of NF-I was localized History to chromosome 17 in 1990 by two teams of investigators (Viskochil et al., 1990; Wallace et al., 1990). Neurofibromatosis (NF) is a common neurocuta- John Merrick, the so-called Elephant Man, was neous disorder that has an incidence of approxi- perhaps the most famous individual to be diag- mately 1 in 4000 (Mulvihill et al., 1990). Although nosed with NF-I although recent reports suggest NF has been postulated to have as many as eight that a more likely diagnosis for Mr Merrick is Pro- different forms (Riccardi & Eichner, 1986), this clas- teus syndrome, an unrelated condition that re- sification system has not been widely adopted. sembles NF-I externally but which arises from Neurofibromatosis is a group of genetic disorders a genetic mutation on a different chromosome including NF type I (NF-I), NF type II (NF-II), and (Ablon, 1995). multiple schwannomatosis, each with distinctly different genetic mutations and pathologic bases. The Phenotype and genotype NF-I gene is nearly ubiquitous in human tissues and so impacts virtually all organ systems. NF-I NF-I is a disorder that has been shrouded in confu- is particularly interesting to neurocognitive sci- sion over the years because the symptoms and clini- entists because of its characteristic phenotypical cal outcome of the disorder vary greatly from patient abnormalities in development of form and func- to patient. Approximately 50% of all cases result tion in brain. NF-II and multiple schwannomatosis from a spontaneous mutation in the NF gene region are essentially disorders of cranial nerves, periph- and so only half of all known cases are familial. eral nerves, and meningeal tissues with no associ- The phenotype of NF-I is highly variable between ated cognitive abnormalities and so these disorders unrelated individuals and even within affected fam- will be excluded from this discussion. ilies (von Deimling et al., 1995), despite the fact The original term neurofibromatosis was derived that the gene produces near-complete penetrance. at the turn of the last century but the disorder is Ethnic, racial, and gender grouping shows no pre- also called von Recklinghausen’s disease because dominance in NF-I. When the gene was localized the condition was described in the late 1800s clini- to chromosome 17 (Viskochil et al., 1990; Wallace cally and scientifically by Friedrich Daniel von Reck- et al., 1990) a new era of refined definition of linghausen (Cawthon et al., 1990; Crump, 1981; the phenotypic profile of NF-I ensued, allowing

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

211 212 Section 2. Effects of cancer and cancer treatment on cognition

segregation of NF-II and multiple schwannomato- Table 15.1. Diagnostic criteria of NF, type I sis patients from study groups. This distinction particularly sharpened clinical definition of the cogni- 1. Six or more cafe-au-lait´ spots greater than 5 mm in tive abnormalities brought about by the NF-I gene, diameter in pre-pubertal children or greater than 15 mm as childhood developmental specialists and neuro- in diameter in post-pubertal individuals scientists recognized the learning disabilities and 2. Two or more neurofibromas of any form or one plexiform neurofibroma behavioral abnormalities common to patients with 3. Freckling in the axillary or inguinal regions NF-I. This cognitive profile is characterized by a 4. Optic glioma high incidence of learning disabilities (LD), behav- 5. Two or more Lisch nodules (iris hamartomas) ioral problems such as attention deficit hyperactiv- 6. A distinctive osseous lesion such as sphenoid dysplasia ity disorder (ADHD), and neurocognitive deficits in or thinning of long bone cortex with or without visual spatial abilities. The cognitive phenotype is pseudoarthrosis considered to be an important predictor of the diag- 7. A first-degree relative with NF-I by the above criteria nosis by some investigators and will be discussed in 8. The presence of two or more criteria constitutes a detail later. definitive diagnosis in an individual. If an individual has The phenotypic variability of NF-I results in a first-degree relative with NF-I, then only one many individuals who live normal lives and expe- additional criterion is required for the diagnosis rience relatively little impact from the disorder, often unaware that they even carry the NF-I gene mutation. The presence of these mildly affected diagnosis of NF-I. The diagnostic criteria for NF- individuals calls into question the accuracy of the I (see Table 15.1) are based on a consensus state- incidence and prevalence estimates of NF-I. The ment developed by the National Institutes of Health presence of cognitive disability in this segment in 1988 (1988) and reaffirmed in 1997 (Gutmann of the population is unknown but may represent et al., 1997) as representing the most frequent clini- theetiologyofasignificantproportionofLDin cal features of NF-I. The diagnosis of NF-I is estab- the “general population.” Chronic, progressive, and lished when two or more features from this list are debilitating morbidity with severe disfigurement as identified in the patient. Therefore, if one family well as multiple types of cancers affect a minority of member carries the diagnosis, only a single crite- patients with NF-I. However in the case of severely rion is needed for the diagnosis in additional family affected individuals these complications generally members. progress in prevalence and severity with advanc- The most promising development in genetic testing age resulting in reduced life span of the group ing for NF-I is DNA sequencing of the NF-I cod- (Riccardi, 1981). The psychological stress of illness ing region on chromosome 17. This technology has in this segment of the population is an important revealed great variability in DNA sequences in the clinical issue, but will not be a topic of this review. NF-I gene region as might be expected in a disor- Initial symptoms of NF evolve with age and early der with great clinical variability. Several hundred diagnosis is often problematic, especially in cases distinct mutations, deletions, and rearrangements of spontaneous mutation. In families where NF-I have been found in DNA samples from individu- is already present, each child born will have a 50% als who meet clinical diagnostic criteria. These find- chance of having the mutation. NF-I can present ings demonstrate the complexity of the disorder on with congenital anomalies that are obvious at birth a molecular level and define the need for extensive or with clinical features conspicuous within the first future research in genotype/phenotype correlation. few years of life. Although biochemical and genetic To date, no specific patterns of mutation in the NF-I testing techniques have been developed, clinical gene are predictive of phenotype, severity, or long- assessment is the most reliable approach to the term outcome for any affected individual. In time Chapter 15. Neurofibromatosis 213

growth factor

growth factor receptor

cell membrane

GTP

GEF

FTP RAF

GTP RAS GTP MAPK ‘‘inactive’’ ‘‘active’’

neurofibromin cell proliferation

Figure 15.1. Ras pathway. Adapted from Packer et al. (2002)

this growing body of information will provide useful mones including human growth hormone and prognostic information to predict relative risks and nerve growth factors. The Ras system serves as outcomes, including neurocognitive morbidity, for the “second messenger” in the signal pathway by individuals who carry the NF-I gene. induced phosphorylation, changing configuration and binding to the inner cell membrane through The NF-I gene cytoskeletal bonds. These bonds are promoted by Localization of the NF-I gene led to the isolation the process of farnesylation via the intracellular of the gene product named “neurofibromin.” The enzyme system farnesyltransferase. Neurofibromin role of the protein neurofibromin was discovered is a constituent component of the reversible phos- through study of chronic monocytic myelogenous phorylation enzyme guanosine 5 -triphosphatase leukemia, a disorder seen in disproportionately high (GTP-ase), accelerating the process to proceed for- frequency in patients with NF-I. This protein prod- ward toward cell activation and/or growth. Muta- uct is now recognized to have a central role in signal tions in the gene coding region produce mutated transduction in the Ras system, which regulates cel- or “truncated” copies of the protein neurofibromin lular growth in Schwann cells as well as numerous leading to defective Ras signaling and the uncon- other cell types (Figure 15.1). trolled schwann cell growth seen in neurofibromas. The Ras “oncogene” is linked to extracellular The enzyme system farnesyltransferase has been receptors that bind various “first messenger” hor- the subject of intense investigation as a potential 214 Section 2. Effects of cancer and cancer treatment on cognition

target to interrupt dysregulated growth and cellular Table 15.2. Embryonic tissues giving rise to mature activation in NF-I. Pharmacologic agents known as tissues as features of NF-I farnesyltransferase inhibitors are currently under clinical trial to inhibit the growth of the benign Neural crest Schwann cell tumors called neurofibromas in NF-I Lisch nodules – pigmented iris hamartoma and these agents may ultimately play a role in the Cafe-au-lait´ – disordered cutaneous migration of treatment of neurofibromas that undergo malignant melanocytes Axillary freckles – disordered cutaneous migration of transformation. melanocytes at limb buds Cardiac conduction bundle – murine models of failed Systemic impact cardiac septal fusion Neuroectoderm Focal growth dysregulation in benign tumors Eye – optic nerve glioma and congenital “glaucoma” Brain – astrocyte/oligodendroglial defects producing brain Neurofibromas are complex benign tumors con- tumor and unidentified bright object taining multiple tissue elements including neural Spinal cord – astrocyte – spinal cord tumors tissue, connective tissue, and vascular components. Spinal root – Schwann cell defects producing radicular Researchers have long recognized the abnormal neurofibroma patterns of growth in these various tissues suggest- Peripheral nerve – Schwann cell defects producing ing that the influence of the NF gene is widespread. peripheral neurofibroma Neurofibromin has now been identified in multiple Ectoderm tissues that are derived from virtually all embryonic Dermis – neural elements producing cutaneous neurofibroma tissue lines including mesoderm, ectoderm, neu- Epidermis roectoderm, and neural crest. Mature tissues known Mesoderm or presumed likely to be affected by NF-I mutations Long bone – limb bud defects, hypertrophy, are listed in Table 15.2. The gene product has been pseudoarthrosis isolated from fetal ectoderm tissues as early as at Spinal bone – vertebral body anomalies, meningiocoels 6 weeks of gestation. This early presentation sug- Muscle – neurofibroma gests that the NF gene is a critical factor in embry- Adipocyte – neurofibroma onic development and therefore is a clue to the Vascular elements – neurofibroma widespread systemic nature of the disorder. Devel- opmental and functional anomalies of the brain in NF-I are best understood within this larger view of ronal migration in normal brain development. The the impact of the gene on embryogenesis. NF-I gene appears to function in roles of both signal transduction and environment sensing (Uhlmann & Gutmann, 2001) as well as programmed apopto- Brain malformation in animal models of NF-I sis. This concept supports the idea that disorders and humans in developing cellular apoptotic signals result in Development of the human central nervous system excess neuronal and astrocytic populations in the involves interactions with virtually all embryonic mature brain in NF-I. Malformations of cortical primitive tissues. The NF-I gene appears to influ- development and white matter are the end result ence central nervous system development during of abnormal apoptotic signaling. The interactions embryogenesis and mutations in this gene produce between gray and white matter precursors appear to malformations of brain in animal models and in be disarrayed, perhaps as a consequence of defec- humans. tive cell-to-cell signaling, resulting in abnormalities Developing gray matter and white matter ele- of oligodendrogliocyte myelin production and mat- ments in the brain interact in the process of neu- uration in NF-I with the consequence that areas of Chapter 15. Neurofibromatosis 215

hyperintense signal are commonly seen on brain common features of NF-I. More significant is the magnetic resonance imaging (MRI). Finally, nests of fact that these various malformations are presumed persistent astrocyte precursors that were not elimi- to result from different pathological mechanisms, nated by apoptosis remain as potential foci of brain supporting the role of neurofibromin in early brain tumor development over time. Numerous clinical development. features of NF-I are explained by these concepts. In contrast, minor malformations in cortical One of the most easily recognized malformations development occur with much greater frequency. in NF-I is aqueductal stenosis, which is seen in Minor malformation of frontal cortex development approximately 15% of patients. The mechanism of has been studied in NF-I as a feature associated with this malformation remains unclear although low- reading disability (Billingsley et al., 2003b). MRI- grade glioma or hamartoma may appear within the based anatomic studies of patients with dyslexia brainstem as age progresses. This common disorder in the general population reveal specific patterns is easily identified by neuroimaging, is frequently of cortical gyral formation in humans within the asymptomatic, and appears to have no distinct inferior frontal gyrus (Leonard et al., 2001). MRI- impact upon cognition. The association between based analysis of reading-disabled patients with malformation and latent brain tumor development NF-I shows a similar occurrence of these patholog- suggests two independent but interrelated pro- ical patterns within the inferior frontal gyrus with cesses that occur in brain development. Benign an approximately 40% rate of occurrence of these macrocephaly, another common clinical feature of minor malformations (Billingsley et al., 2002). NF-I, also implies that the NF-I gene has an impact Common clinical experience in the management on the brain during embryogenesis. Volumetric MRI of NF-I also supports the concept that aberrant cell studies indicate significant increases in gray matter signaling with disordered cell-to-cell neurotrans- volumes in patients with NF-I (Moore et al., 2000). mission occurs. Approximately 40%–60% of chil- Morphologic studies of brain in NF-I also demonstr- dren with NF-I demonstrate behavioral features of ate abnormal development of the corpus callosum ADHD. This frequency is five- to sixfold greater (Kayl et al., 2000). than in the general population. Dopamine reup- Frontal brain structural malformation (includ- take inhibitor stimulant therapy is commonly used ing forebrain fusion and holoprosencephaly) effectively in this population for the management has been modeled in mice using multiple spe- of ADHD. Whether this high incidence of ADHD is cific gene knock-out animals developed and a consequence of neuropharmacologic disorder or studied in conjunction with an additional NF-I frontal/callosal dysmorphism (Kayl et al., 2000), or knock-out. The NF-I gene appears to have a sig- both, remains uncertain at this time. nificant influence on neuronal migration and frontal cortex formation (Zhu et al., 2001) in these Brain tumors in NF-I models. These models predict the presence of the cortical malformations that are reported in NF-I Development of optic nerve glioma may occur in and human clinical descriptions. Gross malforma- association with plexiform neurofibroma and focal tions of cortex have been reported in humans with dysplasia of the tissues of the orbit and periorbital NF-I (Balestri et al., 2003). Three mentally retarded regions. These complex tumors represent the focal patients were evaluated with MRI scans of the brain or mosaic impact of the NF-I gene localized to the that defined right hemispheric transmantle cortical developing cranium. The formation of these tumors dysplasia, periventricular band cortical dysplasia is an example of anomalous cell-to-cell interac- with overlying pachygyria, and polymicrogyria in tion in NF-I during embryogenesis of the eye. Dur- these individuals. These major brain malformations ing brain formation, the eye buds develop from seem a certain cause of mental retardation and the anterior telencephalon. The eye buds migrate seizure in these three individuals but they are not forward and induce the surrounding primitive 216 Section 2. Effects of cancer and cancer treatment on cognition

tal lobes. The majority of these tumors are non- progressive and asymptomatic (Listernick et al., 1997). Tumors limited to the orbit often represent overgrowth of the optic nerve sheath without a progressive astrocytic component. Tumors of the chiasm and optic pathways are typically juvenile pilocytic astrocytomas. The highly stereotyped profile of these tumors appears directly related to the presence of the NF-I gene and its impact on central nervous system development in utero. Optic nerve tumors are also fairly stereotyped in associated symptomatology in patients with NF-I. Tumors of the optic chiasm frequently involve the adjacent hypothalamus causing precocious puberty by unknown endocrine mechanisms. It is interesting to note that tumors of the optic chiasm are usually not related directly to the short stature that is Figure 15.2. Axial MRI image of a 7-year-old boy with NF-I seen in approximately 15% of the NF-I population. showing a large orbital plexiform neurofibroma A significant body of evidence suggests that short stature is a systemic disorder of growth hormone embryonic tissues to form the mature eye. This pro- and growth factor receptors in NF-I. cess occurs in a cascade of events in which mesoder- The majority of optic nerve tumors in NF-I are low mal elements ultimately give rise to the bone of the grade, non-progressive, and do not require interv- orbit, vasculature, and extra-ocular muscles; neural ention. Treatment of symptomatic or progressive crest elements give rise to the iris ciliary body; neu- optic glioma in NF-I is similar to treatment of optic roectodermal elements give rise to the optic nerve glioma in the general population with the excep- and oculomotor nerves; and ectodermal elements tion of the use of cranial radiation (RT). Surgi- give rise to the eyelid. Disruption of one or more cal intervention is rare but may be required if the of the elements of this cascade results in dysplastic tumor is exerting a mass effect within the orbit or formation of the eye, forming either isolated optic withinthesuprasellarregion.Themostcommon nerve glioma or more extensive disfiguring orbital effective treatment is combination chemotherapy plexiform neurofibroma (Figure 15.2). using vincristine and carboplatin, in some cases Astrocytic brain tumors of all types occur in followed by tamoxifen. Cranial radiation is now patients with NF-I with a similar or slightly higher considered to be contraindicated in NF-I. Several incidence than in the general population. The groups have shown that secondary tumor forma- exception is optic pathway glioma, which dev- tion occurs with a higher frequency in the presence elops in approximately 15% of individuals with NF- oftheNFgene(Kortmannet al., 2003). These sec- I. Optic pathway glioma is primarily a tumor of ondary tumors are usually higher-grade infiltrative childhood, usually identified before the age of 6 astrocytes or malignant nerve sheath tumors, both years with a peak incidence occurring around 2 often fatal (Sharif et al., 2006). In addition to sec- years of age. Optic pathway glioma may arise within ondary tumors, children with NF-I also appear to the orbital segment of the optic nerve, within the develop central nervous system vascular malforma- optic chiasm, or within the brain parenchyma in tions at a higher than expected rate after RT lead- the optic pathways, usually limited to the anterior ing to cerebral infarction (Kortmann et al., 2003). As visual pathways and not extending to the occipi- recently as the early 1990s RT was routinely utilized Chapter 15. Neurofibromatosis 217

in the treatment of optic glioma and found to pro- adenylyl cyclase activity in the fruitfly appears to duce a high incidence of hypothalamic dysfunction. occur in NF-I-dependent and NF-I independent No comprehensive studies of the effects of RT spe- mechanisms. The mechanism of NF-I-dependent cific to NF-I are available but clinicians generally activation of the Rut adenylyl cyclase pathway is suspect that the known impact of RT on the devel- essential for Drosophila learning and memory (Guo oping brain may compound the cognitive deficits et al., 2000). known to occur in NF-I. Optic pathway glioma has The implications of these animal studies are no known independent correlation with cognitive far-reaching with respect to learning disabilities deficit in NF-I (De Winter et al., 1999). in humans with NF-I. The functionally disordered GTP-ase of the Ras signaling system has the potential to be present as a common defective second Systemic impact on function messenger in numerous signaling systems through- The impact of the NF-I gene clearly extends far out the body, including the brain. These animal beyond the promotion of the growth of benign nerve studies suggest that abnormal cell signaling in the sheath fibromas. The goal of current clinical trials brain leads to defects in neurotransmission and of farnesyltransferase inhibitors is to interrupt inter- subsequent learning disabilities that respond to cellular transmission of the growth message and pharmacological intervention. so to inhibit the growth of benign neurofibromas. Of perhaps greater interest, however, is the recent Neurocognitive status of children with NF-I demonstration that these agents have a positive impact upon learning deficits in NF-I knock-out From a neuropsychological standpoint, NF-I is an mice (Costa et al., 2002). Knock-out mice that are exceptionally interesting medical disorder. NF-I is haplo-insufficient for the NF-I gene demonstrate associated with much higher incidences of learning visual-spatial memory deficits that model human disability (LD), neuropsychological deficits, behav- NF-I learning disabilities in a limited way. These ioral problems, and brain tumors in comparison mice were studied in a Morris water maze system with the general population. In addition, those and found to have significant visual-spatial learn- with NF-I have a wide range of neuroanatomical ing impairments. The mice were then treated with abnormalities. Neuropsychological studies of chil- a farnesyltransferase inhibitor, which led to signifi- dren and adolescents with NF-I have revealed a cant reductions in learning time and improvement wide range of cognitive sequelae associated with the in overall learning efficiency (Costa et al., 2002). disorder. Visual-spatial deficits and learning disabil- The role of the NF-I gene has also been studied in ities are two of the most commonly reported prob- NF haplo-insufficient Drosophila.TheDrosophila lems, but speech disarticulation, language deficits, NF-I protein is highly conserved showing 60% and motor inco-ordination are also typical features identity with human neurofibromin (Guo et al., (Brewer et al., 1997; Eldridge et al., 1989; Eliason, 2000). The fruit fly depends heavily upon olfaction 1986; Hofman et al., 1994; Moore & Denckla, 1999; for survival, so models of learning must distin- Moore et al., 1994; North et al., 1997). guish between behavioral patterns that include or exclude olfaction. These models include olfactory- Intellectual functioning guided avoidance, olfactory-guided learning, and electric shock avoidance. Studies indicate that It was once widely believed that NF-I was associ- Drosophila NF-I protein acts as both a Ras GTP- ated with a high incidence of mental retardation. ase-activating protein (GAP) and as a regulator of This idea has been widely discounted by numer- the AMP pathway that involves the rutabaga-(rut-) ous studies that have instead found a slight down- encoded adenylyl cyclase. G-protein-activated ward shift of the distribution of IQ (Figure 15.3) and 218 Section 2. Effects of cancer and cancer treatment on cognition

0.50 tual deficiency while others are highly gifted intel- lectually. 0.40 Normal curve Learning and academic achievement profile 0.30 Full-scale IQ Difficulties with academic performance are often reported to be the most significant morbidity asso- 0.20 ciated with childhood NF-I (Coude et al., 2004). The NF Consensus Task Force reported variable rates of 0.10 LD with an average rate of 44.3% and a range of Proportion scor ing in a r ange from 30% to 61% in the 6 studies reporting this infor- 0.00 mation (North et al., 1997). However, not all stud- −4.0 −3.0 −2.0 −1.0 0.0 1.0 2.0 3.0 4.0 ies used the same criteria for defining what consti- Z Score tutes a LD. The Diagnostic and Statistical Manual Figure 15.3. Full-scale IQ distribution in 84 children with of Mental Disorders (DSM-IV; American Psychiatric NF-I compared to the normal curve Association, 2000) does not list “learning disability” per se but describes specific disorders in reading, mathematics, and written expression. The cri- only a slightly elevated incidence of mental retar- terion for a disorder in one of these areas is that aca- dation over population estimates (Eldridge et al., demic achievement, as measured by standardized 1989; Moore et al., 1994; North et al., 1997). Some tests, must be substantially below expectations for early studies reported a discrepancy between ver- the child’s chronological age, intelligence, and age- bal and performance intellectual abilities favoring appropriate education. This is the standard discrep- verbal abilities (Eliason, 1986, 1988; Wadsby et al., ancy model of LD and, while controversial and in 1989). This led to the belief that the LD in NF-I need of change, it is what most school districts now are similar to the classic non-verbal learning dis- use. However, in many studies of NF-I, formal cri- ability (NVLD) syndrome (Rourke, 1989). The NF-I teria for LD have not been applied when arriving at Cognitive Disorders Task Force review of studies for incidence levels. Many children are underachievers, which both verbal and performance IQ data were are in special classes, or have had to repeat a grade. reported found no significant trend for verbal or Others may have behavioral issues such ADHD that performance intellectual advantage or disadvantage interfere with academic success. Thus many chil- (North et al., 1997). The Task Force reviewed 10 stud- dren with NF-I are given the LD label even when ies with a total of 416 patients with NF-I. Of the 6 they may not meet objective diagnostic criteria. studies that reported data on the incidence of men- Visual-spatial deficits are a common characteris- tal retardation (n = 350), the average rate (defined tic of children with NF-I (discussed below) (Eldridge as IQ < 70) was 7.1% with a range of 4.8%–11.2%. et al., 1989; Eliason, 1986), therefore, some have This is higher than the estimated rate in the gen- speculated that they suffer from NVLD. Learn- eral population of 2%–3%, but not as high as once ing disabilities in this population, however, are thought. The average full-scale IQ was 92.9 in the 9 not exclusively non-verbal, as more recent stud- studies with objective standardized measures of IQ ies have shown that verbal deficits are also com- with a range of 88.6–94.8. Whereas the ranges of IQ mon. Mazzocco and colleagues (1995) examined reported in these studies are somewhat similar, it is reading disability in NF-I and found a higher inci- importanttonotethatchildrenwithNF-Icoverthe dence in children with NF-I (53%) compared to entire range of intellectual abilities that is seen in their non-affected siblings (26%). Children with NF- the general population: some show severe intellec- I, in comparison to their siblings without NF-I, had Chapter 15. Neurofibromatosis 219

weaknesses in vocabulary and phonetic abili- Visual-spatial abilities and their impact on ties, reading and mathematics, in addition to academic achievement their visual-spatial deficits (Mazzocco et al., 1995). Cutting et al. (2000) compared the cognitive profiles Dyslexic readers in the general population have of children with NF-I to those of an LD clinic pop- deficits in the rapid processing of visual stimuli ulation. While both groups performed worse than (Eden et al., 1996a, 1996c; Temple et al., 2000). Defi- non-disabled controls on measures of sight read- ciencies in the rapid identification of letters may ing and reading comprehension, the NF-I group also contribute to reading problems in children had more global language impairments compared and adolescents with NF-I. Visual-spatial process- to the LD clinic group. Cutting also found that chil- ing problems, including the rapid identification of dren with NF-I scored significantly lower than an LD objects (Cutting et al., 2000) and the accurate iden- control group on visual-spatial measures, indicating tification of similar lines and angles (Eldridge et al., that children with NF-I have visual-spatial deficits 1989; Moore et al., 1996), have been identified chil- that are not representative of the broader reading- dren and adolescents with NF-I, as well as in poor disabled population (Cutting et al., 2000). readers without NF-I (Eden et al., 1996a). Functional Others have also suggested that the academic neuroimaging studies of the visual-spatial process- profile does not seem to fit the typical types of LD or ing of letters and other stimuli in healthy individu- dyslexia. For example, Descheemaeker et al. (2005) als have shown significant activity in bilateral infe- reported that half of a relatively small sample (n = rior parietal and posterior-superior parietal cortex, 17) of children with NF-I had LD a figure in keep- as well as in lateral frontal and extrastriate cor- ing with other reports. Of those with documented tex (Alivisatos & Petrides, 1997; Booth et al., 2000; LD (n = 8) one-half had spelling deficits but only Greenlee et al., 2000). one had a pure arithmetic deficit (Descheemaeker Children with NF-I have visual-spatial deficits et al., 2005). Brewer and colleagues (1997) used clus- that are not representative of the broader LD pop- ter analysis to document the neurocognitive profile ulation (Cutting et al., 2000), but may be related in a large cohort (N = 105) of children with NF-I. to deficits in reading. Adults with NF-I have also She found that, among 72 children with academic been found to have visual-spatial deficits. Using dis- difficulties, 3 groups emerged. One group had a criminant function analyses, the JLO test accurately normal neurocognitive profile (39%), another had classified adults with NF-I from controls (Pavol general academic deficits (47%), and the third had et al., 2006). Nevertheless, the role of visual-spatial primarily visual-spatial/motor deficits (14%). The deficits in the learning deficits of children with low incidence of visual-spatial deficits was sur- NF-I is far from clear (Brewer et al., 1997). We prising given the often-reported high incidence of have found that the deficit in visual-spatial abili- deficits in this area. However, this study did not ties is somewhat specific rather than general. For include the Judgment of Line Orientation (JLO) test example, the JLO test is impaired in approximately (Lindgren & Benton, 1980), which is often reported 70% of children with NF-I and it bears a strong as the most impaired test of visual perceptual skills relation with academic performance in general in this population (Schrimsher et al., 2003). (Figure 15.4). When taken together, reports of cognitive deficits Although performance by most children with NF- in children and adolescents with NF-I show that this I on the JLO test is impaired (Lindgren & Benton, disorder is characterized by a complex range of both 1980), not all areas of visual-spatial processing are verbal and visual-spatial deficits, and does not lead affected in children with NF-I. For example, the abil- to a strictly non-verbal LD syndrome, insofar as this ity to discriminate among two-dimensional draw- syndrome has been described in other populations ings of similar geometric figures is not impaired rel- (Rourke, 1989). ative to normal control subjects (Figure 15.5). 220 Section 2. Effects of cancer and cancer treatment on cognition

With patient scoring zero on the JLO excluded ( n = 53) 145 JLO Acad. Acl Pearson (r = 0.51, p<0.001) 130 Mean 6.6 97.1 SD 3.9 13.0 115

100

70 Academic achievement score 55 14 7 1013 16 JLO score

Figure 15.4. Correlation between standard score on the Judgment of Line Orientation (JLO) test and general academic achievement

Control subjects

7 NF-1

4 Standard score (10+3)

1 JLO Rec. Discrim. Type of spatial task

Figure 15.5. Visual spatial performance on the Judgment of the Line Orientation (JLO) versus the Recognition-Discrimination test tasks

Schrimsher and colleagues (2003) reported that the tion Test) were a significant predictor of NF- multivariate combination of visual-spatial/motor I diagnostic status (p = 0.000 000 4). Two of tasks was highly discriminative of the NF-I diagnosis these tests are purely visual spatial and two are in that it correctly identified 90% of individuals with visual perceptual/motor. The JLO (a purely visual- clinically identified NF-I. Four visual processing spatial task) by itself, but not the other tests, tests (JLO, Orientation, Block Design, Recognition- was still highly predictive of NF diagnostic status. Discrimination Test, Beery Visual-Motor Integra- Although the diagnosis of NF-I is usually made Chapter 15. Neurofibromatosis 221

on the basis of the clinical features presented in play a role in learning and cognitive difficulties. Table 15.1, additional information may be provided However, the severity of cognitive and learning by performance on tests of visual-spatial abilities, deficits in children with NF-I is not exacerbated by especially when the diagnosis based on clinical the presence of a brain tumor unless cranial radia- characteristics is marginal (e.g., 5, but not 6 cafe-au-´ tion therapy is given (De Winter et al., 1999; Moore lait spots). Schools and entities testing large num- et al., 1994). bersofschool-agedchildrenshouldbeawarethat a profile consisting of visual-spatial deficits, ADHD, MR hyperintensities and learning deficits in the context of intact intellectual abilities is often associated with NF-I. For Areas of hyperintense signal on T2-weighted and, medical personnel caring for children with diag- more conspicuously, on fluid-attenuated inversion nosed NF-I, a comprehensive neurocognitive eval- recovery (FLAIR) MRI sequences are observed in uation should become part of their standard of care. the brains of most children and adolescents with Since visual-spatial functioning influences reading NF-I. These “MR hyperintensities,” as we will refer and academic abilities in general, tests of visual- to them, are benign, do not appear to occupy spatial abilities should be included. A brief screen- space, and may occur in multiple regions in the ing examination is even warranted for suspected same individual. Their most common locations are NF-I should school or medical personnel have sus- the basal ganglia, cerebellum, brainstem, and dien- picions of the diagnosis. cephalon. Uncertainty continues with regard to the makeup and clinical significance of MR hyperintensities in children with NF-I. One report of three Brain structure/function correlates children with NF-I seen at autopsy documented NF-I presents with a unique combination of white that MR hyperintensities consisted of spongiotic matter abnormalities, low-grade, or sometimes tissue with fluid-filled vacuoles, which accounts for malignant, brain tumors, and abnormalities in brain their appearance on MRI (DiPaolo et al., 1995). morphology. These morphological differences can Using magnetization transfer ratio measurement be seen in both gross and regional brain develop- techniques, MR hyperintensities were characterized ment and have been widely investigated for their as areas of hypomyelination or structurally abnor- role in the neuropsychological and learning deficits mal myelin (Margariti et al., 2007). Because of their seen in NF-I. The relationship between brain MR uncertain nature, these areas of MR hyperinten- hyperintensities, brain tumors, and macrocephaly sity have been informally referred to as “unidenti- with neuropsychological functioning in NF-I is an fied bright objects” (UBOs). However, because they area of intense research interest. are bright only on MRI and because they are not space-occupying objects, we prefer the label MR hyperintensities. It has been estimated that between Brain tumors 50% and 70% of children and adolescents with NF- Approximately 15% of children and adolescents I have MR hyperintensities, leading some to pre- with NF-I will be diagnosed with a brain tumor, usu- dict that they are markers for more extensive, albeit ally an optic glioma. Only 20% of optic gliomas are unobservable, white matter abnormalities (Moore symptomatic however. Although optic gliomas in et al., 1996; North et al., 1994). Several studies NF-I are usually benign and are often left untreated, have reported that MR hyperintensities disappear one might surmise that an optic glioma is still a or diminish in size with advancing age (Aoki et al., pathological condition of the central nervous sys- 1989; Itoh et al., 1994; Sevick et al., 1992), strength- tem, and therefore might be associated with other ening the case that they are an anomaly of the nor- less conspicuous brain abnormalities that could mal developmental process of myelination. 222 Section 2. Effects of cancer and cancer treatment on cognition

Correlative neuropsychological studies of MR matter ratio and significantly larger corpus callosa hyperintensities have generally been disappoint- were reported in this group of 52 children with NF-I ing however, because few consistent relations with (Kayl et al., 2000; Moore et al., 2000). Greater volume cognitive functioning have been observed. Early of gray matter and size of the corpus callosum was studies failed to find a significant association positively correlated with the degree of discrepancy between learning disabilities and MR hyperintensi- between IQ and academic achievement in children ties in the brain (Duffner et al., 1989; Dunn & Roos, with NF-I but not controls (Moore et al., 2000). The 1989; Ferner et al., 1993). More recent studies have differences between these studies may be related to reported significant associations between the pres- the selection of brain structures included in the vol- ence (North et al., 1994), number (Denckla et al., umetric calculations (Moore et al., 2000). 1996; Hofman et al., 1994), and location (Moore et al., 1996) of MR hyperintensities and neurocog- Congenital malformation nitive functioning. These conflicting results suggest that MR hyperintensities are not consistent pre- In proficient readers there is a left-greater-than- dictors of cognitive deficits or learning disabilities right superior temporal lobe asymmetry. Measure- across samples of the NF-I population, but may rep- ment of specific regions of the superior tem- resent a neurocognitive burden if found in sufficient poral lobe, the planum temporale (PT) and planum numbers and in certain locations. MR hyperintensi- parietale (PP), has been reported in 24 children ties in the thalamus are associated with lower scores and adolescents with NF-I and an equal number of on tests of visual-spatial and memory ability (Moore controls (Billingsley et al., 2002). Intelligence-based et al., 1996) and with lower IQ (Goh et al., 2004). discrepancy scores of reading and math achieve- When located in the globus pallidus, MR hyperin- ment, which are commonly used to define learning tensities are associated with relatively low attention disabilities, were significantly related to PT asym- scores (Goh et al., 2004). metry in the NF-I group. Specifically, boys with NF- I had an absence of the normal asymmetry seen in proficient readers. In addition, the left PT in boys Macrocephaly with NF-I was smaller than in girls with NF-I and in Macrocephaly occurs in 30%–50% of patients with non-NF-I controls (Billingsley et al., 2002). NF-I (Bale et al., 1991) and is associated with increased clinical and physical severity (Zvulunov Functional imaging studies et al., 1998) but not increased neuropsychological impairment (Ferner et al., 1996). The relation- Structural neuroanatomy is an important approach ship between macrocephaly and learning disabili- to studying disorders such as NF-I because of the ties in NF-I has been a focus of several recent studies high incidence of morphological abnormalities (dis- using quantitative volumetric imaging techniques. cussed above). However, this approach may have Said and colleagues (1996) reported greater over- limitations for determining the etiology of the cog- all brain volume, specifically cortical white mat- nitive effects also commonly observed in indi- ter, using quantitative imaging techniques. Using viduals with NF-I. Structural abnormalities may the magnetization transfer ratio, Margariti and col- only be an indirect indication of function. A more leagues concluded that macrocephaly results from complete understanding of learning disabilities in increased volumes of gray matter and white mat- NF-I requires methods that can associate under- ter in children with NF-I (Margariti et al., 2007). lying neuronal activity with cognitive operations Our group, however, found significantly larger over- in a time-linked fashion. Hemodynamic imaging all brain volumes, due to gray but not white mat- methods, such as functional MRI (fMRI), provide ter differences. In addition, a higher gray-to-white a way to analyze regional brain function that is Chapter 15. Neurofibromatosis 223

temporally linked to cognitive processing. Func- especially in the right hemisphere. These results tional MRI detects blood oxygen level dependent agree with previous morphological studies indicat- (BOLD) responses in the brain during cognitive ing inferior frontal cortex malformations in adults activity. These BOLD responses are temporally with developmental language disorders (Clark & linked to changes in underlying neuronal activity Plante, 1998). resulting from cognitive activity. Functional MRI has As discussed above, visual-spatial impairments been used to study reading and visual-spatial pro- are a hallmark of NF-I. Just as in phonological processing associated with learning disabilities, includ- cessing, children with NF-I appear to have differing dyslexia, in healthy individuals as well as in a ent activation patterns from controls during visual- variety of patient populations with neurological dis- spatial processing. Relative to lateral and inferior orders (Billingsley et al., 2001; Eden et al., 1996b; frontal cortex, children with NF-I activate poste- Paulesu et al., 1996; Shaywitz et al., 1998; Temple rior cortex (occipital, parietal, and middle temporal) et al., 2001). to a greater extent than controls (Billingsley et al., Developmental reading impairments involve 2004). Patterns of activation were associated not problems in learning to relate visual input to only with their accuracy during the activation task phonological representations. Phonological dis- during fMRI, but also with their standardized read- crimination, which requires an individual to identify ing scores in a normal testing environment. distinct sounds that make up words and letters, has been shown to be a core component of learning to read (Fletcher et al., 1994; Stanovich, 1988). Summary Functional MRI studies in other populations have implicated inferior frontal, dorsolateral prefrontal, Neurofibromatosis is more common than other and temporal cortices in phonological processing high profile disorders such as muscular dystro- skills (Pugh et al., 1996; Shaywitz et al., 1998; Temple phy, Tay-Sachs disease, cystic fibrosis, and Hunting- et al., 2001). Previous fMRI investigations of phono- ton’s disease combined (Korf & Rubenstein, 2005) logical processing in poor readers who do not have and yet it remains virtually unknown to the pub- NF-I have shown differential neural responses to lic. The underlying molecular pathogenesis of NF phonological stimuli (Temple et al., 2000, 2001). was discovered almost a century after its first clin- Children with dyslexia were found to have reduced ical descriptions. Advances in the understanding neural activity in left temporal-parietal cortex of the NF gene mutation have led to insights into during a phonological decision task that required specific tumor suppressor gene function as well as them to determine whether two letters rhymed. insights into the impact of growth dysregulation Activity in left inferior frontal cortex, a region that upon embryonic development, including structural has been identified as critical to phonological and functional brain development. processing in neurologically normal individu- The vast size and complexity of the NF gene als (Pugh et al., 1996), was similar in dyslexic results in a broad spectrum of distinct human NF children compared with controls (Temple et al., mutations and so the disorder is expressed with 2001). variable manifestations of physical and behavioral Phonological processing is one of the most basic phenotypes. Individuals with mild symptoms are skills involved in learning to read and the infe- often unaware that they even carry the disordered rior frontal cortex is integral to this skill. Using an gene. This is a particular problem in childhood as fMRI paradigm that involved phonological process- the morbidity of NF-I increases with age and early ing, Billingsley and colleagues (2003a) found that detection of developing problems is important. children with NF-I activate inferior frontal relative The neurocognitive phenotype of NF-I consists to posterior cortex to a greater extent than controls, of average or slightly below average intellectual 224 Section 2. Effects of cancer and cancer treatment on cognition

abilities, difficulties in school achievement, visual- ACKNOWLEDGMENTS spatial processing deficits, ADHD, and frequently low self-esteem. Some children with NF have none The authors wish to acknowledge the support of Cheniere of these neurocognitive features, a fact that demon- Energy, Inc., Houston, Texas, Kirk Gentle, and Chris Shaw strates the variability of the phenotype. The neu- for helping to make this work possible. rocognitive phenotype ranges from severe mental deficiency to superior intellect, and yet any individ- REFERENCES ual with NF may exhibit a LD. The NF population includes children who struggle in special education Ablon J (1995). “The Elephant Man” as “self” and “other”: hoping to graduate from high school and children the psycho-social costs of a misdiagnosis. Soc Sci Med 40: who attain top academic achievement despite com- 1481–1489. petition from their non-NF peers. Whether learning Alivisatos B, Petrides M (1997). Functional activation of the disability is a cause of social dysfunction and failure human brain during mental rotation. Neuropsychologia in personal achievement remains to be studied, but 35: 111–118. one could argue that learning disability is the great- American Psychiatric Association (2000). Diagnostic and est morbidity of NF. Statistical Manual of Mental Disorders (4th edn.) (Test Revision). DSM-IV-TR. Arlington, VA: American Psychi- The reasons for the high incidence of LD in atric Association. the NF population remain largely unknown. MRI- Aoki S, Barkovich AJ, Nishimura K et al. (1989). Neurofi- based morphologic studies of the brains of chil- bromatosis types 1 and 2: cranial MR findings. Radiology dren with NF-I have revealed both gross and 172: 527–534. fine structural abnormalities of brain develop- Bale SJ, Amos CI, Parry DM et al. (1991). Relationship ment similar to abnormalities seen in idiopathic between head circumference and height in normal dyslexia and ADHD. Much of this evidence suggests adults and in the nevoid basal cell carcinoma syndrome both structural and functional disorders of frontal and neurofibromatosis type I. Am J Med Genet 40: 206– cortex. 210. Undoubtedly mutation of the NF-I gene plays Balestri P, Vivarelli R, Grosso S et al. (2003). Malforma- a role in these specific disorders of neural devel- tions of cortical development in neurofibromatosis type 1. Neurology 61: 1799–1801. opment in children with NF-I and thereby indi- Billingsley RL, McAndrews MP, Crawley AP et al. (2001). rectly influences their learning and neurocognitive Functional MRI of phonological and semantic process- profile. Genetically engineered murine and insect ing in temporal lobe epilepsy. Brain 124: 1218–1227. models bearing NF-I mutations provide excel- Billingsley RL, Schrimsher GW, Jackson EF et al. (2002). lent surrogates that mimic the human conditions Significance of planum temporale and planum parietale of learning deficits. NF-I haplo-insufficient mice morphologic features in neurofibromatosis, type I. Arch model deficits in spatial learning (Costa et al., 2002; Neurol 59: 616–622. Silva et al., 1997), while the NF-I haplo-insufficient Billingsley RL, Jackson EF, Slopis JM et al. (2003a). Func- Drosophila models deficits in olfactory learning and tional magnetic resonance imaging of phonologic pro- independent mechanisms of memory. These mod- cessing in neurofibromatosis 1. J Child Neurol 18: 731– els provide behavioral platforms for pharmacolog- 740. Billingsley RL, Slopis JM, Swank PR et al. (2003b). Cortical ical trials with a direct molecular genetic window morphology associated with language function in neu- into correlative studies of signal transduction in the rofibromatosis, type I. Brain Lang 85: 125–139. brain. In this way the ubiquitous nature of the NF Billingsley RL, Jackson EF, Slopis JM et al. (2004). Func- gene provides a rare opportunity to study neural sig- tional MRI of visual-spatial processing in neurofibro- nal transduction in models that correlate a specific matosis, type I. Neuropsychologia 42: 395–404. genotype with stereotyped behavioral and develop- Booth JR, MacWhinney B, Thulborn KR et al. (2000). Devel- mental patterns. opmental and lesion effects in brain activation during Chapter 15. Neurofibromatosis 225

sentence comprehension and mental rotation. Dev Neu- Duffner P,Cohen M, Seidel F et al. (1989). The significance ropsychol 18: 139–169. of MRI abnormalities in children with neurofibromato- Brewer VR, Moore BD, Hiscock M (1997). Learning disabil- sis. Neurology 39: 373–378. ity subtypes in children with neurofibromatosis. J Learn Dunn DW, Roos KL (1989). Magnetic resonance imaging Disabil 30: 521–533. evaluation of learning difficulties and incoordination in Cawthon RM, Weiss R, Xu GF et al. (1990). A major segment neurofibromatosis. Neurofibromatosis 2:1–5. of the neurofibromatosis type 1 gene: cDNA sequence, Eden GF, Stein JF, Wood HM et al. (1996a). Differences in genomic structure, and point mutations [published erra- visuospatial judgement in reading-disabled and normal tum appears in Cell 1990 Aug 10:62(3):following 608]. Cell children. Percept Motor Skills 82: 155–177. 62: 193–201. EdenGF,VanMeterJW,RumseyJMet al. (1996b). Abnor- Clark MM, Plante E (1998). Morphology of the inferior mal processing of visual motion in dyslexia revealed by frontal gyrus in developmentally language-disordered functional brain imaging [see comments]. Nature 382: adults. Brain Lang 61: 288–303. 66–69. Costa RM, Federov NB, Kogan JH et al. (2002). Mechanism Eden GF, VanMeter JW, Rumsey JM et al. (1996c). The visual for the learning deficits in a mouse model of neurofibro- deficit theory of developmental dyslexia. Neuroimage 4: matosis type 1. Nature 415: 526–530. S108–S117. Coude FX, Mignot C, Lyonne S et al. (2004). Academic Eldridge R, Denckla MB, Bien E et al. (1989). Neurofibro- impairment is the most frequent complication of neu- matosis type 1 (Recklinghausen’s disease). Neurologic rofibromatosis type-1 (NF1) in children. Behav Genet 34: and cognitive assessment with sibling controls. Am J Dis 635. Child 143: 833–837. Crump T (1981). Translation of case reports in Ueber die Eliason MJ (1986). Neurofibromatosis: implications for multiplen Fibrome der Haut und ihre Beziehung zu den behavior and learning. Neurofibromatosis 7: 175–179. multiplen Neuromen [On Multiple Fibromas of the Skin Eliason MJ (1988). Neuropsychological patterns: neurofi- and their Relationship to Multiple Neuromas] by F. V. bromatosis compared to developmental learning disor- Recklinghausen. In Riccardi VM, Mulvihill J (eds.) Neu- ders. Neurofibromatosis 1: 17–25. rofibromatosis (von Recklinghausen Disease): Genetics, Ferner RE, Chaudhuri R, Bingham J et al. (1993). MRI Cell Biology, and Biochemistry (Vol. 23, pp. 259–275). in neurofibromatosis 1: the nature and evolution of New York: Raven Press. increased intensity T2 weighted lesions and their rela- Cutting LE, Koth CW, Denckla MB (2000). How children tionship to intellectual impairment. J Neurol Neurosurg with neurofibromatosis type 1 differ from “typical” learn- Psychiatry 56: 492–495. ing disabled clinic attenders: nonverbal learning disabil- Ferner RE, Hughes RA, Weinman J (1996). Intellectual ities revisited. Dev Neuropsychol 17: 29–47. impairment in neurofibromatosis 1. J Neurol Sci 138: De Winter AE, Moore BD, Slopis JM et al. (1999). Brain 125–133. tumors in children with neurofibromatosis: additional Fletcher JM, Shaywitz SE, Shankweiler DP et al. (1994). Cog- neuropsychological morbidity? Neurooncology 1: 275– nitive profiles of reading disability: comparisons of dis- 281. crepancy and low achievement definitions. JEducPsy- Denckla MB, Hofman K, Mazzocco MM et al. (1996). chol 86: 6–23. Relationship between T2-weighted hyperintensities Goh WHS, Khong PL, Leung CSY et al. (2004). T2-weighted (unidentified bright objects) and lower IQs in chil- hyperintensities (unidentified bright objects) in children dren with neurofibromatosis-1. Am J Med Genet 67: with neurofibromatosis 1: Their impact on cognitive 98–102. function. J Child Neurol 19: 853–858. Descheemaeker MJ, Ghesquiere P, Symons H et al. (2005). Greenlee MW, Magnussen S, Reinvang I (2000). Brain Behavioural, academic and neuropsychological profile regions involved in spatial frequency discrimination: evi- of normally gifted neurofibromatosis type 1 children. dence from fMRI. Exp Brain Res 132: 399–403. J Intellect Disabil Res 49: 33–46. Guo HF, Tong JY, Hannan F et al. (2000). A DiPaolo DP, Zimmerman RA, Rorke LB et al. (1995). neurofibromatosis-1-regulated pathway is required Neurofibromatosis type 1: pathologic substrate of for learning in Drosophila. Nature 403: 895–898. high-signal-intensity foci in the brain. Radiology 195: Gutmann DH, Aylsworth A, Carey JC et al. (1997). The diag- 721–724. nostic evaluation and multidisciplinary management of 226 Section 2. Effects of cancer and cancer treatment on cognition

neurofibromatosis 1 and neurofibromatosis 2. JAmMed Moore BD, Slopis JM, Jackson EF et al. (2000). Brain vol- Assoc 278: 51–57. ume in children with neurofibromatosis, type 1: relation Hofman KJ, Harris EL, Bryan RN et al. (1994). Neurofibro- to neuropsychological status. Neurology 54: 914–920. matosis type 1: the cognitive phenotype. J Pediatr 124: Mulvihill JJ, Parry DM, Sherman JL et al. (1990). S1–S8. Neurofibromatosis-1 (Recklinghausen Disease) and Itoh T, Magnaldi S, White RM et al. (1994). Neurofibro- neurofibromatosis-2 (bilateral acoustic neurofibromato- matosis type 1: the evolution of deep gray and white mat- sis) – an update. Ann Int Med 113: 39–52. ter MR abnormalities. AJNR Am J Neuroradiol 15: 1513– National Insitutes of Health (1988). Consensus develop- 1519. ment conference: neurofibromatosis conference state- Kayl AE, Moore B, Slopis JM et al. (2000). Quantitative mor- ment. Arch Neurol 45: 575–578. phology of the corpus callosum in children with neurofi- NorthK,Joy,P,YuilleDet al. (1994). Specific learning dis- bromatosis and attention deficit hyperactivity disorder. ability in children with neurofibromatosis type 1: signif- J Child Neurol 15: 90–96. icance of MRI abnormalities [see comments]. Neurology Korf BR, Rubenstein AE (2005). Neurofibromatosis: A Hand- 44: 878–883. book for Patients, Families, and Health Care Professionals North KN, Riccardi V, Samango-Sprouse C et al. (1997). (2nd edn.). New York: Thieme Medical Publications. Cognitive function and academic performance in neu- Kortmann RD, Timmermann B, Taylor RE et al. (2003). Cur- rofibromatosis. 1: consensus statement from the NF1 rent and future strategies in radiotherapy of childhood Cognitive Disorders Task Force. Neurology 48: 1121– low-grade glioma of the brain. Part II: Treatment-related 1127. late toxicity. Strahlenther Onkol 179: 585. Packer RJ, Gutmann DH, Rubenstein A et al. (2002). Plexi- Leonard CM, Eckert MA, Lombardino LJ et al. (2001). form neurofibromas in NF1: toward biologic-based ther- Anatomical risk factors for phonological dyslexia. Cere- apy. Neurology 58(10): 1461. bral Cortex 11: 148–157. Paulesu E, Frith U, Snowling M et al. (1996). Is developmen- Lindgren SD, Benton AL (1980). Developmental patterns tal dyslexia a disconnection syndrome? Evidence from of visuospatial judgment. J Pediatr Psychol 5: 217– PET scanning. Brain 119 (Pt 1): 143–157. 225. Pavol M, Hiscock M, Massman P et al. (2006). Neuropsy- Listernick R, Louis DN, Packer RJ et al. (1997). Optic path- chological function in adults with von Recklinghausen’s way gliomas in children with neurofibromatosis 1: con- neurofibromatosis. Dev Neuropsychol 29: 509–526. sensus statement from the NF1 Optic Pathway Glioma Pugh KR, Shaywitz BA, Shaywitz SE et al. (1996). Cerebral Task Force. Ann Neurol 41: 143–149. organization of component processes in reading. Brain Margariti PN, Blekas K, Katzioti FG et al. (2007). Magneti- 119: 1221–1238. zation transfer ratio and volumetric analysis of the brain Riccardi VM (1981). Von Recklinghausen neurofibromato- in macrocephalic patients with neurofibromatosis type sis. NEnglJMed305: 1617–1627. 1. Eur Radiol 17: 433–438. Riccardi V, Eichner J (1986). Neurofibromatosis: Phenotype, Mazzocco MM, Turner JE, Denckla MB et al. (1995). Lan- Natural History, and Pathogenesis. Baltimore, MD: Johns guage and reading deficits associated with neurofibro- Hopkins University. matosis type 1: evidence for a not-so-nonverbal learning Rourke BP (1989). Nonverbal Learning Disabilities.New disability. Dev Neuropsychol 11: 503–522. York: Guilford Press. Moore BD, Denckla MB (1999). Neurofibromatosis. In Said SM, Yeh TL, Greenwood RS et al. (1996). MRI mor- Yeates K, Ris M, Taylor H (eds.) Pediatric Neuropsychol- phometric analysis and neuropsychological function in ogy: Research, Theory, and Practice.NewYork:Guilford patients with neurofibromatosis. Neuroreport 7: 1941– Publications, Inc. 1944. Moore BD, Ater JL, Needle MN et al. (1994). Neuropsycho- Schrimsher GW, Billingsley RL, Slopis JM et al. (2003). logical profile of children with neurofibromatosis, brain Visual-spatial performance deficits in children with neu- tumor, or both. J Child Neurol 9: 368–377. rofibromatosis type-1. Am J Med Genetics Part A 120A: Moore BD, Slopis JM, Schomer D et al. (1996). Neuropsy- 326–330. chological significance of areas of high signal intensity Sevick RJ, Barkovich AJ, Edwards MS et al. (1992). Evolution on brain MRIs of children with neurofibromatosis. Neu- of white matter lesions in neurofibromatosis type 1: MR rology 46: 1660–1668. findings. AJR Am J Roentgenol 159: 171–175. Chapter 15. Neurofibromatosis 227

Sharif S, Ferner R, Birch JM et al. (2006). Second primary bromatosis 1 and 2 gene function. Cell Biochem Biophys tumors in neurofibromatosis I patients treated for optic 34: 61–78. glioma: substantial risks after radiotherapy. J Clin Oncol Viskochil D, Buchberg AM, Xu G et al. (1990). Deletions and 24: 2570–2575. a translocation interrupt a cloned gene at the neurofibro- Shaywitz SE, Shaywitz BA, Pugh KR et al. (1998). Functional matosis type 1 locus. Cell 62: 187–192. disruption in the organization of the brain for reading in von Deimling A, Krone W, Menon AG (1995). Neurofibro- dyslexia. Proc Natl Acad Sci USA 95: 2636–2641. matosis type 1: pathology, clinical features and molecu- Silva AJ, Frankland PW, Marowit Z et al. (1997). A mouse lar genetics. Brain Pathol 5: 153–162. model for the learning and memory deficits associated Wadsby M, Lindehammar H, Eeg-Olofsson O (1989). with neurofibromatosis type I. Nat Genet 15: 281–284. Neurofibromatosis in childhood: neuropsychological Stanovich KE (1988). Explaining the differences between aspects. Neurofibromatosis 2: 251–260. the dyslexic and the garden-variety poor reader: the Wallace MR, Marchuk DA, Andersen LB et al. (1990). Type phonological-core variable-difference model. J Learn 1 neurofibromatosis gene: identification of a large tran- Disabil 21: 590–604. script disrupted in three NF1 patients [published erra- Temple E, Poldrack RA, Protopapas A et al. (2000). Disrup- tum appears in Science 1990 Dec 21; 250(4988): 1749]. tion of the neural response to rapid acoustic stimuli in Science 249: 181–186. dyslexia: evidence from functional MRI. Proc Natl Acad Zhu Y, Romero MI, Ghosh P et al. (2001). Ablation of NF1 Sci USA, 97: 13907–13912. function in neurons induces abnormal development of Temple E, Poldrack RA, Salidis J et al. (2001). Disrupted cerebral cortex and reactive gliosis in the brain. Genes neural responses to phonological and orthographic pro- Dev 15: 859–876. cessing in dyslexic children: an fMRI study. Neuroreport Zvulunov A, Weitz R, Metzker A (1998). Neurofibromatosis 12: 299–307. type 1 in childhood: evaluation of clinical and epidemio- Uhlmann EJ, Gutmann DH (2001). Tumor suppressor gene logic features as predictive factors for severity. Clin Pedi- regulation of cell growth – recent insights into neurofi- atr (Phila) 37: 295–299. 16

Hematological malignancies

Melissa Friedman and Mercedes Fernandez

Introduction treatments are associated with significant depletion of cognitive and functional abilities. Development of effective treatments for cancer has significantly improved survival rates for hematological cancer patients. For example, in 1964 the 5- Neurological complications independent of year survival rate for acute lymphoblastic leukemia cognitive deficits (ALL) was 3%; in 1995–2001, it was 86% (Leukemia and Lymphoma Society, 2005; Ries et al., 2005). Neurological complications, independent of cogni- However, survival is often associated with negative complaints, are common in hematological can- tive effects on cognitive functioning which inter- cer patients. Neurological complications have been fere with patients’ current and future functional reported in 11%–65% of hematopoietic stem cell status. transplantation (HSCT) patients (Faraci et al., 2002; In patients with hematological malignancies, risk Gallardo et al., 1996; Graus et al., 1996; Harder et factors for cognitive disorders are present during al., 2002; Sostak et al., 2003) and may be the main all stages of the disease and treatment process. cause of death in 8.5%–26% of recipients (Faraci et Risk factors include cancer treatments, anemia and al., 2002; Gallardo et al., 1996; Snider et al., 1994; fatigue, immune response activity, central nervous Sostak et al., 2003). White matter abnormalities or system (CNS) involvement of the primary malig- focal lesions have been observed on MRI in up to nancy (especially in the case of ALL), disease and 50% of cases (Harder et al., 2002; Sostak et al., 2003). treatment complications affecting the CNS such as Neurological complications may include neo- infection, hemorrhage, degeneration and leukoen- plasms, infections, encephalopathy, seizures, and cephalopathy, and cognitive and psychiatric disor- strokes. Neoplasms of hematopoietic or lymphoid ders that occur in the general population indepen- origin may compress brain structures, the extra- dent of having cancer. dural region, peripheral nerves, or may directly Information on the long-term neuropsychologi- invade the meninges (i.e., leptomeningeal infiltra- cal effects of cancer therapies is an important com- tion). Acute lymphoblastic leukemia is associated ponent of not only the informed consent process, with a particular propensity for leptomeningeal but also the treatment planning process. In the involvement; routine prophylactic CNS treatment risk-benefit analysis for selecting treatment, sur- is frequently instituted in cases of ALL. In addi- vival rates alone may be insufficient because some tion to neoplasms, viral or fungal infections (e.g.,

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

228 Chapter 16. Hematological malignancies 229

aspergillus) may occur as a result of treatment- treatment may allow for reversal of neurological induced immunosuppression and may lead to symptoms (Atkinson et al., 1984; Shah, 1999). encephalitis, meningitis or peripheral neuropathies Yet, discontinuing or lowering ciclosporin dose, or (Recht & Mrugala, 2003). Metabolic encephalopathy implementing new medications to manage neuro- has been reported in 3%–37% of HSCT patients. Pro- logical side-effects, may place patients at increased gressive, treatment-induced leukoencephalopathy risk for other life-threatening complications (Uckan has been reported in 1%–2% of patients (Antonini et al., 2005). et al., 1998; Antunes et al., 2000; Graus et al., 1996). Seizures occur in approximately 10% of HSCT Cognitive deficits patients, and cerebrovascular infarctions or hemor- rhages in 6% (Gallardo et al., 1996). Studies have provided valuable information that Identified risk factors for neurological complica- quantifies the impact of hematological cancer tions in HSCT patients include older age, prior treat- and treatment on patients’ cognitive functioning, ment with intrathecal methotrexate (MTX), long- and can lead to improvement in interventions and term use of ciclosporin, use of total body irradiation outcomes. Nevertheless, methodological challenges (TBI) in the conditioning regimen, corticosteroid of studies in this population warrant caution in their medication (Faraci et al., 2002; Sostak et al., 2003), interpretation. Most studies do not use random and allogeneic transplantation. Among allogeneic assignment of patients to treatment groups, and recipients, those with a donor who has a match- treatments are often inextricably confounded with ing human leukocyte antigen (HLA) display the disease variables, such as severity, co-morbidity lowest risk for complications (De Brabander et al., or underlying disease process. Small sample sizes 2000). Neurological complications have been asso- may lead to underestimation of group differences ciated with acute and severe graft verses host dis- due to low statistical power, or overestimation of ease (GVHD) (Faraci et al., 2002; Sostak et al., 2003), group differences when multiple statistical tests or as well as chronic GVHD which has been linked to comparisons are conducted. Cross-sectional stud- hypertension and small vessel disease (Padovan et ies that compare outcome between groups after al., 1998). treatment do not allow the identification of pre- Methotrexate, used as a systemic or intrathe- treatment group differences, and do not allow for cal cancer treatment for hematological malignancy, the identification of intra-individual factors that is associated with symptoms ranging from fatigue may account for post-treatment findings. In longitu- and dizziness to encephalopathy, with encephalo- dinal studies, subject attrition carries the confound pathic symptoms including hemiparesis, ataxia, and that patients evaluated at follow-up represent a dif- seizures. Methotrexate neurotoxicity can become ferent group than those who do not complete the chronic, lasting months to years, and may include study. Finally, many patients receive multiple treat- leukoencephalopathy, which can lead to coma and ments or experience multiple disease complica- death (Vezmer et al., 2003). tions, which makes it difficult to attribute any iden- Ciclosporin, administered as an immunosuppres- tified cognitive impairments to a single cause. sive agent to minimize GVHD, has been associated with a range of neurological complications Cognitive effects of treatment and symptoms. These include cerebellar symptoms, confusion (Atkinson et al., 1984), tremor, EEG and Most patients with hematological malignancies are MRI abnormalities (Shah, 1999), posterior leukoen- exposed at various phases in their illness to neu- cephalopathy evidenced by severe oculogyric cri- rotoxic cancer therapies, including chemotherapy, sis (Antunes et al., 1999), and seizures leading to radiation, HSCT, and biological therapies, all of death (Velu et al., 1985). Discontinuing ciclosporin which have been associated with adverse effects 230 Section 2. Effects of cancer and cancer treatment on cognition

on cognition (Harder et al., 2005; for reviews, see apy. A subset of these patients displayed cognitive Anderson-Hanley et al., 2003; Armstrong, 2001; Lee impairments on tests assessing psychomotor func- et al., 2004; Meyers & Valentine, 1995; Mulhern & tions, and the ability to copy and to later recall a Butler, 2004). In addition, the reader is referred to complex figure. However, in contrast to previous Chapters 7–9 in this volume discussing treatment findings (Ahles et al., 2002), these investigators did effects. not identify significant differences between patients with a history of systemic chemotherapy as compared to those with local radiation only. Neither did Chemotherapy they observe differences between patients who had In cancer patients, systemic chemotherapy has been received only one course of systemic chemother- associated with cognitive dysfunction not only in apy as compared to those who had received mul- the scientific medical literature but also in the pop- tiple courses. The investigators concluded that ular media, which has referred to such dysfunction chemotherapy was probably not the only contribu- as chemobrain and chemofog. Systemic chemother- tor to the observed cognitive deficits in this hemato- apy agents are generally introduced through oral logical cancer population. or intravenous routes. In addition, patients with In children, cognitive deficits observed at a given ALL may receive intrathecal chemotherapy or cra- time point should be viewed in the context of neu- nial radiotherapy (RT) in addition to systemic ropsychological development. Although some stud- chemotherapy, to prevent CNS metastatic disease. ies do not document chemotherapy-related changes In adults, chemotherapy is associated with cog- on neuropsychological tests (Rodgers et al., 2003) nitive effects on psychomotor function and mem- or MRI measures of hippocampal volume (Hill ory which are subtle but interfere with everyday et al., 2004), other studies associate chemother- functioning, and may still be present 10 years after apy with modest cognitive deficits (Espy et al., chemotherapy is completed (Ahles & Saykin, 2001; 2001; Kaemingk et al., 2004; Kingma et al., 2002). Ahles et al., 2002). In a cross-sectional compari- In children with ALL, receiving MTX (intrathecal son of neuropsychological performance in breast and/or high-dose IV administration) without RT cancer and lymphoma patients receiving systemic is associated with slowed processing speed, but chemotherapy versus local radiation therapy, sys- little or no effect on accuracy or on attentional and temic chemotherapy was associated with lower information-processing tasks (Buizer et al., 2005; performance in a subset of patients, particularly Mennes et al., 2005). Kaemingk et al. (2004) docu- on measures of psychomotor functioning and ver- mented math weaknesses in survivors of ALL who bal memory. Despite the group differences, how- had completed treatment with systemic chemother- ever, both treatment groups performed within nor- apy and intrathecal prophylaxis including cytosine mal limits overall, suggesting that the effects of arabinoside, hydrocortisone and MTX, to prevent chemotherapy on cognition were subtle. Group dif- CNS metastases. Survivors performed below nor- ferences were not accounted for by depression, anx- mal limits on one math test, and lower than their iety or fatigue (Ahles et al., 2002), but may have matched controls but within normal limits on four been related to disease factors necessitating sys- other math tests. Of note, illness-related school temic chemotherapy as opposed to localized treat- absences may contribute to or even account for the ment. modest deficits in these patients. Another study (Harder et al., 2005) evaluated neu- The modest findings identified in research stud- ropsychological functioning of 183 hematological ies may underestimate the true effect of chemother- cancer patients, 101 of whom were scheduled to apy on patients’ lives. The quiet and controlled undergo HSCT, and nearly all (n = 173) of whom neuropsychological testing environment does not had received at least one course of chemother- mimic the distractions and multiple cognitive Chapter 16. Hematological malignancies 231

demands often present in everyday situations. Addi- et al., 2006). Nevertheless, the effect of RT is dif- tionally, the neuropsychological tests used may not ficult to quantify because it is often administered be sensitive enough to detect cognitive impairments in combination with intrathecal MTX (Waber et al., affecting patients’ functioning. In children, cogni- 1995), and there may be a synergistic effect of treat- tive deficits may manifest themselves at a time ments. In fact, RT may not induce negative cogni- point in academic development beyond the test- tive effects (Mulhern et al., 1992), and any negative ing. Finally, it may be that only a subset of patients impact of RT in ALL treatment should be evaluated in these studies is vulnerable to chemotherapy- relative to the improved protection that it provides induced deficits. Research on lymphoma and breast against CNS relapse (Langer et al., 2002). Also, con- cancer patients suggests that the ε4 allele of sistent with an emerging literature on genetic vul- the apolipoprotein E (APOE) gene, which predis- nerability to treatment-induced neurotoxicity (e.g., poses people to Alzheimer’s disease (Richard & Ahles et al., 2003), particular genotypes may predis- Amouyel, 2001), may also predispose some patients pose ALL patients to RT-induced cognitive deficits to chemotherapy-induced cognitive deficits (Ahles (Krajinovic et al., 2005). et al., 2003).

Radiation therapy Hematopoietic stem cell transplantation

Cranial radiotherapy (CRT), like intrathecal MTX, is Hematopoetic stem cell transplantation (HSCT) used as prophylaxis in the treatment of ALL to pre- refers to the two-part process of administering vent CNS metastases. Most studies evaluating the a conditioning regimen of intensive high-dose neuropsychological effects of RT in hematological chemotherapy or radiation, followed by infusing the patients are conducted on children with ALL. Some patient with stem cells obtained from a donor’s research suggests that treatment in young children marrow or peripheral blood, for the purposes of (i.e., before 36 months of age) is associated with hematopoietic rescue. HSCT is based on the ratio- greater cognitive deficits than treatment in older nale that eradication of cancer cells is more likely if children (i.e., after 36 months of age) (Waber et al., the chemotherapy and radiation doses are not lim- 2001). ited by their lethal effects on the patient’s blood Until the 1980s, 24 Gy was the standard RT dose production system. In HSCT, patients are adminis- for children with ALL. Due to the recognition of tered treatment regimens (conditioning regimens) cognitive deficits associated with this, lower doses in intensive doses, which leave patients severely (e.g., 12–18 Gy) are now used (Oeffinger & Hudson, vulnerable to infection, anemia, and hemorrhage 2004). At these lower doses of RT, cognitive deficits (due to lack of platelets). After administration of have been detected. One study assessed the intel- such otherwise lethal treatment doses, patients are lectual, academic, attention and memory perfor- infused with a donor’s blood stem cells as a method mance in children at least 5 years after their diag- of restoring normal blood cell production. nosis of ALL (Spiegler et al., 2006), all of whom Timing of HSCT during the course of a patient’s were treated with a uniform chemotherapy pro- disease varies according to medical and practical tocol and intrathecal therapy. Those whose CNS considerations, including the disease, prognosis, prophylaxis consisted additionally of RT performed medical status, history of complications, availability more poorly on most cognitive measures compared of a compatible donor, and age. Although occasion- with those receiving high- or very-high-dose intra- ally patients receive HSCT prior to any other cancer venous MTX; the authors concluded that avoidance treatment, patients frequently receive HSCT during of RT in the treatment strategy is associated with first remission, first relapse, second remission, and good long-term neurocognitive outcomes (Spiegler so on. Therefore, many patients who undergo HSCT 232 Section 2. Effects of cancer and cancer treatment on cognition

have been exposed to potentially neurotoxic treat- Biological treatments ments before their transplant. Biological response modifiers such as interferon In adults, cross-sectional and prospective stud- alpha (IFN-α), commonly used to treat chronic ies indicate that many HSCT patients display pre- myelogenous leukemia, and interleukin-2 (IL-2) transplant cognitive deficits in the areas of mem- have been associated with a higher risk for cognitive ory, complex attention, and psychomotor speed impairment than other forms of treatment (Mey- (Andrykowski et al., 1992; Friedman, 2001; Harder ers & Abbruzzese, 1992). Patients treated with IFN-α et al., 2005, 2006; Peper et al., 2000), which display poorer cognitive speed and mood than those worsen during the hospital stay (Ahles et al., 1996; treated with chemotherapy (Pavol et al., 1995). Neu- Meyers et al., 1994) but are not significantly dif- rotoxic effects are seen at low doses of IFN-α,and ferent from baseline at or beyond a 1-year post- cognitive and mood functioning are correlated with HSCT follow-up (Friedman, 2001; Harder et al., length of time on treatment (Meyers, 1999). 2006). Despite post-HSCT improvements, cognitive Treatment with IFN-α has been associated with abilities may not return to normal levels; memory memory and executive dysfunction and slowed pro- impairments have been documented 8 months after cessing speed that cannot be accounted for by the transplantation (Meyers et al., 1994), and executive frequently co-occurring cytokine-induced depres- and information processing deficits up to 8 years sion. These deficits are severe enough to interfere afterwards (Harder et al., 2002). Patients also experi- with occupational and daily functioning, and may ence effects on quality of life indices, including sleep be exacerbated by a high cumulative IFN-α dose and energy deficits noted 18 months after trans- or concurrent chemotherapy. The memory deficits plantation (Andrykowski et al., 1997), and inability are associated with executive and information- to work (Stalfelt & Zettervall, 1997). processing dysfunction, rather than hippocampal In children, prospective studies indicate cognitive damage (Scheibel et al., 2004). This pattern is con- deficits associated with HSCT. Kramer et al. (1997) sistent with frontal-subcortical pathology as seen found that IQ and behavioral adaptation declined in Parkinson’s disease, and, in fact, some patients from pre-transplant to a 1-year follow-up evalu- on IFN-α treatment have been reported to display ation with no additional decline detected after 3 extrapyramidal signs including rigidity, tremor, and years. Phipps et al. (2000) did not identify decline masked facies (Meyers et al., 1991). in IQ or achievement measures between the pre- Patients receiving IFN-α experience neurovegeta- transplant baseline and the 1- and 3-year follow- tive symptoms such as appetite change and fatigue up evaluations, but found that achievement scores around the 2nd week of treatment, with mood and were generally one-half to two-thirds of a standard cognitive disturbance appearing around the 8th to deviation below the normative mean, and this was 12th weeks of treatment (Capuron et al., 2002). Neu- not attributable to prior therapy. Data suggest that ropsychological impairments have been observed younger children (under 36 months of age) at the from 3 weeks to 2 years after discontinuation of time of transplantation display greater vulnerabil- the treatment, and the deficits may not always be ity to cognitive impairment than older children (6 reversible (Meyers et al., 1991). years or older) (Kramer et al., 1997; Phipps et al., 2000). Younger children may be more vulnerable to the effects of psychosocial variables, such as iso- Anemia and iron deficiency lation and hospitalization, on neuropsychological functioning and behavioral adaptation. Addition- Anemia is a concern for hematological cancer ally, brain development may be more susceptible to patients due to the effects of both the cancer treatment side-effects at earlier stages of develop- itself and the treatment on the blood produc- ment than at later stages. tion system. Anemia is associated with increased Chapter 16. Hematological malignancies 233

fatigue and cognitive decline particularly in func- Metallinos-Katsaras et al., 2004), although deficits tions related to attention and memory. Jacobsen are not always reversible. For instance, Walter (1994) et al. (2004) examined the relationship between described two studies, one conducted in Chile and hemoglobin levels, fatigue, and cognitive function- the other in Costa Rica, demonstrating the negative ing in adult cancer patients undergoing repeated cognitive effects of anemia in infants. Fewer anemic chemotherapy administrations. In patients whose children than controls successfully completed tasks hemoglobin declined to below normal levels, the assessing language comprehension and balance. amount of hemoglobin decline was related to the Follow-up studies at 5 years of age revealed that level of fatigue and to deficits in attention, exec- those who were anemic as infants scored lower on utive functioning, and visual memory. In elderly neuropsychological tests than controls, suggest- cancer patients, chemotherapy-related anemia may ing that anemia during critical periods of neural adversely affect not only cognitive but also func- development may have enduring negative effects. tional status, and successful anemia treatment with recombinant human erythropoeitin (rHuEPO) sup- Immune response plementation over the course of chemotherapy may help to preserve functional independence (Man- The normal immune response to illness consists of cuso et al., 2006). inflammatory cytokine activity, which may underlie In non-cancer patients, reports of mental sta- neuropsychological changes in cancer patients (Lee tus changes associated with anemia date back to et al., 2004). Cytokines are chemicals secreted in the 1830s (Stivelman, 2000). Thalassemia, an inher- the body which act as messengers between immune ited form of anemia associated with faulty synthe- cells. Proinflammatory cytokines, including tumor sis of hemoglobin, has been associated with tran- necrosis factor alpha (TNF-α), IL-1 β,andIL-6,are sient ischemic attacks, silent infarctions, and in rare proteins that augment the body’s immune response cases stroke (Armstrong, 2005). In end-stage renal by helping to speed the elimination of pathogens disease, anemia has been associated with changes and the resolution of the inflammatory challenge. not only in cognitive functioning (Martin-Lester, They communicate with white blood cells, causing 1997; Temple et al., 1995) but also in measures of them to become activated and to respond to infec- event related brain potentials (ERPs) (Brown et al., tion and inflammation. They have also been linked 1991; Marsh et al., 1991; Nissenson, 1992), which are to stress reactions, hormone and neurotransmitter direct electrophysiological measures of brain func- activity, and “sickness behaviors” such as increased tioning. Furthermore, successful treatment of ane- sleep, decreased appetite and decreased sexual mia with rHuEPO in renal patients is associated drive (Watkins, 2000). Proinflammatory cytokines with improved neuropsychological performance on and their receptors have been detected in vari- measures of attention and working memory (Marsh ous areas of the brain, including the hippocam- et al., 1991) and with increased amplitudes and pus and hypothalamus. They may enter the brain decreased latencies of the P300 component of the from outside of the CNS, or may be synthesized ERP (Singh et al., 2006). In healthy adults, artificially and released from within the CNS. Hematopoietic induced anemia has been associated with increased cytokines, such as IL-3, IL-5 and colony-stimulating latency of the P300 (Weiskopf et al., 2005); it has also factors, are involved in altering the hematopoietic been linked to slowed performance on the Digit- response, and may play a role in hematological Symbol Substitution Test, which may be reversed by malignancies (see Kronfol & Remick, 2000 for a transfusing individuals with either fresh or stored review). autologous erythrocytes (Weiskopf et al., 2006). The immune response alone, independent of can- In children, cognitive deficits may improve cer treatment, may have an impact on cognition. once the anemia resolves (Bruner et al., 1996; In a group of patients with acute myelogenous 234 Section 2. Effects of cancer and cancer treatment on cognition

leukemia or myelodysplastic syndrome, cognitive tion and information processing may help improve impairment and fatigue were observed before the these abilities (Butler & Copeland, 2002). initiation of treatment, and poorer cognitive per- Regarding psychiatric medications, paroxetine formance was associated with higher levels of cir- has been found effective at reducing symptoms of culating cytokines (Meyers et al., 2005). Increased depression, anxiety, cognitive dysfunction, and pain levels of proinflammatory cytokines and poorer in melanoma patients receiving IFN-α therapy, in a performance status rating have been observed double-blind, placebo-controlled randomized trial in untreated Hodgkin’s disease patients (Seymour (Capuron et al., 2002). Methylphenidate has been et al., 1997). found effective at reducing social and attentional In healthy humans, experimental activation of deficits in children with ALL and brain tumors the immune response produces memory distur- (Mulhern et al., 2004). Treatment issues are bance, anxiety and depression, even in the absence addressed in greater detail in Chapter 18. of subjective symptoms of sickness, and these neuropsychological changes are correlated with cytokine secretion levels (Reichenberg et al., 2001). Summary In neonates, cytokine secretion levels have been associated with impaired cerebral metabolism dur- Cognitive deficits have been documented in hema- ing the neonatal period and with abnormal neu- tological cancer patients. Such deficits may be rodevelopment at 30 months of age (Bartha et al., attributable to cancer treatment, underlying dis- 2004). ease factors such as anemia, or immune response mechanisms. Additionally, co-occurring psychiatric or neurological conditions independent of the can- Treatment for cognitive and emotional cer may contribute to cognitive deficits. The studies dysfunction reviewed suggest that cognitive deficits attributable to the cancer or its treatment are often subtle as Due to the multiple contributing etiological factors detected on neuropsychological testing, but severe in cancer-related symptomatology, an individ- enough to interfere with everyday or occupational ualized, multidisciplinary treatment approach functioning. However, not all patients experience addressing cognitive, emotional, and physical neuropsychological deficits, and any deficits should symptoms is recommended. be evaluated in the context of the increased oppor- Psychosocial interventions often play an impor- tunity for survival and sometimes cure that many tant role in remediation of cognitive deficits or treatments offer. Additionally, many interventions overall functioning. Stress management techniques, exist to minimize the impact of neuropsychological breathing exercises, and aerobic exercise may be disturbances on the daily lives of patients. useful for improving fatigue and quality of life (Kim & Kim, 2005; Wilson et al., 2006). Patient and caregiver education about cancer- and treatment- REFERENCES related cognitive symptoms are also recommended. Children with cognitive and behavioral problems Ahles TA, Saykin A (2001). Cognitive effects of standard- following cancer treatment should be identified for dose chemotherapy in patients with cancer. Cancer special education services whenever possible, espe- Invest 19(8): 812–820. cially since the rate of utilization of special educa- Ahles TA, Tope DM, Furstenberg C, Hann D, Mills L (1996). tion services is not always consistent with the need Psychologic and neuropsychologic impact of autologous in this population (Buizer et al., 2006). In childhood bone marrow transplantation. J Clin Oncol 14: 1457– cancer survivors, exercises for remediating atten- 1462. Chapter 16. Hematological malignancies 235

Ahles TA, Saykin AJ, Furstenberg CT et al. (2002). Neuropsy- Bruner AB, Joffe A, Duggan AK, Casella JF, Brandt J (1996). chologic impact of standard dose systemic chemother- Randomised study of cognitive effects of iron supple- apy in long term survivors of breast cancer and lym- mentation in non-anaemic iron-deﬁcient adolescent phoma. J Clin Oncol 20: 485–493. girls. Lancet 348: 992–996. Ahles TA, Saykin AJ, Noll WW et al. (2003). The relationship Buizer AI, de Sonneville LMJ, Van Den Heuvel-Eibrink MM, of APOE genotype to neuropsychological performance in Veerman AJP (2005). Chemotherapy and attentional dys- long-term cancer survivors treated with standard dose function in survivors of childhood acute lymphoblas- chemotherapy. Psychooncology 12: 612–619. tic leukemia: effect of treatment intensity. Pediatr Blood Anderson-Hanley C, Sherman ML, Riggs R, Agocha VB, Cancer 45: 281–290. Compas BE (2003). Neuropsychological effects of treat- Buizer AI, de Sonneville LMJ, Van Den Heuvel-Eibrink MM, ments for adults with cancer: a meta-analysis and Veerman AJP (2006). Behavioral and educational lim- review of the literature. J Int Neuropsychol Soc 9: 967– itations after chemotherapy for childhood acute lym- 982. phoblastic leukemia or Wilms tumor. Cancer 106: 2067– Andrykowski MA, Schmitt FA, Gregg ME, Brady MJ, 2075. Lamb DG, Henslee-Downey PJ (1992). Neuropsychologic Butler RW, Copeland DR (2002). Attentional processes and impairment in adult bone marrow transplant candi- their remediation in children treated for cancer: a lit- dates. Cancer 70: 2288–2297. erature review and the development of a therapeutic Andrykowski MA, Carpenter JS, Greiner CB et al. (1997). approach. J Int Neuropsychol Soc 8: 115–124. Energy level and sleep quality following bone mar- Capuron L, Gumnick JF, Musselman DL, Lawson DH row transplantation. Bone Marrow Transplant 20: 669– et al. (2002). Neurobehavioral effects of interferon 679. alpha in cancer patients: phenomenology and paroxe- Antonini G, Ceschin V, Morino S et al. (1998). Early neu- tine responsiveness of symptom dimensions. Neuropsy- rologic complications following allogeneic bone marrow chopharmacology 26: 643–652. transplant for leukemia: a prospective study. Neurology De Brabander C. Cornelissen J, Smitt, P, Vecht CJ, van 50: 1441–1445. den Bent MJ (2000). Increased incidence of neurologi- Antunes NL, Small TN, George D, Farid B, Lis E (1999). cal complications in patients receiving an allogenic bone Posterior leukoencephalopathy syndrome may not be marrow transplantation from alternative donors. J Neu- reversible. Pediatr Neurol 20: 241–243. rol Neurosurg Psychiatry 68: 36–40. Antunes NL, Boulad FL, Prasad V, Rosenblum M, Lis E, Espy KA, Moore IM, Kaufmann PM, Kramer JH, Matthay K, Souweidane M (2000). Rolandic encephalopathy and Hutter JJ (2001). Chemotherapeutic CNS prophylaxis and epilepsia partialis continua following bone marrow neuropsychologic change in children with acute lym- transplant. Bone Marrow Transplant 26: 917–919. phoblastic leukemia: a prospective study. J Pediatr Psy- Armstrong FD (2001). Acute and long term neurodevel- chol 26:1–9. opmental outcomes in children following bone marrow Faraci M, Lanino E, Dini G et al. (2002). Severe neurologic transplantation. Front Biosci 6: 6–12. complications after hematopoietic stem cell transplan- Armstrong FD (2005). Thalassemia and learning: neu- tation in children. Neurology 59: 1895–1904. rocognitive functioning in children. Ann N Y Acad Sci Friedman MA (2001). Neuropsychological functioning in 1054: 283–289. hematopoietic stem cell transplantation. Dissertation Atkinson K, Biggs J, Darveniza P, Boland J, Concannon A, Abstracts. Houston, TX: University of Houston. Dodds A (1984). Cyclosporine associated central nervous Gallardo D, Ferra C, Berlanga JJ et al. (1996). Neu- system toxicity after allogeneic bone marrow transplan- rologic complications after allogeneic bone marrow tation. Transplantation 38: 34–37. transplantation. Bone Marrow Transplant 18: 1135– Bartha AI, Foster-Barber A, Miller SP et al. (2004). Neonatal 1139. encephalopathy: association of cytokines with MR spec- GrausF,SaizA,SierraJet al. (1996). Neurologic com- troscopy and outcome. Pediatr Res 56(6): 960–966. plications of autologous and allogeneic bone marrow Brown WS, Marsh JT, Wolcott D et al. (1991). Cognitive transplantation in patients with leukemia. Neurology 46: function, mood and P3 latency: effects of the ameliora- 1004–1009. tion of anemia in dialysis patients. Neuropsychologia 29: Harder H, Cornelissen JJ, Van Gool AR et al. (2002). Cog- 35–45. nitive functioning and quality of life in long-term adult 236 Section 2. Effects of cancer and cancer treatment on cognition

survivors of bone marrow transplantation. Cancer 95: Lee BN, Dantzer R, Langley KE et al. (2004). A cytokine- 183–192. based neuroimmunologic mechanism of cancer-related Harder H, van Gool AR, Cornelissen JJ et al. (2005). Assess- symptoms. Neuroimmunomodulation 11: 279–292. ment of pre-treatment cognitive performance in adult Leukemia and Lymphoma Society (2005). Facts 2005. Avail- bone marrow or haematopoietic stem cell transplanta- able online at www.leukemia-lymphoma.org, accessed 4 tion patients: a comparative study. Eur J Cancer 41: 1007– April 2008. 1016. MancusoA,MigliorinoM,DeSantisS,SaponieroA,De Harder H, Duivenvoorden HJ, van Good AR et al. (2006). Marinis F (2006). Correlation between anemia and func- Neurocognitive functions and quality of life in haema- tional/cognitive capacity in elderly lung cancer patients tological patients receiving haematopoietic stem cell treated with chemotherapy. Ann Oncol 17(1): 146–150. grafts: a one-year follow-up pilot study. J Clin Exp Neu- Marsh JT, Brown WS, Wolcott D et al. (1991). rHuEPO treat- ropsychol 28: 283–293. ment improves brain and cognitive function of anemic Hill DE, Ciesielski KT, Hart BL, Jung RE (2004). MRI mor- dialysis patients. Kidney Int 39: 155–163. phometric and neuropsychological correlates of long- Martin-Lester M (1997). Cognitive function in dialysis term memory in survivors of childhood leukemia. Pedi- patients. ANNA J 24: 359–365. atr Blood Cancer 42: 611–617. Mennes M, Stiers P, Vandenbussche E et al. (2005). Jacobsen PB, Garland LL, Booth-Jones M et al. (2004). Attention and information processing in survivors of Relationship of hemoglobin levels to fatigue and cog- childhood acute lymphoblastic leukemia treated with nitive functioning among cancer patients receiving chemotherapy only. Pediatr Blood Cancer 44: 478–486. chemotherapy. J Pain Symptom Manage 28: 7–18. Metallinos-Katsaras E, Valassi-Adamn E, Dewey KG, Lon- Kaemingk KL, Carey ME, Moore IM, Herzer M, Hutter nerdal B, Stamoulakatou A, Pollitt E (2004). Effect of iron JJ (2004). Math weaknesses in survivors of acute lym- supplementation on cognition in Greek preschoolers. phoblastic leukemia compared to healthy children. Child Eur J Clin Nutri 58: 1532–1542. Neuropsychol 10: 14–23. Meyers CA (1999). Mood and cognitive disorders in can- Kim S-D, Kim H-S (2005). Effects of a relaxation breathing cer patients receiving cytokine therapy. In Dantzer R (ed.) exercise on fatigue in haemopoietic stem cell transplan- Cytokines, Stress and Depression. New York: Kluwer Aca- tation patieints. J Clin Nurs 14: 51–55. demic/Plenum Publishers. Kingma A, Van Dommelen RI, Mooyaart EL, Wilmink JT, Meyers CA, Abbruzzese JL (1992). Cognitive functioning in Deelman BG, Kamps WA (2002). No major cognitive cancer patients: effect of previous treatment. Neurology impairment in young children with acute lymphoblastic 42: 434–436. leukemia using chemotherapy only: a prospective longi- Meyers CA, Valentine AD (1995). Neurological and psy- tudinal study. J Pediatr Hematol-Oncol 24: 106–114. chiatric adverse effects of immunological therapy. CNS Krajinovic M, Robaey P, Chiasson S et al. (2005). Polymor- Drugs 3: 56–68. phisms of genes controlling homocysteine levels and IQ Meyers CA, Scheibel RS, Forman AD (1991). Persistent score following the treatment for childhood ALL. Phar- neurotoxicity of systematically administered interferon macogenomics 6: 293–302. alpha. Neurology 41: 672–676. Kramer JH, Crittenden MR, DeSantes K, Cowan MJ (1997). Meyers CA, Weitzner M, Byrne K, Valentine A, Champlin Cognitive and adaptive behavior one and three years RE, Przepiorka D (1994). Evaluation of the neurobehav- following bone marrow transplantation. Bone Marrow ioral functioning of patients before, during, and after Transplant 19: 607–613. bone marrow transplantation. J Clin Oncol 12: 820–826. Kronfol Z, Remick DG (2000). Cytokines and the brain: Meyers CA, Albitar M, Estey E (2005). Cognitive impair- implications for clinical psychiatry. Am J Psychiatry 157: ment, fatigue, and cytokine levels in patients with acute 683–694. myelogenous leukemia or myelodysplastic syndrome. Langer T, Martus P, Ottensmeier H, Hertzberg H, Beck Cancer 104: 788–793. JD, Meier W (2002). CNS late-effects after ALL therapy Mulhern RK, Butler RW (2004). Neurocognitive sequelae of in childhood. Part III: Neuropsychological performance childhood cancers and their treatment. Pediatr Rehabit in long-term survivors of childhood ALL: impairments 7: 1–14. of concentration, attention, and memory. Med Pediatr Mulhern RK, Ochs J, Fairclough D (1992) Deterioration Oncol 38: 320–328. of intellect among children surviving leukemia: IQ test Chapter 16. Hematological malignancies 237

changes modify estimates of treatment toxicity. J Consult Seymour JF, Talpaz M, Hagemeister FB, Cabanillas F, Clin Psychol 60: 477–480. Kurzrock R (1997). Clinical correlates of elevated Mulhern RK, Khan RB, Kaplan S et al. (2004). Short serum levels of interleukin-6 in patients with untreated term efficacy of methylphenidate: a randomized double- Hodgkin’s disease. Am J Med 102: 21–28. blind, placebo-controlled trial among survivors of child- Shah AK (1999). Cyclosporine A neurotoxicity among bone hood cancer. J Clin Oncol 22: 4795–4803. marrow transplant recipients. Clin Neuropharmacol 22: Nissenson AR (1992). Epoetin and cognitive function. Am J 67–73. Kidney Dis 20: 21–24. Singh NP,Sahni V,Wadhwa A et al. (2006). Effect of improve- Oeffinger KC, Hudson MM (2004). Long term complica- ment in anemia on electroneurophysiological markers tions following childhood and adolescent cancer: foun- (P300) of cognitive dysfunction in chronic kidney dis- dations for providing risk-based health care for survivors. ease. Hemodial Int 10(3): 267–273. Erratum in Hemodial CA Cancer J Clin 54: 208–236. Int 2006; 10(4): 408. PadovanCS,YousryTA,SchleuningM,HollerE,KolbH- Snider S, Bashir R, Bierman P (1994). Neurologic com- J, Straube A (1998). Neurological and neuroradiological plications after high-dose chemotherapy and autolo- findings in long-term survivors of allogeneic bone mar- gous bone marrow transplantation for Hodgkin’sdisease. row transplantation. Ann Neurol 43: 627–633. Neurology 44: 681–684. Pavol MA, Meyers CA, Rexer JL, Valentine AD, Mattis PJ, Tal- Sostak P, Padovan CS, Yousry TA, Ledderose G, Kolb HJ, paz M (1995). Pattern of neurobehavioral deficits associ- Straube A (2003). Prospective evaluation of neurological ated with interferon alfa therapy for leukemia. Neurology complications after allogeneic bone marrow transplan- 45: 947–950. tation. Neurology 60: 842–848. Peper M, Steinvorth S, Schraube P et al. (2000). Neurobe- Spiegler BJ, Kennedy K, Maze R et al. (2006). Comparison havioral toxicity of total body irradiation: a follow-up in of long-term neurocognitive outcomes in young children long-term survivors. Int J Radiat Oncol Biol Phys 46(2): with acute lymphoblastic leukemia treated with cranial 303–311. radiation or high-dose or very high-dose intravenous Phipps S, Dunavant M, Srivastava DK, Bowman L, Mulhern methotrexate. J Clin Oncol 24: 3858–3864. RK (2000). Cognitive and academic functioning in sur- Stalfelt AM, Zettervall O (1997). Quality of life in young vivors of pediatric bone marrow transplantation. JClin patients with chronic myelocytic leukaemia during Oncol 18 (5): 1004–1011. intensive treatment including interferon. Leukemia Res Recht L, Mrugala M (2003). Neurologic complications 21(8): 775–783. of hematologic neoplasms. Neurol Clin 21: 87– Stivelman JC (2000). Benefits of anaemia treatment on 105. cognitive function. NephrolDialTransplant15(3): 29– Reichenberg A, Yirmiya R, Schuld A et al. (2001). Cytokine- 35. associated emotional and cognitive disturbances in Temple RM, Deary IJ, Winney RJ (1995). Recombinant humans. Arch Gen Psychiatry 58: 445–452. erythropoietin improves cognitive function in patients Richard F, Amouyel P (2001). Genetic susceptibility factors maintained on chronic ambulatory peritoneal dialysis. for Alzheimer’s disease. Eur J Pharmacol 412: 1–12. NephrolDialTransplant10: 1733–1738. Ries LAG, Eisner MP, Kosary CL et al. (2005). SEER Can- Uckan D, Cetin M, Yigitkanli I et al. (2005). Life-threatening cer Statistics Review, 1975–2002. Bethesda, MD: National neurological complications after bone marrow trans- Cancer Institute. http://seer.cancer.gov/csr/1975 2002/, plantation in children. Bone Marrow Transplant 35: 71– based on November 2004 SEER data submission, posted 76. to the SEER website, 2005. Velu T, Debusscher L, Stryckmans PA (1985). Cyclosporin- Rodgers J, Marckus R, Kearns P, Windebuck K (2003). associated fatal convulsions. Lancet Jan 26; 1(8422): 219. Attentional ability among survivors of leukaemia treated Vezmer S, Becker A, Bode, U, Jaehde U (2003). Biochem- without cranial irradiation. Arch Dis Child 88: 147– ical and clinical aspects of methotrexate neurotoxicity. 150. Chemotherapy 49: 92–104. Scheibel RS, Valentine AD, O’Brien S, Meyers CA (2004). Waber DP, Tarbell NJ, Fairclough D et al. (1995). Cognitive Cognitive dysfunction and depression during treatment sequelae of treatment in childhood acute lymphoblas- with interferon-alpha and chemotherapy. J Neuropsychi- tic leukemia: cranial radiation requires an accomplice. atry Clin Neurosci 16: 185–191. J Clin Oncol 13: 2490–2496. 238 Section 2. Effects of cancer and cancer treatment on cognition

Waber DP, Shapiro BL, Carpentieri SC et al. (2001). Excel- Weiskopf RB, Toy P, Hopf HW et al. (2005). Acute iso- lent therapeutic efficacy and minimal late neurotoxicity volemic anemia impairs central processing as determin- in children treated with 18 grays of cranial radiation ther- ed by P300 latency. Clin Neurophysiol 116: 1028–1032. apy for high-risk acute lymphoblastic leukemia. Cancer Weiskopf RB, Feiner J, Hopf H et al. (2006). Fresh blood and 92: 15–22. aged stored blood are equally efficacious in immediately Walter T (1994). Effect of iron-deficiency anaemia on cog- reversing anemia-induced brain oxygenation deficits in nitive skills in infancy and childhood. Baillieres Clin humans. Anesthesiology 104: 911–920. Haematol 7: 815–827. Wilson RW, Taliaferro LA, Jacobsen PB (2006). Pilot study Watkins LR (2000). The pain of being sick: implications of a self-administered stress management and exercise of immune-to-brain communication for understanding intervention during chemotherapy for cancer. Support pain. Annu Rev Psychol 51: 29–57. Care Cancer 14: 928–935. 17

Paraneoplastic disorders

Edward Dropcho

Introduction 1%–3%. Other tumors over-represented among adults with paraneoplastic syndromes include Neurological paraneoplastic disorders refer to breast carcinoma, ovarian carcinoma, Hodgkin’s non-metastatic disorders that are not attributable lymphoma, thymoma, and testicular germ cell to the toxicity of cancer therapy, cerebrovascu- tumors. Except for opsoclonus-myoclonus associ- lar disease, coagulopathy, infection, or toxic and ated with neuroblastoma, paraneoplastic disorders metabolic causes. Paraneoplastic disorders can in children are rare. affect any part(s) of the central (CNS) or peripheral Paraneoplastic disorders are far less common (PNS) nervous systems (Table 17.1). Patients can be than nervous system metastases and are relatively roughly grouped into those with pure or relatively rare compared to other non-metastatic neurological pure clinical involvement of one part of the nervous complications of systemic cancer, but they are clin- system, such as cerebellar degeneration or sensory ically important for several reasons. First, in most neuronopathy, and those with signs and symp- patients with paraneoplastic disorders, the neu- toms of a diffuse and multifocal “paraneoplastic rological symptoms are the presenting feature of encephalomyelitis” (Dropcho, 2002; Graus et al., an otherwise undiagnosed tumor. Physicians must 2004). Several syndromes should always raise the therefore be able to identify the disorder as para- possibility of a paraneoplastic etiology, including neoplastic and to initiate the appropriate search for Lambert–Eaton myasthenic syndrome, subacute the tumor. Second, paraneoplastic disorders often cerebellar degeneration, severe sensory neuronopa- cause severe and permanent neurological morbid- thy, limbic encephalopathy, and opsoclonus- ity. Third, prompt recognition of a paraneoplas- myoclonus. None of the clinical syndromes, how- tic disorder maximizes the likelihood of successful ever, have an absolute association with neoplasia, tumor treatment and a favorable neurological out- and each can occur in patients without tumors. come. For any paraneoplastic neurological disorder, Most neurological paraneoplastic disorders are there is a clear over-representation of one or a believed to be autoimmune diseases. The cen- few particular neoplasms. Overall, small cell lung tral theory of autoimmunity postulates that tumor carcinoma is the tumor most often associated with cells express “onconeural” antigen(s) that are iden- paraneoplastic phenomena in adults, although the tical or antigenically related to molecules nor- actual incidence of paraneoplastic disorders among mally expressed by neurons, and that in rare patients with this tumor is probably no more than instances an autoimmune response initially arising

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

239 240 Section 2. Effects of cancer and cancer treatment on cognition

Table 17.1. Neurologic paraneoplastic disorders more widespread or pan-neuronal reactivity and are associated with a variety of clinical neurologi- Central nervous system Peripheral nervous system cal syndromes, or with multifocal encephalomyelitis. The most prevalent such antibodies are anti-Hu Multifocal encephalomyelitis Sensory neuronopathy Cerebellar degeneration Nerve vasculitis and anti-CV2 (Pittock et al., 2004). Limbic encephalitis Sensorimotor polyneuropathy There are good but not perfect correlations Opsoclonus-myoclonus Motor neuropathy among particular paraneoplastic syndromes, anti- Extrapyramidal syndrome Neuromyotonia neuronal antibody specificities, and associated Brainstem encephalitis Autonomic insufficiency tumor types. Antineuronal antibodies are useful Myelopathy Lambert–Eaton syndrome diagnostic tools because, when present, they greatly Motorneuron disease Inflammatory myopathy increase the index of suspicion for a paraneoplas- Stiff person syndrome Necrotizing myopathy tic condition, and the type of antibody can help Optic neuritis guide the search for the underlying tumor. Antineu- Retinal degeneration ronal antibody assays do, however, have important practical clinical limitations. First, a given clinical against the tumor “spills over” to attack neurons syndrome (e.g., limbic encephalitis) may be associ- expressing the same or related antigen(s) (Dropcho, ated with one of several autoantibodies; conversely, 2002; Roberts & Darnell, 2004). If true, this theory a given autoantibody (e.g., anti-Hu) may be asso- should be supported by several lines of evidence: ciated with a variety of clinical presentations. Sec- (1) the neuropathology should be consistent with ond, for several of the syndromes, a few patients an immune or inflammatory process; (2) affected have high-titer antineuronal autoantibodies and yet patients should have specific antibody or cellular never develop a demonstrable tumor. The presence immune autoreactivity; (3) tumor cells in affected of antibodies does not absolutely indicate an under- patients should express the onconeural antigen(s); lying neoplasm. Third, several of the autoantibodies (4) there should be a demonstrable antitumor are present at low titers in tumor patients without immune response; (5) immunosuppressive treat- any accompanying clinical neurological manifes- ment should produce a beneficial clinical effect; tations. Fourth, patients with a suspected para- (6) the clinical and neuropathological features neoplastic syndrome may not have demonstrable should be reproducible in an experimental model. antineuronal antibodies, or may have “atypical” or Many patients with neurological paraneoplastic incompletely characterized antibodies not detected disorders have one or more of a steadily growing in commercially available assays. A negative anti- list of circulating antineuronal antibodies (Dropcho, body assay, therefore, does not rule out the possi- 2002). The neuronal molecular targets of several of bility of a paraneoplastic disorder and the presence these autoantibodies have been cloned and char- of an underlying neoplasm. acterized. Protein antigens reacting with antineu- For a few neurological paraneoplastic syndromes ronal antibodies are known to be expressed by the antineuronal autoantibodies are directly tumors from affected patients, supporting the gen- involved in causing clinical disease (Dropcho, 2002). eral theory of an autoimmune response arising Prime examples are Lambert–Eaton myasthenic against shared onconeural antigens. Some parane- syndrome caused by antibodies against P/Q-type oplastic antibodies have selective neuronal reactiv- voltage-gated calcium channels at the pre-synaptic ity and are found only in patients with a particular neuromuscular junction, and neuromyotonia clinical syndrome, such as anti-recoverin antibodies caused by antibodies against voltage-gated potas- in patients with retinal degeneration, and anti- sium channels at pre-synaptic nerve terminals. Yo antibodies in patients with cerebellar degen- Antibodies may also directly mediate neuronal dys- eration. Most paraneoplastic autoantibodies show function or injury for some CNS syndromes. Chapter 17. Paraneoplastic disorders 241

Examples include anti-recoverin antibodies in immunosuppressive or immunomodulatory treat- carcinoma-associated retinal degeneration, anti- ment (Dropcho, 2005). Several factors make it bodies against P/Q-type voltage-gated calcium difficult to interpret the published literature regard- channels or glutamate receptors in some patients ing immunotherapy for paraneoplastic disorders: with paraneoplastic cerebellar degeneration, and 1. These syndromes are relatively rare – for some anti-voltage-gated potassium channel antibodies of the syndromes there is only a handful of well- in a subset of patients with paraneoplastic or characterized published cases non-paraneoplastic limbic encephalitis (see below). 2. Most reports are anecdotal and nearly all single- For most paraneoplastic syndromes associated institution or multi-institution series are retro- with antineuronal antibodies, the antibodies are spective probably an epiphenomenon or they play a minor 3. There is a reporting bias, in that studies on indirect role in causing neuronal injury. Recent patients who respond to treatment are more studies of two of the most common CNS para- likely to be published than those on patients who neoplastic syndromes, i.e., encephalomyelitis/ do not respond sensory neuronopathy associated with small cell 4. For some syndromes there are pharmaco- lung cancer and cerebellar degeneration associ- logical treatments that improve neurological ated with breast or ovarian carcinoma, implicate symptoms independent of tumor treatment or cell-mediated immune effectors in causing neu- immunotherapy. Examples include Lambert– ronal injury. For these disorders it is postulated Eaton syndrome treated with pyridostigmine or that onconeural antigens released by apoptotic 3,4-diaminopyridine, or stiff person syndrome tumor cells are presented to T lymphocytes in treated with diazepam and baclofen draining peripheral lymph nodes, initiating a Th1 5. Patients with paraneoplastic encephalomyelitis, helper response that eventually gains access to cerebellar degeneration, and other syndromes the CNS and attacks neurons expressing the anti- often stabilize spontaneously (although at a level gens (Roberts & Darnell, 2004). There are many of severe neurological disability), so that it is dif- unanswered questions regarding exactly how this ficult to interpret reports of “neurological stabi- happens. Presently there is no fully successful lization” with immunotherapy animal model for any cell-mediated paraneoplastic 6. Patients often receive concomitant tumor treat- syndrome affecting the CNS. ment and immunotherapy, making it difficult The clinical neurological outcome of patients to discern the impact of each therapy on the with paraneoplastic syndromes varies considerably neurological outcome. For many syndromes among different disorders and among patients with immunotherapy is more likely to be effective a given disorder. With very few exceptions neurolog- when the tumor is also treated successfully. ical paraneoplastic syndromes do not remit spon- Factors that interact in influencing the response taneously. For several syndromes the neurological to immunotherapy include the neuroanatomical outcome is linked to the associated tumor type and site (central versus peripheral), the cellular location antineuronal antibody type. Successful treatment of of the onconeural target antigen(s) (neuronal cell the underlying tumor can bring about significant surface versus intracellular), and the proven or pre- neurological improvement, at least for some syn- sumed mechanism(s) of neuronal injury (antibody- dromes and for some but not all patients. Unfortu- mediated versus cell-mediated). In general, syn- nately, many patients are left with severe and per- dromes affecting the PNS are more likely to improve manent neurological disability despite response or with tumor treatment and/or immunosuppressive apparent cure of the associated tumor. treatment than are CNS syndromes. Lambert–Eaton If paraneoplastic disorders are autoimmune myasthenic syndrome and other syndromes caused diseases they should theoretically respond to by autoantibodies reacting with ion channels or 242 Section 2. Effects of cancer and cancer treatment on cognition

cell surface receptors are likely to respond to many if not most central syndromes it is likely that immunotherapy, probably because the antibodies patients have already suffered neuronal death or do not usually cause axonal degeneration or neu- irreversible injury by the time the diagnosis of a ronal cell death. paraneoplastic disorder is made. Unfortunately, the two most prevalent paraneo- There is theoretical concern that if paraneoplastic plastic CNS syndromes in adults, i.e., encephalo- disorders arise from an immune response directed myelitis/sensory neuronopathy associated with against the tumor, attempts to treat the neurological small cell lung cancer and cerebellar degeneration disorder with immunosuppression may adversely associated with breast or ovarian carcinoma, have affect the evolution of the tumor. At this time, there a poor prognosis. Fewer than 10% of these patients is no definite evidence that patients given immuno- show significant neurological improvement despite suppressive treatment have a worse tumor outcome aggressive tumor treatment and a variety of than those who are not (Keime-Guibert et al., 1999; immunosuppressive therapies (Dropcho, 2002, Rojas et al., 2000). 2005). Patients with other CNS syndromes includ- Two paraneoplastic disorders cause significant ing opsoclonus-myoclonus, limbic encephalitis, cognitive dysfunction in cancer patients and will or stiff person syndrome have a somewhat higher be discussed in detail. These are paraneoplastic likelihood of neurological improvement, suggesting limbic encephalitis associated with a variety of that the immune-mediated neuronal dysfunction neoplasms in adults, and the syndrome of para- or injury is less severe or of a sort more likely to be neoplastic opsoclonus-myoclonus occurring in reversible. children with neuroblastoma. Even for the “unfavorable” syndromes such as encephalomyelitis and cerebellar degeneration, there are a few patients who do show a meaning- Paraneoplastic limbic encephalitis ful neurological response to immunotherapy. For these few responders, the only factors that some- Limbic encephalitis as a clinicopathological entity times correlate with neurological improvement are was first described by Brierley in 1960 (Brierley et al., successful tumor treatment, and the duration and 1960), and its frequent association with neoplasia severity of neurological deficits prior to diagnosis was documented over the next several years (Corsel- and initiation of therapy. For patients who have lis et al., 1968; Henson et al., 1965). Paraneoplastic already stabilized at a plateau of severe neurological limbic encephalitis (PLE) may occur either as part of disability for more than several weeks, subsequent a multifocal encephalomyelitis, or less commonly as improvement with any intervention is not impos- an isolated clinicopathological syndrome. sible but extremely unlikely. The decision whether Approximately 50%–60% of reported patients to use immunosuppressive therapies must therefore with PLE have small cell lung carcinoma (Alamo- be based on the particular syndrome and on the witch et al., 1997; Gultekin et al., 2000; Lawn et al., individual patient’s circumstances. 2003). Other associated neoplasms include testicu- There are several potential explanations for the lar germ cell tumors (Dalmau et al., 2004; Rosen- disappointingly poor response to immunotherapy feld et al., 2001), thymoma (Ances et al., 2005; in many patients. The continuing presence of even Antoine et al., 1995; Fujii et al., 2001; Rickman a small tumor burden seems to provide an “anti- et al., 2000), Hodgkin’s lymphoma (Deodhare et al., genic drive” for further neuronal injury. It is also 1996; Duyckaerts et al., 1985), non-Hodgkin’s lym- likely that current immunotherapies do not ade- phoma (Mihara et al., 2005; Thuerl et al., 2003), non- quately gain access to the CNS, and do not effec- small cell lung cancer (Bakheit et al., 1990; Benke tively abrogate an ongoing autoimmune response et al., 2004), breast carcinoma (Fakhoury et al., that is “sequestered” in the CNS. Unfortunately, for 1999; Lawn et al., 2003; Sutton et al., 2000), ovarian Chapter 17. Paraneoplastic disorders 243

teratoma (Aydiner et al., 1998; Nokura et al., 1997; spheres, basal ganglia, diencephalon, brainstem, Taylor et al., 1999; Vitaliani et al., 2005), endomet- cerebellum, gray matter of the spinal cord, dorsal rial carcinoma (Petit et al., 1997), colon carcinoma root ganglia, and autonomic ganglia (Brierley et al., (Tsukamotoet al., 1993), renal carcinoma (Bell et al., 1960; Ingenito et al., 1990; Kinirons et al., 2003). The 1998; Kararizou et al., 2005), and prostate carci- pathological changes may be more widespread than noma (Modrego et al., 2002; Stern & Hulette, 1999). would have been predicted based on patients’ signs Because of the tumor associations, PLE occurs most and symptoms. commonly in middle-aged or older adults, although it may occur in adolescents or young adults when associated with Hodgkin’s lymphoma, thymoma, Clinical features ovarian teratoma, or testicular tumors (Lee et al., 2003; Okamura et al., 1997; Rosenbaum et al., 1998). In the majority of patients with PLE the neurological symptoms are the presenting feature of the associated neoplasm, often preceding discovery of Pathology the tumor by several months or longer. Excep- Patients with subacute neurological symptoms tional patients have been reported to develop lim- referable to the temporal lobes or limbic system may bic encephalitis after apparent cures of a previously have a brain biopsy to rule out herpes encephalitis diagnosed tumor (Kodama et al., 1991; Lacomis or other infection, and are subsequently diagnosed et al., 1990); in these patients the association with PLE. Biopsy specimens in these patients show between limbic encephalitis and the previous tumor a variable degree of non-specific changes including may be fortuitous. neuronal loss, astrogliosis, or perivascular and lep- Paraneoplastic limbic encephalitis generally has tomeningeal mononuclear cell infiltrates (Dalmau a subacute onset evolving over days to weeks. et al., 2004; Deodhare et al., 1996; Gultekin et al., Patients typically present either with an amnes- 2000; Ingenito et al., 1990; Rosenbaum et al., 1998). tic syndrome or with psychiatric symptoms; most The most consistent and severe neuropatholog- patients eventually develop features of both (Alam- ical abnormalities in autopsied cases of PLE are owitch et al., 1997; Gultekin et al., 2000; Lawn extensive neuronal loss, gliosis, and microglial nod- et al., 2003). The memory loss includes short-term ules in the hippocampus and amygdala (Brierley anterograde amnesia and a variable period of retro- et al., 1960; Corsellis et al., 1968; Dalmau et al., 2004; grade amnesia. Denial of the deficit and confabula- Gultekin et al., 2000). Similar but less severe changes tion are common. The psychiatric disorder usually are often present in the parahippocampal gyrus, includes some combination of depression, anxiety, cingulate gyrus, insular cortex, orbital frontal cor- emotional lability, and personality change. Hallu- tex, basal ganglia, and diencephalon. Perivascular cinations and paranoid delusions may occur. Gen- lymphocytic cuffing and leptomeningeal mononuc- eralized or partial complex seizures occur in most lear cell infiltrates are patchy and variable. In some patients, may be the initial neurological feature, and patients with clinically “pure” PLE the pathological can be medically intractable. changes at autopsy are entirely confined to the lim- Less common manifestations of limbic or dien- bic system (Bakheit et al., 1990; Duyckaerts et al., cephalic dysfunction include abnormal sleep-wake 1985; Farrugia et al., 2005; Fujii et al., 2001). Most cycles, disturbed temperature regulation, labile patients with PLE and small cell lung cancer, and blood pressure, inappropriate secretion of antidi- many patients with other associated tumors, have uretic hormone, and elements of the Kluver–Bucy multifocal encephalomyelitis, with patchy neuronal syndrome, such as hyperphagia and hypersexuality loss or inflammatory infiltrates in any or all areas (Aydiner et al., 1998; Dalmau et al., 2004; Overeem of the nervous system, including the cerebral hemi- et al., 2004; Rosenbaum et al., 1998). 244 Section 2. Effects of cancer and cancer treatment on cognition

In addition to the limbic encephalopathy, most At some time during the course of illness, the patients with small cell lung carcinoma and many cerebrospinal fluid (CSF) in about two-thirds of patients with other tumors develop manifestations patients with PLE shows a mild lymphocytic pleo- of a more generalized, multifocal paraneoplastic cytosis and/or slightly elevated protein (Alamowitch encephalomyelitis. These include varied combina- et al., 1997; Dalmau et al., 2004; Gultekin et al., 2000; tions of signs and symptoms referable to the extra- Lawn et al., 2003). Some patients additionally have limbic cerebral cortex, basal ganglia, brainstem, oligoclonal bands and/or an elevated CSF IgG index. cerebellum, dorsal root ganglia, spinal cord, and Normal CSF does not rule out PLE. autonomic system (Alamowitch et al., 1997; Gul- Approximately 75% of patients with PLE have an tekin et al., 2000; Hirayama et al., 2003; Kinirons abnormal electroencephalogram (EEG) during the et al., 2003; Lawn et al., 2003; Rickman et al., course of their illness (Gultekin et al., 2000; Lawn 2000). Patients with small cell lung carcinoma may et al., 2003). The most common EEG abnormality develop PLE concurrent with peripheral neuropa- is slowing, either diffuse or localized to the frontal thy or Lambert–Eaton myasthenic syndrome. Other or temporal regions. There may be superimposed extralimbic clinical features have particular associ- paroxysmal sharp waves and spikes with or without ations with certain tumors and antineuronal anti- clinical seizures. bodies (see below). Fluorodeoxyglucose positron emission tomography (PET scanning) may demonstrate unilateral or bilateral hippocampal hypermetabolism in patients Diagnostic studies with PLE (Ances et al., 2005; Fakhoury et al., 1999; Brain magnetic resonance imaging is abnormal in at Na et al., 2001; Provenzale et al., 1998; Scheid et al., least two-thirds of patients with PLE, showing areas 2004b). The hypermetabolic areas do not necessar- of abnormal T2-weighted and/or fluid-attenuated ily correspond to lesions seen on MR scans. The inversion recovery (FLAIR) signal in the mesial findings on PET scanning do not distinguish PLE temporal lobe and amygdala bilaterally, and less from other causes of limbic encephalitis. commonly in the hypothalamus and basal frontal cortex (Alamowitch et al., 1997; Gultekin et al., Autoimmunity 2000; Kodama et al., 1991; Lawn et al., 2003). The lesions enhance with gadolinium in a minority of Most but not all patients with PLE have one cases. Some patients additionally have extratempo- or more circulating antineuronal autoantibodies ral cortical or subcortical lesions (Ances et al., 2005; (Table 17.2). Several antineuronal antibodies asso- Hirayama et al., 2003; Lawn et al., 2003; Rickman ciated with PLE have pan-neuronal reactivity and et al., 2000; Rosenbaum et al., 1998). There is a single are present in patients with a variety of clinical neu- report of bilateral T2-weighted hyperintensity in the rological syndromes. The most common of these posterior thalamus (“pulvinar sign”) in a patient are polyclonal IgG anti-Hu antibodies found in with non-Hodgkin’s lymphoma (Mihara et al., 2005). patients with various clinical manifestations of mul- Patients with anti-Ma2 antibodies (see below) may tifocal encephalomyelitis, reflecting involvement of have MR lesions in the thalamus, hypothalamus, the cerebral hemispheres, limbic system, cerebel- and/or brainstem in addition to the mesial temporal lum, brainstem, spinal cord, dorsal root ganglia, lobes (Bennett et al., 1999; Dalmau et al., 2004). In and autonomic ganglia (Graus et al., 2001; Gul- some patients the MR lesions subsequently resolve tekin et al., 2000; Sillevis Smitt et al., 2002; Vernino with or without concomitant clinical improvement, et al., 2002). More than 90% of patients with anti- sometimes eventuating in temporal lobe atrophy Hu antibodies and paraneoplastic encephalomyel- (Benke et al., 2004; Dirr et al., 1990; Kodama et al., itis have small cell lung carcinoma, with reports of 1991; Rosenbaum et al., 1998). other tumors including non-small cell lung cancer, Chapter 17. Paraneoplastic disorders 245

Table 17.2. Antineuronal antibodies in paraneoplastic limbic encephalitis

Autoantibody Associated tumor(s) Antibody reactivity anti-Hu (ANNA-1) Small cell lung carcinoma, others Pan-neuronal nuclear > cytoplasmic (Alamowitch et al., 1997; Graus et al., 2001; staining; 35- to 40-kDa Gultekin et al., 2000; Sillevis Smitt et al., 2002) RNA-binding proteins Anti-CV2 (CRMP-5) Small cell lung carcinoma, Cytoplasm of neurons; 66-kDa CV2 (Kinirons et al., 2003; Lawn et al., 2003; Rickman thymoma, others protein, 62-kDa CRMP-5 protein et al., 2000; Yu et al., 2001) Anti-Ma2 (anti-Ta) Germ cell tumors, breast Pan-neuronal nuclei and nucleoli; (Dalmau et al., 2004; Rosenfeld et al., 2001; carcinoma, non-small cell lung 40- to 42-kDa Ma1 and Ma2 Voltz et al., 1999) carcinoma proteins Anti-VGKC Thymoma, small cell lung Voltage-gated potassium channels (Buckley et al., 2001; Pozo-Rosich et al., 2003; carcinoma Vernino & Lennon, 2004) Novel neuropil antibodies Ovarian teratoma, others Hippocampal dendrites and synapses (Ances et al., 2005; Vitaliani et al., 2005) Anti-amphiphysin Breast, small cell lung carcinoma Neuropil; 125-kDa synaptic (Antoine et al., 1999; Dorresteijn et al., 2002; vesicle-associated protein Pittock et al., 2005) ANNA-3 Small cell lung carcinoma Nuclei of Purkinje cells and dentate (Chan et al., 2001) neurons; 170-kDa protein Anti-VGCC Small cell lung carcinoma, others P/Q-type and N-type calcium (Lawn et al., 2003) channels PCA-2 Small cell lung carcinoma Neuronal cytoplasm; 280-kDa protein (Vernino & Lennon, 2000) Anti-Zic Small cell lung carcinoma 35- to 55-kDa zinc ﬁnger proteins (Bataller et al., 2004) “Atypical” Thymoma, small cell lung Varied (Antoine et al., 1995; Fujii et al., 2001; Scheid carcinoma, others et al., 2004a, b; Tsukamoto et al., 1993) Antibody-negative Various

neuroblastoma, carcinoma of the breast or prostate, et al., 1997; Gultekin et al., 2000). Anti-Hu-positive or thymoma. patients usually show additional signs and symp- Limbic encephalitis is an early and prominent toms of multifocal paraneoplastic encephalomyel- feature in 10%–20% of patients with paraneoplas- itis, whereas patients with small cell lung cancer but tic encephalomyelitis and anti-Hu antibodies; most without anti-Hu antibodies are more likely to have of these patients develop other multifocal signs and “pure” limbic system involvement. symptoms during the course of their illness (Dal- Anti-Hu antibodies are a valuable clinical marker mau et al., 1992; Graus et al., 2001; Gultekin et al., for PLE or paraneoplastic encephalomyelitis, but it 2000; Sillevis Smitt et al., 2002). Among patients is currently thought that cellular immune effectors with PLE and small cell lung carcinoma, approx- and not the anti-Hu antibodies are the mediators of imately one-half have anti-Hu antibodies, a few neuronal injury. Evidence to support cell-mediated patients have other antibodies, and the remain- autoimmune neuronal injury includes the presence der have no identiﬁable antibodies (Alamowitch of CD8+ T lymphocytes clustered around neurons 246 Section 2. Effects of cancer and cancer treatment on cognition

in the brain and dorsal root ganglia (Bernal et al., The mechanisms of autoimmune neuronal injury 2002), and the presence of oligoclonal T lympho- in anti-Ma2-associated PLE are not known. Intersti- cytes in the blood and dorsal root ganglia (Plon- tial and perivascular infiltration of T lymphocytes quet et al., 2002). T lymphocytes from patients’ in affected brain areas suggests cellular immune peripheral blood recognize and respond to peptides effectors (Dalmau et al., 2004). In one attempted derived from the HuD onconeural antigen (Plon- animal model, adoptive transfer of rat T lympho- quet et al., 2003; Rousseau et al., 2005). cytes specific for the Ma1 onconeural protein Several other antineuronal antibodies associated caused meningeal and perivascular inflammatory with PLE or with other CNS syndromes have a pan- infiltrates in recipient rats, but the recipient ani- neuronal or widespread reactivity. These include mals did not develop neuronal loss or clinical dis- anti-CV2 (CRMP-5) antibodies, anti-amphiphysin ease (Pelkofer et al., 2004). antibodies, PCA-2 antibodies, anti-voltage-gated Some patients with limbic encephalitis have cir- calcium channel antibodies, and ANNA-3 anti- culating antibodies against voltage-gated potas- bodies (Table 17.2). Any of these antibodies can be sium channels (VGKC). To date, most of the small present in patients with PLE, but none of them has number of reported patients with limbic encephal- a particular association with PLE versus other neu- itis and anti-VGKC antibodies do not have an iden- rological syndromes. Small cell lung cancer is by far tifiable neoplasm (Thieben et al., 2004; Vincent the tumor most commonly associated with these et al., 2004). In some patients the limbic encephal- antibodies. It is not unusual for patients with small itis occurs as a paraneoplastic syndrome, usually cell lung carcinoma and PLE (or other CNS syn- in association with thymoma (Buckley et al., 2001; dromes) to have more than one type of antineuronal Vernino & Lennon, 2004) or small cell lung car- antibody (Pittock et al., 2004). At this time it is not cinoma (Pozo-Rosich et al., 2003). Voltage-gated known whether any of these antibodies is directly potassium channels comprise hetero-oligomers of involved in causing neuronal injury. different subunits. Subtypes of VGKCs are widely A few antineuronal antibodies have a specific distributed throughout the brain and PNS. Anti- linkage to PLE and are not commonly associated VGKC antibodies are also found in patients with with other neurological syndromes. Anti-Ma2 (anti- paraneoplastic or non-paraneoplastic neuromyo- Ta) antibodies mainly occur in young men with tes- tonia (Hart et al., 2002), and in patients with ticular germ cell tumors (Dalmau et al., 2004; Rosen- the syndrome of “Morvan’s fibrillary chorea” fea- feld et al., 2001; Voltz et al., 1999). There are a turing neuromyotonia, hyperhidrosis and other few reported cases of anti-Ma2 (anti-Ta) antibodies dysautonomia, insomnia, hallucinations, and lim- in women with breast carcinoma or non-small cell bic encephalopathy (Lee et al., 1998; Liguori et al., lung carcinoma (Sahashi et al., 2003; Sutton et al., 2001). Unlike many other paraneoplastic antineu- 1993). Some patients with anti-Ma2 antibodies have ronal antibodies that are believed to be markers a clinically “pure” limbic encephalitis, while the of autoimmunity but do not directly mediate neu- majority present with a combined syndrome reflect- ronal injury, anti-VGKC antibodies may directly ing involvement of the limbic system, diencephalon cause neuronal dysfunction. In an experimental (e.g., sleep disorder or autonomic dysfunction), and model of neuromyotonia, patients’ anti-VGKC anti- brainstem (especially ocular motor disturbance) bodies were shown to cross-link the receptors (Bennett et al., 1999; Dalmau et al., 2004; Overeem and reduce potassium channel currents (Tomim- et al., 2004; Waragi et al., 2006). Patients whose itsu et al., 2004). Less is known about the effects antibodies react with the Ma1 protein in addi- of anti-VGKC antibodies on CNS neuronal function to the Ma2 protein tend to have more severe tion. Presumably, differences in the fine specificity cerebellar and brainstem dysfunction (Dalmau of reactivity of anti-VGKC antibodies account for the et al., 2004). heterogeneous clinical presentation among patients Chapter 17. Paraneoplastic disorders 247

with limbic encephalitis, neuromyotonia, and Mor- tumors (Behin & Delattre, 2004). This condition gen- van’s syndrome (Kleopa et al., 2006). erally presents with global cognitive dysfunction The newest autoantibodies associated with PLE and gait apraxia, rather than the selective memory are “novel neuropil antibodies,” which stain the loss seen in prototypic cases of limbic encephalitis, dendritic network and synaptic-enriched regions in but there is some overlap. Limbic encephalitis asso- the neuropil of the hippocampus (Ances et al., 2005; ciated with human Herpes virus type 6 has occurred Vitaliani et al., 2005). Associated neoplasms include in patients following allogeneic stem cell transplan- ovarian teratoma and thymic tumors. Some patients tation (Ogata et al., 2006; Wainwright et al., 2001). present with a “typical” limbic encephalopathy, Varicella zoster virus may also cause a selective lim- while others have a more severe clinical course bic encephalitis in immunocompromised patients with acute psychosis, seizures, lethargy, and cen- (Tattevin et al., 2001). tral hypoventilation requiring extended ventilatory For patients without a previous cancer diagno- support. The latter patients usually have evidence sis who present with limbic encephalitis, the level for multifocal extralimbic involvement based on MR of suspicion for a paraneoplastic etiology depends imaging, PET scans, or autopsy. on the patient’s age, gender, risk factors (especially Several reported patients with PLE have “atyp- cigarette smoking), and the presence of antineu- ical” or incompletely characterized antineuronal ronal antibodies. The most common alternative antibodies. Most reports are of single patients, with diagnoses are Herpes simplex encephalitis, a pri- one of a variety of tumors including thymoma, mary psychiatric disorder, or non-paraneoplastic small cell lung cancer, and breast carcinoma (Fujii limbic encephalitis. Patients with PLE who have et al., 2001; Scheid et al., 2004a). It is important to early and prominent affective symptoms or halluci- keep in mind that some patients with PLE, regard- nations are often initially diagnosed as having a pri- less of tumor association, do not have identiﬁable mary psychiatric condition, especially when accom- autoantibodies. panying “hard” neurological ﬁndings are absent, missed, or misinterpreted. There is increasing recognition of patients whose Differential diagnosis clinical presentations are indistinguishable from Differential diagnosis in patients with suspected those of PLE, but in whom no tumor is ever dis- PLE partly depends on whether there is a known covered, even at autopsy (Bien et al., 2000; Kohler cancer diagnosis and on the tumor histology. et al., 1988; Mori et al., 2002). Some patients with Among patients with a prior cancer diagnosis who non-paraneoplastic limbic encephalitis have anti- develop cognitive dysfunction, the level of suspi- VGKC antibodies (Ances et al., 2005; Buckley et al., cion for a paraneoplastic disorder is much higher 2001; Fauser et al., 2005; Pozo-Rosich et al., 2003; for patients with small cell lung carcinoma, thy- Thieben et al., 2004; Vincent et al., 2004). Patients moma, Hodgkin’s lymphoma, and testicular germ with anti-VGKC antibodies often improve with cell tumors than for patients with other tumors. plasma exchange, intravenous immunoglobulin, or Tumor metastases and neurotoxicity of cancer treat- corticosteroids. There are reports of a few patients ments are far more common than paraneoplas- with limbic encephalitis and anti-Hu, anti-Ma2, tic disorders and should always be considered, as or other antineuronal antibodies in whom no should metabolic derangements and CNS infec- tumor was ever discovered (Ances et al., 2005; tion. Methotrexate, procarbazine, ifosfamide, and Dalmau et al., 2004; Gultekin et al., 2000). No other chemotherapeutic drugs can cause a diffuse clinical features or laboratory studies (including or multifocal encephalopathy (Dropcho, 2004). Dif- CSF, EEG, MR imaging, or PET) reliably distinguish fuse cerebral injury may occur following cranial paraneoplastic from non-paraneoplastic limbic radiation therapy for primary or metastatic brain encephalitis. 248 Section 2. Effects of cancer and cancer treatment on cognition

Patient management The course of PLE is variable and rather unpredictable. A few patients with clinically “pure” PLE Any middle-aged patient with a history of cigarette show spontaneous remission of the neurological smoking who develops limbic encephalitis should condition prior to any treatment (Sillevis Smitt be suspected of harboring a small cell lung carci- et al., 2002). Paraneoplastic limbic encephalitis is noma. Chest CT or MR scanning is clearly more sen- rather unusual among CNS paraneoplastic disor- sitive than a “plain” chest X-ray in detecting a small ders in that a significant proportion of patients neoplasm. If present, anti-Hu, anti-CV2, or other have major neurological improvement after suc- serum antineuronal antibodies (Table 17.2) are a cessful treatment of the associated tumor. Approx- highly specific marker for small cell lung carcinoma imately one-half of patients with PLE and small (rarely another tumor). Total body PET scanning cell lung cancer improve after tumor treatment may detect lung or other neoplasms in patients who (Alamowitch et al., 1997; Bak et al., 2001; Dal- are suspected of having paraneoplastic syndromes mau et al., 1992; Dorresteijn et al., 2002; Gul- and yet have unrevealing or equivocal chest CT or tekin et al., 2000; Kaniecki & Morris, 1993). Patients MR scans (Linke et al., 2004; Younes-Mhenni et al., with small cell lung cancer but without antineu- 2004). If a patient’s initial evaluation for an occult ronal antibodies are more likely to improve than tumor is unrevealing, which is not at all uncom- those with anti-Hu antibodies. Among patients with mon,theworkupshouldberepeatedatregular anti-Hu antibodies in whom limbic encephalitis is intervals. a component of multifocal encephalomyelitis, the In young adults or non-smokers presenting with “limbic” features may improve after tumor treat- limbic encephalitis, the most common neoplasms ment, whereas the other neurological features rarely to consider are thymoma, Hodgkin’s lymphoma, do so. testicular germ cell tumor, and ovarian teratoma. Among patients with PLE, testicular germ cell These patients should have a thorough physical tumors, and anti-Ma2 antibodies who receive examination and CT or MR scanning of the chest tumor treatment and/or immunosuppressive ther- and abdomen. Anti-VGKC antibodies should raise apy, approximately 25% have neurological improve- suspicion for an associated thymoma, although ment and about another 25% have neurological sta- many if not most patients with limbic encephalitis bilization (Dalmau et al., 2004; Landolfi & Nadkarni, and anti-VGKC antibodies do not have an under- 2003). Successful tumor treatment is correlated with lying tumor. Young men should also have testic- a better neurological outcome. Some patients have ular ultrasound, which can show a small tumor improved memory and cognition but continue to even after negative clinical examinations (Winger- have chronic intractable seizures. chuk et al., 1998). Elevated serum alpha-fetoprotein Approximately one-half of reported patients with or human chorionic gonadotropin in young men PLE and thymoma, including those with anti-VGKC may indicate a non-seminomatous germ cell tumor. or anti-CV2 antibodies, have significant neurologi- Serum anti-Ma2 antibodies are a marker for testicu- cal improvement following successful tumor treat- lar germ cell tumors, although a negative assay does ment, with or without immunotherapy (Ances et al., not rule out a tumor (Dalmau et al., 2004). There 2005; Antoine et al., 1995; Buckley et al., 2001; Gul- are reports of young men with brainstem or limbic tekin et al., 2000). There are also reports of partial encephalitis, anti-Ma2 antibodies, and negative or or complete reversal of PLE after treatment of the equivocal testicular ultrasound, in whom orchiec- underlying Hodgkin’s lymphoma (Deodhare et al., tomy revealed a microscopic intratubular germ cell 1996), ovarian teratoma (Ances et al., 2005; Nokura tumor (Dalmau et al., 2004). Young women should et al., 1997; Vitaliani et al., 2005), renal carcinoma additionally have a pelvic examination and imaging (Bell et al., 1998), or ovarian carcinoma (Bloch et al., to look for an ovarian teratoma. 2004). Chapter 17. Paraneoplastic disorders 249

As outlined above, the responsiveness of PLE is probably an overestimate of the true frequency to immunosuppressive therapy is difficult to of paraneoplastic opsoclonus (POM) due to report- judge precisely from the literature. Many reported ing bias; POM occurs in approximately 2%–3% of patients received tumor treatment concurrently children with neuroblastoma (Gambini et al., 2003; with immunosuppressive therapy including oral Rudnick et al., 2001). The median age at onset of or intravenous pulse corticosteroids, cyclophos- POM is 18–24 months. In nearly all cases it is the phamide, intravenous immunoglobulin, or plasma- neurological syndrome that leads to discovery of pheresis (Alamowitch et al., 1997; Gultekin et al., an otherwise occult neuroblastoma. Paraneoplas- 2000; Vitaliani et al., 2005). There are reports of def- tic opsoclonus can rarely occur months to years inite responses to immunotherapy. These include after successful neuroblastoma treatment, without patients with thymoma (and anti-VGKC or anti-CV2 evidence of tumor recurrence. antibodies) (Buckley et al., 2001; Rickman et al., Paraneoplastic opsoclonus-myoclonus typically 2000), testicular tumors (anti-Ma2 antibodies) has an abrupt onset. The cardinal feature, opso- (Dalmau et al., 2004; Scheid et al., 2003), small cell clonus, is continuous multidirectional rapid eye lung cancer (anti-VGKC antibodies) (Pozo-Rosich movements (saccadic oscillations) without an et al., 2003), patients with ovarian teratoma (Lee intersaccadic interval. In addition to opsoclonus, et al., 2003; Taylor et al., 1999), and patients with children have some combination of myoclonus, novel neuropil antibodies (Ances et al., 2005), ataxia, and altered sensorium (Gambini et al., who improved following some combination of 2003; Rudnick et al., 2001; Telander et al., 1989). corticosteroids, intravenous immunoglobulin, or Myoclonic jerks are arrhythmic, multifocal, spon- plasmapheresis. taneous and/or stimulus triggered, and vary in Roughly one-half of patients with PLE fail to severity from mild to incapacitating. Ataxia may improve with tumor treatment, with or without involve limbs, trunk, and gait, and again may be immunosuppressive therapy, regardless of tumor mild to severe. Nearly all children are irritable in the association (Dalmau et al., 2004; Gultekin et al., acute phase. Other signs and symptoms can include 2000). These patients are usually left with moder- nausea and vomiting, dysarthria, facial diplegia, ate or severe neurological disability. Less commonly, hearing loss, and upper motor neuron findings. A patients may become progressively demented with high percentage of children have residual long-term eventual obtundation and fatal coma. motor, neurocognitive, and behavioral problems (see below). The majority of children with POM have mild elevation of CSF protein, lymphocytic pleocytosis, Paraneoplastic opsoclonus-myoclonus oligoclonal IgG bands, and an elevated IgG index. Brain MR scans are usually normal; there are indi- Clinical features vidual case reports of cerebellar vermal lesions The syndrome of “myoclonic encephalopathy of in the acute phase (Telander et al., 1989) or of infancy” or opsoclonus-myoclonus was first clearly eventual diffuse cerebellar atrophy (Hayward et al., described by Kinsbourne in 1962 (Kinsbourne, 2001). The main differential diagnosis of POM is 1962). The association between opsoclonus- opsoclonus-myoclonus occurring during an acute myoclonus and neuroblastoma in children was viral infection (e.g., enterovirus or Coxsackie virus) identified in two publications in 1968 (Dyken & (Kuban et al., 1983; Tabarki et al., 1998) or as a Kolar, 1968; Solomon & Chutorian, 1968). Since that syndrome following infection with agents including time approximately one-half of the reported cases Epstein–Barr virus (Sheth et al., 1995) or Streptococ- of opsoclonus-myoclonus syndrome in children cus (Candler et al., 2006). There are no clinical, neu- occurred in association with neuroblastoma. This roimaging, or CSF findings that reliably differentiate 250 Section 2. Effects of cancer and cancer treatment on cognition

POM from “post-infectious” or idiopathic neoplasms (Konczak et al., 2005; Ravizza et al., 2006; opsoclonus-myoclonus syndrome in children Ronning et al., 2005). or adults (Boltshauser et al., 1979; Pohl et al., 1996). Autoimmunity

The pathological and biological features of Pathology neuroblastomas in children with POM are gen- The pathological substrate of childhood POM erally favorable. The tumors of a disproportionately remains unclear, as there are no distinctive or uni- high percentage of children with POM are classified formly present lesions. Some of the very few pub- as ganglioneuroblastoma and are in a favorable lished autopsied cases show partial loss of cere- histological group. Patients with POM are less likely bellar Purkinje cells (Moe & Nellhaus, 1970; Ziter to have advanced-stage neuroblastoma at diagnosis et al., 1979). Among adult patients with paraneo- compared to neuroblastoma patients in general plastic opsoclonus associated with small cell lung (Rudnick et al., 2001; Russo et al., 1997). Perivascu- cancer, breast carcinoma or other tumors, autop- lar and interstitial infiltrates of B lymphocytes and T sies have shown mild to severe dropout of Purkinje lymphocytes are more frequent and more intense in cells, and/or patchy neuronal loss and perivascular POM patients than in neuroblastoma patients with- mononuclear cell infiltrates in the inferior olivary out POM (Cooper et al., 2001; Gambini et al., 2003; nuclei and other areas of the brainstem (Ander- Mitchell & Snodgrass, 1990; Telander et al., 1989). son et al., 1988; Wong et al., 2001). In a signifi- Amplification of the N-myc oncogene is relatively cant proportion of autopsied children and adults rare (Gambini et al., 2003). The occurrence of POM with POM there are no identifiable histopathologi- in a child with neuroblastoma generally carries a cal abnormalities in either the cerebellum or brain- good prognosis for survival independent of patient’s stem (Hersh et al., 1994; Ridley et al., 1987). Dis- age, tumor site, or tumor stage (Rudnick et al., 2001; turbance of the tonic inhibitory control of saccadic Russo et al., 1997), although a good oncological out- burst neurons by the “omnipause” neurons in the come is not universal (Hiyama et al., 1994). These pontine reticular formation has been postulated observations indirectly support the theory that as the key pathophysiological event in producing POM occurs when an anti-neuroblastoma immune opsoclonus, but the pons may not show any his- response causes tumor regression or differentiation, tological changes (Ridley et al., 1987). An alterna- but simultaneously attacks cross-reacting neuronal tive model postulates that opsoclonus is the result antigens. It has been postulated that some cases of cerebellar Purkinje cell injury, which disinhibits of “idiopathic” or “post-infectious” opsoclonus- the fastigial oculomotor region (Wong et al., 2001). myoclonus represent instances of an occult neu- One of the many unanswered questions in para- roblastoma that is obliterated by a cross-reacting neoplastic childhood POM is why affected children antitumor/antineuronal immune response. develop cognitive and behavioral problems if the Some patients with POM have serum auto- autoimmune response is seemingly directed against antibodies that react with shared neuronal- the brainstem and cerebellum. One possible expla- neuroblastoma antigens. A small percentage of nation is that POM is actually a diffuse or multifocal patients have anti-Hu antibodies (Fisher et al., encephalitis that also involves supratentorial struc- 1994; Hayward et al., 2001; Korfei et al., 2005). Other tures. Another possibility is that cerebellar injury patients have one of a number of antibodies with is the source of the neurocognitive deficits in chil- heterogeneous reactivities on immunocytochem- dren with POM; this derives from recent studies of ical staining and immunoblots (Antunes et al., neurocognitive sequelae in children with cerebellar 2000; Blaes et al., 2005; Connolly et al., 1997; Korfei Chapter 17. Paraneoplastic disorders 251

et al., 2005). The identity of the onconeural anti- produces rapid and dramatic neurological improve- gens remains to be shown. There is no universally ment in at least two-thirds of children (Hammer observed antibody or antigen in published studies. et al., 1995). Oral or intravenous (IV) corticos- Some of the antibodies are also present in children teroids are also used but are probably less effective with non-paraneoplastic opsoclonus-myoclonus, than ACTH (Hammer et al., 1995; Rostasy et al., and in children with neuroblastoma but no neu- 2005). Patients may also show improvement with rological symptoms. In one study IgG antibodies IV Ig, either given alone or in combination with from POM patients exerted an antiproliferative and ACTH/corticosteroids (Fisher et al., 1994; Petruzzi cytotoxic effect on neuroblastoma cells in vitro & Alarcon, 1995; Rudnick et al., 2001; Veneselli (Korfei et al., 2005). Other evidence supporting et al., 1998). Plasmapheresis (Yiu et al., 2001) or an autoimmune etiology for POM includes an rituximab (Pranzatelli et al., 2005d; Tersak et al., abnormally increased number of B lymphocytes 2005) may be effective in patients refractory to other and T lymphocyte subsets in the CSF (Pranzatelli therapies. In one study, the addition of intravenous et al., 2004b), and elevated CSF neopterin, which is cyclophosphamide to ACTH/IV Ig did not improve a marker for cellular immune activation (Pranzatelli the short-term or long-term neurological outcome et al., 2004a). These CSF abnormalities do not for patients over that associated with ACTH/IV distinguish between patients with paraneoplastic Ig alone (Pranzatelli et al., 2005c). In one small or non-paraneoplastic opsoclonus-myoclonus. series chronic mycophenolate allowed reduced To date there is no successful animal model for corticosteroid doses and produced a reduction in POM, and the actual immunopathogenetic mechani- activated T lymphocytes and T cell cytokines in the sm(s) of neuronal injury or dysfunction are unclear. CSF (Pranzatelli et al., 2005b). At least one-half of children with POM have a protracted or ﬂuctuating course. Exacerbations of Patient management neurological symptoms may occur when ACTH or All children who present with opsoclonus should corticosteroids are tapered or discontinued, or dur- have a workup for neuroblastoma including chest ing febrile illnesses (Hammer et al., 1995; Mitchell radiograph, abdominal CT scan, and a 24-h urine & Snodgrass, 1990; Mitchell et al., 2005; Telander collection for vanillylmandelic acid, homovanillic et al., 1989). Unfortunately, at least two-thirds of acid, and metanephrine (Telander et al., 1989). children are left with some combination of resid- Nuclear medicine imaging with the norepinephrine ual motor deﬁcits, speech delay, learning disability, analog metaiodobenzylguanidine (MIBG) may impulsive or aggressive behavior, and sleep distur- demonstrate a tumor in the absence of a radio- bance (Hayward et al., 2001; Mitchell et al., 2002; graphic lesion (Parisi et al., 1993; Shapiro et al., Papero et al., 1995; Pranzatelli et al., 2005a). An ini- 1994). If initially unrevealing, the workup should tially good neurological response to tumor treat- be repeated at regular intervals. In at least 75% ment and/or immunotherapy does not necessar- of children with POM the neuroblastoma is diag- ily predict a better long-term neurological outcome nosed within 6 months after onset of neurological (Hammer et al., 1995; Hayward et al., 2001; Rud- symptoms (Rudnick et al., 2001; Russo et al., 1997). nick et al., 2001; Russo et al., 1997). In one retro- The majority of children with POM have a local- spective study the minority of patients with POM ized neuroblastoma and undergo tumor resection. who received chemotherapy had a more favorable Some of these children show post-operative neuro- long-term neurological outcome (Russo et al., 1997), logical improvement without any other treatment but this was not seen in other series. A long-term (Hayward et al., 2001). Whether given before or after longitundinal neurodevelopmental study showed surgical resection, adrenocorticotropin (ACTH) that patients who had a monophasic course had a 252 Section 2. Effects of cancer and cancer treatment on cognition

better long-term outcome than those with exacer- Benke T, Wagner M, Pallua AK et al. (2004). Long-term cog- bations or relapses (Mitchell et al., 2005). nitive and MRI findings in a patient with paraneoplastic limbic encephalitis. J Neurooncol 66: 217–224. Bennett JL, Galetta SL, Frohman LP et al. (1999). Neuro- REFERENCES ophthalmologic manifestations of a paraneoplastic syndrome and testicular carcinoma. Neurology 52: 864–867. Alamowitch S, Graus F, Uchuya M et al. (1997). Limbic Bernal F, Graus F, Pifarre A et al. (2002). Immunohisto- encephalitis and small cell lung cancer: clinical and chemical analysis of anti-Hu-associated paraneoplastic immunological features. Brain 120: 923–928. encephalomyelitis. Acta Neuropathol 103: 509–515. Ances BM, Vitaliani R, Taylor RA et al. (2005). Treatment- Bien CG, Schulze-Bonjage A, Deckert M et al. (2000). Lim- responsive limbic encephalitis identified by neuropil bic encephalitis not associated with neoplasm as a cause antibodies: MRI and PET correlates. Brain 128: 1764– of temporal lobe epilepsy. Neurology 55: 1823–1828. 1777. BlaesF,FuhlhuberV,KorfeiMet al. (2005). Surface- Anderson NE, Budde-Steffen C, Rosenblum MK binding autoantibodies to cerebellar neurons in opso- et al. (1988). Opsoclonus, myoclonus, ataxia, and clonus syndrome. Ann Neurol 58: 313–317. encephalopathy in adults with cancer: a distinct para- Bloch MH, Hwang WC, Baehring JM, Chambers SK (2004). neoplastic syndrome. Medicine 67: 100–109. Paraneoplastic limbic encephalitis: ovarian cancer pre- Antoine JC, Honnorat J, Anterion CT et al. (1995). Limbic senting as an amnestic syndrome. Obstet Gynecol 104: encephalitis and immunological perturbations in two 1174–1177. patients with thymoma. J Neurol Neurosurg Psychiatry Boltshauser E, Deonna T, Hirt HR (1979). Myoclonic 58: 706–710. encephalopathy of infants, or “dancing eyes syndrome.” Antoine JC, Absi L, Honnorat J et al. (1999). Anti- Helv Paediatr Acta 34: 119–133. amphiphysin antibodies are associated with various Brierley JB, Corsellis JA, Hierons R, Nevin S (1960). Sub- paraneoplastic neurological syndromes and tumors. acute encephalitis of later adult life mainly affecting the Arch Neurol 56: 172–177. limbic areas. Brain 83: 357–368. Antunes NL, Khakoo Y, Matthay KK et al. (2000). Antineu- Buckley C, Oger J, Clover L et al. (2001). Potassium chan- ronal antibodies in patients with neuroblastoma and nel antibodies in two patients with reversible limbic paraneoplastic opsoclonus-myoclonus. J Pediatr Hema- encephalitis. Ann Neurol 50: 73–78. tol Oncol 22: 315–320. Candler PM, Dale RC, Griffin S et al. (2006). Post- Aydiner A, Gurvit H, Baral I (1998). Paraneoplastic limbic streptococcal opsoclonus-myoclonus syndrome associ- encephalitis with immature ovarian teratoma. J Neuro- ated with anti-neuroleukin antibodies. J Neurol Neuro- oncol 37: 63–66. surg Psychiatry 77: 507–512. Bak TH, Antoun N, Balan KK, Hodges JR (2001). Mem- Chan KH, Vernino S, Lennon VA (2001). ANNA-3 anti- ory lost, memory regained: neuropsychological findings neuronal nuclear antibody: marker of lung cancer- and neuroimaging in two cases of paraneoplastic limbic related autoimmunity. Ann Neurol 50: 301–311. encephalitis with radically different outcomes. J Neurol Connolly AM, Pestronk A, Mehta S et al. (1997). Serum Neurosurg Psychiatry 71: 40–47. autoantibodies in childhood opsoclonus-myoclonus Bakheit AM, Kennedy PG, Behan PO (1990). Paraneoplas- syndrome: an analysis of antigenic targets in neural tis- tic limbic encephalitis: clinicopathologic correlations. sues. J Pediatr 130: 878–884. J Neurol Neurosurg Psychiatry 53: 1084–1088. Cooper R, Khakoo Y, Matthay KK et al. (2001). Opsoclonus- Bataller L, Wade DF, Graus F et al. (2004). Antibodies to Zic4 myoclonus-ataxia syndrome in neuroblastoma: histo- in paraneoplastic neurological disorders and small cell pathologic features. Med Pediatr Oncol 36: 623–629. lung cancer. Neurology 62: 778–782. Corsellis JA, Goldberg GJ, Norton AR (1968). “Limbic Behin A, Delattre JY (2004). Complications of radiation encephalitis” and its association with carcinoma. Brain therapy on the brain and spinal cord. Semin Neurol 24(4): 91: 481–496. 405–417. Dalmau J, Graus F, Rosenblum MK, Posner JB (1992). Bell BB, Tognoni PG, Bihrle R (1998). Limbic encephalitis as Anti-Hu-associated paraneoplastic encephalomyeli- a paraneoplastic manifestation of renal cell carcinoma. tis/sensory neuronopathy: a clinical study of 71 patients. JUrol160: 828. Medicine 71: 59–72. Chapter 17. Paraneoplastic disorders 253

Dalmau J, Graus F, Villarejo A et al. (2004). Clinical analy- Graus F, Delattre JY, Antoine JC et al. (2004). Recom- sis of anti-Ma2-associated encephalitis. Brain 127: 1831– mended diagnostic criteria for paraneoplastic neurolog- 1844. ical syndromes. J Neurol Neurosurg Psychiatry 75: 1135– Deodhare S, O’Connor P, Ghazarian D, Bilbao JM (1996). 1140. Paraneoplastic limbic encephalitis in Hodgkin disease. Gultekin SH, Rosenfeld MR, Voltz R et al. (2000). Para- Can J Neurol Sci 23: 138–140. neoplastic limbic encephalitis: neurological symptoms, Dirr LY, Elster AD, Donofrio PD, Smith M (1990). Evolution immunological ﬁndings and tumour association in 50 of brain MRI abnormalities in limbic encephalitis. Neu- patients. Brain 123: 1481–1494. rology 40: 1304–1306. Hammer MS, Larsen MB, Stack CV (1995). Outcome of chil- Dorresteijn LD, Kappelle AC, Renier WO, Gijtenbeek JM dren with opsoclonus-myoclonus regardless of etiology. (2002). Anti-amphiphysin associated limbic encephali- Pediatr Neurol 13: 21–24. tis: a paraneoplastic presentation of small-cell lung car- Hart IK, Maddison P, Newsom-Davis J et al. (2002). Pheno- cinoma. J Neurol 249: 1307–1308. typic variants of autoimmune peripheral nerve hyperex- Dropcho EJ (2002). Remote neurologic manifestations of citability. Brain 125: 1887–1895. cancer. Neurol Clin 20 (1): 85–122. Hayward K, Jeremy RJ, Jenkins S et al. (2001). Long-term Dropcho EJ (2004). Neurotoxicity of cancer chemotherapy. neurobehavioral outcome in children with neuroblas- Semin Neurol 24(4): 419–426. toma and opsoclonus-myoclonus-ataxia syndrome: rela- Dropcho EJ (2005). Immunotherapy for paraneoplastic tionship to MRI ﬁndings and anti-neuronal antibodies. neurological disorders. Expert Opin Biol Ther 5(10): J Pediatr 139: 552–559. 1339–1348. Henson RA, Hoffman HL, Urich H (1965). Encephalomyeli- Duyckaerts C, Derouesne C, Signoret JL et al. (1985). Bilat- tis with carcinoma. Brain 88: 449–464. eral and limited amygdalohippocampal lesions causing Hersh B, Dalmau J, Dangond F et al. (1994). Paraneoplas- a pure amnestic syndrome. Ann Neurol 18: 314–319. tic opsoclonus-myoclonus associated with anti-Hu anti- Dyken P, Kolar O (1968). Dancing eyes, dancing feet: infan- body. Neurology 44: 1754–1755. tile polymyoclonus. Brain 91: 305–320. Hirayama K, Taguchi Y, Sato M, Tsukamoto T (2003). Lim- Fakhoury T, Abou-Khalil B, Kessler RM (1999). Limbic bic encephalitis presenting with topographical disorien- encephalitis and hyperactive foci on PET scan. Seizure 8: tation and amnesia. J Neurol Neurosurg Psychiatry 74: 427–430. 110–112. Farrugia ME, Conway R, Simpson DJ, Kurian KM (2005). Hiyama E, Yokoyama T, Ichikawa T et al. (1994). Poor out- Paraneoplastic limbic encephalitis. Clin Neurol Neuro- come in patients with advanced stage neuroblastoma surg 107: 128–131. and coincident opsomyoclonus syndrome. Cancer 74: Fauser S, Talazko J, Wagner K et al. (2005). FDG-PET and 1821–1826. MRI in potassium channel antibody-associated non- Ingenito GG, Berger JR, David NJ, Norenberg MD (1990). paraneoplastic limbic encephalitis. Acta Neurol Scand Limbic encephalitis associated with thymoma. Neurol- 111: 338–343. ogy 40: 382. Fisher PG, Wechsler DS, Singer HS (1994). Anti-Hu anti- Kaniecki R, Morris JC (1993). Reversible paraneoplastic lim- body in a neuroblastoma-associated paraneoplastic syn- bic encephalitis. Neurology 43: 2418–2419. drome. Pediatr Neurol 10: 309–312. KararizouE,MarkouI,ZalonisIet al. (2005). Paraneoplastic Fujii N, Furuta A, Yamaguchi H et al. (2001). Limbic limbic encephalitis presenting as acute viral encephali- encephalitis associated with recurrent thymoma: a post- tis. J Neurooncology 75: 229–232. mortem study. Neurology 57: 344–347. Keime-Guibert F, Graus F, Broet P et al. (1999). Clin- Gambini C, Conte M, Bernini G et al. (2003). Neuro- ical outcome of patients with anti-Hu-associated blastic tumors associated with opsoclonus-myoclonus encephalomyelitis after treatment of the tumor. Neurol- syndrome: histological, immunohistochemical and ogy 53: 1719–1723. molecular features of 15 Italian cases. Virchows Arch 442: Kinirons P, Fulton A, Keoghan M et al. (2003). Paraneo- 555–462. plastic limbic encephalitis and chorea associated with GrausF,Keime-GuibertF,ReneRet al. (2001). Anti-Hu- CRMP-5 neuronal antibody. Neurology 61: 1623–1624. associated paraneoplastic encephalomyelitis: analysis of Kinsbourne M (1962). Myoclonic encephalopathy of 200 patients. Brain 124: 1138–1148. infants. J Neurol Neurosurg Psychiatry 25: 271–276. 254 Section 2. Effects of cancer and cancer treatment on cognition

Kleopa KA, Elman LB, Lang B et al. (2006). Neuromyotonia Mitchell WG, Snodgrass SR (1990). Opsoclonus-ataxia due and limbic encephalitis sera target mature Shaker-type to childhood neural crest tumors: a chronic neurologic K+ channels: subunit specificity correlates with clinical syndrome. J Child Neurol 5: 153–158. manifestations. Brain 129: 1570–1584. Mitchell WG, Davalos Y, Brumm VL et al. (2002). Kodama T, Numaguchi Y, Gellad FE et al. (1991). Magnetic Opsoclonus-ataxia caused by childhood neuroblastoma: resonance imaging of limbic encephalitis. Neuroradiol- developmental and neurologic sequelae. Pediatrics 109: ogy 33: 520–523. 86–98. Kohler J, Hufschmidt A, Hermle L et al. (1988). Limbic Mitchell WG, Brumm VL, Azen CG et al. (2005). Longitu- encephalitis: two cases. J Neurooncology 20: 177–178. dinal neurodevelopmental evaluation of children with Konczak J, Schoch B, Dimitrova A et al. (2005). Func- opsoclonus-ataxia. Pediatrics 116: 901–907. tional recovery of children and adolescents after cerebel- Modrego PJ, Cay A, Pina A, Monge A (2002). Paraneoplas- lartumourresection.Brain 128: 1428–1441. tic subacute encephalitis caused by adenocarcinoma of Korfei M, Fuhlhuber V, Schmidt T et al. (2005). Functional prostate: a case report. Acta Neurol Scand 105: 351–353. characterisation of autoantibodies from patients with Moe PG, Nellhaus G (1970). Infantile polymyoclonia- pediatric opsoclonus-myoclonus syndrome. J Neuroim- opsoclonus syndrome and neural crest tumors. Neurol- munol 170: 150–157. ogy 20: 756–764. Kuban KC, Ephros MA, Freeman RL et al. (1983). Syndrome Mori M, Kuwabara S, Yoshiyama M et al. (2002). Suc- of opsoclonus-myoclonus caused by Coxsackie B3 infec- cessful immune treatment for non-paraneoplastic lim- tion. Ann Neurol 13: 69–71. bic encephalitis. J Neurol Sci 201: 85–88. Lacomis D, Khoshbin S., Schick RM (1990). MR imaging Na DL, Hahm DS, Park JM, Kim SE (2001). Hyper- of paraneoplastic limbic encephalitis. J Comput Assist metabolism of the medial temporal lobe in limbic Tomogr 14: 115–117. encephalitis on FDG-PET scan: a case report. Eur Neurol Landolfi JC, Nadkarni M (2003). Paraneoplastic lim- 45: 187–189. bic encephalitis and possible narcolepsy in a patient Nokura K, Yamamoto H, Okawara Y et al. (1997). Reversible with testicular cancer: case study. Neurooncology 5: limbic encephalitis caused by ovarian teratoma. Acta 214–216. Neurol Scand 95: 367–373. Lawn ND, Westmoreland BF, Kiely MJ et al. (2003). Clini- Ogata M, Kikuchi H, Satou T et al. (2006). Human cal, magnetic resonance imaging, and electroencephalo- herpesvirus 6 DNA in plasma after allogeneic stem graphic findings in paraneoplastic limbic encephalitis. cell transplantation: incidence and clinical significance. Mayo Clin Proc 78: 1363–1368. J Infect Dis 193: 68–79. Lee AC, Ou Y, Lee WK, Wong YC (2003). Paraneoplastic lim- Okamura H, Oomori N, Uchitomi Y (1997). An acutely con- bic encephalitis masquerading as chronic behavioural fused 15-year-old girl. Lancet 350: 488. disturbance in an adolescent girl. Acta Paediatr 92: 506– Overeem S, Dalmau J, Bataller L et al. (2004). Hypocretin-1 509. CSF levels in anti-Ma2 associated encephalitis. Neurol- Lee EK, Maselli RA, Agius MA (1998). Morvan’s fibrillary ogy 62: 138–140. chorea: a paraneoplastic manifestation of thymoma. Papero PJ, Pranzatelli MR, Margolis LJ et al. (1995). Neu- J Neurol Neurosurg Psychiatry 65: 857–862. robehavioral and psychological functioning of children Liguori R, Vincent A, Clover L et al. (2001). Morvan’s syn- with opsoclonus-myoclonus syndrome. Dev Med Child drome: peripheral and central nervous system and car- Neurol 37: 915–932. diac involvement with antibodies to voltage-gated potas- Parisi MT, Hattner RS, Matthay KK et al. (1993). Optimized sium channels. Brain 124: 2417–2426. diagnostic strategy for neuroblastoma in opsoclonus- Linke R, Schroeder M, Helmberger T, Voltz R (2004). myoclonus. J Nucl Med 34: 1922–1926. Antibody-positive paraneoplastic neurologic syn- Pelkofer H, Schubart AS, Hoftberger R et al. (2004). Mod- dromes: value of CT and PET for tumor diagnosis. elling paraneoplastic CNS disease: T-cells specific for Neurology 63: 282–286. the onconeuronal antigen PNMA1 mediate autoimmune Mihara M, Sugase S, Konaka K et al. (2005). The “pulv- encephalitis in the rat. Brain 127: 1822–1830. inar sign” in a case of paraneoplastic limbic encephalitis Petit T, Janser JC, Achour NR et al. (1997). Paraneoplastic associated with non-Hodgkin’s lymphoma. J Neurol Neu- temporal lobe epilepsy and anti-Yo autoantibody. Ann rosurg Psychiatry 76: 882–884. Oncol 8: 919. Chapter 17. Paraneoplastic disorders 255

Petruzzi MJ, Alarcon PA (1995). Neuroblastoma-associated Ravizza SM, McCormick CA, Schlerf JE et al. (2006). Cere- opsoclonus-myoclonus treated with intravenously bellar damage produces selective deficits in verbal work- administered immune globulin G. J Pediatr 127: 328– ing memory. Brain 129: 306–320. 329. Rickman OB, Parisi JE, Yu Z et al. (2000). Fulminant autoim- Pittock SJ, Kryzer TJ, Lennon VA (2004). Paraneoplastic mune cortical encephalitis associated with thymoma antibodies coexist and predict cancer, not neurological treated with plasma exchange. Mayo Clin Proc 75: 1321– syndrome. Ann Neurol 56: 715–719. 1326. Pittock SJ, Lucchinetti CF, Parisi JE et al. (2005). Ridley A, Kennard C, Scholtz CL et al. (1987). Omnipause Amphiphysin autoimmunity: paraneoplastic accompa- neurons in two cases of opsoclonus associated with oat niments. Ann Neurol 58: 96–107. cell carcinoma of the lung. Brain 110: 1699–1709. Plonquet A, Gherardi RK, Creange A et al. (2002). Oligo- Roberts WK, Darnell RB (2004). Neuroimmunology of the clonal T-cells in blood and target tissues of patients with paraneoplastic neurological degenerations. Curr Opin anti-Hu syndrome. J Neuroimmunol 122: 100–105. Immunol 16: 616–622. Plonquet A, Garcia-Pons F, Fernandez E et al. (2003). Pep- RojasI,GrausF,Keime-GuibertFet al. (2000). Long- tides derived from the onconeural HuD protein can elicit term clinical outcome of paraneoplastic cerebellar cytotoxic responses in HHD mouse and human. J Neu- degeneration and anti-Yo antibodies. Neurology 55: roimmunol 142: 93–100. 713–715. Pohl KR, Pritchard J, Wilson J (1996). Neurologic sequelae Ronning C, Sundet K, Due-Tonnessen B et al. (2005). Persis- of the dancing eye syndrome. Eur J Pediatr 155: 237–244. tent cognitive dysfunction secondary to cerebellar injury Pozo-Rosich P, Clover L, Saiz A et al. (2003). Voltage-gated in patients treated for posterior fossa tumors in child- potassium channel antibodies in limbic encephalitis. hood. Pediatr Neurosurg 41: 15–21. Ann Neurol 54: 530–533. Rosenbaum T, Gartner J, Korholz D et al. (1998). Paraneo- Pranzatelli MR, Hyland K, Tate ED et al. (2004a). Evi- plastic limbic encephalitis in two teenage girls. Neurope- dence of cellular immune activation in children with diatrics 29: 159–162. opsoclonus-myoclonus: cerebrospinal fluid neopterin. Rosenfeld MR, Eichen JG, Wade DF et al. (2001). Molecular J Child Neurol 19: 919–924. and clinical diversity in paraneoplastic immunity to Ma Pranzatelli MR, Travelstead AL, Tate ED et al. (2004b). B- proteins. Ann Neurol 50: 339–348. and T-cell markers in opsoclonus-myoclonus syndrome: Rostasy K, Behnisch W, Kulozik A et al. (2005). High- immunophenotyping of CSF lymphocytes. Neurology 62: dose pulsatile dexamethasone therapy in children with 1526–1532. opsoclonus-myoclonus syndrome [abstract]. Ann Neurol Pranzatelli MR, Tate ED, Dukart WS et al. (2005a). Sleep 58 [Suppl. 9]: S109. disturbance and rage attacks in opsoclonus-myoclonus Rousseau A, Benyahia B, Dalmau J et al. (2005). T-cell syndrome: response to trazodone. J Pediatr 147: response to Hu-D peptides in patients with anti-Hu syn- 372–378. drome. J Neurooncology 71: 231–236. Pranzatelli MR, Tate ED, Travelstead AL et al. (2005b). Rudnick E, Khakoo Y, Antunes NL et al. (2001). Mycophenolate reduces CSF T-cell activation and is Opsoclonus-myoclonus-ataxia syndrome in neuroblas- a steroid sparer in opsoclonus-myoclonus syndrome toma: clinical outcome and antineuronal antibodies: a (abstract). Ann Neurol 58 [Suppl. 9]: S111. report from the Children’s Cancer Group. Med Pediatr Pranzatelli MR, Tate ED, Travelstead AL et al. (2005c). Oncol 36: 612–622. Cyclophosphamide therapy in pediatric opsoclonus- Russo C, Cohn SL, Petruzzi MJ, Alarcon PA (1997). Long- myoclonus syndrome [abstract]. Ann Neurol 58 [Suppl. term neurologic outcome in children with opsoclonus- 9]: S90. myoclonus associated with neuroblastoma: a report Pranzatelli MR, Tate ED, Travelstead AL, Longee D (2005d). from the Pediatric Oncology Group. Med Pediatr Oncol Immunologic and clinical responses to rituximab in a 29: 284–288. child with opsoclonus-myoclonus syndrome. Pediatrics Sahashi K, Sakai K, Mano K, Hirose G (2003). Anti- 115: e115–119. Ma2 antibody related paraneoplastic limbic/brain stem Provenzale JM, Barboriak DP, Coleman RE (1998). Limbic encephalitis associated with breast cancer expressing encephalitis: comparison of FDG PET and MR imaging Ma1, Ma2, and Ma3 mRNAs. J Neurol Neurosurg Psychia- findings. AJR Am J Roentgenol 170: 1659–1660. try 74: 1332–1335. 256 Section 2. Effects of cancer and cancer treatment on cognition

Scheid R, Voltz R, Guthke T et al. (2003). Neuropsychi- opsoclonus-myoclonus syndrome [abstract]. Ann Neurol atric findings in anti-Ma2-positive paraneoplastic limbic 58 [Suppl. 9]: S111. encephalitis. Neurology 61: 1159–1160. Thieben MJ, Lennon VA, Boeve BF et al. (2004). Potentially Scheid R, Honnorat J, Delmont E et al. (2004a). A new reversible autoimmune limbic encephalitis with neu- anti-neuronal antibody in a case of paraneoplastic lim- ronal potassium channel antibody. Neurology 62: 1177– bic encephalitis associated with breast cancer. J Neurol 1182. Neurosurg Psychiatry 75: 338–340. Thuerl C, Muller K, Laubenberger J et al. (2003). MR imag- Scheid R, Lincke T, Voltz R et al. (2004b). Serial FDG ing of autopsy-proven paraneoplastic limbic encephali- positron emission tomography and magnetic resonance tis in non-Hodgkin lymphoma. AJNR Am J Neuroradiol imaging of paraneoplastic limbic encephalitis. Arch Neu- 24: 507–510. rol 61: 1785–1789. Tomimitsu H, Arimura K, Nagado T et al. (2004). Mecha- Shapiro B, Shulkin BL, Hutchinson RJ et al. (1994). Locating nism of action of voltage-gated K+ channel antibodies in neuroblastoma in the opsoclonus-myoclonus syndrome. acquired neuromyotonia. Ann Neurol 56: 440–444. J Nucl Biol Med 38: 545–555. Tsukamoto T, Mochizuki R, Mochizuki H et al. (1993). Sheth RD, Horwitz SJ, Aronoff S et al. (1995). Opsoclonus Paraneoplastic cerebellar degeneration and limbic myoclonus syndrome secondary to Epstein-Barr virus encephalitis in a patient with adenocarcinoma of the infection. J Child Neurol 10: 297–299. colon. J Neurol Neurosurg Psychiatry 56: 713–716. Sillevis Smitt P, Grefkens J, de Leeuw B et al. (2002). Survival Veneselli E, Conte M, Biancheri R et al. (1998). Effect and outcome in 73 anti-Hu positive patients with para- of steroid and high-dose immunoglobulin therapy on neoplastic encephalomyelitis/sensory neuronopathy. opsoclonus-myoclonus syndrome occurring in neuro- J Neurol 249: 745–753. blastoma. Med Pediatr Oncol 30: 15–17. Solomon GE, Chutorian AM (1968). Opsoclonus and occult Vernino S, Lennon VA (2000). New Purkinje cell anti- neuroblastoma. NEnglJMed279: 475–477. body (PCA-2): marker of lung cancer-related neurologi- Stern RC, Hulette CM (1999). Paraneoplastic limbic cal autoimmunity. Ann Neurol 47: 297–305. encephalitis associated with small cell carcinoma of the Vernino S, Lennon VA (2004). Autoantibody profiles and prostate. Mod Pathol 12: 814–818. neurological correlations of thymoma. Clin Cancer Res Sutton I, Winer J, Rowlands D, Dalmau J (2000). Limbic 10: 7270–7275. encephalitis and antibodies to Ma2: a paraneoplastic Vernino S, Eggenberger ER, Rogers LR, Lennon VA (2002). presentation of breast cancer. J Neurol Neurosurg Psychi- Paraneoplastic neurological autoimmunity associated atry 69: 266–268. with ANNA-1 autoantibody and thymoma. Neurology 59: Sutton RC, Lipper MH, Brashear HR (1993). Limbic 929–932. encephalitis occurring in association with Alzheimer’s Vincent A, Buckley C, Schott J et al. (2004). Potassium chan- disease. J Neurol Neurosurg Psychiatry 56: 808–811. nel antibody-associated encephalopathy: a potentially Tabarki B, Palmer P, Lebon P, Sebire G (1998). Spontaneous immunotherapy-responsive form of limbic encephalitis. recovery of opsoclonus-myoclonus syndrome caused by Brain 127: 701–712. enterovirus infection. J Neurol Neurosurg Psychiatry 64: Vitaliani R, Mason W, Ances B et al. (2005). Paraneoplastic 406–422. encephalitis, psychiatric symptoms, and hypoventilation Tattevin P, Schortgen F, Broucker T et al. (2001). Varicella- in ovarian teratoma. Ann Neurol 58: 594–604. zoster virus limbic encephalitis in an immunocompro- Voltz RD, Gultekin HS, Rosenfeld MR et al. (1999). A sero- mised patient. Scand J Infect Dis 33: 786–788. logic marker of paraneoplastic limbic and brain-stem Taylor RB, Mason W, Kong K, Wennberg R (1999). encephalitis in patients with testicular cancer. New Engl Reversible paraneoplastic encephalomyelitis associated JMed340: 1788–1795. with a benign ovarian teratoma. Can J Neurol Sci 26: 317– Wainwright MS, Martin PL, Morse RP et al. (2001). Human 320. herpesvirus 6 limbic encephalitis after stem cell trans- Telander RL, Smithson WA, Groover RV (1989). Clinical out- plantation. Ann Neurol 50: 612–619. come in children with acute cerebellar encephalopathy Waragi M, Chiba A, Uchibori A et al. (2006). Anti-Ma2 asso- and neuroblastoma. J Pediatr Surg 24: 11–14. ciated paraneoplastic neurological syndrome present- Tersak JM, Safier RA, Schor NF (2005). Rituximab in ing as encephalitis and progressive muscular atrophy. the treatment of refractory neuroblastoma-associated J Neurol Neurosurg Psychiatry 77: 111–113. Chapter 17. Paraneoplastic disorders 257

Wingerchuk DM, Noseworthy JH, Kimmel DW (1998). Para- Younes-Mhenni S, Janier MF, Cinotti L et al. (2004). FDG- neoplastic encephalomyelitis and seminoma: impor- PET improves tumour detection in patients with para- tance of testicular ultrasonography. Neurology 51: 1504– neoplastic neurological syndromes. Brain 127: 2331– 1507. 2338. Wong AM, Musallam S, Tomlinson RD et al. (2001). Opso- Yu Z, Kryzer TJ, Griesmann GE et al. (2001). CRMP-5 neur- clonus in three dimensions: oculographic, neuropatho- onal autoantibody: marker of lung cancer and thymoma- logic and modelling correlates. J Neurol Sci 189: 71–81. related autoimmunity. Ann Neurol 49: 146–154. Yiu VW, Kovithavongs T, McGonigle LF, Ferreira P (2001). Ziter FA, Bray PF, Cancilla PA (1979). Neuropathologic ﬁnd- Plasmapheresis as an effective treatment for opsoclonus- ings in a patient with neuroblastoma and myoclonic myoclonus syndrome. Pediatr Neurol 24: 72–74. encephalopathy. Arch Neurol 36: 51. 18

Symptomatic therapies and supportive care issues

Alan Valentine and Eduardo Bruera

Introduction tus and behavior, interventions, and potential areas for future research. The supportive care of cancer patients routinely involves management of multiple symptoms as well as neuropsychiatric disorders associated with cog- Delirium nitive dysfunction, notably delirium, but also in some instances depression. Cancer therapies them- The American Psychiatric Association deﬁnes delir- selves (especially medications) can cause or exac- ium as a syndrome characterized by rapid onset of erbate cognitive dysfunction. Cognitive disorders impaired cognition, and altered consciousness, and of all types are the second most common psychi- it is presumed to be due to one or more physical atric disorders experienced by cancer patients after or disease-related factors (Table 18.1, American Psy- mood disorders (Derogatis et al., 1983). Patients in chiatric Association, 2000; DSM-IV TR). Other def- particular settings and stages of the disease con- initions include these criteria but also emphasize tinuum are at particular risk for cognitive impair- altered ability to attend and changes in sleep-wake ment, with potential implications for prognosis. The cycle and behavior (psychomotor agitation or retar- boundaries between these disorders are not always dation) (Table 18.2, basic psychopathology of delir- distinct, which complicates accurate diagnosis and ium). Delirium is a serious complication of med- treatment. Co-morbidity is common. The stigma ical illness, especially for the hospitalized patient associated with mental illness and the physical bur- and in geriatric custodial care settings. It is asso- dens of caring for affected patients place family ciated with increased length and cost of hospital members and other caregivers at increased risk for stay and with increased morbidity and mortality physical and emotional distress. The common cog- (Caraceni et al., 2000; Franco et al., 2001; Leslie et al., nitive disorders seen in the oncology setting often 2005). The prevalence of delirium in oncology varies respond well to treatment. In other cases, palliation greatly and ranges from 7%–50% in various inpa- of symptoms is possible and individual patients may tient settings to >85% at the end of life (Fann et al., respond to creative and unconventional medication 2002; Lawlor et al., 2000a; Ljubisavljevic & Kelly, interventions. Here we discuss common neuropsy- 2003; Massie et al., 1983; Prieto et al., 2002). Delir- chiatric syndromes and clinical settings associated ium is often overlooked and is easily misdiagnosed. with cognitive dysfunction and altered mental sta- In intensive care and peri-operative settings, the

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

258 Chapter 18. Symptomatic therapies and supportive care issues 259

Table 18.1. DSM-IV criteria for delirium Table 18.2. Basic psychopathology of delirium (Lipowski, 1990) 1. Disturbance of consciousness (i.e., reduced clarity of r awareness of the environment) with reduced ability to Impaired awareness of self and surroundings (also focus, sustain, or shift attention referred to as “reduced level of consciousness”) r 2. Change in cognition (i.e., memory deficit, Impairment of directed thinking r disorientation, language disturbance) or development of Disorder of attention, with hypo- or hyper-alertness r a perceptual disturbance that is not better accounted for Impairment of memory r by a pre-existing or evolving dementia Diminished perceptual discrimination, with a tendency 3. Development of the disturbance over a short period of toward misperceptions, i.e., illusions and hallucinations r time (usually hours–days) with fluctuating course during Impairment of spatiotemporal orientation (may be the day absent in a mild case) r 4. Evidence from this history, physical examination, or Disturbance of psychomotor behavior, with hyper- or laboratory findings that the disturbance is caused by hypoactivity, both verbal and non-verbal r physiological consequences of a medical condition Disordered sleep-wake cycle, usually marked by drowsiness and naps during the day, insomnia at night, or both r Source: Reprinted with permission from the Diagnostic and Unpredictable fluctuations in alertness and in severity of Statistical Manual of Mental Disorders (4th edn. Text Revi- cognitive impairment during the day and overall exacer- sion) (Copyright 2000) (American Psychiatric Association, bation of symptoms at night and upon awakening r 2000). Acute onset and relatively brief duration (hours to several weeks) r Laboratory evidence of widespread cerebral dysfunction, misleading term “ICU psychosis” is often used. This especially diffuse changes (slowing or fast activity of implies a functional emotional reaction to the stress background activity on the EEG) of the intensive care setting and discourages or triv- ializes the need for aggressive search for reversible Source: Lipowski ZJ (1990). Delirium: Acute Confusional causes of altered mental status. Because of the fluc- States. New York: Oxford University Press with permission. tuating course and variable presentations of delirium it may be confused with other neuropsychi- tive delirium is a potential physical threat to self atric disorders including depression, dementia, and and others, and will often require treatment with primary thought disorders, leading to inappropri- antipsychotic drugs and possibly physical restraint. ate treatment. The course of delirium often includes Hypoactive delirium presents with slowing of lucid intervals characterized by periods in which the thought processes, speech, and behavior. Affect is patient appears cognitively and behaviorally intact. minimally reactive. Level of arousal is decreased and It may appear that the delirium has resolved only to attention to the surrounding environment is dimin- have the patient return to previous or new abnormal ished; this can be difficult to observe during a brief behaviors. examination. This diagnosis is easily missed, as the Delirium may be classified by the level and inten- initial and prevailing impression may be of a patient sity of associated psychomotor activity. Hyperac- who is severely depressed or possibly demented. tive, hypoactive, and mixed forms of the syndrome A mixed form of delirium is characterized by fea- have been described (O’Keeffe & Lavan, 1999; Ross tures of both the hyperactive and hypoactive forms. et al., 1991). The hyperactive form with obvious The behaviors of a patient in a mixed-form delirium motor agitation, autonomic instability, prominent may vary from day to day or within a day, often with- delusions or hallucinations, and affective lability is out warning and without any predictability in the usually fairly easy to recognize. The classic model variability. for hyperactive delirium is alcohol withdrawal delir- Delirium risk factors in general hospital and geri- ium (delirium tremens). The patient in a hyperac- atric settings are probably relevant in oncology. 260 Section 2. Effects of cancer and cancer treatment on cognition

These include advanced age, baseline cognitive methotrexate, cytosine arabinoside, ifosfamide and function, medications (especially anticholinergic cyclophosphamide (Breitbart & Cohen, 1998). Bio- drugs and central nervous system depressants), logical response modifiers especially interleukin-2 and electrolyte and metabolic dyscrasias (Inouye, and interferon alpha are, especially if used in com- 1998, 2000). Recent investigations in cancer patients bination, associated with delirium and sometimes have produced conflicting data regarding the use with a prolonged encephalopathic state (Meyers & of certain common supportive care drugs (benzodi- Valentine, 1995). azepines, opioid analgesics) as risk factors for delir- Metabolic disturbances are common causes of ium (Gaudreau et al., 2005; Ljubisavljevic & Kelly, delirium in this setting. Malignancy-associated 2003). hypercalcemia is often implicated, though hypoxia The cognitive impairment of delirium is such that (often related to respiratory failure or anemia or the patient often (but not always) has no memory both) is likely a more common cause. Electrolyte of related events afterwards. One sign of improve- disturbances (e.g., significant hyper- or hypona- ment in delirium is the patient’s realization that per- tremia) of any origin and metabolic impairments ceptual disturbances (illusions and hallucinations) associated with liver or renal failure (hepatic and are not real. The patient cannot attend, or does renal encephalopathies) are seen frequently. Sys- so inconsistently. Short-term memory (and some- temic and central nervous system infections are times long-term memory) is impaired. Frequently other common causes. Especially in critically ill the patient is disoriented. Thought processes are patients, it may not be possible to identify a single illogical. Speech may be affected. cause for delirium. The etiology frequently is multifactorial. Figure 18.1 summarizes the most common causes of delirium. Common causes of delirium in cancer patients

Medications may be the most common precipitants Assessment of delirium of delirium in hospitalized patients, and the list of drugs with any potential to cause altered men- The diagnosis of delirium is clinical and based on tal status is extremely long (Brown, 2000; Carter history and observation of the patient, review of the et al., 1996). Geriatric patients are at potentially high medical record and nursing notes, physical exami- risk for drug-induced delirium because of altered nation, and laboratory studies. The delirious patient (slowed) drug metabolism and because of their rel- is often if not always an unreliable historian. It is atively high rates of baseline cognitive impairment. necessary to rely on family members, nursing staff, In peri-operative settings, delirium thought and other care providers for details of the onset and secondary to effects of anesthesia is common. characteristics of the clinical presentation. The his- Other medications that are routinely associated tory of use of medications, use or discontinuation of with altered mental status include opioid anal- alcohol and illicit drugs, baseline cognitive function gesics, benzodiazepine sedative-hypnotics, cortico- and other events (e.g., falls) is critical in considera- steroids, antiarrhythmics and sympathomimetics tion of possible evolving delirium. (Marcantonio et al., 1994). Anticholinergic drugs Observation and clinical interview of the patient have long been associated with a characteristic will allow assessment of the level of arousal, psy- hyperactive delirium and anticholinergic processes chomotor activity, range and stability of affect, are the best-studied causes of delirium (Trzepacz, distractibility, and possible illusions, hallucination, 1996; Tune et al., 1981). and delusions. The formal mental status exam- Many antineoplastic therapies have been asso- ination should test attention, orientation, mem- ciated with delirium, but there are relatively few ory, abstract thinking, and speed and dynamics of that commonly cause the syndrome. These include thought (Lipowski, 1990). Chapter 18. Symptomatic therapies and supportive care issues 261

Tumor b y products and cy tokines

O ther drugs O pioids and/or withdraw al

Renal/liver Delirium Sepsis failure

Metabolic (calcium, s odiu m , Deh y d ration ox y g en) CNS CNS tumor inv olve m ent

Figure 18.1. Common causes of delirium

Delirium rating scales Careful physical examination provides information crucial to the search for precipitating and The Mini-Mental State Examination (MMSE) may potentially dangerous causes of altered mental sta- be the most frequently used screen of cognitive tus, and will often inﬂuence the choice of diagnos- function and is often employed to assess deliri- tic tests. Current and recent vital signs should be ous patients (Folstein et al., 1975). However, the assessed. Cardiovascular and pulmonary examina- MMSE does not distinguish delirium from demen- tions should be performed. The neurological exam- tia and patients may score poorly on the test for ination should be emphasized, with assessment for a number of reasons not directly related to delir- lateralizing signs and increased intracranial pres- ium (Anthony et al., 1982). The Confusion Assess- sure. ment Method (CAM) developed by Inouye and colleagues is frequently used and is designed to allow to non-psychiatric clinicians to screen for the presence Laboratory assessment and other tests of delirium (Inouye et al., 1990). Even frequently used instruments such as the CAM have a signiﬁ- Laboratory assessment of possible delirium cant associated error rate and patients who score includes serum chemistries: electrolytes, creatinine, above threshold levels on these scales should go blood urea nitrogen, calcium, magnesium, liver on to careful clinical evaluation. Other instruments function assays, as well as urinalysis and complete including the Delirium Rating Scale (DRS) and the blood count with differential and platelet count. Memorial Delirium Assessment Scale (MDAS) are Chest X-ray and electrocardiogram (ECG) should used more to assess the severity of delirium and be reviewed or obtained. In appropriate patients, can be used to follow patients over time and assess serum drug levels should be obtained; these may their response to therapy, and also as research include immunosuppressants (e.g., ciclosporin), instruments (Breitbart et al., 1997; Trzepacz et al., anticonvulsants, cardiac drugs (e.g., digoxin), and 1988; Trzepacz, 1994). Physicians, nurses, and other psychotropics (e.g., lithium, tricyclic antidepres- health care professionals with palliative care expe- sants). Especially in the absence of reliable history, rience can be trained on the use of the MDAS in a urine toxicology screens of prescription and illicit single 2-h review (Fadul, 2007). drugs should be obtained. Vitamin B12 and serum 262 Section 2. Effects of cancer and cancer treatment on cognition

folate levels should be checked in patients known frequently and assured of his or her safety. The pres- or suspected to be alcohol-dependent. ence of family members is often a source of com- Other tests should be considered in certain set- fort to the patient. The converse however may not tings, but are probably not routinely necessary. be true: family members are often unprepared for or Neuroimaging (computed axial tomography, mag- frightened of behaviors associated with delirium. In netic resonance imaging) may be helpful if phys- such cases and when care requires that the patient ical exam reveals focal neurological signs, and in be closely monitored, it may be best for all con- the absence of other obvious causes of delirium, cerned to engage a professional sitter. or when delirium persists despite appropriate treat- Patients in hyperactive delirium are potential ment. The electroencephalogram (EEG) will almost threats to themselves or others (e.g., pulled IV lines, always reveal diffuse, non-speciﬁc slowing in the self-extubation, violent or impulsive response to delirious patient. While EEG evaluation is prob- perceived threat). The use of physical restraints, ably not routinely necessary, it should be obtained while occasionally necessary to aid early manage- in cases of suspected seizures. It may also be ment of a patient in a hyperactive delirium, should useful in attempts to distinguish delirium from be avoided if possible and should never be done other causes of similar behavior (e.g., dementia, without pharmacological management. severe depression, “functional” psychiatric disorders) (Boutros & Struve, 2002). Pharmacotherapy

Medication management of delirium (especially Management of delirium hyperactive and mixed forms) is usually intended to Ideally delirium is managed by being prevented. provide symptomatic control of motor agitation and In some settings it is reasonable to anticipate the distressing psychotic symptoms, while attempts are development of delirium. This is especially true made to identify and correct underlying causes of in peri-operative and intensive care settings, the the disorder. Antipsychotic drugs are the mainstays elderly, patients with known cognitive impairment of such treatment. Of these, haloperidol is gener- or central nervous system malignancy, patients ally the agent of choice. The drug may be admin- known to be dependent on alcohol or illicit drugs, istered by multiple routes (including intravenously, and those on complicated medication regimens at which is common and effective, though not for- baseline. Data from the history, physical exami- mally approved), and has comparatively modest nation, and laboratory work-up should guide an anticholinergic and anti-alpha-adrenergic effects. It aggressive search for potentially reversible or treat- has been associated with QTc prolongation and ven- able abnormalities that might be responsible for the tricular arrhythmia (torsades de pointes) when used acute altered mental status. at high intravenous doses. Atypical antipsychotic Behavioral management of delirium includes drugs (e.g., risperidone, olanzapine, quetiapine, attention to the physical environment. Ideally the ziprasidone) have generally favorable safety pro- patient is not subjected to excessive light or noise, or ﬁles, though lack of available intravenous adminis- other distractions. At the same time, sensory under- tration makes their use potentially problematic in stimulation is probably not helpful. Artiﬁcial or nat- critical care and other special settings (Cassem et al., ural light should be provided during the day in 2004; Del Fabbro, 2006a; Mazzocato et al., 2000; an attempt to help maintain a normal sleep-wake Valentine & Bickham, 2005). cycle. Particularly in cases of hypoactive delirium, Benzodiazepines (BZPs) are treatments of choice some physical activity should be encouraged. Low- for alcohol withdrawal delirium because their agon- level background sound and light should be main- ist effects on gamma-aminobutyric acid (GABA) tained at night. The patient should be reoriented transmission and function directly address one of Chapter 18. Symptomatic therapies and supportive care issues 263

the primary pathological mechanisms responsible Table 18.3. DSM-IV major depressive episode for alcohol withdrawal (Mayo-Smith & Mayo-Smith, r 1997). The drugs do not have antipsychotic prop- Depressed mood r Diminished interest or pleasure in activities erties and concomitant use of an antipsychotic r may be required. The sedative-hypnotic effects of Significant weight loss/gain or decrease/increase in appetite BZPs are such that they are often used in inten- r Insomnia or hypersomnia sive care and palliative care settings as first-line r Psychomotor agitation or retardation treatments for delirium not associated with alco- r Fatigue or loss of energy r hol withdrawal. This approach to management of Feelings of worthlessness or excessive guilt r delirium is problematic. Used alone BZPs will usu- Diminished ability to think or concentrate, or indecisive- ally not improve cognitive dysfunction associated ness r with delirium and, especially in patients with base- Recurrent thought of death or suicidal ideation line cognitive impairment (e.g., dementia, elderly patients with mild cognitive impairment), they are Source: Reprinted with permission from the Diagnostic and likely to exacerbate cognitive dysfunction and may Statistical Manual of Mental Disorders (4th edn., Text Revi- sion) (Copyright 2000). (American Psychiatric Association, increase psychomotor agitation (Breitbart et al., 2000) 1996). Medication management of hypoactive delirium is less well described in the literature. The patient significant cognitive impairment. Though major in a hypoactive delirium is often not obviously psy- depression and significant adjustment disorders are chotic or in emotional distress, and psychomotor the most common psychiatric disorders seen in retardation is such that the patient is usually not cancer (Derogatis et al., 1983), the pseudodemen- a direct threat to self or others. However, hypoac- tia construct is likely of less importance in oncol- tive delirium is not innocuous and is associated with ogy than in general geriatric medicine, where it is a medical morbidity and adverse outcomes (Kelly risk factor for development of primary irreversible et al., 2001; O’Keeffe & Lavan, 1999). Because most dementia (Dobie, 2002; Reifler, 2000; Reischies & antipsychotic drugs are variably sedating, their use Neu, 2000; Visser et al., 2000). In our experience, it in this setting would seem counterintuitive. Classi- is uncommon to encounter a patient with depres- cal antipsychotics still may be effective against some sion and significant cognitive impairment without symptoms (Platt et al., 1994). Similarly, anticholin- the presence of disease- or treatment-related fac- ergic effects of psychostimulants have the poten- tors that could also be contributing to the dysfunc- tial to exacerbate some presentations of delirium, tion, though there are instances in which the diag- but these drugs have been reported effective against nosis is in doubt (Pereira & Bruera, 2001). The issue the hypoactive form of the syndrome (Gagnon et al., may become more important over time as clinicians 2005; Morita et al., 2000). encounter and treat increasing numbers of cancer survivors, who are vulnerable to new-onset mood disorders but who have also received neurotoxic Depression therapies. Anxiety is frequently a co-morbid condition Cognitive impairment of variable severity is a rec- with depression, especially in elderly patients. The ognized component of mood disorders, especially chronically anxious patient’s complaints of cogni- major depression (Table 18.3, American Psychi- tive problems may be a function of impaired atten- atric Association DSM-IV TR, 2000). In this setting tion with subsequent memory registration difficul- and especially in the elderly, the terms depressive ties. In such instances primary treatment of anxiety dementia or pseudodementia are used to describe may greatly improve day-to-day function. 264 Section 2. Effects of cancer and cancer treatment on cognition

Disease- and treatment-related factors feron, interleukin-2, thalidomide), and brain radiation (late delayed radiation toxicity) (Crossen et al., Multiple disease- and treatment-related factors may 1994; New, 2001; Verstappen et al., 2003). Toxicity is be associated with cognitive impairment in the can- in part a function of dose and duration of treatment, cer setting. Often these problems can be palliated, if and route of administration. not completely reversed. There has been considerable debate about cogni- Malignancy-associated metabolic derangements tive dysfunction associated with hormonal antineo- include hypernatremia, hypercalcemia, and other plastics (e.g., anti-estrogens, aromatase inhibitors). derangements associated with hepatic and renal Literature has been mixed regarding the frequency failure. Acutely and severe dyscrasias are associated and severity of objective impairment (Ahles & with delirium. Chronic and less severe dysfunction Saykin, 2002; Schagen et al., 2002; Wefel et al., 2004) is associated with mild encephalopathy. Appropri- but these agents are common sources of patient ate primary medical therapies including hydration, complaints of “chemobrain,” and not infrequently pharmacotherapy, and dialysis often will improve of reactive depression given that patients are likely cognitive function. to be on the drugs for years. Infectious processes will most often present as delirium that usually will respond to treatment. Viral Supportive care drugs encephalitis may leave the patient with chronic cognitive impairment, with or without affected senso- Any drug with central nervous system (CNS) depres- rium. sant properties has potential to cause cognitive impairment, with or without altered sensorium. In the oncology setting, these would include Cancer/treatment-related causes of dementia opioid analgesics, benzodiazepine anxiolytics and and cognitive impairment hypnotics, phenothiazine anti-emetics, anticonvulsants, and some antidepressants. Primary and metastatic brain tumors and some systemic cancers (i.e., small cell lung cancer) are often associated with cognitive impairment, which is Palliative care settings usually progressive. Leptomeningeal carcinomato- sis and paraneoplastic syndrome may also present Delirium is the most frequent neuropsychiatric with impaired cognition and a clear sensorium. Dis- finding in the last days of life. It has been found in ease progression and complications may result in approximately 85%–90% of cancer patients in the nutritional deficiencies, respiratory insufficiency, or last hours to days before death (Bruera et al., 1992c; anemia that contribute to cognitive impairment. Lawlor et al., 2000a; Massie et al., 1983). Delirium is Please see Chapters 7–17 in this book for more also associated with increased morbidity and mor- indepth discussion of cancer and cancer treatment- tality rates. In studies of advanced cancer patients, related cognitive dysfunction. those with delirium had a median survival of 21 days as compared to 39 days in those without (Caraceni et al., 2000). Antineoplastic therapies Although delirium is one of the most frequent rea- Several chemotherapy agents and other therapies sons for admission to palliative care units (Lawlor, may cause temporary (occasionally permanent) 2002) it is frequently under-diagnosed. When objec- cognitive impairment. These include antimetabo- tive assessment of cognitive function was not per- lites (e.g., methotrexate, ifosfamide, cytosine arab- formed in patients admitted to a palliative care unit, inoside), biological response modifiers (e.g., inter- episodes of delirium went undetected by physicians Chapter 18. Symptomatic therapies and supportive care issues 265

and nurses in 23% and 20% of cases respectively Table 18.4. Symptom intensity scores upon referral to (Bruera et al., 1992c). In palliative care patient’s a multidisciplinary palliative care clinic (n = 135)a delirium is frequently misdiagnosed as depression, dementia, or just simply sedation. Median ESAS Intensity In patients admitted to palliative care programs (Inter Quartile Range) delirium may be reversed through a suitable thera- Pain 6 (4, 8) peutic approach in almost 50% of patients (Gagnon Fatigue 7 (4.5, 8) Nausea 1 (0, 5) et al., 2000; Lawlor et al., 2000b; Sarhill et al., 2001). Drowsiness 5 (1.5, 8) Since delirium is extremely frequent and Anxiety 5 (1, 8) reversible in almost 50% of cases it is very impor- Depression 4 (1, 7) tant to use assessment tools for the screening of Anorexia 6 (4, 8) patients who do not appear to be in delirium. Dyspnea 3 (0, 6) Maintaining a very high index of suspicion and Sleep 6 (4, 8) using validated screening instruments will result Well-being 6 (5, 8) in earlier detection of this devastating syndrome. a This will allow clinicians to conduct immediate Edmonton Symptom Assessment Scores (0; Best, 10; investigation of the possible causes, appropriate Worst) pharmacological management of symptoms such as agitation, hallucinations, and delusions, and to have an established protocol for the evaluation appropriate counseling of caregivers. and management of this complex syndrome. Distressed or ill-informed caregivers can exacerbate a patient’s distress. Agitated behavior is particularly distressing for families and caregivers Symptomatic pharmacotherapy (Breitbart et al., 2002). Agitated behavior may be interpreted as a sign of pain and/or suffering. Patients with advanced cancer develop a number Discussions with family should include a simple of devastating physical and psychosocial symptoms. explanation of delirium, its increased frequency in Table 18.4 summarizes the median (inter quar- advanced illness, potential causes and various clini- tile range) Edmonton Symptom Assessment Scales cal presentations, and the efforts made to manage (0 = best, 10 = worst) for 135 patients referred to it. In some cases it might be appropriate to con- an outpatient palliative care clinic. Since the vast duct specific delirium tests with the family present majority of patients develop multiple symptoms so as to demonstrate the level of cognitive impair- they frequently require multiple drug interventions, ment and disinhibition. This will help families bet- including opioid analgesics, adjuvant drugs with ter understand why patients may present increased central effects such as anticonvulsants or antide- symptomatic and/or emotional expression as a con- pressants, anti-emetics, and psychoactive medica- sequence of disinhibition. tions. Fatigue and sedation can occur as a conse- The importance of providing a safe and quiet quence of these persistent symptoms but also as a environment with limited visual and auditory stim- consequence of the pharmacological interventions ulation, surrounded by familiar objects and/or used for their management. sounds, and avoiding discussions and confronta- A temporary impairment in neuropsychiatric tions should be discussed with caregivers and family tests and increased sedation are frequently members. observed when patients are started on opioid Delirium is one of the most common causes of analgesics or after they undergo a significant dose conflict between health care professionals and fam- increase (Bruera et al., 1989). These effects can be ilies in the palliative care setting. It is very important reversed by methylphenidate (Bruera et al., 1992a, 266 Section 2. Effects of cancer and cancer treatment on cognition

1992b). Methylphenidate and other psychostim- Antidepressants with “activating” side-effects ulants have been used in the management of (e.g., bupropion, fluoxetine) may be useful in some depression and hypoactive delirium associated cases. Some authors have found that antidepres- with advanced cancer (Gagnon et al., 2005; Homsi sants with adrenergic activity are particularly useful et al., 2000; Morita et al., 2000; Olin & Masand, against the cognitive side-effects of interferon 1996; Rozans et al., 2002; Sood et al., 2006). It has therapy (Capuron et al., 2002; Raison et al., 2005). been found to improve both sedation as well as performance on simple tasks such as finger tapping and arithmetic (Bruera et al., 1992a, 1992b). In Future research addition, methylphenidate can allow for increased opioid dose in patients with severe pain who have Cognitive dysfunction is a problem that requires dose-limiting sedation (Bruera et al., 1992b). A more research attention from supportive care spe- preliminary study observed that patients receiving cialists working in oncology. Potential questions patient-controlled methylphenidate for the man- include but are not limited to the following. agement of cancer fatigue experienced significant relief (Bruera et al., 2003a). However, a randomized controlled trial found that this improvement was Detection and screening not superior to placebo (Bruera et al., 2006). Patients receiving chronic opioid therapy fre- The utility of screening oncology populations at quently develop signs of opioid-induced neurotoxi- high risk for delirium or other disease-related cog- city such as delirium, generalized myoclonus, seda- nitive dysfunction is not well established. Outcomes tion, or hyperalgesia (Del Fabbro et al., 2006a). In research using appropriate endpoints (e.g., quality these patients the most important approach is to of life, complication rates, length of survival, cost) identify opioids as the likely cause and use opi- would help to determine whether screening is an oid rotation to mimimize symptoms (Mercadante & effective use of health care resources. The role of Bruera, 2006). screening in outpatient/community cancer center Fatigue is the most common symptom in patients settings (and appropriate instruments for this pur- with advanced cancer. It is a multidimensional syn- pose) is largely unaddressed. drome that is frequently associated with cachexia, depression, physical symptoms, and drugs such as Treatment opioids and benzodiazapenes (Del Fabbro et al., 2006b). Unfortunately, there is no approved phar- Aside from treatment of potential alcohol with- macological treatment for cancer fatigue. A prelimi- drawal delirium, the utility, if any, of prophylactic nary study found that donepezil was able to improve pharmacotherapy in patients at risk for delirium has both sedation and fatigue in patients with advanced not been evaluated. The appropriate role of psy- cancer receiving opioids (Bruera et al., 2003b). Ran- chostimulants and other agents (e.g., cholinesterase domized controlled trials on the role of donepezil inhibitors, neuroprotective agents) in management for both sedation and fatigue are currently being of disease- and treatment-induced chronic cog- conducted. nitive dysfunction is not well established and is Other psychostimulants such as dextro- worthy of additional investigation. In a setting amphetamine and modafinil have been occa- where haloperidol is the only antipsychotic drug sionally used to treat fatigue and decreased arousal routinely administered intravenously, there is a in a number of settings including oncology (Ballon need to develop new agents that may be given by & Feifel, 2006). There is limited experience with this route, and to investigate new delivery methods these drugs in advanced cancer. for atypical antipsychotics. Chapter 18. Symptomatic therapies and supportive care issues 267

Communication/education Ballon JS, Feifel MD (2006). A systematic review of modafinil: potential clinical uses and mechanisms of Research into the efficacy of various medical edu- action. J Clin Psychiatry 67(4): 554–566. cation methods related to cancer-related cognitive Boutros NN, Struve F (2002). Electrophysiological assess- impairment would support the goal of improving ment of neuropsychiatric disorders. Sem Clin Neuropsy- screening and the treatment of at-risk patients. The chiatry 7(1): 30–41. impact of patient and caregiver education on this Breitbart W, Cohen KR (1998). Delirium. In Holland JC (ed.) subject is largely unknown. Psycho-oncology (pp. 564–573). New York: Oxford Univer- sity Press. Breitbart W, Marotta R, Platt MM et al. (1996). A double-blind trial of haloperidol, chlorpromazine, and Conclusion lorazepam in the treatment of delirium in hospitalized AIDS patients. Am J Psychiatry 153(2): 231–237. Symptom management is the primary goal of sup- Breitbart W, Rosenfeld B, Roth A et al. (1997). The memorial delirium assessment scale. J Pain Symptom Manage portive care for cancer patients. The management 13(3): 128–137. of various presentations of cancer-related cognitive Breitbart W, Gibson C, Tremblay A (2002). The delirium impairment will become increasingly important as experience: delirium recall and delirium-related distress the general population ages and patients live longer in hospitalized patients with cancer, their spouses/ care- before being diagnosed, or with the disease as a givers, and their nurses. Psychosomatics 43: 183–194. chronic illness. Clinicians should be familiar with Brown TM (2000). Drug-induced delirium. Sem Clin Neu- various presentations of delirium, its many causes ropsychiatry 5(2): 113–124. and differential diagnosis, treatment, and the set- Bruera E, MacMillan K, Hanson J, MacDonald RN (1989). tings in which it is likely to occur. A high index of The Edmonton staging system for cancer pain: prelimi- suspicion is required when evaluating the cogni- nary report. Pain 37: 203–209. tively impaired cancer patient, and creative phar- Bruera E, Miller MJ, Macmillan K, Kuehn N (1992a). Neu- ropsychological effects of methylphenidate in patients macotherapy approaches may be required. Ongoing receiving a continuous infusion of narcotics for cancer and future research efforts into the causes, identifi- pain. Pain 48(2): 163–166. cation, and treatment of cognitive dysfunction have Bruera E, Fainsinger R, MacEachern T, Hanson J (1992b). the potential to improve quality of life for cancer The use of methylphenidate in patients with incident patients and caregivers, and possibly to improve the cancer pain receiving regular opiates. A preliminary efficacy of primary cancer therapies. report. Pain 50(1): 75–77. Bruera E, Miller L, McCallion J, Macmillan K, Krefting L, Hanson J (1992c). Cognitive failure in patients with terminal cancer: a prospective study. J Pain Symptom Man- REFERENCES age 7: 192–195. Bruera E, Driver L, Barnes EA et al. (2003a). Patient- Ahles TA, Saykin AJ (2002). Breast cancer chemotherapy- controlled methylphenidate for the management of related cognitive dysfunction. Clin Breast Cancer 3 fatigue in patients with advanced cancer: a preliminary [Suppl. 3]: S84–S90. report. J Clin Oncol 21(23): 4439–4443. American Psychiatric Association (2000). Diagnostic and Bruera E, Strasser F, Shen L et al. (2003b). The effect of Statistical Manual of Mental Disorders (4th edn. Text donepezil on sedation and other symptoms in patients Revision edn. Washington, DC: American Psychiatric receiving opioids for cancer pain: a pilot study. JPain Association. Symptom Manage 26(5): 1049–1054. Anthony JC, LeResche L, Niaz U et al. (1982). Limits of the Bruera E, Valero V, Driver L et al. (2006). Patient-controlled mini-mental state as a screening test for dementia and methylphenidate for cancer fatigue: a double-blind, ran- delirium among hospital patients. Psychol Med 12(2): domized, placebo-controlled trial. J Clin Oncol 24(13): 397–408. 2073–2078. 268 Section 2. Effects of cancer and cancer treatment on cognition

Capuron L, Hauser P, Hinze-Selch D et al. (2002). Treat- Gagnon B, Low G, Schreier G (2005). Methylphenidate ment of cytokine-induced depression. Brain Behav hydrochloride improves cognitive function in patients Immunol 16(5): 575–580. with advanced cancer and hypoactive delirium: a CaraceniA,NanniO,MaltoniMet al. (2000). Impact of prospective clinical study. J Psychiatry Neurosci 30(2): delirium on the short term prognosis of advanced can- 100–107. cer patients. Italian Multicenter Study Group on Pallia- Gaudreau JD, Gagnon P, Roy MA et al. (2005). Association tive Care. Cancer 89(5): 1145–1149. between psychoactive medications and delirium in hos- Carter GL, Dawson AH, Lopert R (1996). Drug-induced pitalized patients: a critical review. Psychosomatics 46(4): delirium. Incidence, management and prevention. Drug 302–316. Safety 15(4): 291–301. Homsi J, Walsh D, Nelson KA (2000). Psychostimulants in Cassem NH, Murray GB, Lafayette JM, Stern TA (2004). supportive care. Support Care Cancer 8(5): 385–397. Delirious patients. In Stern T (ed.) Massachusetts General Inouye SK (1998). Delirium in hospitalized older patients: Hospital Handbook of General Hospital Psychiatry (5th recognition and risk factors. J Geriatric Psychiatry Neurol edn.). Philadelphia, PA: Mosby. 11(3): 118–125. Crossen JR, Garwood D, Glatstein E, Neuwelt EA (1994). Inouye SK (2000). Prevention of delirium in hospitalized Neurobehavioral sequelae of cranial irradiation in older patients: risk factors and targeted intervention adults: a review of radiation-induced encephalopathy. J strategies. Ann Med 32(4): 257–263. Clin Oncol 12(3): 627–642. Inouye SK, van Dyck CH, Alessi CA et al. (1990). Clarify- Del Fabbro E, Dalal S, Bruera E (2006a). Symptom control in ing confusion: the confusion assessment method. A new palliative care – Part III: Dyspnea. Dyspnea and delirium. method for detection of delirium. Ann Int Med 113(12): J Palliat Med 9(2): 422–436. 941–948. Del Fabbro E, Dalal S, Bruera E (2006b). Symptom control Kelly KG, Zisselman M, Cutillo-Schmitter T et al. (2001). in palliative care – Part II: Cachexia/anorexia and fatigue. Severity and course of delirium in medically hospitalized J Palliat Med 9(2): 409–421. nursing facility residents. Am J Geriatric Psychiatry 9(1): Derogatis LR, Morrow GR, Fetting J (1983). The prevalence 72–77. of psychiatric disorders among cancer patients. JAm Lawlor PG (2002). The panorama of opioid-related cogni- Med Assoc 249: 751–757. tive dysfunction in patients with cancer: a critical litera- Dobie DJ (2002). Depression, dementia, and pseudode- ture appraisal. Cancer 94(6): 1836–1853. mentia. Semin Clin Neuropsychiatry 7(3): 170–186. Lawlor PG, Gagnon B, Mancini IL et al. (2000a). Occur- Fadul N, Kaur G, Zhang T, Palmer JL, Bruera E (2007). rence, causes, and outcome of delirium in patients with Evaluation of the Memorial Delirium Assessment Scale advanced cancer: a prospective study. Arch Int Med (MDAS) for the screening of delirium by means of sim- 160(6): 786–794. ulated cases by palliative care health professionals. Sup- Lawlor PG, Gagnon B, Mancini IL et al. (2000b). Delirium port Cancer Care 15(11): 1271–1276. as a predictor of survival in older patients with advanced Fann JR, Roth-Roemer S, Burington BE et al. (2002). Delir- cancer. Arch Int Med 160: 2866–2868. ium in patients undergoing hematopoietic stem cell Leslie DL, Zhang Y, Holford TR et al. (2005). Premature transplantation. Cancer 95(9): 1971–1981. death associated with delirium at 1-year follow-up. Arch Folstein MF, Folstein SE, McHugh PR (1975). Mini-mental Intern Med 165(14): 1657–1662. state. A practical method for grading the cognitive state Lipowski ZJ (1990). Delirium: Acute Confusional States. of patients for the clinician. J Psychiatr Res 12(3): 189– New York: Oxford University Press. 198. Ljubisavljevic V, Kelly B (2003). Risk factors for develop- Franco K, Litaker D, Locala J, Bronson D (2001). The cost ment of delirium among oncology patients. Gen Hosp of delirium in the surgical patient. Psychosomatics 42(1): Psychiatry 25(5): 345–352. 68–73. Marcantonio ER, Juarez G, Goldman L et al. (1994). The Gagnon B, Allard P, Masse B, DeSerres M (2000). Delir- relationship of postoperative delirium with psychoactive ium in terminal cancer: a prospective study using daily medications. J Am Med Assoc 272(19): 1518–1522. screening, early diagnosis, and continuous monitoring. J Massie MJ, Holland J, Glass E (1983). Delirium in terminally Pain Symptom Manage 19: 412–426. ill cancer patients. Am J Psychiatry 140(8): 1048–1050. Chapter 18. Symptomatic therapies and supportive care issues 269

Mayo-Smith MF, Mayo-Smith MF (1997). Pharmacological Reischies FM, Neu P (2000). Comorbidity of mild cognitive management of alcohol withdrawal. A meta-analysis and disorder and depression – a neuropsychological analysis. evidence-based practice guideline. American Society of Eur Arch Psychiatry Clin Neurosci 250(4): 186–193. Addiction Medicine Working Group on Pharmacological Ross CA, Peyser CE, Shapiro I, Folstein MF (1991). Delirium: Management of Alcohol Withdrawal. J Am Med Assoc 278 phenomenologic and etiologic subtypes. Int Psychogeri- (2): 144–151. atr 3(2): 135–147. Mazzocato C, Stiefel F, Buclin T, Berney A (2000). Psy- Rozans M, Dreisbach A, Lertora JJ, Kahn MJ (2002). Pallia- chopharmacology in supportive care of cancer: a review tive uses of methylphenidate in patients with cancer: a for the clinician: II Neuroleptics. Support Care Cancer review. J Clin Oncol 20(1): 335–339. (8)2: 89–97. Sarhill N, Walsh D, Nelson KA et al. (2001). Assessment of Mercadante S, Bruera E (2006). Opioid switching: a system- delirium in advanced cancer: the use of the bedside con- atic and critical review. Cancer Treat Rev 32: 304–315. fusion scale. Am J Hospice Palliat Care 18: 335–341. Meyers CA, Valentine AD (1995). Neurological and psy- Schagen SB, Muller MJ, Boogerd W (2002). Late effects of chiatric adverse effects of immunological therapy. CNS adjuvant chemotherapy on cognitive function: a follow- Drugs 3: 56–68. up study in breast cancer patients. Ann Oncol 9(13): Morita T, Otani H, Tsunoda J et al. (2000). Successful pal- 1387–1397. liation of hypoactive delirium due to multi-organ failure Sood A, Barton DL, Loprinzi CL (2006). Use of by oral methylphenidate. Support Care Cancer 8(2): 134– methylphenidate in patients with cancer. Am J Hos- 137. pice Palliat Care 23(1): 35–40. New P (2001). Radiation injury to the nervous system. Curr Trzepacz PT (1994). A review of delirium assessment instru- Opinion Neurol 14(6): 725–734. ments. Gen Hosp Psychiatry 16(6): 397–405. O’Keeffe ST, Lavan JN (1999). Clinical significance of delir- Trzepacz PT (1996). Delirium. Advances in diagnosis, ium subtypes in older people. Age Ageing 28(2): 115– pathophysiology, and treatment. Psychiatrics Clin North 119. Am 19(3): 429–448. Olin J, Masand P (1996). Psychostimulants for depression in Trzepacz PT, Baker RW, Greenhouse J (1988). A symptom hospitalized cancer patients. Psychosomatics 37(1): 57– rating scale for delirium. Psychiat Res 23(1): 89–97. 62. Tune LE, Damlouji NF, Holland A et al. (1981). Associa- Pereira J, Bruera E (2001). Depression with psychomotor tion of postoperative delirium with raised serum levels retardation: diagnostic challenges and the use of psy- of anticholinergic drugs. Lancet 2(8248): 651–653. chostimulants. J Palliat Med 4(1): 15–21. Valentine AD, Bickham JG (2005). Delirium and substance Platt MM, Breitbart W, Smith M et al. (1994). Efficacy of withdrawal. In Shaw A (ed.) Acute Care of the Cancer neuroleptics for hypoactive delirium. J Neuropsychiatry Patient (pp. 545–557). New York: Marcel Dekker. Clin Neurosci 6(1): 66–67. Verstappen CC, Heimans JJ, Hoekman K, Postma TJ (2003). Prieto JM, Blanch J, Atala J et al. (2002). Psychiatric morbid- Neurotoxic complications of chemotherapy in patients ity and impact on hospital length of stay among hemato- with cancer: clinical signs and optimal management. logic cancer patients receiving stem-cell transplantation. Drugs 63(15): 1549–1563. J Clin Oncol 20(7): 1907–1917. Visser PJ, Verhey FR, Ponds RW et al. (2000). Distinction Raison CL, Demetrashvili M, Capuron L, Miller AH (2005). between preclinical Alzheimer’s disease and depression. Neuropsychiatric adverse effects of interferon-alpha: J Am Geriatr Soc 48(5): 479–484. recognition and management. CNS Drugs 19(2): 105– Wefel JS, Lenzi R, Theriault RL, Davis RN, Meyers CA 123. (2004). The cognitive sequelae of standard-dose adjuvant Reifler BV (2000). A case of mistaken identity: pseudode- chemotherapy in women with breast carcinoma: results mentia is really predementia. J Am Geriatr Soc 48(5): 593– of a prospective, randomized, longitudinal trial. Cancer 594. 100(11): 2292–2299. 19

Animal models and cancer-related symptoms

Adrian Dunn

Animal models have frequently been useful in mal models have relied on chronic stress paradigms. developing treatments for a variety of diseases. However, tests for depression would include proce- Their use permits researchers to ask questions dures such as the forced swim test or the tail suspen- aboutmechanismsthatwouldbedifficultorunethi- sion test used to assess depression-like or antide- cal in humans, and also permits the testing of poten- pressant activity (see below). Cognitive function in tial treatments. Models that work in rodents are usu- animals has been addressed almost exclusively in ally preferred because we know much about their tests of memory, except in non-human primates. physiology and behavior, and because rodents are However, other symptoms such as depression and relatively inexpensive. Moreover, a host of experi- fatigue that are common in cancer patients can also mental manipulations have been developed for use affect cognitive function. Thus this chapter will also in these species. Work with primates that may be address animal models of these symptoms, along more valid is substantially more expensive, and the with classical cognitive tasks. numbers of subjects that can be used are normally Details of animal models of behavior can be found very limited. The development of animal models for in a number of reviews (e.g., Weiss & Kilts, 1998) behavioral symptoms presents a special challenge, and some books; for example, the recently pub- because although many such models and tests exist, lished The Behavior of the Laboratory Rat (Whishaw they address poorly the symptoms that are of most & Kolb, 2005), Animal Models of Human Emotion concern to cancer patients. This chapter will pro- and Cognition (Haug & Whalen, 1999), and in a book vide a selective overview of animal models and tests, focused on transgenic mice, What’s Wrong with My and provide examples of what has been achieved Mouse? (Crawley, 2000). in other areas. The limitations of the use of the non-human animal models and tests will also be Why use animal models? addressed. It is important at the outset to note the distinc- The rationales for using animal models of disease tion between an animal model and an animal test. A states are to identify or determine potential underly- model is a procedure used to induce a state in the ing mechanisms of the disease, and to test potential animal that resembles the disease under study. In treatments. Their use is frequently justified either this context, a test is a procedure that reveals symp- because certain kinds of experiments are difficult to toms that resemble aspects of the human disease. perform in humans, or because they would be con- For example, in the case of depression, most ani- sidered unethical. Most often the justification is that

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

270 Chapter 19. Animal models and cancer-related symptoms 271

the measurement of some important variable is too disease studied. Treatments that are effective in the intrusive to measure in humans, or because a pro- model are less likely to work in the clinical situation posed experimental therapy carries unknown risks. ifthemethodusedtoinducethemodelisunrelated In the most common examples, a drug known to to the basic mechanism(s) of the disease. McKinney affect a metabolic process may be tested for its effi- and Bunney (1969) proposed a set of criteria for ani- cacy, and/or to reveal unforeseen unacceptable or mal models of human mental health disorders. The toxic side-effects. model should resemble the condition it represents An obvious problem with the use of animal in its etiology, biochemistry, symptomatology, and models is that the drug/treatment may not work treatment. in humans the way it does in animals. Conversely, An example of the successful use of an animal treatments that work in the animal model may model is the development of treatments for atten- induce unforeseen side-effects in humans. It is tion deficit hyperactivity disorder (ADHD). It was fair to say that animal models work best when observed that methylphenidate, an amphetamine- the method for inducing the model is related to like drug that stimulates release of the neuro- the underlying cause of the disease. This may not transmitter dopamine (DA), induced a reversal of be too difficult when the underlying cause of the the hyperactivity observed in 6-hydroxydopamine- disease is known or suspected, but may be partic- treated rats (Shaywitz et al., 1976). This finding ularly difficult in the case of behavioral symptoms, was paradoxical, because amphetamine-like drugs especially symptoms like those of depression and normally increase locomotor activity. Thus it was psychosis. Do we really know that a rat or a mouse reasoned that methylphenidate might counter the can experience depression as humans do? And, hyperactivity observed in children with ADHD. In what is psychotic behavior in a rat? Can such symp- practice, methylphenidate treatment has subse- toms be modeled in primates? In such cases, the quently proven to be very effective for the treat- model can be aimed to have face value, i.e., that ment of ADHD in teenagers (Shaywitz et al., it appears to the investigator that the behavior of 2001). In this example, the model was almost cer- the rat resembles a human behavior. An example tainly useful because ADHD in children involves would be the forced swim test developed by Porsolt some abnormality in brain catecholaminergic func- to assess depression-like behavior. In this test, tion. However, not all models have proven so it is conceived that the animal learns whether a useful. For example, the olfactory bulbectomy behavioral strategy is useful or not (see below). A model mimics depression in several tests, and “depressed” rat or mouse gives up more rapidly the effects respond appropriately to many treat- and floats or stays immobile for a longer time. ments effective in depressed patients (Cryan & Most often, the test is based on a pharmacological Mombereau, 2004; Cryan et al., 2002; Kelly et al., validation, often exclusively. By this is meant that 1997; Song & Leonard, 2005; Willner, 1984). Nev- drugs (or other treatments that affect the human ertheless, although the model is of considerable disease) similarly affect animals in the model. All scientific interest, it has failed to provide insight drugs that are clinically effective should work in the into the mechanism of depression and has so far model, whereas drugs that have no (useful) clinical failed to inspire useful clinical therapies (Willner, effect should not work in the model. Moreover, the 1984). dose–response relationships should be similar in The most obvious models for cancer patients humans and the animal models. would be animals bearing tumors. There have been relatively few studies of this type. In an early Animal models study, Chance et al. (1983) studied the anorexia associated with inoculation of rats with Walker 256 For a model to be useful, the methods used to create carcinosarcoma cells and noted increases in brain it should bear some relationship to the underlying tryptophan and serotonin (5-hydroxytryptamine, 272 Section 2. Effects of cancer and cancer treatment on cognition

5-HT) metabolism. The neurochemical changes Porsolt, 2000; Willner, 2005). However, many of were reversed when the tumor was resected these models may not be directly relevant to the (Chance et al., 1988). They suggested that the clinical situation (Matthews et al., 2005). Most such anorexia was associated with changes in hypotha- models have been based on chronic stress (Anis- lamic 5-HT. Chuluyan et al. (2000) studied the man & Zacharko, 1982). Willner (1984) discussed at neurochemical and pituitary–adrenal effects of length the problems with most of the early mod- inoculating mice with the murine lymphoma cell els. He proposed that predictive validity should line AW5E. There were no significant changes after be assessed by whether a model correctly identi- 6 or 8 days, but a sustained increase in hypotha- fies antidepressant treatments of pharmacologically lamic norepinephrine (NE) and 5-HT metabolism diverse types, without making errors of omission or appeared 10 days after injection. On the last day commission, and whether potency in the model cor- tested (day 14), plasma corticosterone was slightly relates with clinical potency (Willner, 1984). elevated, as were the catabolites of DA, NE and 5-HT and concentrations of tryptophan in the Stress brain. The changes in catecholamines were most evident in the hypothalamus. Such changes in brain Clinical depression is frequently associated with life catecholamines and 5-HT, and in corticosterone are stress (Anisman & Zacharko, 1982), although stress characteristic of stress. In a subsequent study, DBA2 has not been identified as a factor in all cases of mice were inoculated with L1210 mouse leukemia clinical depression (Willner, 1984; van Praag, 2004). cells and their behavioral activity was assessed Hans Selye (who has the dubious distinction of in the tail suspension test (TST) and the Porsolt being known as the father of stress) defined the forced swim test (FST), as well as activity in the important distinction between what he called stress, open field (OF) (Dunn et al., 2004). No consistent which he defined as a state induced by adverse differences from controls were observed in the circumstances or treatments, and the stressor,the behavior of these animals on days 8 or 11. How- agent responsible for inducing the stress. The most ever, on day 15, mice inoculated with L1210 cells frequently used stress paradigms in rodents are such (5000 or 50 000 cells/mouse) exhibited significant treatments as chronic electric footshock or restraint. increases in immobility in the TST and the FST (see By chronic is meant once or twice daily for 1– below). The inoculated mice exhibited increases 2 weeks or more. Needless to say these are not in plasma corticosterone, as well as significant the most common stressors for humans. “Physical” increases in the metabolism of NE, but not DA, stressors, such as high or low temperature or food in the hypothalamus, as well as increases in tryp- deprivation, are seldom used in animal models, tophan and 5-HT metabolism in the cortex and because any changes in body temperature and/or hypothalamus. Vegas et al. (2004) recently reported metabolism are likely to confound the physiolog- increases in immobility, and changes in social inter- ical measures made. However, common stressors action, as well as increases in brain DA and 5-HT in the human condition, such as bereavement, do metabolism in mice inoculated with B16 melanoma not appear to affect rodents similarly (although they cells. may do so in primates). Another model commonly used is the chronic mild stress model (CMS) which uses a series of different stressors over an extended Models of depression time period (4–6 weeks). However, the CMS model promoted by Willner and others has proven to be The most recent reviews indicate that there are difficult to replicate in some laboratories, and even some half-dozen rodent models of depression in Willner’s own laboratory when he moved it from (Cryan & Mombereau, 2004; Cryan et al., 2002; London to Swansea (Willner, 1997). Others have Chapter 19. Animal models and cancer-related symptoms 273

promoted other “cafeteria-style” models in which divert metabolism towards increasing glucose con- a variety of different stressors are applied sequen- centrations in the circulation, and have a host of tially (e.g., Katz, 1981). Yet another model is learned other physiological actions, including driving lym- helplessness (Maier & Seligman, 1976; Overmier & phocytes from the thymus and spleen into the gen- Seligman, 1967). In this model (initially performed eral circulation, presumably to fight pathogens and in dogs) chronic stressors are used to invoke a to facilitate blood clotting. As its name implies, state of despair, in which the stressed dogs become the HPA axis exists at three levels. In the brain immobile and helpless, although whether this is it involves the secretion of corticotropin-releasing a “learning effect” has been questioned. It has factor (CRF) from hypothalamic neurons, which, been proposed that the learning merely reflects after secretion into the portal blood supply, acts on a decrease in activity associated with the stress- the pituitary to elicit the secretion of adrenocor- related reduction in brain NE (Weiss & Kilts, 1998). ticotropic hormone (adrenocorticotropin, ACTH) Yet another possibility is to use chronic social stress. into the general circulation which in turn acts When rats are confined in a limited space such as on the adrenal cortex stimulating glucocorticoid a cage, a dominance hierarchy is created, in which secretion. one animal rules over the others. The immediate This much is classical physiology, but more recent subordinates are the most stressed in this kind of sit- studies have indicated that the peripheral dual- uation. Social models have been developed by var- ity between the HPA and catecholamine compo- ious groups, most notably the Blanchards with the nents also occurs in the brain. Thus noradrener- visible burrow system in rats and mice (Blanchard gic systems within the brain that have cell bodies et al., 2001, 2003). in the brainstem pour out NE, so that it bathes The underlying physiology of stress is similar most if not all of the brain, and certain amounts in animals and humans. The stress response has of endorphins. The global noradrenergic activation both central and peripheral components. Periph- in the brain appears to arouse the brain such that erally, the principal components are activation of it pays special (selective) attention to the novel the adrenal gland and the sympathetic nervous sys- factors in the environment that have caused the tem (SNS). Activation of the SNS results in the stress. Moreover, CRF is not confined to activation of secretion of substantial quantities of NE and cer- the pituitary. Hypothalamic and extrahypothalamic tain peptides (such as substance P and opioid pep- CRF-containing neurons exert their own effects on tides, e.g., the enkephalins), which can act locally the brain, eliciting fear- and anxiety-like responses, and systemically. The adrenal gland also contributes and enhancing other brain functions, presumably to circulating concentrations of NE, but also adds to address coping with the effects of the stressor epinephrine, and still more peptides (again includ- (Dunn & Berridge, 1990). The link between stress ing some opioid peptides). The major effects of and depression becomes obvious when it is rec- catecholaminergic activation are the mobilization ognized that depression appears to be associated of glucose stored as glycogen (principally in the with increased activity of brain NE (e.g., Wong liver, but also in other organs including the brain), et al., 2000), and hyperactivity of the HPA axis and the redirection of blood away from the vis- (Carroll, 1978; Sachar, 1967). cera and towards voluntary muscle (for fighting or fleeing). The adrenal cortex joins the fray as Animal tests of depression the final component of the so-called hypothalamo– pituitary–adrenocortical (HPA) axis, rapidly increas- The Porsolt forced swim test ing circulating glucocorticoids: corticosterone in rodents, and cortisol in humans and many other Currently, the most frequently used test is the FST animals. The glucocorticoids, as their name implies, developed by Porsolt for the rat (Porsolt et al., 274 Section 2. Effects of cancer and cancer treatment on cognition

1977b) and the mouse (Porsolt et al., 1977a). The The tail suspension test test involves placing a rat or a mouse in a tall The TST is conceptually similar to the FST. A mouse cylinder of water from which it cannot escape, so is suspended by its tail and the latency to cease that it must either swim or float. Rats are initially struggling and the duration of passive immobility placed in the cylinder of water for 15 min. The rats are scored. The duration of immobility displayed are put back into the cylinder 24 h later for 5 min. by the suspended animals is reduced by antide- An observer records the activity of the rat, especially pressant treatments (Cryan & Mombereau, 2004; the time at which the rat stops struggling and Cryan et al., 2002). Although a version of the test for floats (i.e., the latency to float), and the duration rats has been reported (Chermat et al., 1986), most of the floating, immobile except for the minimal investigators have had difficulty using the test in paw movements necessary to keep the nose above rats. The TST has the advantage of eliminating the water. The concept was that the “depressed” rat confounding hypothermic effects of the swimming would give up struggling or swimming earlier, (Cryan & Mombereau, 2004). It has been suggested and float for longer. It has been suggested that that the FST and TST differ in the biological the first day session is necessary for the rats to substrates underlying the observed behaviors learn that escape is impossible (Borsini & Meli, (Cryan & Mombereau, 2004). 1988). Drugs or other treatments are administered immediately after this session and then again 5 h and1hbeforethetestonthesecondday(Porsolt et al., 1977b). However, Porsolt subsequently Cognitive tests showed that many antidepressant treatments tested positively with only one or two of the three In general cognitive tests developed for rodents may treatments (Porsolt et al., 1978). Antidepressant be difficult to relate to the cognitive impairments treatments increase the latency to float, and reduce that cancer patients experience. Summarized below the time spent immobile (floating), whereas chronic are the kinds of tests that are available. stress and certain other treatments considered depressogenic increase the time spent floating. This Tests of learning and memory test has been shown to work for many antidepressant treatments, including atypical antidepressants, Many different tests of learning and memory have such as the monoamine oxidase inhibitors (MAOIs) been devised (Alkon et al., 1991). To review each of and electroconvulsive therapy (ECT), although the them in detail is beyond the scope of this chapter. selective serotonin reuptake inhibitors (SSRIs) are In general the tests can be divided into appetitive less effective and some are ineffective (Cryan & and aversive tests. In aversive tests, the animal is Mombereau, 2004; Porsolt, 2000). Thus the test has conditioned to learn that a particular stimulus or been validated pharmacologically, even though action is associated with a punishment, most often the drug treatments work after one, two or three an electric shock. The advantage of such tests is that treatments, in contrast to the chronic treatments they are typically rapidly learned and the memory required in depressed patients. In the test Porsolt is retained for a long time. For example, a com- developed for mice, they were forced to swim only monly used task is one-trial passive avoidance in once, usually with the antidepressant treatments which a rat (or more often a mouse) is placed in applied shortly before the test. Immobilization is a novel apparatus and if it steps through an open- scored during the last 4 min of the 6-min test (Por- ing it immediately receives a brief footshock. This solt et al., 1977b). However, some subsequent stud- test is normally learned in one trial and the memory ies have employed a 2-day test in mice, like that in persists for months (Geller et al., 1970). A common rats. active avoidance task is the shuttle box in which an Chapter 19. Animal models and cancer-related symptoms 275

animal (normally a rat) is trained so that when a light. If the change is only to a different color, this lightisilluminateditshouldmovetoanotherpartof would constitute an intradimensional shift. How- the apparatus within a few seconds, to avoid an elec- ever, once the rat has learned that the color contin- tric footshock. Once again in normal animals this gency can be changed, the investigator can play a conditioning is learned rapidly and the memory is cruel trick. The contingency can be switched from retained for a long time. the color to the shape of the illuminated light, so Perhaps the most popular task, the Morris water that a light that was formerly circular is now in the maze employs a different kind of aversive stimu- form of a triangle, a square, or a cross. This would lus (Morris, 1984; Schimanski & Nguyen, 2004). The constitute an extradimensional shift. Rats find it water maze is a circular tank of water that contains relatively easy to learn the intradimensional shift, a platform which is slightly below the surface of the but the extradimensional shift is much more diffi- water onto which the rat can climb and thus avoid cult. A series of studies in the 1970s showed that swimming. The water is made opaque with a suit- certain peptides, especially ACTH and fragments of able paint or dye, so that the rat cannot see the plat- this molecule, facilitated the learning of the intradi- form, and the rat is expected to remember the loca- mensional shift, but would impair the learning of tion of the platform. the extradimensional shift (Sandman et al., 1974). Appetitive tasks normally rely on a food reward, Interestingly, the same peptides that had such oppo- for example a food pellet, a sugar pill, or perhaps site effects in rats had similar effects on intra- and a sweet drink (a sugar solution or sweetened milk). extradimensional shifts in humans (Sandman et al., Normally the association is made with a particular 1975). location, for example making the correct turn in a “T-maze.” The most popular task is the radial arm maze (Alkon et al., 1991; Olton, 1987). The radial arm Tests for fatigue maze has eight arms, any one of which can be baited with a food reward. Various contingencies can be Fatigue has been modeled in various ways, but there set up, such as not entering the same arm twice is some question about whether the models nec- before locating the correct arm; or remembering essarily reflect fatigue as it is observed in cancer the location of the reward with respect to the start- patients. First of all, fatigue has different meanings ing position; or, the position of the arm within the to different specialists. According to Dalakas et al. room. (1998), to a physiologist fatigue is a decrease in the capacity to perform work; to a pathologist, fatigue is an indicator of a neuromuscular or metabolic disor- Other cognitive tasks der; and to a psychologist, fatigue is a symptom of The number of cognitive tests for rodents other depression associated with decreased motivation to than memory tests is quite limited. One such task engage in mental and physical activities. However, involves changing stimuli for conditioning. For patients report decreased physical performance and example, rats and mice can be trained to press a bar muscle weakness, decreased motivation and sad- to obtain a food reward or to avoid a shock. Once the ness, and a lack of concentration and decreased task is being performed with a high degree of accu- ability for problem solving (Glaus et al., 1996). racy, the “rules” can be changed. Thus, for example, The animal tests in the literature mostly reflect a food reward that was formerly available when a locomotor activity, such as running in a running green light was illuminated may no longer be avail- wheel or on a treadmill. Takagi et al. (1972) stud- able when the light is green, but only when the light ied the activity of mice in a battery of tests on a is red. The rat can then be assessed on its ability to treadmill and balancing on an oscillating shaker. learn the new contingency, a different color of the They observed “fatigue” (failure to keep up with the 276 Section 2. Effects of cancer and cancer treatment on cognition

treadmill) that could be prevented by amphetamine, related to the proximity of the model to the under- caffeine and similar stimulants. Chao et al. (1992) lying causes of the disease. In many cases this may observed reduced voluntary running and delayed simply reflect luck in the choice of the “right” model. initiation of grooming after swimming associated The modeling of psychiatric symptoms in small with infection with Corynebacterium parvum anti- animals is more difficult, because it is not at all gen or Toxoplasma gondii. Ottenweller et al. (1998) clear that rodents exhibit affective states and emo- observed decreased activity in the running wheel tions like humans. Thus while a number of tests for and grooming associated with infection with Bru- depression exist, their validity rests largely on the cella abortus.Daviset al. (1997) observed markedly effects of drugs or other therapies known to ame- decreased treadmill running following treatment liorate depression in humans. This approach/model with poly I:poly C (a synthetic double-stranded RNA was very effective in identifying useful therapies for known to be an effective inducer of interferon- ADHD, but has not so far been particularly useful for α/β (IFN-α/β). This reduced running correlated with the treatment of such psychiatric diseases as mania, the appearance of IFN-α/β and was prevented by depression, and psychosis. This may be because antibodies to IFN-α/β. Consistent with this, IFN-α these diseases are associated with the increased reduced open field activity and swimming in mice complexity of the primate brain. Thus primate mod- (Dunn & Crnic, 1993). We have also observed that els may be more effective for these kinds of signs. IFN-α administered intracerebroventricularly (icv) The neuropsychological symptoms of cancer to rats increased immobility time in the Porsolt FST patients undergoing therapy have additional com- (Dunn & Swiergiel, 2004). Kaur and Kulkarni (2000) plexities. Dealing with the symptoms of the cancer subjected rats to daily forced swims in the Porsolt itself is one matter, but many of the most bother- swim test for 7 days. Antidepressant treatments pro- some symptoms may be associated with the ther- vided symptomatic relief. Ayada et al. (2002) studied apies used to treat the cancer: the radiation ther- gnawing behavior of mice in a narrow plastic cylin- apy, the chemotherapy, and/or the cytokine therapy der, which they proposed as a simple system for the (e.g., IFN-α). To make matters worse, it is even possi- study of muscle activity, fatigue, and stress. ble that the unwanted symptoms may reflect inter- Unfortunately, all of these tests were focused on actions between the cancer itself and the therapies locomotor activity as an indicator of fatigue. Sev- used to treat it. Nevertheless, animal models may be eral of them also used infection with pathogens able to provide indications of the effects of the can- that are known to induce sickness behavior (Kent cer and its therapies that may provide clues for novel et al., 1992), thought to be mediated by certain therapies. cytokines, such as interleukin-1 (Dunn & Swiergiel, 2005). However, the fatigue experienced by cancer patients appears to be a motivational problem, ACKNOWLEDGMENTS which may be better assessed with operant tasks, in which an animal has to press a bar or touch an object in order to obtain a reward, or avoid a pun- The author’s research cited in this article was supported by the U.S. National Institute of Neurological Disorders and ishment. Stroke (NS35370).

Conclusions REFERENCES Animal models have proven to be very useful for developing therapies for a number of human dis- Alkon DL, Amaral DG, Bear MF et al. (1991). Learning and eases. However, the success is probably closely memory. Brain Res Rev 16: 193–220. Chapter 19. Animal models and cancer-related symptoms 277

Anisman H, Zacharko RM (1982). Depression: the predis- DavisJM,WeaverJA,KohutML,ColbertLH,GhaffarA, posing influence of stress. Behav Brain Sci 5: 89–137. Mayer EP (1997). Immune system activation and fatigue Ayada K, Tadano T, Endo Y (2002). Gnawing behavior of a during treadmill running: role of interferon. Med Sci mouse in a narrow cylinder: a simple system for the study Sports Exerc 30: 863–868. of muscle activity, fatigue, and stress. Physiol Behav 77: Dunn AJ, Berridge CW (1990). Physiological and behavioral 161–166. responses to corticotropin-releasing factor administra- Blanchard RJ, Yudko E, Dulloog L, Blanchard DC (2001). tion: is CRF a mediator of anxiety or stress responses? Defense changes in stress nonresponsive subordinate Brain Res Rev 15: 71–100. malesinavisibleburrowsystem.Physiol Behav 72: 635– Dunn AJ, Swiergiel AH (2004). Interferon effects on behav- 642. ior, corticosterone, catecholamines and indoleamines in Blanchard DC, Griebel G, Blanchard RJ (2003). The Mouse mice and rats. Neuropsychopharmacology 29[Suppl. 1]: Defense Test Battery: pharmacological and behavioral S120. assays for anxiety and panic. Eur J Pharmacol 463: 97– Dunn AJ, Swiergiel AH (2005). Effects of interleukin-1 and 116. endotoxin in the forced swim and tail suspension tests in Borsini F, Meli A (1988). Is the forced swimming test a mice. Pharmacol Biochem Behav 81: 688–693. suitable model for revealing antidepressant activity? Psy- Dunn AJ, Swiergiel AH, Cork R, Newman RA (2004). Behav- chopharmacology 94: 147–160. ior, neurochemical and endocrine responses to leukemia Carroll BJ (1978). Neuroendocrine function in psychi- in mice. In 2004 Abstract Viewer/Itinerary Planner (Vol. atric disorders. In Psychopharmacology: A Generation of 462.13). Washington DC: Society for Neuroscience. Progress (pp. 487–497). New York: Raven Press. Dunn AL, Crnic LS (1993). Repeated injections of Chance W, Von Meyenfeldt M, Fischer J (1983). Changes in interferon-α A/D in Balb/c mice: behavioral effects. brain amines associated with cancer anorexia. Neurosci Brain Behav Immun 7: 104–111. Biobehav Rev 7: 471–479. Geller A, Robustelli F, Jarvik ME (1970). Incubation and the Chance WT, Cao L, Nelson JL, Foley-Nelson T, Fischer Kamin effect. JExpPsychol85: 61–65. JE (1988). Reversal of neurochemical aberrations after Glaus A, Crow R, Hammond S (1996). A qualitative study tumor resection in rats. Am J Surg 155: 124–129. to explore the concept of fatigue/tiredness in cancer Chao CC, DeLaHunt M, Hu S, Close K, Peterson PK (1992). patients and in healthy individuals. Support Care Cancer Immunologically mediated fatigue: a murine model. Clin 4: 82–96. Immunol Immunopathol 64: 161–165. Haug M, Whalen RE (1999). Animal Models of Human Emo- Chermat R, Thierry B, Steru L, Simon P (1986). Adaptation tion and Cognition (p. 341). Washington DC: American of the tail suspension test to the rat. J Pharmacol 17: 348– Psychological Association. 350. Katz RJ (1981). Animal models and human depressive dis- Chuluyan HC, Wolcott RM, Chervenak R, Dunn AJ orders. Neurosci Biobehav Rev 5: 231–246. (2000). Catecholamine, indoleamine and corticosteroid Kaur G, Kulkarni SK (2000). Comparative study of antide- responses in mice bearing tumors. Neuroimmunomodu- pressants and herbal psychotropic drugs in a mouse lation 8: 107–113. model of chronic fatigue. JChronicFatigueSyndr6: 23– Crawley JN (2000). What’s Wrong with My Mouse? Behav- 34. ioral Phenotyping of Transgenic and Knockout Mice Kelly JP, Wrynn AS, Leonard BE (1997). The olfactory bul- (pp. 1–368). Wilmington, DE: Wiley-Liss. bectomized rat as a model of depression: an update. Cryan JF, Mombereau C (2004). In search of a depressed Pharmacol Ther 74: 299–316. mouse: utility of models for studying depression-related Kent S, Bluthe´ R-M, Kelley KW, Dantzer R (1992). Sickness behavior in genetically modified mice. Mol Psychiatry behavior as a new target for drug development. Trends 9: 1050–1062. Pharmacol Sci 13: 24–28. Cryan JF, Markou A, Lucki I (2002). Assessing antidepres- Maier SF, Seligman MEP (1976). Learned helplessness: the- sant activity in rodents: recent developments and future ory and evidence. JExpPsychol1: 3–46. needs. Trends Pharmacol Sci 23: 238–245. Matthews K, Christmas D, Swan J, Sorrell E (2005). Dalakas MC, Mock V, Hawkins MJ (1998). Fatigue: defi- Animal models of depression: navigating through nitions, mechanisms, and paradigms for study. Semin the clinical fog. Neurosci Biobehav Rev 29: 503– Oncol 25: S48–S53. 513. 278 Section 2. Effects of cancer and cancer treatment on cognition

McKinney WT, Bunney WE (1969). Animal models of Shaywitz BA, Klopper JH, Yager RD, Gordon JW (1976). depression. Arch Gen Psychiatry 21: 240–248. Paradoxical response to amphetamine in developing rats Morris R (1984). Developments of a water-maze procedure treated with 6-hydroxydopamine. Nature 261: 153–155. for studying spatial learning in the rat. J Neurosci Meth- Shaywitz BA, Fletcher JM, Shaywitz SE (2001). Attention ods 11: 47–60. deﬁcit hyperactivity disorder. Curr Treat Options Neurol Olton DS (1987). The radial arm maze as a tool in behav- 3: 229–236. ioral pharmacology. Physiol Behav 40: 793–797. Song C, Leonard BE (2005). The olfactory bulbectomized Ottenweller JE, Natelson BH, Gause WC et al. (1998). Mouse rat as a model of depression. Neurosci Biobehav Rev 29: running activity is lowered by Brucella abortus treat- 627–647. ment: a potential model to study chronic fatigue. Physiol Takagi T, Saito H, Lee C-H, Hayashi T (1972). Pharmacolog- Behav 63: 795–801. ical studies on fatigue I. Jpn J Pharmacol 22: 17–26. Overmier JB, Seligman MEP (1967). Effects of inescapable van Praag HM (2004). Can stress cause depression? Prog shock on subsequent escape and avoidance responding. Neuropsychopharmacol Biol Psychiatry 28(5): 891–907. J Comp Physiol Psychol 63: 28–33. Vegas O, Beitia G, Sanchez-Martin JR, Arregi A, Azpiroz A Porsolt RD (2000). Animal models of depression: utility for (2004). Behavioral and neurochemical responses in mice transgenic research. Rev Neurosci 11: 53–58. bearing tumors submitted to social stress. Behav Brain Porsolt RD, Bertin A, Jalfre M (1977a). Behavioural despair Res 155: 125–134. in mice: a primary screening test for antidepressants. Weiss JM, Kilts CD (1998). Animal models of depres- Arch Int Pharmacodyn 229: 327–336. sion and schizophrenia. In Schatzberg AF, Nemeroff CB Porsolt RD, Le Pichon M, Jalfre M (1977b). Depression: (eds.) The American Psychiatric Press Textbook of Psy- a new animal model sensitive to antidepressant treat- chopharmacology (2nd edn.) (pp. 89–131). Washington ments. Nature 266: 730–732. DC: American Psychiatric Press. Porsolt RD, Anton G, Blavet N, Jalfre M (1978). Behavioural Whishaw IQ, Kolb B (2005). The Behavior of the Laboratory despair in rats: a new model sensitive to antidepressant Rat. A Handbook with Tests. Oxford: Oxford University treatments. Eur J Pharmacol 47: 379–391. Press. Sachar EJ (1967). Corticosteroids in depressive illness: II. A Willner P (1984). The validity of animal models of depres- longitudinal psychoendocrine study. Arch Gen Psychia- sion. Psychopharmacology (Berl) 83: 1–16. try 17: 554–567. Willner P (1997). Validity, reliability and utility of the Sandman CA, Beckwith BE, Gittis MM (1974). Melanocyte- chronic mild stress model of depression: a 10-year review stimulating hormone (MSH) and overtraining effects on and evaluation. Psychopharmacology 134: 319–329. extradimensional shift (EDS) learning. Physiol Behav 13: Willner P (2005). Chronic mild stress (CMS) revisited: con- 1631–1666. sistency and behavioural-neurobiological concordance Sandman CA, George JM, Nolan JD, van Riezen H, Kastin in the effects of CMS. Psychopharmacology (Berl) 52: 90– AJ (1975). Enhancement of attention in man with 110. ACTH/MSH4–10. Physiol Behav 15: 427–431. Wong ML, Kling MA, Munson PJ et al. (2000). Pronounced Schimanski LA, Nguyen PV (2004). Multidisciplinary and sustained central hypernoradrenergic function in approaches for investigating the mechanisms of major depression with melancholic features: relation to hippocampus-dependent memory: a focus on inbred hypercortisolism and corticotropin-releasing hormone. mouse strains. Neurosci Biobehav Rev 28: 463–483. Proc Natl Acad Sci USA 97: 325–330. SECTION 3

Interventions and implications for clinical trials

Behavioral strategies and rehabilitation

Dona E. C. Locke, Jane H. Cerhan, and James F. Malec

Significant cognitive impairment frequently accom- In contrast, there is a long tradition of apply- panies cancer and cancer treatment for both adult ing cognitive rehabilitation interventions in cases of and pediatric patients. As detailed in previous chap- acquired brain injury, specifically traumatic brain ters, common cognitive complaints include cogni- injury (TBI) and cerebral vascular accident (CVA). tive slowing, deficits in attention, and memory inef- A number of studies have addressed the efficacy ficiencies. In people with brain tumors, cognitive of various interventions for cognitive impairment, impairments cause difficulty returning to work or as summarized in recent reviews by Cicerone et al. school more often than physical impairments, and (2000, 2005). Further development, investigation, caregivers cite cognitive problems as the most dif- and application of cognitive rehabilitation was also ficult problems to manage (Meyers & Boake, 1993). recommended by an NIH Consensus Conference For these reasons, evidence-based cognitive reha- (NIH Consensus Statement Online 1998 October 26– bilitation has the potential to meet prevalent needs 28). More recently, similar cognitive rehabilitation for improving functional abilities in patients with interventions have been under investigation in neu- cancer. rodegenerative conditions such as Alzheimer’s dis- There is considerable focus in the literature on ease (Clare et al., 2005; De Vreese et al., 2001) and the general concept of quality of life in patients mild cognitive impairment (MCI; Chandler & Smith, with many different kinds of cancer. Interventions personal communication, 2005). to improve quality of life are often focused on The lack of empirical evaluation of behavioral symptoms of the cancer (e.g., physical limitations, interventions for cognitive difficulties specifically in decreased activity), physical symptoms related to cancer patients encourages reflection and consid- treatment (e.g., fatigue, nausea, pain), health behav- eration of the best direction, focus, and method- ior change, or distress related to the diagnosis of ologies for this research. In the sections below we cancer (e.g., Clark et al., 2003; Kuhn et al., 2005; will: (1) define cognitive rehabilitation and pro- Ronson & Body, 2002). However, very few studies vide a framework for categorizing cognitive rehabil- specifically target cognitive symptoms in cancer itation techniques; (2) briefly summarize selected patients and in fact, many studies exclude patients behavioral strategies and supporting evidence for with cognitive impairment. Thus, the empirical lit- use with patients with TBI/CVA; (3) describe how erature on behavioral interventions for cognitive similar behavioral strategies have been applied to impairments in cancer patients is very limited. patients with a degenerative condition; (4) describe

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

281 282 Section 3. Interventions and implications for clinical trials

Regulator of procedure Self Other

Internal i z ed memory s trateg ies D omain -specif ic le arning (e.g., pegw ord, methodof lo ci, (e.g., lear ni ng personal im ag ery ) telephone nu mber thro u g h Internal method of v ani shing cu es)

Self-d irect ed use o f Ot her-direct ed s y stem s of

Type of procedure externa l memory aids memo ry aids (e.g., p agi ng (e. g ., mem ory notebo ok ) sy stem) External

Figure 20.1. Dimensions of rehabilitation. Adapted from Malec & Cicerone (2006)

the limited empirical information on cognitive reha- employing one of two approaches: strategy training bilitation with cancer patients; and (5) describe how or restitution training (Cicerone et al., 2005). clinicians may currently apply behavioral interven- The goal of strategy training techniques is to com- tions to cognitive symptoms in cancer patients by pensate for cognitive deficits using spared abili- adapting techniques that have been used with other ties or external devices. Examples of strategy train- patient populations. ing techniques include using a memory notebook to compensate for memory deficits (Sohlberg & Mateer, 1989) or skill training in Alzheimer’spatients Definition of cognitive rehabilitation using their intact procedural memory while bypass- ing impaired episodic memory (De Vreese et al., Cognitive rehabilitation has been defined as “a sys- 2001). The goal of restitution training is to restore tematic, functionally oriented service of therapeu- the underlying impaired function, such as through tic activities that is based on assessment and under- massed practice of sustained attention via com- standing of the patient’s brain-behavior deficits” puterized tasks. While both approaches may have (Cicerone et al., 2000, p. 1596). Cognitive rehabili- merit, strategy training enjoys more research sup- tation tends to be a highly individualized undertak- port than restitution training as a cognitive rehabil- ing in which interventions are closely matched to itation model, especially when the deficits involve the patient’s specific cognitive impairment profile, attention or memory (Cicerone et al., 2000, 2005). severity of impairment, and functional goals. Cog- Thus, we will focus primarily on strategy training nitive rehabilitation techniques may be categorized cognitive rehabilitation techniques in this chapter. by the specific type and severity of cognitive impair- Malec and Cicerone (2006) suggest that strategy ment that is targeted (e.g., intervention for severe training cognitive rehabilitation techniques may be memory impairment). More broadly, however, the further classified along two dimensions (see Figure wide array of techniques can also be classified as 20.1). One dimension describes the degree to which Chapter 20. Behavioral strategies and rehabilitation 283

the procedure that enhances cognitive functioning to encode new information or procedural memory. is internal or external to the person (e.g., a mem- These latter two approaches may be appropriate for ory mnemonic versus a memory notebook). The patients with more severe cognitive impairments. other dimension is the degree to which the proce- These dimensions are not meant to be inflexible dure is regulated by the person or by the environ- or dogmatic, but rather are meant to provide some ment, including other people. Figure 20.1 provides guidance to clinicians when choosing a cognitive examples of cognitive rehabilitation interventions rehabilitation strategy for the individual patient. specific to the memory domain defined by these dimensions. In general, a patient’s type and sever- Cognitive rehabilitation in TBI/CVA ity of cognitive impairment as well as the patient’s preserved cognitive abilities, preferences, goals, and Cicerone and colleagues (2000, 2005) recently amount of family or social support dictate where the updated their review of evidence-based cognitive specific cognitive rehabilitation technique should rehabilitation. Based on their findings, they offer fall along these two dimensions. Comprehensive Practice Standards, Practice Guidelines, and Prac- neuropsychological evaluation plays an important tice Options for cognitive rehabilitation of patients role in directing the choice of cognitive rehabilita- with TBI and stroke. These types of recommenda- tion technique due to the need to understand areas tions are directly linked to the available empirical of deficit as well as areas of preserved functioning. evidence. Practice Standards are based on substan- For example, cognitive rehabilitation strategies tive evidence of effectiveness based on the pres- that are internal procedures (e.g., a memory ence of at least one well-designed and adequately mnemonic) and regulated by the patient (e.g., self- powered randomized controlled trial along with applied in the appropriate situation) may depend additional supportive evidence from less rigor- on some degree of preserved working memory as ous trials, such as non-randomized cohort stud- well as relatively intact executive functions such as ies. Practice Guidelines are offered based on the ability to organize and execute such a strategy. probable evidence of effectiveness obtained from That is, internal procedures that are self-regulated non-randomized or case–control studies. Practice by the patient may be most appropriate for patients Options are based on evidence of possible effective- with relatively mild cognitive impairment and rel- ness from uncontrolled case series or case reports. atively intact cognition in many areas. In contrast, The recommendations of Cicerone and associates learning an external procedure (e.g., a memory for cognitive rehabilitation of attention, executive notebook) that is regulated by the patient (e.g., self- functioning, memory and learning – the cogni- applied in the appropriate situation) theoretically tive domains most commonly affected in cancer mayrelymoreonprocedurallearning.Thus,an patients – are summarized in Table 20.1. In the external procedure that is regulated by the patient remainder of this section, we will provide more may be helpful for patients with either a relatively detailed description of specific interventions to significant cognitive impairment or mild executive serve as examples. functioning difficulty. Domain-specific new learning (an internal procedure) through other-directed Strategy training for attention deficits overlearning or other specialized procedure may rely entirely on procedural memory and require less Fasotti et al. (2000) reported an empirical trial self-monitoring than the above techniques. Exter- of compensation for slowed information process- nal prompting for implementation of the new learning and the experience of “information overload” ing may still be required. External compensation in daily tasks. Patients in the intervention group procedures that are entirely other-directed (such received strategy training in the form of Time Pres- as a paging system) may require minimal ability sure Management (TPM). This approach involves 284 Section 3. Interventions and implications for clinical trials

Table 20.1. Recommendations from Cicerone et al. such as reiterating information, asking for repetition (2000, 2005) of information, asking for clarification of information, or asking for a short pause. Patients in the TPM Attention deficits group utilized more of these strategies after receiv- 1. Practice Standard: strategy training (i.e., ing training in TPM than they had prior to train- compensation strategy) ing, and patients in the TPM group utilized more Executive functioning deficits of these strategies than the generic concentration 2. Practice Guideline: formal training in problem solving training group. In addition, TPM appeared to gen- with application to everyday situations eralize spontaneously over a broader range of tasks 3. Practice Option: strategies including verbal self-instruction, self-questioning, and self-monitoring than generic advice. to promote self-regulation Memory and learning deficits Strategy training for memory and learning 4. Practice Standard: internalized or external procedures inefficiencies: the memory notebook for treatment of mild memory impairment 5. Practice Guideline: external procedures with direct Sohlberg and Mateer (1989) offered empirical sup- application to functional activities for treatment of port for compensatory strategy training using a moderate to severe memory impairment memory notebook for patients with unreliable memory following acquired brain injury. A memory notebook is a portable paper notebook with teaching patients strategies for coping with men- one or more sections that is personalized to the tal slowness utilizing relatively preserved cognitive specific functional needs of the patient. The pur- skills. Examples of specific coping strategies include pose is to compensate for memory difficulties by the following, depending on the patient’s relative teaching patients to store information in the note- cognitive strengths: optimizing planning and organ- book in a manner that facilitates access as well as ization, rehearsing task requirements, or modifi- structuring it to provide cues for tasks to be com- cation of task environment. In general, TPM is pleted. Thus, this cognitive rehabilitation interven- a four-step cognitive strategy: (1) recognize time tion involves an external procedure that is self reg- pressure in the task at hand, (2) prevent as much ulated by the patient (Figure 20.1). Because the time pressure as possible, (3) deal with time pres- procedure is external, it serves to cue and prompt sure as quickly and effectively as possible, and (4) self-regulation by the patient. As such it may be use- self-monitor while using these strategies. Thus, in ful for patients with more severe memory problems terms of our general cognitive rehabilitation model, with some executive functioning deficits. TPM is an internal procedure that is self regulated Possible notebook sections could include: orien- (Figure 20.1) and would be appropriate for patients tation, memory log, calendar, things to do, trans- with relatively mild cognitive impairment and rela- portation, feelings log, names, and today at work. tively intact executive functioning. Patients are trained to use the notebook, for The control group in Fasotti et al. (2000) received example to remember to complete daily tasks and to four generic suggestions regarding concentration document completion of these tasks. Training in the and remembering as much as possible. Participants memory notebook system consists of three steps: who received TPM showed significantly greater use acquisition, application, and adaptation. The acqui- of self-management strategies and better perfor- sition stage consists of introducing the concept of mance on behavioral tasks requiring attention and the notebook to the patient and beginning con- memory (e.g., completing a task according to spe- struction of the patient’s individualized notebook. cific multistep instructions). Specifically, patients in The application stage involves training and practice the TPM group were more likely to utilize strategies in using the notebook daily. The adaptation stage Chapter 20. Behavioral strategies and rehabilitation 285

involves practicing the use of the notebook in order also improved in psychosocial functioning follow- to make it habitual. ing training in problem solving. These findings are Ownsworth and McFarland (1999) further showed consistent with those of an earlier trial of sim- that pairing self-instruction training with use of ilar training in formal problem-solving strategies a memory notebook enhances usage of the note- and application of those strategies to everyday sit- book. The self-instruction training provided a sys- uations and functional activities in patients with tematic method to train patients on how to use acquired brain injury (Von Cramen & Mathes-Von a memory notebook to compensate for memory Cramen, 1991). Patients in the latter trial were clas- problems. The self-instruction specifically involved sified as poor problem solvers on the basis of for- teaching patients to cue themselves using the fol- mal tests of problem solving. Training in prob- lowing abbreviation: WSTC, where W = what you lem solving emphasized teaching patients to reduce are going to do; S = select a strategy for the task; complex problems into manageable steps (prob- T = try out the strategy; C = check out how the strat- lem orientation, problem definition and formula- egy is working. In this study, self-instruction training tion, generation of alternative, decision making, and improved subjects’ ability to spontaneously access solution verification). their diaries and make entries in appropriate everyday situations. Cognitive rehabilitation in degenerative conditions Formal problem solving for executive functioning deficits Most patients with TBI or CVA are on a stable Ownsworth et al. (2000) reported results of the or improving course when they are referred for effects of formal training in problem solving on the cognitive rehabilitation. Some patients with can- application of other compensation strategies and cer may have a similarly stable course. However, on psychosocial functioning in patients with brain some cancer patients are on a declining course. injury. The program involved 16 weekly 90-min This brings into question the applicability of the group sessions. The sessions involved patient edu- research on patients with TBI or stroke to some cation, sharing of patient-developed coping strate- patients with cancer. For patients with cancer gies, and teaching of specific strategies recom- where decline is expected, behavioral interven- mended in the literature. Formal problem solving tions designed or adapted to treat cognitive dys- was included in the specific strategies taught to the function in patients with degenerative conditions patients to help them decide which of the other such as MCI or Alzheimer’s disease may be more compensation techniques should be applied to a appropriate. problem situation. In many cases, formal problem In general, limited information suggests that solvingisaninternal procedure that is self regulated strategy training, or training to compensate for cog- by the patient (Figure 20.1). However, for patients nitive deficits, may be helpful in neurodegenera- with greater cognitive impairment, this type of strat- tive conditions (e.g., De Vreese et al., 2001). In this egy can be taught in an overlearning procedural section, we will offer detailed examples of compen- paradigm (other regulated) andpromptedbyexter- satory cognitive rehabilitation techniques that have nal materials such as a diagram of the steps of prob- been used in people with Alzheimer’s disease. Since lem solving (external procedure). the primary dysfunction in Alzheimer’s disease is The results of Ownsworth et al. (2000) suggested memory, the interventions are focused on that type significant improvements in a patient’s anticipatory of deficit. The reader is referred to De Vreese et al. awareness of deficits, selection of coping strategies, (2001) for a more comprehensive review of cognitive and effectiveness of selected strategies. Patients rehabilitation techniques in Alzheimer’s disease. 286 Section 3. Interventions and implications for clinical trials

External memory aids schedule on a dry erase board) depending on the severity of the patient’s impairment. The subject External memory aids aim to help patients with sample was small [N = 7), but mean rates of repet- dementia compensate for memory deficits by pro- itive question asking declined from 21.9 per day at viding cues for memory, support for organization baseline to 11.2 per day at the end of 12 weeks of and retrieval of information, and prompting ini- treatment to 8.6 per day at the 6 months follow-up. tiation of activities. For example, Bourgeois et al. (1997) investigated a program for training care- Cognitive rehabilitation versus givers in behavioral management strategies aimed mental stimulation at reducing repetitive question-asking in home- dwelling patients with mild to moderate Alzheimer’s Other investigations of learning and cueing strate- disease. Behavioral systems were individualized for gies that may help patients with Alzheimer’s dis- each patient, but generally consisted of a written ease have identified three promising techniques – cue to help the patients remember the answer to spaced retrieval, dual cognitive support, and pro- their repetitive question, and a series of instruc- cedural memory training (Loewenstein et al., 2004). tions to the caregiver including: (1) deliver the cue- These three cognitive rehabilitation strategies are ing system to the patient, (2) refrain from extra- examples of internal procedures that are other reg- neous verbal directions or explanations other than ulated because of the patient’s limited capacity for to look at or read the cue, (3) praise the patient self-regulation. for using the written cue, and (4) walk away if an Spaced retrieval involves learning associations argument occurs related to the cueing. Cards with over multiple trials at progressively longer time written cues contained pictures or clock drawings intervals and with corrective feedback of any learn- to facilitate understanding. Thus, this type of inter- ing errors. Dual cognitive support uses cues and vention is an external procedure that is other regu- techniques to enhance the saliency of the infor- lated and designed for patients with severe cogni- mation to be remembered at both learning and tive impairment resulting in a very limited capacity retrieval. Procedural memory training involves elic- for self-regulation. iting complex motor behaviors during learning in For example, if a patient repetitively asks about an order to activate the presumed intact procedural upcoming outing, an index card with a brief written memory system in Alzheimer’s disease patients. For reminder of the outing and its time is provided to example, Loewenstein et al. (2004) compared the the patient along with a brief explanation, such as, efficacy of: (1) a cognitive rehabilitation condition “Today we are going to (location) at (time). Here is a utilizing the above three techniques with a mem- card that tells you this so you can remember when ory notebook and (2) a mental stimulation condi- and where we are going. If you forget, look at this tion using computer games involving memory, con- card.” It is recommended that the patient carry the centration, and problem-solving skills in patients card in a pocket. If the patient repeats the question with mild Alzheimer’s dementia. All patients were about the outing, the caregiver is instructed to say, stable on cholinesterase inhibitor medications dur- “Look at your card” and walk away. If the caregiver ing the intervention. All patients received 24 indi- sees the patient reviewing the card without cueing, vidual training sessions of 45 min each over a period the caregiver is instructed to praise the patient and of 12–16 weeks. say, “Good, you are looking at that card that says The cognitive rehabilitation condition utilized a we’ll be going to (location) at (time).” variety of internal and external procedures that Cues were either simplified [e.g., just a simple ori- varied in the degree of self or other regula- entation statement such as “I have lived at (address) tion. This broad-spectrum multi-modal approach for 30 years”] or broadened (e.g., listing a day’s appeared to provide some benefit to most patients. Chapter 20. Behavioral strategies and rehabilitation 287

Ideally interventions could also be more specifi- Among the few published studies of cognitive cally selected for patient needs. The mental stimu- rehabilitation for patients with cancer, one of the lation condition involved computer games requir- most elegant is the ongoing rehabilitation research ing patients to match pairs of letters, numbers, program with pediatric patients designed by But- or designs from memory, exercises such as “hang- ler and Copeland (2002). These clinician investiga- man,” tasks requiring patients to rearrange sets of tors combine a variety of training techniques in their letters to generate as many words as possible, and Cognitive Remediation Program or “CRP.” The pro- asking patients to freely recall information from the gram is described as a combination of: (1) drill- recent or remote past. Results showed that patients oriented practice, (2) learning skills and strategy in the cognitive rehabilitation condition, but not acquisition, and (3) cognitive-behavioral therapy. the mental stimulation condition, improved signif- These methods combine three disciplines: brain icantly in orientation, learning of new face–name injury rehabilitation, special education/educational associations, speed of processing, and specific func- psychology, and clinical psychology. tional abilities. Patients in the cognitive rehabil- The restitution component, utilizing Sohlberg itation condition maintained these gains at the and Mateer’s (1989; Sohlberg et al., 2000) Attention 3 months follow-up while the patients receiving the Process Training, involves massed practice of sus- mental stimulation condition continued to decline. tained, selective, divided, and executive attention skills. The strategy training component includes teaching a range of metacognitive strategies individ- Cognitive rehabilitation for patients ualized to the patient’s particular needs and abil- with cancer ities. Cognitive-behavioral therapy is also used as another form of strategy training aimed at teach- In this section, we will extend the discussion of ing patients to resist distraction. A manual for the cognitive rehabilitation interventions by reviewing complete program is available from the investiga- research on interventions used specifically with tors. This combination of strategies is offered across cancer patients. Several published studies report a approximately 50 h of treatment over a 6-month benefit of general inpatient rehabilitation to adult time period. patients with cancer and brain tumors (Cole et al., Butler and Copeland (2002) reported preliminary 2000; Huang et al., 1998; Marciniak et al., 2001; results of 21 patients receiving the intervention with O’Dell et al., 1998). However, relatively few can- a comparison group of 10 patients who did not cer patients are referred for rehabilitation services receive any intervention. Patients who received the despite significant need (Davies et al., 2003; Kirsh- intervention showed significant improvement on blum et al., 2001; Movsas et al., 2003). Meyers neuropsychological measures of simple attention, and Boake (1993) note that at least 34% of non- sustained vigilance attention, and memory. These CNS cancer patients and essentially all CNS can- findings are encouraging, but the authors note that cer patients develop cognitive deficits during treat- further research is needed to investigate the eco- ment, but that rehabilitation hospitals do not rou- logical validity of this intervention program. The tinely provide cognitive rehabilitation services to program developers report that the program is cur- cancer patients and cancer hospitals do not rou- rently being evaluated in a multi-site phase III clini- tinely provide cognitive or vocational rehabilitation cal trial. services. Empirical studies supporting the effective- Sherer et al. (1997) have offered the only pub- ness of specific rehabilitation interventions in this lished investigation of cognitive rehabilitation for population could enhance referring oncology prac- adult patients. Their trial involved intensive cogni- titioners’ confidence in rehabilitation services, and tive rehabilitation of a small sample of adults with thus increase referrals. primary brain tumors (N = 13). All patients were 288 Section 3. Interventions and implications for clinical trials

post-treatment and relatively stable in their dis- problem-solving therapy showed decreased distress ease. Rehabilitation was conducted in an intensive and increased quality of life. At 1 year, patients who day treatment model that has been empirically sup- received problem-solving therapy with a significant ported for patients with TBI. Patients underwent other were the least distressed and had the high- daily cognitive compensation strategy training for est overall quality of life. No cognitive evaluation or an average of 2.5 months in the context of a general measurement of cognitive symptoms was included day treatment program to improve self-awareness in this trial. of deficits, behavioral self-regulation, adjustment, At the Mayo Clinic, we are conducting a trial and social and vocational participation. The goal of a brief cognitive rehabilitation intervention for was to teach patients techniques to compensate patients with brain tumors. Our primary aim is for cognitive difficulties in order to increase inde- to determine the feasibility and tolerability of a pendence and productivity. The specific interven- combined, tailored, cognitive rehabilitation and tion was multidisciplinary in nature and involved problem-solving therapy intervention with a sample strategies from all quadrants of the model in Figure of patients with brain tumors. The eventual goal 20.1 depending on the need of the specific patient. of investigating this type of compensation-focused The results of that small trial were positive, with six intervention is to increase independence and pro- patients increasing their independence and eight ductivity in patients with brain tumors who are patients increasing their productivity; these gains experiencing cognitive difficulties. were maintained at the 8 months follow-up. In the intervention, patients with primary brain Nezu and colleagues’ (1998, 2003) investigation tumors of any kind who have mild to moderate of problem-solving therapy is also relevant. Their cognitive impairment are given a total of 12 ses- behavioral intervention is primarily aimed to reduce sions of compensation-focused intervention during psychological distress, not cognitive dysfunction, the time they are receiving radiation treatment. A in cancer patients. However, the technique they support person (e.g., spouse or companion) attends describe is very similar to formal problem-solving all training sessions with the patient in order to therapy recommended as a cognitive rehabilitation act as a training partner and to have the knowl- treatment for TBI patients with executive dysfunc- edge to reinforce these strategies outside the treat- tion and therefore might have similar application to ment sessions. Appendices A and B provide more a cancer patient population. details of the interventions, which are adapted from Nezu et al. (1998) outline a 10-session program evidence-based interventions in other populations. to teach patients with cancer effective problem- The memory rehabilitation component involves solving abilities in order to lessen emotional dis- six training sessions over a 2-week period. The spe- tress and improve quality of life. Sessions involve cific intervention involves the development and uti- training in positive problem orientation and in the lization of the type of memory notebook that has four rational problem-solving tasks using didactics, been empirically supported for use in patients with in-session practice, and homework assignments. In TBI (Sohlberg & Mateer, 1989). The problem-solving their book, Nezu et al. (1998) provide a detailed component involves six sessions concurrent with manual for the program. the memory rehabilitation sessions. This part of the Nezu et al. (2003) reported results of a large ran- intervention is an abbreviated adaptation of Nezu domized controlled efficacy trial of problem-solving et al.’s intervention that was empirically supported therapy in distressed adults with cancer. A total of for use with adults with cancer as described above 132 patients were randomized to 1 of 3 groups: (1) (Nezu et al., 1998, 2003). problem-solving therapy, (2) problem-solving ther- It should be apparent from the brevity of this apy with a significant other, and (3) wait list con- review that the literature on cognitive rehabilita- trol. At post-treatment, both groups who received tion in cancer patients is in its infancy. Additional Chapter 20. Behavioral strategies and rehabilitation 289

Table 20.2. Speciﬁc research questions. CVA, cerebral Clinical application of cognitive vascular accidents; TBI, traumatic brain injury rehabilitation strategies in cancer patients

• Is it feasible to provide cognitive rehabilitation as In the sections above, we have tried to make the case patients undergo treatment for cancer? that cognitive rehabilitation with cancer patients • Do the cognitive impairments in cancer patients differ requires innovation on the part of practitioners, qualitatively from the cognitive impairments in TBI/CVA including adapting the knowledge gained from the patients in nature, localization of dysfunction, and literature on other populations including those with response to treatment? • How does the feasibility and effectiveness of cognitive TBI, CVA, and Alzheimer’s disease. In some cases rehabilitation vary with respect to cancer type and among patients with cancer, modifications to the severity? cognitive rehabilitation intervention will be appro- • Would brief, compensation-oriented cognitive priate because the patient’s disease status, recovery rehabilitation be feasible and effective for cancer curve, or prognosis differs from that of the original patients? validation group. • For patients where cure is expected, should cognitive Patients with TBI or CVA are expected to have rehabilitation be postponed until remission is achieved? their worst cognitive deficits at the time of injury • How does the etiology of the cognitive impairment (e.g., and to subsequently improve and become stable. tumor effects vs. radiation effects vs. chemotherapy However, patients with neurodegenerative condi- effects) impact the behavioral intervention plan? tions are expected to have a worsening of cognitive deficits over time as disease progresses. In patients with cancer, impairments may be improving or empirical investigations that identify beneficial stable, similar to TBI/CVA, or progressive, similar cognitive rehabilitation interventions are urgently to a neurodegenerative condition. Prognosis among needed. In Table 20.2, we list some research ques- patients with cancer varies, with some patients tions that could be addressed as this field of inquiry cured of disease and others eventually succumbing moves forward. over varying lengths of time. Thus, the appropriate While a randomized controlled trial may be the cognitive rehabilitation technique for the individual gold standard for the most rigorous research trial cancer patient may be one that has been used with of a medical or rehabilitation intervention, basic patients with an acquired brain injury or acognitive questions of feasibility and appropriate outcome rehabilitation technique that has been shown useful measures would appropriately precede randomized in cases with progressive deficits as in a degenera- controlled trials at this stage. Furthermore, random- tive illness. Practitioner flexibility will be required to ized controlled trials are best applied with interven- apply cognitive rehabilitation strategies to cognitive tions that can be strictly standardized for applica- deficits associated with cancer. tion to very specific deficits. Such strictly controlled The primary impetus for cognitive rehabilitation scenarios are rare in rehabilitation practice where in cancer patients is a complaint from the patient, a patients vary greatly in terms of cognitive profiles, significant other, or from the treating physician that goals, and resources such as stamina and social cognitive status is of concern. We first recommend support; and interventions are often combined or neuropsychological assessment to fully characterize adapted. Useful guidelines for researchers designing the specific type of impairment, severity of impair- cognitive rehabilitation experiments are provided ment, and areas of retained ability. Fully under- by Levine and Downey-Lamb (2002) and in a spe- standing and objectively quantifying this informa- cial supplement to the American Journal of Physi- tion is important for selecting an appropriate cog- cal Medicine and Rehabilitation (Millis & Johnston, nitive rehabilitation approach. For example, for 2003). patients with relatively mild cognitive impairment 290 Section 3. Interventions and implications for clinical trials

Table 20.3. Guided application of cognitive rehabilitation for patients with cancer

Patient or significant other complains of cognitive impairment or behavioral problems • Neuropsychological evaluation to specify type and severity of cognitive behavioral disorder and for recommendations for intervention Is the patient’s cognitive impairment relatively mild? • If yes, first consider an internal compensation technique such as a memory mnemonic, Time Pressure Management training, or problem-solving training as appropriate to the type of cognitive impairment • If no, consider an external compensation technique such as a memory notebook or cue card reminders What is the patient’s prognosis? • <6 months or declining course: – External compensation technique – Education and coaching for significant other to support, prompt, and cue patient • >6 months and stable course: – Self-directed compensation techniques focused on impaired cognitive domains determined by neuropsychometric evaluation What is the status of the patient’s ability to monitor and regulate his/her own behavior? • If poor, consider external regulation of cognitive rehabilitation techniques • If intact, consider internal regulation of cognitive rehabilitation techniques Is cognitive impairment complicated by lack of self-awareness? • If yes and prognosis <6 months or declining course: – Education and coaching for significant other on managing and coping with patient’s cognitive and behavioral problems (i.e., external coping technique that is externally regulated) • If yes and prognosis >6 months with stable course: – Intensive day rehabilitation for brain injury What are patient’s goals? • Enhanced daily functioning and improved quality of life – Environmental modifications only if goals can be achieved without rehabilitation – Cognitive rehabilitation if environmental modifications are inadequate or internal control is desired by patient • Educational or vocational reintegration – Consider a cognitive rehabilitation strategy that is internally regulated by the patient – In addition to cognitive rehabilitation, specific functional rehabilitation including on-site evaluation, environmental modifications, and coaching with gradual increase in time spent at work or school Is cognitive impairment complicated by fatigue? • If yes: – Pharmacological treatment for fatigue – Coaching for pacing activities – Appropriate pacing of rehabilitation intervention Is cognitive impairment complicated by behavioral problems (e.g., disinhibition, abulia)? • If yes: – Pharmacological interventions for behavioral management – Consider externally regulated cognitive rehabilitation techniques and education and coaching for significant other on managing and coping with patient’s behavioral disorder Is cognitive impairment complicated by depression or other emotional disorder? • If yes: – Pharmacological intervention – Cognitive behavior therapy Chapter 20. Behavioral strategies and rehabilitation 291

and relatively intact executive functioning (i.e., abil- In general, we recommend that clinicians offer ity to organize and self-monitor), utilizing an inter- compensation-oriented cognitive rehabilitation nal compensation strategy that can be applied by approaches to patients with cancer, since this type the patients themselves may be a place to start with of technique is most supported by the empiri- cognitive rehabilitation. If cognitive impairment is cal literature in other areas. In addition, within more mild to moderate, but the patient’s execu- compensation-oriented techniques, we recom- tive functioning is relatively intact, an external com- mend considering the dimensions outlined in pensation strategy may be necessary, but could be Figure 20.1 and reiterated throughout this chapter. applied by the patient without external prompt- Choosing specific techniques along the dimensions ing. For more severely impaired patients, or patients outlined in Figure 20.1 will depend on what is for whom behavioral problems are also significant, appropriate to the patient’s impairments, preserved an external compensation technique may need to abilities, cancer prognosis, goals for cognitive be externally regulated. Sutor et al. (2001) suggest rehabilitation, preferences, and available social some of these types of behavioral management support. strategies based on their work with patients with In Table 20.3, we present a series of questions to dementia. guide the clinical application of cognitive rehabili- Another critical consideration, given the differ- tation in patients with cancer. These guidelines are ences between patients with cancer and those with based on clinical experience as well as on empir- TBI or CVA, is the timing of intervention. Many ical evidence from investigations involving other patients with cancer notice cognitive difficulty dur- patient populations. We are optimistic that progress ing the period of diagnosis and treatment. Although in cognitive rehabilitation for persons with cancer delaying cognitive rehabilitation until cancer treat- will continue, and that the quality of life for patients ment is complete may reduce the stress and com- and their families will be enhanced. plexity of care for the patient, this must be weighed against the potential benefit of early intervention. In this sense, the goal of cognitive rehabilitation for Appendix A some types of cancer patients, especially those for whom cure is not expected, may be similar to the Cognitive rehabilitation intervention protocol goal of cognitive rehabilitation in a neurodegener- This intervention is modeled after the techniques ative condition; that is, maximizing functional abil- described by Sohlberg and Mateer (1989). In each ities as much as possible in the face of expected session, some general strategies are used with each decline. patient in order to maximize learning. The patients In TBI/CVA rehabilitation, treatment can be very themselves are encouraged to write in their calen- intensive (multiple hours daily) and may last for dars (instead of the therapist or the support per- several months. This model may be necessary and son). Indirect cueing (i.e., a reminder without spe- appropriate for a select group of patients with can- cific direction) is used whenever possible. Positive cer or brain tumors who have multiple cognitive feedback is used as much as possible. deficits and impaired self-awareness of deficits but who also have relatively low-grade disease with the Session 1 expectation of disease remission or cure. For many r other cancer patients, however, decreased stamina, Explain the purpose of a memory notebook (to more limited deficits, busy cancer treatment sched- compensate for memory, attention, or organiza- ules as well as prognosis issues suggest that tion problems by using this external device) r less intensive, briefer interventions may be more Determine if the patient already uses a calendar appropriate. system 292 Section 3. Interventions and implications for clinical trials

r Introduce our calendar and orient to the format At session 6 – seek commitment from the patient (one page per day, specific times, action list area, and caregiver to using the calendar post-treatment date at top) r Have patient and caregiver verbalize difficulties Appendix B they notice r Strategize using the calendar for one difficulty Problem-solving therapy intervention protocol (e.g., remembering medications, entering doctor’s appointments, remembering other planned out- This intervention is modeled after the techniques ings, reducing repeated asking or questions) and described by Nezu et al. (1998, 2003). The interven- agree to check off items as they are completed tion is six 50-min sessions over a 2-week period. The patient concurrently receives six sessions of cogni- Session 2 tive rehabilitation from another provider. r Review orientation to the calendar with patient Session 1 leading the way (where the date is located, one day r per page, specific times versus action list area) Explain problem-solving model of stress and a r Review patient’s acquisition of using the calendar brain tumor as a major life stressor r for their identified area of difficulty Present goals of problem-solving therapy r r Review checking off of completed items Present four components of a positive problem- r Agree to expand the calendar for use with other solving orientation areas (therapist should help indicate what should go in the calendar – upcoming medical appoint- Session 2 ments, medication schedule, upcoming family r events, etc.) More detailed development of positive problem- solving orientation using ABC method of con- structive thinking, reversed advocacy role play, Session 3 and using feelings as cues r r Check with patient and caregiver regarding spon- Briefly, present the steps of problem solving: taneous use of the calendar outside the session defining the problem, generating alternative solu- r Review calendar for the previous day, today, and tions, decide on a solution strategy, implement next day for: (1) entries and (2) notation of the solution, review the outcome of that imple- completion mentation r r Ask three questions for the patient to answer Present categories and specific potential tumor- (what is today’s date, when is your next medical related problems (e.g., side-effects, psychological appointment, and did you take your medications distress, marital and family, medical interaction, yesterday?) sexual)

Sessions 4–6 Session 3 r r Review use of calendar and indications of comple- Detailed application of problem-solving steps to tion since the last session example problem r r Ask three patient-speciﬁc questions they would Patient and support person choose a problem in need the calendar to answer the session to which they can apply the steps r r Ask patient and caregiver for additional difﬁcul- Patient and support person will practice the steps ties and see if they can be incorporated into using on chosen problem outside of the session before the calendar session 4 Chapter 20. Behavioral strategies and rehabilitation 293

Sessions 4–6 ing after closed head injury. Neuropsychol Rehabil 10: r 47–65. Review patient’s and support person’s use of the Huang ME, Cifu DX, Keyser-Marus L (1998). Functional strategies r outcome after brain tumor and acute stroke: a compar- Review positive problem-solving orientation or ative analysis. Arc Phys Med Rehabil 79: 1386–1390. any of the steps of problem solving if necessary Kirshblum S, O’Dell MW, Ho C, Barr K (2001). Rehabilita- r Troubleshoot any problems using the steps tion of persons with central nervous system cancer. Can- r Continued practice of the techniques in session cer Suppl 92: 1029–1038. r Continued reﬁnement of patient’s and support Kuhn GK, Boesen E, Ross L, Johansen C (2005). Evaluation person’s application of the steps and outcome of behavioural changes in the rehabilitation of cancer patients: a review. Eur J Cancer 41: 216– 224. Levine B, Downey-Lamb MM (2002). Design and evalua- REFERENCES tion of rehabilitation experiments. In Eslinger PJ (ed.) Neuropsychological Interventions: Clinical Research and Bourgeois MS, Burgio LD, Schulz R, Beach S, Palmer Practice. New York: The Guilford Press. B (1997). Modifying repetitive verbalizations of Loewenstein DA, Acevedo A, Czaja SJ, Duara R (2004). Cog- community-dwelling patients with AD. Gerontologist 37: nitive rehabilitation of mildly impaired Alzheimer’s dis- 30–39. ease patients on cholinesterase inhibitors. J Geriatr Psy- Butler RW, Copeland DR (2002). Attentional processes and chiatry 12: 395–402. their remediation in children treated for cancer: a lit- Malec JF, Cicerone KD (2006). Cognitive rehabilitation. In erature review and the development of a therapeutic Evans RW (ed.) Neurology and Trauma (2nd edn.) (pp. approach. J Intl Neuropsychol Soc 8: 115–124. 238–261). New York: Oxford University Press. Cicerone KD, Dahlberg C, Kalmar K et al. (2000). Evidence- Marciniak CM, Sliwa JA, Heinemann AW, Semik PE (2001). based cognitive rehabilitation: recommendations for Functional outcomes of persons with brain tumors after clinical practice. Arch Phys Med Rehabil 81: 1596–1615. inpatient rehabilitation. Arch Phys Med Rehabil 82: 457– Cicerone KD, Dahlberg C, Malec JF et al. (2005). Evidence- 463. based cognitive rehabilitation: updated review of the lit- Meyers CA, Boake C (1993). Neurobehavioral disorders in erature from 1998–2002. Arch Phys Med Rehabil 86: 1681– brain tumor patients: rehabilitation strategies. Cancer 1692. Bull 45: 362–364. Clare L, Woods RT, Moniz Cook ED, Orrell M, Spector A Millis S, Johnston M (eds.) (2003). Clinical trials in medical (2005). Cognitive rehabilitation and cognitive training for rehabilitation: enhancing rigor and relevance. Am J Phys early-stage Alzheimer’s disease and vascular dementia. Med Rehabil 82 [Suppl. 1]. Cochrane Library 2. Movsas SB, Chang VT, Tunkel RS, Vipul SV, Ryan LS, Millis Clark MM, Bostwick JM, Rummans TA (2003). Group and SR (2003). Rehabilitation needs of an inpatient medical individual treatment strategies for distress in cancer oncology unit. Arch Phys Med Rehabil 84: 1642–1646. patients. Mayo Clin Proc 78: 1538–1543. Nezu AM, Nezu CM, Friedman SH, Faddis S, Houts PS Cole RP, Scialla SJ, Bednarz L (2000). Functional recovery (1998). Helping Cancer Patients Cope. Washington DC: in cancer rehabilitation. Arch Phys Med Rehabil 81: 623– American Psychological Association. 627. Nezu AM, Nezu CM, Felgoise SH, McClure KS, Houts PS Davies E, Hall S, Clark, C. (2003). Two year survival after (2003). Project genesis: assessing the efﬁcacy of prob- malignant cerebral glioma: patient and relative reports of lem solving therapy for distressed adult cancer patients. handicap, psychiatric symptoms, and rehabilitation. Dis- J Consult Clin Psychol 71: 1036–1048. abil Rehabil 25: 259–266. NIH Consensus Statement Online 1998 October 26–28. De Vreese LP, Neri M, Fioravanti M, Belloi L, Zanetti O Rehabilitation of Persons With Traumatic Brain Injury. (2001). Memory rehabilitation in Alzheimer’s disease: a NIH Consensus Statement 1998 Oct 26–28; 16(1): 1–41. review of progress. Int J Geriatr Psychiatry, 16: 794–806. O’Dell MW, Barr K, Spanier D, Warnick RE (1998). Func- Fasotti L, Kovacs F, Eling PATM, Brouwer WH (2000). Time tional outcome of inpatient rehabilitation in persons pressure management as a compensatory strategy train- with brain tumors. Arch Phys Med Rehabil 79: 1530–1534. 294 Section 3. Interventions and implications for clinical trials

Ownsworth TL, McFarland K (1999). Memory remediation Sohlberg MM, Mateer CA (1989). Training use of com- in long-term acquired brain injury: two approaches in pensatory memory books: a three stage behavioral diary training. Brain Injury 13: 605–626. approach. J Clin Exp Neuropsychol 11: 871–891. Ownsworth TL, McFarland K, Young RM (2000). Self- Sohlberg MM, McLaughlin KA, Pavese A, Heidrich A, Pos- awareness and psychosocial functioning following ner MI (2000). Evaluation of attention process training acquired brain injury: an evaluation of a group support and brain injury education in persons with acquired programme. Neuropsychol Rehabil 10: 465–484. brain injury. J Clin Exp Neuropsychol 22: 656–676. Ronson A, Body JJ (2002). Psychosocial rehabilitation of Sutor B, Rummans TA, Smith GE (2001). Management of cancer patients after curative therapy. Support Care Can- behavioral disturbances in nursing home patients with cer 10: 281–291. dementia. Mayo Clin Proc 76: 540–550. Sherer M, Meyers CA, Bergloff P (1997). Efﬁcacy of post- Von Cramen DY, Mathes-Von Cramen MN (1991). Problem acute brain injury rehabilitation for patients with pri- solving deﬁcits in brain injured patients: a therapeutic mary malignant brain tumors. Cancer 80: 250–257. approach. Neuropsychol Rehabil 1: 45–64. 21

Support services

Bebe Guill and Renee H. Raynor

Introduction tional, and practical assistance as needed, and by appropriately fostering a sense of hope or opti- Whether the primary treatment approach to mism. Here, we refer to a wide range of strategies cognitive impairment in cancer is remedia- designed to improve emotional and social adjust- tion/rehabilitation, pharmacotherapy, or a combi- ment and functioning, increase coping, assist with nation, one must not underestimate the importance decision-making, and minimize distress. Each indi- of comprehensive support services throughout the vidual patient’s and family’s need for support will illness continuum. be unique, and will depend on a number of factors, Cognitive deficits related to cancer may be pri- including the amount of stress present in the family mary, related to the disease entity itself, or may be prior to the illness, the amount of support available secondary, related to the various methods used to from friends and family, and the patient’s emotional treat the cancer. Such deficits may also be related and medical response to treatment. to direct and indirect co-morbidities of the can- Effective support takes place in a variety of con- cer and treatments. Among the most common co- texts and from a variety of sources – both formal and morbidities are mood disturbances (e.g., depres- informal. Best support practices have been identi- sion, anxiety) and fatigue. Fatigue is the most fied as those that ensure continuity of care, involve widely reported deleterious symptom in adult can- all members of the treatment team, and match sup- cer patients (Valentine & Meyers, 2001). Addition- port services to the unique needs of each patient ally, in childhood cancer survivors, fatigue and and family at every phase of the illness continuum “aches and pains” are reported as most problem- (Clinical Practice Guidelines in Oncology, 2008). atic relative to other symptoms (Zebrack & Chesler, Support services should be considered a vital com- 2002). In this chapter, we will consider cognitive ponent, not an optional extra, of care for persons deficits, mood disturbances or emotional distress, with cancer. and fatigue as highly inter-related symptoms of can- Sensitivity to the various support needs of cer and will discuss support services that may apply patients is the responsibility of all members of the to one or more of these conditions in isolation or in interdisciplinary treatment team – not just those combination. in the mental health fields. Physicians and nurses Support has many definitions, but is generally are important primary sources of support; how- understood as strengthening the patient’s and fam- ever, successful management of the cancer patient’s ily’s resources by providing emotional, informa- health care is best accomplished by a concerted

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

295 296 Section 3. Interventions and implications for clinical trials

interdisciplinary effort, extending far beyond the their care (Sanson-Fisher et al., 2000). The pri- core medical team. This interdisciplinary team mary oncology team of oncologist, nurse, and social approach requires that team members not only worker should formally assess and regularly mon- have expertise in their respective disciplines, but itor the emotional, informational, and practical must also be able to contribute to the group effort needs of patients with cancer and their families. For on behalf of the patient and the family. patients with cognitive impairment, it is essential to Numerous studies have demonstrated the impor- check the extent of support available to the patient tance of effective support in improved quality of and family, including ascertaining needs for assis- life for patients with cancer. The extent to which a tance with practical issues such as transportation, person with cancer has support and feels supported childcare, work, or school, and exploring how the has been identified as a major factor in adjustment patient and family are coping with any issues related to the disease. Support services such as psycho- to cognition. logical therapies have been shown to improve Taking the initiative in inquiring about residual emotional adjustment and social functioning, symptoms and concerns is important for health and reduce both treatment- and disease-related care professionals. Indeed, studies have shown distress in patients with cancer. The efficacy of both that many patients will not raise their concerns supportive and cognitive-behavioral therapies in unless this is explicitly invited (Bertakis et al., 1991; the treatment of depressive disorders in cancer Maguire, 1999). In their study of untreated distress patients, both in individual and in group therapies, in cancer patients, Carlson et al. (2004) reported that has also been demonstrated (Devine & Westlake, approximately half of all patients who met distress 1995; Sheard & Maguire, 1996, 1999). A meta- criteria had not accessed psychosocial support ser- analysis of 116 intervention studies found that vices, despite being aware of their availability. When patients with cancer receiving psycho-educational queried as to their reason for not seeking such ser- or psychosocial interventions showed much lower vices, 44.1% of these patients (who were character- rates of anxiety, depression, mood disorders, ized as distressed) reported a self-perception of not nausea, vomiting and pain, and significantly greater needing any help. knowledge about disease and treatment, than the Two assessment tools have proved to be both clin- control group (Devine & Westlake, 1995). ically sensitive and easy to use for early identifi- Distinct needs for support for both patients and cation of emotional and quality-of-life issues. The their families result from cognitive impairment Beck Depression Inventory-II (BDI-II) is a 21-item related to cancer and its treatment. By understand- inventory that conforms to the depression criteria ing the importance and appropriate use of support of the Diagnostic and Statistical Manual of Men- services, health professionals can help to reduce tal Health Disorders (4thedn.)(DSM-IV).Itisa patient and family distress, restore a sense of con- widely used instrument for detecting depression trol to patients and caregivers, and improve health- in adults and adolescents and takes approximately related quality of life. In this chapter, we consider 5 minutes to complete. The BDI-II has been demon- the use of needs assessment, goals for support ser- strated to be useful in screening for depression in vices across the illness continuum, relevant con- cognitively impaired cancer patients in the clinical texts/settings, modalities for delivery of support ser- setting (Allen et al., 2003). A second instrument, the vices, and identification of appropriate resources. National Comprehensive Cancer Network’s Distress Thermometer and Problem List, uses a scale of 0–10 that can be completed quickly and consists of a list Needs assessment of problems that the patient reads, indicating pos- A high proportion of people with cancer may have sible reasons for the distress. Used in conjunction unmet needs, despite expressing satisfaction with with clinical assessment by the primary oncology Chapter 21. Support services 297

team, such tools can be a valuable means for eval- have cognitive deficits due to the cancer itself, uating depression, anxiety, and other quality-of-life surgical interventions, medications, early radia- symptoms in the clinical setting (NCCN Standards tion/chemotherapy treatment or associated psy- of Care for Distress Management, 2008) and repre- chological distress related to fear of treatment or sent a time- and cost-effective method for inform- death. Psycho-education has been shown to be the ing decisions about what support services will be most appropriate modality of support in the early most appropriate. stage of the disease, just after diagnosis and pre- Being prepared to make recommendations about treatment, when the patient’s need for informa- support services and where these may be avail- tion appears to be at its highest (Carlson et al., able is also important (Ell et al., 1989). Patients 2004). frequently request information from their primary Additional needs for intervention in this early oncology team about support and educational ser- phase may involve supportive psychotherapy to vices. Matthews et al. (2004) surveyed a large mul- help cope with diagnosis, to prepare the patient tidisciplinary sample of oncology health care pro- for the treatment road ahead, and/or to begin to fessionals and learned that 94% of them had been process the many changes that the diagnosis and asked about at least one cancer-related support treatment will bring. They may already have appre- service by their patients. Of the referrals sought ciable cognitive deficits for which they would bene- by patients in this study, 72% of inquiries were fit from early cognitive rehabilitation, occupational related to information and education about can- therapy, speech therapy, and/or physical therapy. cer, 65% were about support groups, and 52% were Emotional lability may be high as a result of situ- about hospice referrals. Licensed mental health pro- ational stress or due to metabolic or medication- fessionals and certified pastoral caregivers experi- related psychiatric symptoms. This sense of emo- enced in psychosocial aspects of cancer and cog- tional upheaval can leave a patient feeling out of nitive impairment should be readily available as control and desperate. Proposing a rehabilitative staff members or by referral (Zebrack & Chesler, approach for acutely post-operative brain tumor 2002). patients, Gabanelli (2005) points out that, in some cases, medical staff members are not familiar with the common cognitive and emotional sequelae of Support needs across the disease continuum cancer and may misinterpret psychological distress Support needs and targeted outcomes of services as evidence of mental deterioration. Such a reaction will vary depending upon the patient’s stage of ill- may leave the patient feeling shamed or patronized. ness/recovery, type of treatment, and the degree The newly diagnosed patient may benefit from of cognitive deficits experienced. Services should early exposure to support groups to gain support address primary cognitive deficits, psychologi- and insight from fellow patients. Consultation and cal/psychiatric distress related to the disease and/or guidance with a spiritual or religious advisor may the cognitive deficits, and coping/adjustment chal- provide additional help and perspective. Addition- lenges related to the losses/changes in the patient’s ally, the patient may seek out complementary and life. alternative therapies to treat not only the cancer, but also its associated impacts on quality of life. Goals for support may include crisis intervention, stabil- Newly diagnosed/pre-treatment ization of the patient’s mood, mobilizing support The newly diagnosed patient may still be in “shock” resources including family, friends, and appropriate at the diagnosis of cancer, in denial that they are referrals, and addressing practical concerns such as having cognitive symptoms, or relieved to have an lodging, childcare, or navigating the challenges of “explanation” for their symptoms. He or she may insurance and financial concerns. 298 Section 3. Interventions and implications for clinical trials

Active treatment there may be a tendency to expect distress levels to reduce once a treatment plan has begun and During the treatment for cancer, the patient may the “shock” of the cancer diagnosis has passed, experience fatigue and cognitive deficits as a result evidence suggests that many patients remain dis- of radiation therapy, chemotherapy, hormonal ther- tressed throughout the treatment phase and into apies, or adjuvant medications. This may be the first long-term follow-up. Carlson et al. (2004) reported time the patient realizes the systemic impact of can- that a significant percentage of a large heteroge- cer treatment. Cumulative effects of months of ther- neous cancer patient population reported clini- apy may make deficits more pronounced or intol- cally relevant levels of distress across the disease erable. Prolonged changes in cognition, mood, or continuum. Symptoms may range from “situa- behavior need to be evaluated and treated. tional” or “reactive” depression to major depres- Rehabilitation services can be critically impor- sion, a condition characterized by a constellation tant for maintaining maximum functioning dur- of signs and symptoms meeting criteria for clin- ing the treatment phase for cancer patients. His- ical diagnosis according to accepted psychiatric torically, cancer patients have not been referred to standards. Angelino and Treisman (2001) refer to intensive rehabilitation centers because of concerns “situational” depression as demoralization and about reduced life expectancy, pain, and multiple describe it as a normal psychological reaction to medical complications (DeLisa, 2001). For patients life stresses that may not imply brain pathology. with neurological injury (stroke, traumatic brain They characterize major depression, with its impact injury, or spinal cord injury) and patients with med- on the patient’s mood and self view, along with its ically related debilitation, rehabilitation services anhedonia and neurovegetative symptoms, as a have long been valued by patients and health pro- manifestation of organic brain dysfunction. They fessionals as the treatment of choice for optimizing recommend antidepressant therapy for major function and independence. While cancer patients depression in cancer patients and suggest the often have similar needs and capabilities to benefit usefulness of psychotherapy in the treatment of from rehabilitative therapies, the anticancer treat- demoralization. Best clinical practice in the man- ment has many times remained the sole treatment agement of depression suggests that a combination focus (Cheville, 2001). of psychotherapy and antidepressant drug therapy However, the benefits of rehabilitation services for has shown better treatment outcomes than use of cancer patients have been demonstrated in multiple either modality alone (Sutherland et al., 2003). In large-scale, well-designed studies. DeLisa (2001) a meta-analysis of 45 randomized trials of cancer reviewed relevant literature and concluded that patients receiving psychological interventions, an there was both a pressing need for rehabilitation average of 12% showed significant improvement in in cancer patients and convincing evidence of the measures of emotional adjustment, 10% in social value it may have in their functional status and qual- functioning, 14% in treatment- and disease-related ity of life. Garrard et al. (2004) reported on a sample symptoms, and 14% in overall quality of life, com- of 21 patients with primary and non-primary cen- pared to patients not undergoing psychological tral nervous system (CNS) malignancies who under- therapies (Meyer & Mark, 1995). Goals for sup- went inpatient rehabilitation. They found that these portive therapy may include balancing hope with patients were responsive to rehabilitative interven- realistic expectations, addressing fear of recurrence tions and showed functional improvement regard- and death, dealing with treatment burnout, and less of the type of malignancy. addressing the desire for normalcy. In addition to deficits associated with treatment, Additionally, patients may struggle with identity many patients will struggle with emotional diffi- and self-esteem concerns because of changes in culties during the active phase of treatment. While their family role (e.g., no longer the breadwinner, no Chapter 21. Support services 299

longer the caregiver). Family members and friends after months or years of aggressive and toxic can- may need education on how to identify cognitive cer treatments. Rarely, however, does life after can- and/or mood changes and may need help adjust- cer return to pre-diagnosis “normal.” Recognition ing and coping with these changes in their loved that life may never be the same can be troubling for one. many patients and families. Further needs for intervention during the active Fatigue is an almost universal side-effect of treatment phase may involve practical life concerns. cancer and its related treatment. In many cases, For example, the patient may need assistance as it persists well after treatment is discontinued he or she starts to realize that they may not be (Valentine & Meyers, 2001). Disturbances in cog- able to continue work/school during the treatment nition and mood are frequently seen in post- phase. Many patients have unrealistic expectations treatment cancer patients, likely a result of aggres- for recovery following diagnosis and early interven- sive systemic antineoplastic treatments aimed at tions and may need help recognizing and accept- combating the cancer, but which may also dam- ing that they may not return to work right away. In age vulnerable CNS tissues (Meyers, 2000). A recent a recent examination of employment pathways in evaluation of 10-year survivors of childhood medul- cancer patients, it was reported that 41% of males loblastoma showed significant impairment in cog- and 39% of females who were working at diagno- nitive abilities and psychosocial domains, including sis stopped during cancer treatment (Short et al., employment, driving, education, independent liv- 2005). In a separate review examining return to work ing, and dating (Maddrey et al., 2005). practices in cancer patients, Spelten et al. (2003) Patients may question why they still feel poorly or reported a median number of 278 days on sick leave still have trouble with their thinking and function- at 12 months follow-up, with a range of 3–652 days. ing. They may struggle with loss of identity due to Fatigue levels predicted the return to work in this an inability to function in the same capacity as they study. did prior to the cancer diagnosis. Behavioral distur- It may be impossible for the patient to appreci- bances may start to wear on family members when ate the large impact that fatigue will have on their treatment is over but the patient does not return to functioning until they are well into active treatment. his or her baseline personality. As a result, caregivers At that point, help in navigating the paperwork and family members may struggle with the loss required to apply for disability may be required. of their previous relationship with the patient and Similarly, younger patients may need help arranging this may have implications for the future of family hospital-based or in-home schooling. roles. The patient may not be able to return to work Further practical needs during the active treat- or may be unable to continue working because of ment phase might involve assistance with the cancer-related disability. Short et al. (2005) reported bureaucracy of health insurance companies, decid- that approximately 13% of a large cohort of can- ing and executing health care and legal powers of cer survivors had stopped working because of rea- attorney, and managing the financial burden that sons related to their cancer within 4 years of diagno- extended cancer therapy invariably causes. sis. They also found that among survivors who went back to work during the first year, 11% quit due to cancer-related reasons in the next 3 years. Clearly, Post-treatment stable disease the impact of cancer on work and career is some- In many ways, the end of active cancer treatment, times long-lasting. with the outcome of stable disease or remission, Goals for supportive care during this phase in is a joyous time. Patients may finally have hope of the cancer patient’s life may involve education survival and may begin to return to pre-diagnosis about the long-term cognitive effects of cancer activities. Many will begin to feel healthy again and its treatments and providing support as the 300 Section 3. Interventions and implications for clinical trials

patient learns to adjust to his or her disability and life, pharmacological interventions, such as the use seeks a new normal. The patient may need con- of psychostimulants, have proven useful. Gagnon crete help with disability or vocational rehabilita- et al. (2005) showed improvement in cognitive and tion issues. Caregivers may need support in accept- psychomotor activities in advanced cancer patients ing the patient’s new post-cancer role and what that with hypoactive delirium after treatment with low means for their future relationship. The goals for doses of methylphenidate. Similarly, Rozans et al. intervention may shift from rehabilitation to com- (2002) found the use of methylphenidate beneficial pensation and acceptance. Modalities for support in combating opioid-induced somnolence and in services in this phase may include vocational reha- improving cognitive functioning in late-stage can- bilitation services, cancer survivor support groups, cer patients. and individual/family psychotherapy. Hospice agencies are the professional groups most often involved in providing care to patients and families at the end of life. Involvement of a hos- End of life/palliative care pice may begin soon after diagnosis or may be initi- The support needs of the cancer patient who is fac- ated later in the disease continuum, when treatment ing the end of life either because of treatment fail- fails or when decisions are made regarding preserv- ure or a conscious decision to discontinue treating quality of life instead of continuing potentially ment are numerous. They may have physical needs harsh, invasive anticancer therapies. There is some- to be addressed by medical intervention (e.g., pain, times resistance among patients or families to begin fatigue) and they may benefit from rehabilitation hospice care due to concerns that the patient will therapies (e.g., physical, occupational, and speech) just be “waiting to die.” On the contrary, the very to maximize their functioning and communication foundation of hospice care is built on the desire to in the last portion of their life. They may also ben- help patients live out their remaining life with as efit from supportive psychotherapy, either in indi- much vitality and dignity as is possible. Manage- vidual or group format, to process the many feelings ment of symptoms such as pain and mood distur- associated with end-of-life issues. Spiritual and exis- bance is a specialty of hospice care. Hospice staff tential issues may become paramount to the indi- are dedicated to helping patients remain as alert, vidual in this stage of the disease continuum and engaged, and communicative as possible at the end provision of support in these areas of concern may of life, so that they can have quality time with fam- be best provided by members of the clergy or by hos- ily and friends. In this regard, hospice staff, work- pice staff. ing in conjunction with and as an extension of the Historically, when a patient enters this stage of his patient’s primary treatment team, can assess and or her disease trajectory, interventions to preserve maximize the patient’s cognitive abilities in order cognition and mood have been seen as superflu- to provide optimum quality of life in the remaining ous or of lesser importance than the medical ele- days of life. ments of comfort care (e.g., pain control). Garrard et al. (2004) make a convincing argument for com- Support settings/contexts bining the approaches of rehabilitation and palliation in the care of advanced cancer patients in Cancer survivors and their families face specific order to maintain symptom management and to challenges as they re-negotiate roles and relation- address psychosocial needs related to end of life, ships that are necessary for successful integration while also maximizing the patient’s level of func- back into school, work, family, and community set- tioning during the remaining days of life. In an effort tings. Effective support by the primary treatment to maximize cognitive and physical functioning in team can promote optimal coping and adjustment cancer patients, especially those near the end of during these transition periods. Chapter 21. Support services 301

Community re-entry/integration re-entry programs should initially include educating the teachers and/or counselors about the At some point during or after successful cancer cognitive effects associated with cancer and its treatment, the patient may attempt to move back treatment, and the specific signs, symptoms, and into their previous life roles. When there are sig- special needs associated with the patient’s treat- nificant changes in cognitive ability, the degree ment and treatment outcome (Leigh & Miles, 2002). to which patients can realistically do so may be Additional components typically include educating affected. The reactions of others to their differ- the child’s peers about cancer and its effects by ent cognitive capabilities may be unexpected and means of age-appropriate didactic materials and can cause discomfort for the patient, their family, class discussions, and by dispelling myths relating co-workers, friends, and acquaintances. It is vitally to the disease. important that patients have support in determin- Ongoing liaison with the school at regular inter- ing if and how they should resume previously held vals is important for many reasons. First, the neu- roles. Appropriate support can uncover unrealistic ropsychological status of the child may not remain expectations, manage anxiety, ease awkwardness, stable after treatment has been completed. The prevent embarrassment, and monitor progress. In onset of some deficits is delayed and others are the event that the transition is unsuccessful or not evident until the ability is normally expected incomplete, relevant support can identify appro- (Armstrong et al., 1999). In addition, most children priate modifications or help the patient begin to have multiple teachers who will change with each accept that the resumption of the previous role is school year. not attainable. For the child with cognitive deficits, ongoing support requiring the attention of the interdisciplinary treatment team may include identifying an appro- School settings priate educational course (which may be different During the early treatment period, the child with than what was expected or previously engaged in); cancer may require some hospital and/or home- negotiating in concert with parents and the school bound instruction. However, school attendance for special services and/or considerations for the should be the goal as soon as the pediatric oncol- child; and referral to specialists who can help the ogist considers the child physically able to attend. student develop strategies for coping in the educa- The continuation of schooling provides the child tional environment. with hope and stability, because attending school is Specialists such as pediatric psychologists, what normally developing children do daily (Deasy- speech, occupational, and physical therapists, hos- Spinetta & Spinetta, 1980). pitalsocialworkers,andeducationalconsultants It has been recommended that all pediatric oncol- and advocates may help the child in a variety of ogy centers have a structured school re-entry pro- ways. These include: helping the child develop gram for students who have undergone treatment compensatory strategies to minimize effects of for cancer (Deasy-Spinetta & Spinetta, 1980). Com- lowered energy levels and/or cognitive deficits; munication with the student’s teachers and coun- providing technological assistance and training; selors is paramount in such programs. One of the offering special education support; advocating for greatest dangers in not communicating with the special needs within the educational system; or patient’s teachers is that if the child has subsequent focusing on social skill development. struggles in the classroom, these can be falsely Special education services for children aged attributed to attitude problems, daydreaming, a 3–21 years who attend public school and have a lack of motivation, or emotional maladjustment documented need are mandated (in the US) by the (Butler & Mulhern, 2005). Components of school federal Individuals with Disabilities Education Act 302 Section 3. Interventions and implications for clinical trials

(IDEA). Many children undergoing treatment for Once the patient’s range of ability and extent cancer as well as those who have completed treat- of cognitive deficits have been assessed, support ment will be classified as “other health impaired” should include a liaison with the workplace to by their local school systems as a means of access- ensure informed supervision and the on-site pres- ing resources for their special needs. The Individu- ence of a supervisor who knows how the conse- alized Education Plan (IEP) is a legal document that quences of cognitive impairment affect the worker’s records the determination by school personnel, par- ability to perform duties successfully. Modifica- ents, and members of the health care team of what tions of the job responsibilities, such as alterna- services the child needs and how those services tive work schedules or job restructuring, may also will be provided. These services can include occu- be required. A neuropsychological evaluation can pational therapy, physical therapy, speech therapy, be helpful in making these determinations about a special teacher for visually impaired or hearing- abilities and in assisting the employer to identify impaired children, placement in a special educa- and implement possible job role modifications to tion resource room for all or part of the school accommodate their employee. On-the-job behav- day, and/or an aide to assist the child. Classroom ioral counseling by a job coach and education of accommodations might include things such as oral co-workers are also recommended. Practical issues, tests, reduced workload, extra time for tests or such as transportation to and from the workplace (if other tasks, and the use of tape recorders or note driving is an issue), should also be addressed. The takers. benefit of support is strongest when combined with The IDEA does not apply to colleges and univer- supported employment involving a job coach, an sities, but other laws (for example, the Americans interdisciplinary team, and appropriate on- and off- with Disabilities Act and the Rehabilitation Act of site support (Hall & Cope, 1995). 1973) require these institutions to provide special services to students with disabilities, including cog- Family settings nitive impairment. The experience of cancer and cognitive impairment affects not only the patient but also the family (Blue- Work settings glass, 1991). Research with adult cancer survivors Although many cancer survivors will return to work and their families suggests that families are vul- after their diagnosis, the rate of employment among nerable to distress in the patient (Compas et al., survivors is lower than among people without a his- 1994; Nijboer et al., 2000; Welch et al., 1996; Ybema tory of cancer. For working-age cancer survivors, et al., 2001), that children of patients with cancer interruption of employment and diminished capac- may be in particular need of support, and that dis- ity to work are serious consequences of cancer and tress levels of partners of cancer patients may be treatment that can have economic, psychological, considerable, sometimes even higher than that of and social implications. Cognitive deficits, includ- the patients themselves, but they receive less sup- ing difficulty concentrating, learning new things, port (Baider & Denour, 1999; Cliff & McDonagh, analyzing data, and keeping up with the pace set by 2000; Northouse et al., 2000). One study reports others (Bradley & Bednarek, 2002), are cited as rea- that approximately 20%–30% of spouses of cancer sons for reduced work effort and adverse economic patients suffer from psychological impairment and outcomes (Chirikos et al., 2002). Job discrimination, mood disturbance (Blanchard et al., 1997). In addi- difficulties in obtaining work, and subsequent diffi- tion to increased symptoms of depression, anxi- culties in obtaining health and life insurance have ety, and psychosomatic symptoms, research sug- been reported in studies with childhood cancer sur- gests that family caregivers experience restriction of vivors (Langeveld et al., 2002). roles and activities, strain in marital relationships, Chapter 21. Support services 303

severe sleep problems, and diminished physical limited and conflicting results. Some studies sug- health (Carter & Chang, 2000; Johnson, 1988; Nort- gest that parents of survivors show adequate lev- house, 1988; Oberst et al., 1989). In their study com- els of adjustment (Frank et al., 2001), while others paring the familial impact of mental illness (includ- argue that parents of childhood cancer survivors ing dementia) to other common chronic conditions show high rates of continued distress (Sloper, 2000; (including cancer), Holmes and Deb (2003) found Van Dongen-Melman et al., 1995). Studies of par- that brain-related conditions impose the most sig- ent adjustment and stress have identified such nificant risk to the psychological well-being of fam- areas as grief, uncertainty, and the experience of ily members. post-traumatic stress disorder as being important Caregivers’ depression and perceived burden, (Bonner et al., 2006; Brown et al., 2003; Kazak et al., including financial burden, have been shown to 1997, 2004; Stewart & Mishel, 2000; Van Dongen- increase as patients’ functional status declines Melman et al., 1995, 1998). Other data have sug- (Covinsky, 1994). Alternatively, economic distress gested that parental stress and distress are ongoing, not only directly increases the chance that fam- often lasting well beyond the cancer survivor’schild- ily members will experience emotional distress, but hood and continuing into adulthood (Ressler et al., it also appears to reduce the family’s ability as a 2003; Svavarsdottir, 2005). whole to cope psychologically with chronic illness The ongoing feelings of stress and distress that (Holmes & Deb, 2003). Therefore, the family’s abil- parents of childhood cancer survivors experience ity to provide needed care for the patient may be many years after treatment make them a critical impaired (Cassileth et al., 1985; Given et al., 1993; group to assist. Indeed, as Butler and Mulhern Nijboer et al., 1999). (2005) have observed, “the family environment Studies of childhood cancer survivors have exam- might be of equal or greater importance in the treat- ined many salient issues, including cognition, social ment and recovery of a chronic life-threatening dis- functioning, and post-traumatic stress syndrome. ease compared with an acute event such as a trau- Residual effects of the disease and its treatment, matic brain injury.” coupled with the potential for newly emerging late Care and support for the family caregiver in con- effects over time and/or disease recurrence, may junction with treatment of the patient has been continue to be a source of stress not only for the increasingly acknowledged as essential (Svavarsdot- patient but also for the family. Since parents serve tir, 2005). Clinicians are advised that they should not as the primary caregivers and decision-makers for assume that the family is able to offer the patient their children, parental adaptation to the cancer the support they need and should routinely assess experience will almost certainly have an impact on and monitor the support needs of all members of the cancer patient’s adjustment and quality of life, the patient’s family. Providers should be especially as well as those of any other siblings or family mem- watchful when their patients with cognitive deficits bers. Many families, in fact, report that they never come from families with limited financial resources return to where their family was before diagnosis, and inadequate insurance coverage. but instead are forced to find a new “normal” (Van Dongen-Melman et al., 1995). Recognizing that the Modalities of support consequences of chronic or life-threatening illness in childhood concern not only the child, but also the The delivery of support services to cancer patients family (Kazak et al., 1997), investigators have begun may be through any number of modalities, depend- to focus attention on the impact on parents. ing upon the nature of the support, the point along The body of research on parents of childhood can- the disease continuum at which the support is cer survivors has, to date, focused almost exclusively provided, and the preference/learning style of the on psychosocial adjustment, and has yielded both patient or caregiver receiving the services. Support 304 Section 3. Interventions and implications for clinical trials

services may be provided by any member of an apy intervention experienced a reduction in symp- interdisciplinary cancer treatment team, by special- tom severity scores as compared to patients receiv- ists (e.g., rehabilitation professionals, mental health ing conventional care alone (Given et al., 2004). professionals, complementary and alternative heal- In patients participating in a short course of can- ers) or by members of the community (e.g., fam- cer counseling sessions with a humanist framework, ily, friends, faith groups, peers). Frequently, in the participants reported that the program was helpful treatment of cancer patients, the initial support is in expressing feelings, examining and understand- provided by members of hospitals/medical institu- ing emotional responses, confronting the fear of tions; as the patient is further along in their treat- death, and working through powerful thoughts and ment/recovery, the emphasis may shift more toward feelings (Boulton et al., 2001). Patients in this pro- a family and community focus. Ideally, support- gram reported overwhelmingly positive attitudes ive care is provided to patients and families in this about being in counseling, despite the sometimes manner with a smooth transition back and forth difficult subject matter and intense emotions they between the various members of these support were asked to consider. The benefits of individual teams as conditions and needs for support change psychotherapy for cancer patients were further evi- along the disease trajectory (see Table 21.1). This denced in a European study evaluating the effi- section addresses psychosocial, rehabilitative, and cacy of short-term face-to-face counseling in self- complementary/alternative modalities. referred cancer patients (Boudioni et al., 2000). The great majority of patients (>90%) who returned the evaluation of this service reported that their emo- Psychosocial support tional health was better at the end of the counsel- Many studies have shown that psychosocial inter- ing sessions. Almost all of the reporting participants ventions can have a positive impact on the psy- expressed a positive view of the service, with more chological distress experienced in cancer patients. than 95% of them stating that they would return for A meta-analysis of controlled outcome studies further counseling if they needed help in the future demonstrated that psychosocial interventions have and would recommend the counseling services to a positive impact on quality of life in adult cancer others. patients and that many different forms of intervention are beneficial (Rehse & Pukrop, 2003). Inter- Support groups estingly, the study found that even more than the In their groundbreaking 1981 study, Spiegel et al. specific format of the psychosocial intervention, the (1981) reported that a group of women with breast duration of the intervention emerged as the most cancer, attending a weekly support group, had lower relevant variable. Some of the more common for- mood disturbance scores, fewer maladaptive cop- mats are individual psychotherapy or counseling, ing responses, and lower phobia scores. This was support groups (professionally facilitated or peer one of the first empirical studies that revealed strong led), and psycho-educational activities. The type of evidence for the benefit of support groups in pro- psychosocial intervention selected must be made viding psychological benefit. While there may be based on available resources, patient preference, multiple benefits for cancer patients sharing their and the nature of the psychosocial stressors being experiences with other patients in similar circum- targeted by treatment. stances, one concept is that the presence of peer support may help to reduce the stigma associated Individual psychotherapy with the diagnosis of cancer and may help to over- In one randomized trial, cancer patients with rel- come social isolation (Weis, 2003). In their review atively high levels of symptom severity scores who of several studies on cancer peer support groups, participated in a cognitive-behavioral psychother- Campbell et al. (2004) found consistent and positive Chapter 21. Support services 305

Table 21.1. Modalities of support across the disease continuum

Modality Professional staff Stage of disease Example goals

Psychosocial support Neuropsychologist Newly diagnosed Coping with “shock” of –Individual Social Worker diagnosis; crisis psychotherapy Psychiatrist management –Couples/family Support Group Leader Active treatment Dealing with rigors of therapy Chaplain treatment; adjusting to –Group therapy Pastoral Care Specialist changes in life –Support groups Post-treatment – Learning to live with late stable effects of treatment End of life/palliative Death and dying issues; existential concerns Rehabilitation Physiatrist Newly diagnosed Post-operative –Inpatient unit Neuropsychologist physical/cognitive –Outpatient center Physical Therapist deﬁcits –Home health Occupational Therapist Active treatment Managing fatigue; Speech Therapist symptom management; Recreational Therapist cognitive rehabilitation Social Worker Post-treatment – Recovery from treatment; stable facing long-term effects of treatment End of life/palliative Maximizing/preserving functioning (e.g., treating hypoactive delirium); transitioning to hospice Community integration Voc Rehab specialists Newly diagnosed Arranging for short-term –Work/school Education specialists disability/sick leave –Family Neuropsychologist Active treatment Setting up long-term –Social/recreation Community members disability or managing Clergy or religious leaders symptoms at work Post-treatment – Returning to work/school; stable resuming social life; renegotiating family roles End of life/palliative Preparing for death; ﬁnal arrangements/legal planning; saying goodbye 306 Section 3. Interventions and implications for clinical trials

benefits from peer support, regardless of the spe- knowledge and self-confidence (Braden et al., 1998) cific manner in which it was delivered and irrespec- and in decreasing depression and anxiety (Johnson, tive of the theoretical model on which it was based. 1982). Each individual participating in a support group may take away different benefits from the experi- Rehabilitation programs ence, including varying levels of practical information, emotional support, sense of community, and Comprehensive rehabilitation programs for mul- feelings of altruism. Support groups are a very effi- tiple advanced disease populations (e.g., stroke, cient method of providing individualized support traumatic brain injury, spinal cord injury) have been services to a number of cancer patients at one time in existence for years. Few would argue the benefits and in an environment that feels safe, friendly, and of these programs in helping patients to achieve the less intimidating than might individual face-to-face highest functional status possible, given the limita- counseling sessions. tions created by the disease. These multidisciplinary programs focus on the treatment and management Online support groups of medical symptoms and complications (physia- The appeal of online cancer support groups is trist), improving mobility and level of independence understandable. Online support avoids some of (physical and occupational therapists), enhanc- the barriers of traditional support groups, such as ing communication (speech language pathologists), inconvenient meeting times, lack of meeting places, improving cognitive functioning (neuropsycholo- and medical complications that make travel to a gists), and expanding social activities and outlets meeting site challenging. Indeed, there has been a (recreational therapists). Many participants in such large increase in recent years with respect to the programs make remarkable gains in abilities and number of cancer support groups that are held achieve levels of increased autonomy and satis- exclusively online. In their review of the relevant, faction that would not be possible without this but fairly scant, outcome literature on online sup- time- and energy-intensive, formalized approach port groups, Klemm et al. (2003) found that most to restoring impaired patients to the highest level of the studies were small in sample size and were of functioning attainable. See Chapter 20 for an overly homogeneous with respect to gender and dis- indepth discussion of rehabilitation techniques. ease site. Nonetheless, the authors report that their Even though the same principles for rehabili- review of the available literature did seem to suggest tation as applied to these aforementioned medi- a benefit from online support groups for patients cal populations are appropriate for cancer patients, with cancer. They found that participants in these and even though aggressive cancer therapies have studies used the forum primarily to gather infor- increased survival rates while causing correspond- mation and to give and receive emotional support. ing increases in debilitation, very little attention has Interestingly, the authors did not find that com- been given to rehabilitation of the cancer patient puter inexperience was a barrier to successful use of (Cheville, 2001; Kirshblum et al., 2001). Often-cited online groups. reasons for not referring cancer patients to rehabilitation programs are life expectancy, pain, and med- Psycho-educational approaches ical co-morbidities. Yet, as Cheville (2001) points Psycho-educational programs have both psycho- out, significant infirmity and poor prognosis for dis- logical and educational components. Such pro- ease improvement in cancer patients should be dis- grams aim to enhance understanding and knowl- tinguished from the possibility of improvement in edge about cancer and associated issues, including functional abilities. DeLisa (2001) maintains, “reha- symptom management, psychosocial support, and bilitation for patients with cancer should be no dif- resource identification. Psycho-educational pro- ferent from rehabilitation for those of other diag- grams have been shown to be effective in increasing nostic conditions, such as cerebrovascular disease, Chapter 21. Support services 307

spinal cord injury, or brain injury.” Many rehabili- matic brain injury patients matched by age, gender, tation strategies used in other patient populations and baseline functional status (O’Dell et al., 1998). have been successful in improving functional status Interestingly, the length of stay for the brain tumor in cancer patients. These include rehabilitation of patientsinthisstudywasonaverageshorterthan motor deficits, sensory deficits, cerebellar dysfunc- that for the traumatic brain injury patients. Finally, tion, and deconditioning (Cheville, 2001). Marciniak et al. (2001) found that the functional Clinical pathways to cognitive remediation of gains made by brain tumor patients in an inpa- patients with stroke and brain injury are well docu- tient rehabilitation program did not differ accord- mented, but have only recently started to be applied ing to whether the tumor was primary or metastatic. to cancer patients to improve functioning related A more important finding from this study was that to deficits due to both neurotoxicity of antineo- patients receiving concurrent radiation therapy dur- plastic agents/treatments and direct central ner- ing the course of rehabilitation made greater func- vous system malignancies. Rehabilitation of cogni- tional gains than did those patients not receiving tive deficits related to cancer and cancer treatment radiation therapy. This result has implications for can follow the same model that is used with other health care providers who might avoid sending can- neurorehabilitative programs where both restora- cer patients to intensive rehabilitation programs for tive and compensatory strategies are employed. fear that concurrent antineoplastic therapies might Restorative efforts attempt to improve the impaired make them less amenable to rehabilitation gains. function, whereas compensatory training helps the patient learn to “work around” the deficient area of Complementary and alternative approaches functioning. Attentional deficits may be improved by helping patients learn to manage distractions Interest in complementary and alternative and control their environment as much as possible. medicine has grown tremendously in recent years; Memory difficulties may be compensated for by cancer patients in particular have been drawn to using a memory notebook or by rehearsal and unconventional treatments in higher and higher overlearning strategies. Executive dysfunction may numbers (Ernst & Cassileth, 1998). In a recent study, be addressed by teaching patients to superimpose Bernstein and Grasso (2001) determined that 80% of organized structure onto tasks and by repetitively adult cancer patients in a private non-profit South stressing the practice of breaking tasks down into Florida hospital used some form of complementary their simplest component steps. Communication and alternative medicine treatment, including may be enhanced by using deliberative speech prac- vitamins, herbal products, relaxation techniques, tices and by learning alternative ways to communi- massages, and home remedies. In their 2004 evalu- cate for profoundly aphasic patients (e.g., message ation of a complementary and alternative medicine boards, gesturing). program at the Stanford Center for Integrative The efficacy of comprehensive rehabilitation ser- Medicine, Rosenbaum et al. (2004) found that over vices for cognitive deficits in cancer patients can be 90% of the patients in the program reported benefit best illustrated by the results of several studies in from its services. Those services with the most par- brain tumor patients. Sherer et al. (1997) demon- ticipants in the program were massage, yoga, and strated that six patients with primary malignant qigong. Other unconventional methods for treating brain tumors undergoing an inpatient rehabilitation the medical complications and psychological dis- program showed evidence of increased indepen- tress that accompany cancer are gaining favor and dence during the time from the start of participation showing efficacy. In a randomized controlled trial to discharge. In another study with a larger sample employing a mindfulness-based stress reduction size of 40, daily gains in functional abilities in brain program in cancer outpatients, participants showed tumor patients completing an inpatient rehabilita- a significant reduction in mood disturbances and tion program were similar to those made by trau- stress (Speca et al., 2000). 308 Section 3. Interventions and implications for clinical trials

As depression, anxiety, and fatigue are known agents with all members of their treatment team in to impair cognitive functioning, it is reasonable order to avoid any drug–herbal interactions or any to expect that complementary and alternative possible interference in the efficacy of the primary medicine therapies that are associated with stress antineoplastic therapies. reduction, improved mood, and reduction in fatigue Thebottomlineregardingtheroleofcomple- levels may indirectly result in stronger cognitive mentary and alternative medicine therapies in the abilities. Such interventions might include mas- treatment of cancer-related cognitive deficits is that sage, deep muscle relaxation, guided visual imagery, while many of these approaches may show promise, mindfulness, meditation, hypnosis, biofeedback, their empirical efficacy, and, in some cases, their yoga, reflexology, and qigong. These therapies are safety in cancer patients are not fully understood most effectively used in combination with conven- and need to be further investigated. It is impera- tional therapies for emotional distress and fatigue, tive that medical professionals working with cancer such as psychotherapy, antidepressants/anxiolytics, patients be educated and aware of complementary activity-rest cycles, sleep hygiene models, and and alternative medicine approaches and prepared hematopoietic agents. to counsel their patients on the risks and benefits Complementary and alternative medicine thera- of these treatments in the management of cancer pies have gained popularity in the prevention and and its symptoms: physical, psychological, and cog- treatment of dementia. Approaches to preserving nitive. It is clear from some emerging literature that or enhancing cognitive functioning, or to mini- many of these unconventional approaches are safe mizing associated functional decline were recently and effective in improving cancer patients’ quality reviewed in an article by Sierpina et al. (2005). The of life. authors grouped these complementary and alternative medicine approaches into three categories: ACKNOWLEDGMENTS (1) mind/body therapies (relaxation, meditation, guided imagery, hypnosis, biofeedback, cognitive- The authors wish to thank the following individuals in The behavioral therapies, and psycho-educational Preston Robert Tisch Brain Tumor Center at Duke Univer- approaches); (2) lifestyle changes/social support sity Medical Center for their assistance in researching and editing this book chapter:. (environment, recreation/education, creative Bart Brigidi, Ph.D, expression, music); and (3) nutrients/botanicals Assistant Research Professor (acetylcholinesterase inhibitors, acetylcholine pre- Lisa Fornnarino cursors, antioxidants, anti-inflammatory agents, Neuropsychological Technician hormonal agents). The authors conclude that Pam Clair mind/body therapies and lifestyle changes/social Staff Assistant supports are extremely safe interventions that may Karen Carter show promise in minimizing the negative impact Clinical Research Associate of cognitive decline on quality of life, although Stephen T. Keir, MPH, DrPH they acknowledge that further controlled stud- Assistant Clinical Research Professor ies on outcome and efficacy are needed. With Quality of Life Research Program respect to nutritional and herbal therapies, they suggest promising links between use of some RESOURCES complementary and alternative medicine agents www.cancer.gov (National Cancer Institute) and improved cognitive functioning in dementia www.cancer.org (American Cancer Society) patients, but emphasize that few of these findings www.cancercare.org (sponsors teleconferences, web- are yet supported by empirical evidence. It is of based conferences, has free social work services, etc.) utmost importance that patients discuss the poten- www.canceradvocacy.org (National Coalition for Cancer tial use of complementary and alternative medicine Survivorship) Chapter 21. Support services 309

www.cancersupportivecare.com (dedicated to improving Blanchard CG, Albrecht TL, Ruckdeschel JC (1997). The cri- quality of life and reducing morbidity in cancer patients) sis of cancer: psychological impact on family caregivers. http://nccam.nih.gov (National Center for Complemen- Oncology 11: 189–194. tary and Alternative Medicine) Blueglass K (1991). Care of the cancer patient’s family. In www.ed.gov (Rehabilitation Services Administration) Watson M (ed.) Cancer Patient Care: Psychosocial Treat- www.disabilityresources.org (vocational rehabilitation ment Methods (pp. 159–189) Cambridge: British Psycho- resources) logical Society and Cambridge University Press. www.hospicefoundation.org (end of life information and Bonner MJ, Hardy KK et al. (2006). Development and val- hospice locator guide) idation of the parent experience of illness (PECI) ques- www.chionline.org (information on hospice care specific tionnaire. J Pediatr Psychol 31: 310–321. to pediatric patients) Boudioni M, Mossman J et al. (2000). An evaluation of a www.biausa.org (Brain Injury Association of America) cancer counselling service. Eur J Cancer Care 9: 212– www.epilepsyfoundation.org (information on seizures 220. and disability discrimination policies) Boulton M, Boudioni M et al. (2001). “Dividing the deso- www.candlelighters.org (Candlelighters Childhood Can- lation”: clients views on the benefits of a cancer coun- cer Foundation) selling service. Psychooncology 10: 124–136. www.acor.org (Association of Cancer Online Resources, Braden CJ, Mishel MH et al. (1998). Self-help intervention Inc. – includes listings of internet support groups) project: women receiving breast cancer treatment. Can- www.oncolink.upenn.edu (Oncolink) cer Pract 6(2): 87–98. www.patientcenters.com (Comprehensive listing of Bradley CH, Bednarek HL (2002). Employment patterns of resources including first person patient experiences) long-term cancer survivors. Psychooncology 11: 188–198. www.nichcy.org (National Dissemination Center for Chil- Brown RT, Madan-Swain A et al. (2003). Posttraumatic dren with Disabilities) stress symptoms in adolescent survivors of child- www.livestrong.org (Lance Armstrong Foundation – hood cancer and their mothers. J Trauma Stress 16: information and support for survivors) 309–318. Butler RW, Mulhern RK (2005). Neurocognitive interventions for children and adolescents surviving cancer. REFERENCES J Pediatr Psychol 30(1): 65–78. Campbell H, Phaneuf M et al. (2004). Cancer peer support Allen DH, Tepper S, Carter K et al. (2003). Use of Beck programs – do they work? Patient Educ Counsel 55: 3–15. Depression Inventory II: evaluating quality of life symp- Carlson LE, Angen M, Cullum J et al. (2004). High levels toms in glioma patients [Abstract] Society for Neuro- of untreated distress and fatigue in cancer patients. Br J Oncology Eighth Annual Meeting, 13–16 November, Cancer 90(12): 2297–2304. 2003. Keystone, Colorado. Neurooncology 5: 329. Carter PA, Chang BL (2000). Sleep and depression in cancer Angelino AF, Treisman GJ (2001). Major depression and caregivers. Cancer Nurs 23: 410–415. demoralization in cancer patients: diagnostic and treat- Cassileth BR, Lusk EJ et al. (1985). A psychological analysis ment considerations. Support Care Cancer 9(5): 344–349. of cancer patients and their next-of-kin. Cancer 55: 72– Armstrong FD, Blumberg MJ et al. (1999). Neurobehavioral 76. issues in childhood cancer. School Psychol Rev 28: 194– Cheville A (2001). Rehabilitation of patients with advanced 203. cancer. Rehabilitation in the New Millennium. Cancer 92 Baider L, Denour AK (1999). Psychological distress and can- [4 Suppl.]: 1039–1048. cer couples: a leveling effect. New Trends Exp Clin Psychi- Chirikos TN, Russell-Jacobs A et al. (2002). Functional atry 15: 197–203. impairment and the economic consequences of female Bernstein BJ, Grasso T (2001). Prevalence of complemen- breast cancer. Women Health 36(1): 1–20. tary and alternative medicine use in cancer patients. Cliff AM, McDonagh RP (2000). Psychosocial morbidity in Oncology (Huntington) 15 (10): 1267– 1272; discussion prostate cancer II; a comparison of patients and part- 1272–1278, 1283. ners. Br J Urol Int 86: 834–839. Bertakis KD, Roter D, Putnam SM (1991). The relationship Clinical Practice Guidelines in Oncology. Version 1.205 of physician medical interview style to patient satisfac- Available at: http://www.nccn.org. Accessed 11 April, tion. J Fam Pract 32(2): 175–81. 2008. 310 Section 3. Interventions and implications for clinical trials

Compas BE, Worsham NL et al. (1994). When mom or dad Holmes AM, Deb P (2003). The effect of chronic illness on has cancer: markers of psychological distress in cancer the psychological health of family members. J Mental patients, spouses, and children. Health Psychol 13: 507– Health Policy Econ 6(1): 13–22. 515. Johnson J (1982). The effect of a patient education course Covinsky KE (1994). The impact of serious illness on on persons with a chronic illness. Cancer Nurs 5(2): 117– patients’ families. SUPPORT investigators study to 123. understand prognoses and preferences for outcomes Johnson J (1988). Cancer: a family disruption. Recent and risks of treatment. J Am Med Assoc 272(23): 1839– Results Cancer Res 108: 306–310. 1844. Kazak AE, Barakat LP et al. (1997). Posttraumatic stress, Deasy-Spinetta P, Spinetta JJ (1980). The child with can- family functioning, and social support in survivors of cer in school: teacher’s appraisal. Am J Pediatr Hematol childhood leukemia and their mothers and fathers. Oncol 2: 89. J Consult Clin Psychol 65: 120–129. DeLisa JA (2001). A history of cancer rehabilitation. Can- Kazak AE, Alderfer M et al. (2004). Posttraumatic stress dis- cer Rehabilitation in the New Millennium. Cancer 92 [4 order (PTSD) and posttraumatic stress symptoms (PTSS) Suppl.]: 970–974. in families of adolescent childhood cancer survivors. Devine EC, Westlake SK (1995). The effects of psychoeduca- J Pediatr Psychol 29: 211–219. tional care provided to adults with cancer: meta-analysis Kirshblum S, O’Dell M et al. (2001). Rehabilitation of per- of 116 studies. Oncol Nurs Forum 22(9): 1369–1381. sons with central nervous system tumors. Cancer Suppl Ell KO, Mantell JE, Hamovitch MB, Nishomoto RH (1989). 92(4): 1029–1038. Social support, sense of control, and coping among Klemm P, Bunnell D et al. (2003). Online cancer support patients with breast, lung, or colorectal cancer. JPsy- groups: a review of the research literature. Computers chosocial Oncol 7: 63–89. Infomatics Nursing 21(3): 136–142. Ernst E, Cassileth BR (1998). The prevalence of comple- Langeveld NE, Stam H et al. (2002). Quality of life in young mentary/alternative medicine in cancer: a systematic adult survivors of childhood cancer. Support Care Cancer review. Cancer 83(4): 777–782. 10(8): 579–600. Frank NC, Brown RT et al. (2001). Predictors of affective Leigh LD, Miles MA (2002). Educational issues for chil- responses of mothers and fathers of children with can- dren with cancer. In Pizzo PA, Poplack DG (eds.) Prin- cer. Psychooncology 10: 293–304. ciples and Practice of Pediatric Oncology (pp. 1463–1476). Gabanelli P (2005). A rehabilitative approach to the patient Philadelphia, PA: Lippincott Williams & Wilkins. with brain cancer. Neurol Sci 26: S51–S52. Maddrey AM, Bergeron JA, Lombardo ER (2005). Neuropsy- Gagnon B, Low G, Schreier G (2005). Methylphenidate chological performance and quality of life of 10 year hydrochloride improves cognitive function in patients survivors of childhood medulloblastoma. J Neurooncol with advanced cancer and hypoactive delirium: a 72(3): 245–253. prospective clinical study. J Psychiatry Neurosci 30(2): Maguire P (1999). Improving communication with cancer 100–107. patients. Eur J Cancer 35: 1415–1422. Garrard P, Farnham C, Thompson AJ, Playford ED (2004). Marciniak C, Sliwa J et al. (2001). Functional outcomes of Rehabilitation of the cancer patient: experience in a persons with brain tumors after inpatient rehabilitation. neurological unit. Neurorehabil Neural Repair 18(2): Arch Phys Med Rehabil 82: 457–463. 76–79. Matthews B, Baker F, Spillers RL (2004). Oncology profes- Given CW, Stommel M et al. (1993). The inﬂuence of cancer sionals and patient requests for cancer support services. patients’ symptoms and functional states on patients’ Support Care Cancer 12(10): 731–738. depression and family caregivers’ reaction and depres- Meyer TJ, Mark MM (1995). Effects of psychosocial inter- sion. Health Psychol 12: 277–285. ventions with adult cancer patients: a meta-analysis of Given C, Given B et al. (2004). Effect of a cognitive behav- randomized experiments. Health Psychol 14(2): 101–108. ioral intervention on reducing symptom severity during Meyers CA (2000). Neurocognitive dysfunction in cancer chemotherapy. J Clin Oncol 22(3): 507–516. patients. Oncology 14(1): 75–85. Hall IM, Cope DN (1995). The beneﬁt of rehabilitation in NCCN Standards of Care for Distress Management, DIS- traumatic brain injury: a literature review. J Head Trauma 3. Version 1.205 Available at: http://www.nccn.org. Rehabil 10: 1–13. Accessed 9 April, 2008. Chapter 21. Support services 311

Nijboer C, Triemstra M et al. (1999). Determinate of care- Sierpina V, Sierpina M et al. (2005). Complementary and giving experiences and mental health of partners of can- integrative approaches to dementia. South Med J 98(6): cer patients. Cancer 86: 577–588. 636–645. Nijboer C, Gtriemstra M et al. (2000). Patterns of caregiver Sloper P (2000). Predictors of distress in parents of children experiences among partners of cancer patients. Geron- with cancer: a prospective study. J Pediatr Psychol 25: 79– tologist 40: 738–746. 92. Northouse LL (1988). Social support in patients’ and Speca M, Carlson L et al. (2000). A randomized, wait- husbands’ adjustment to breast cancer. Nurs Res 37: list controlled clinical trial: the effect of a mindfulness 91–95. meditation-based stress reduction program on mood Northouse LL et al. (2000). Couples’ patterns of adjustment and symptoms of stress in cancer outpatients. Psychoso- to colon cancer. Soc Sci Med 50: 271–284. matic Med 62: 613–622. Oberst MT, Thomas SE et al. (1989). Caregiving demands Spelten ER, Verbeek JH, Uitterhoeve AL et al. (2003). Can- and appraisal of stress among family caregivers. Cancer cer, fatigue and the return of patients to work – a Nurs 12: 209–215. prospective cohort study. Eur J Cancer 39(11): 1562– O’Dell M, Barr K et al. (1998). Functional outcome of inpa- 1567. tient rehabilitation in persons with brain tumors. Arch Spiegel D, Bloom JR et al. (1981). Group support for Phys Med Rehabil 79: 1530–1534. patients with metastatic cancer. A randomized outcome Rehse B, Pukrop R (2003). Effects of psychosocial interven- study. Arch Gen Psychiatry 38(5): 527–533. tions on quality of life in adult cancer patients: meta Stewart JL, Mishel MH (2000). Uncertainty in childhood ill- analysis of 37 published controlled outcome studies. ness: a synthesis of the parent and child literature. Sch Patient Educ Counsel 50: 179–186. Inq Nurs Pract 14: 299–319. Ressler IB, Cash J et al. (2003). Continued parental atten- Sutherland JE, Sutherland SJ, Hoehns JD (2003). Achieving dance at a clinic for adult survivors of childhood cancer. the best outcome in treatment of depression. J Fam Prac J Pediatr Hematol Oncol 25: 868–873. 52(3): 201–209. Rosenbaum E, Gautier H et al. (2004). Cancer supportive Svavarsdottir EK (2005). Caring for a child with cancer: a care, improving the quality of life for cancer patients. A longitudinal perspective. JAdvNurs50: 153–161. program evaluation report. Support Care Cancer 12: 298– Valentine AD, Meyers CA (2001). Cognitive and mood dis- 301. turbance as causes and symptoms of fatigue in cancer Rozans M, Dreisbach A, Lertora JJ, Kahn MJ (2002). Pallia- patients. Cancer Supplement 92(6): 1694–1698. tive uses of methylphenidate in patients with cancer: a Van Dongen-Melman JEWM, Pruyn JFA et al. (1995). Late review. J Clin Oncol 20(1): 335–339. psychosocial consequences for parents of children who Sanson-Fisher R, Girgis A, Boyes A et al. (2000). The unmet survived cancer. J Pediatr Psychol 20: 567–586. supportive care needs of patients with cancer. Support- Van Dongen-Melman JEWM, Van Zuuren FJ et al. (1998). ive Care Review Group. Cancer 88(1): 226–237. Experiences of parents of childhood cancer survivors: Sheard T, Maguire P (1996). The effect of psychological a qualitative analysis. Patient Educ Couns 34: 185– interventions on anxiety and depression in oncology; 200. results of two meta-analyses. Paper presented at Third Weis J (2003). Support groups for cancer patients. Support World Congress of Psycho-Oncology. New York. Care Cancer 11: 763–768. Sheard T, Maguire P (1999). The effect of psychologi- Welch AS, Wadsworth ME et al. (1996). Adjustment cal interventions on anxiety and depression in cancer of children and adolescents to parental cancer. Par- patients: results of two meta-analyses. Br J Cancer 80(11): ents’ and children’s perspectives. Cancer 77: 1409– 1770–1780. 1418. Sherer M, Meyers CA, Bergloff P (1997). Efﬁcacy of post- Ybema JF, Kuijer RG et al. (2001). Depression and percep- acute brain injury rehabilitation for patients with pri- tions of inequity among couples facing cancer. JPsy- mary malignant brain tumors. Cancer 80(2): 250–257. chosocial Oncol 27: 3–13. Short PF, Vasey JJ, Tunceli K (2005). Employment pathways Zebrack BJ, Chesler MA (2002). Quality of life in child- in a large cohort of adult cancer survivors. Cancer 103(6): hood cancer survivors. Psychooncology 11(2): 132– 1292–1301. 141. 22

Pharmacological interventions for the treatment of radiation-induced brain injury

Edward G. Shaw, Jerome Butler, L. Douglas Case, Ralph d’Agostino, Jr., John Gleason, Jr., Edward Ip, Mike E. Robbins, Paul Saconn, and Stephen R. Rapp

Introduction mood cluster including anxiety, depression, and sadness, and an expressive language cluster includ- Neoplasms of the central nervous system (CNS) are ing difficulty reading, writing, and finding the right a pathologically diverse group of benign and malig- words (Gleason et al., 2006). In long-term survivors, nant tumors for which a variety of management three symptom clusters are more common, includ- strategies, including observation, surgery, radiation ing a physical function cluster (decreased energy, therapy (RT), and/or chemotherapy, are employed. fatigue, and frustration), mood cluster (anger, anx- Shown in Table 22.1 are the primary and metastatic iety, confusion, and depression), and a cognition brain tumors treated with RT, and the usual radia- cluster (difficulty concentrating, reading, remem- tion doses employed for each (Shaw, 2000). Regard- bering, and finding the right words) (Saconn et al., less of the type of brain tumor treated, radiation- 2006). The severity of these symptoms changes over treated patients will experience acute side-effects of the lifespan of the brain tumor patient, from ini- therapy and be at risk for late sequelae. Chapter 7 tial diagnosis to treatment, into the post-treatment outlined the biological basis of radiation-induced follow-up or survivorship period. Although quality CNS injury. This chapter will focus on the treatment of life interventions for cancer patients tend to focus and prevention of radiation-induced brain injury, on single symptoms such as fatigue (Dodd et al., with an emphasis on pharmacological therapies. 2001, 2004), as with other types of brain patholo- gies (stroke, trauma, neurodegenerative diseases), a multidisciplinary approach, including pharmacological, behavioral, and rehabilitative therapies, is Symptoms and symptom clusters in brain needed to optimize quality of life in the brain tumor tumor patients patient.

The symptoms of primary and metastatic brain tumors are dependent on tumor location (Table Management of radiation-induced brain 22.2) (Shaw, 2000). Besides location-dependent injury symptoms, patients with brain tumors may experience symptoms related to their physical, emotional, Acute reactions and cognitive functions. Often, these symptoms occur in clusters. In newly diagnosed brain tumor The most common acute reactions associated with patients, two symptom clusters typically occur: a brain radiation include fatigue, hair loss, and skin

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

312 Chapter 22. Pharmacologic interventions for the treatment of radiation-induced brain injury 313

Table 22.1. Usual treatment volumes and doses for primary and metastatic brain tumors treated with radiation therapy

Pathological type Treatment volume Total dose (Gy/no. fractions)

Glioblastoma (WHO IV) 60/30 Initial field Edema and enhancing tumor 46/23 Boost field Enhancing tumor 14/7 Anaplastic astrocytoma, 59.4/33 oligoastrocytoma (WHO III) Initial field Edema and enhancing tumor 50.4/28 Boost field Enhancing tumor 9/5 Astrocytoma (WHO II) Edema (and enhancing tumor if 50.4/28 to 59.4/33 present) Pilocytic astrocytoma (WHO I) Enhancing tumor 50.4/28 to 55.8/31 Pituitary adenoma Enhancing tumor 45/25 to 50.4/28 Meningioma (WHO I) Enhancing tumor 52.2/29 Medulloblastoma and anaplastic 55.8/31 ependymoma Initial volume Entire brain and spine 36/20 Boost volume Enhancing tumor 19.8/11 to 23.4/13 Ependymoma Enhancing tumor 50.4/28 to 59.4/33 Brain metastases Whole brain 30/10 to 50.4/28 erythema. The onset of fatigue is generally within cal trial of d-methylphenidate (d-MPH) in adults several weeks of the first radiation treatment. It is undergoing curative or palliative partial- or whole- usually mild to moderate in severity. Typically, the brain RT. Patients received d-MPH or placebo dur- fatigue persists for 1–2 months after the comple- ing brain RT and for 8 weeks afterward. The prophy- tion of treatment but may be chronic (lasting ≥3 lactic use of d-MPH in this study was not associated months) in some patients. One characteristic of the with an improved overall or brain-specific quality fatigue associated with RT is a lack of improve- of life, or reduction in fatigue, as measured by the ment by rest. Methylphenidate (RitalinR ), a CNS Functional Assessment of Cancer Therapy (FACT) stimulant, can be used to treat the fatigue that including both the brain and fatigue subscales (But- usually occurs in patients receiving whole-brain ler et al., 2005; Weitzner et al., 1995a; Yellen et al., radiation. It also improves the depression and cog- 1997). nitive dysfunction that often accompanies fatigue Hair loss occurs in the same time frame as fatigue, in these patients (Weitzner et al., 1995b). The usual about 2–3 weeks into a course of fractionated dose of methylphenidate is 10 mg twice per day, whole- or partial-brain radiation. Complete or near- escalating to 20–30 mg twice per day in 1-week complete hair regrowth is the rule, though it may increments as tolerated. Methylphenidate can also take 6 months to a year. There are no known inter- be used in children, usually at half the dose rec- ventions to prevent radiation-induced hair loss, ommended for adults (Mulhern et al., 2004). The nor are there any effective treatments to acceler- dose-limiting toxicities are usually anxiety, insom- ate or maximize hair regrowth. Skin erythema is nia, and tachycardia. Butler et al. (2005) recently managed symptomatically with anti-inflammatory reported the results of a Phase III prospective, ran- and moisturizing creams, typically 1% hydrocorti- domized, double-blind, placebo-controlled clini- sone or AquaphorR , which are applied two to four 314 Section 3. Interventions and implications for clinical trials

Table 22.2. Common brain tumor symptoms Early delayed reactions

Aphasia (including dysphasia) (T) There are no known interventions or therapies to Ataxia (including truncal and limb) (BS, CB) prevent or treat early delayed reactions involving Bowel/bladder continence problems (F) the brain, which are thought to occur because of Deafness (BS, CPA, CN VIII) transient demyelination. Somnolence syndrome in Dementia (T) children is an example of an early delayed reac- Diplopia (BS, CS, CN III, IV, VI) tion involving the brain. The symptoms of som- Dizziness (BS, CB, CPA, CN VIII) nolence syndrome include somnolence, irritability, Dysarthria (BS, CN IX/X) anorexia, and sometimes an exacerbation of under- Dysphagia (BS, CN IX/X) lying tumor-associated symptoms or signs. Prior to Facial numbness (BS, CN VII, CPA) or concomitant with the treatment of somnolence Facial pain (BS, CS, CN VII) syndrome, a careful history and exam as well as re- Headache (G, 3/4V, HC) Memory impairment (T) imaging of the brain (usually with magnetic reso- Nausea (G, CB, BS, 3/4V, HC) nance imaging) should be done, to rule out a bleed, Neck pain (CB, BS) stroke, or tumor recurrence. Steroids are usually Personality changes (including initiated with the onset of somnolence syndrome, mood/mentation/concentration) (G, F,T) 2–4 mg dexamethasone bid to qid orally (higher Seizures (G) doses perhaps with intravenous administration if Sensory changes (including numbness, tingling, the symptoms are severe or life-threatening), with a paresthesias) (P,BS) rapid taper if no symptomatic improvement occurs Visual ﬁeld deﬁcits (including blindness) (T, P,S/PIT, SS) (which is usually the case), or a slower taper over 1– Vomiting (G, CB, BS 3/4V, HC) 2 months if the symptoms are responsive to steroids. Weakness (F,BS) Somnolence syndrome is transient, though it may Key : 3/4V, 3rd or 4th ventricle; BS, brainstem; CB, cerebel- last for several weeks to months, and does not pre- lum; CN, cranial nerve; CPA, cerebellopontine angle; CS, dict for subsequent radiation-induced brain injury cavernous sinus; F,frontal lobe; G, general cerebral (includ- (Halperin et al., 1994). ing any intracranial location); HC, symptoms associated with hydrocephalus; P,parietal lobe; PIT, pituitary gland; S, sellar; SS, suprasellar; T, temporal lobe. Late delayed reactions

Although edema and necrosis of the white matter are usually classiﬁed as late delayed reactions, brain edema can occur as an early or late effect of times daily or as needed for patient comfort. Moist radiation. The treatment of radiation-induced brain desquamation behind the ears and in the external edema is more of an art than a science and typically auditory canals may develop following whole brain involves the use of steroids (Wen & Marks, 2002). radiation. Treatment usually involves skin creams Oral dexamethasone is usually used in initial doses and Cortisporin otic suspension (i.e., neomycin, of2–4mgbidformildsymptomsand4–6mgqidfor polymyxin B, and hydrocortisone). Rarely, debride- moderate to severe symptoms. Oral doses in excess ment of the external auditory canals by an oto- of 10 mg qid (40 mg daily) usually do not increase laryngologist may be necessary. Radiation-induced the likelihood of clinical beneﬁt. The initial dexam- otitis media may also occur. Symptomatic treat- ethasone dose is usually maintained for 2–4 weeks, ment with oral decongestants is usually adequate. with a slow taper (2–4 mg per day reduction every 5– Occasionally, a tympanic membrane tube may be 7 days) as tolerated thereafter. For patients with life- necessary. threatening edema, intravenous dexamethasone Chapter 22. Pharmacologic interventions for the treatment of radiation-induced brain injury 315

Months Months Pre-RT 3 post-RT 6 post-RT

Figure 22.1. MRI scans of a 30-year-old man with a WHO grade II oligodendroglioma of the right posterior frontal lobe. The images on the left (T2-weighted images on top, T1-weighted images with contrast on bottom) were obtained before 64.8 Gy in 36 fractions radiation therapy (RT) was given to a localized treatment field (tumor plus margin 1–2 cm). Three months following RT, the patient developed headaches and left-sided weakness and had a repeat MRI scan (center images), which demonstrated radiation necrosis. Note increase in enhancement and surrounding edema in the absence of much mass effect. The patient was treated with a 3-month course of dexamethasone. A follow-up MRI scan 3 months later showed resolution of the imaging changes associated with the radionecrosis (right images). The patient’s MR imaging remained stable for over a decade, by which time he had tumor progression is used, 10–25 mg as a bolus followed by 4–10 mg gias. Treatment is symptomatic. The physical qid. If these patients do not respond to dexam- manifestations of chronic dexamethasone admin- ethasone, intravenous mannitol may be required. istration use can take months to resolve after its Patients on dexamethasone should receive gastritis discontinuation. prophylaxis (with H2 blockers or proton pump Necrosis of the brain can be difficult to clini- inhibitors) and appropriate treatments for hyper- cally and radiographically differentiate from tumor glycemia (oral hypoglycemic agents or insulin) recurrence (Figure 22.1) (Forsyth et al., 1995). Since and oral thrush (fluconazole 200 mg day 1 then cerebral radiation necrosis is always accompanied 100 mg daily for 6 days) should they arise. Pro- by varying degrees of edema, the initial man- phylaxis for pneumocystis pneumonia using one agement of clinically suspected or pathologically double-strength trimethoprim/sulfamethoxazole proven radionecrosis is with steroids, as previ- (BactrimR ) tablet daily three times per week is com- ously described. Several adjunctive medical treat- monly used in children as well as adults also taking ments for brain radiation necrosis have been anec- temozolomide (TemodarR ) chemotherapy (Stupp dotally described as being helpful to arrest or et al., 2002). Patients taking dexamethasone reverse the process, such as hyperbaric oxygen, chronically (1 month or longer) usually become warfarin (CoumadinR ), pentoxifylline (TrentalR ), cushingoid, characterized by fatigue, weight gain, and antioxidant vitamins such as vitamin E (Leber facial swelling, central obesity, muscle wasting et al., 1998; Liu et al., 2001; Nieder et al., 2005). (particularly in the extremities), striae, and arthral- One recent report of 50 patients with biopsy-proven 316 Section 3. Interventions and implications for clinical trials

(a) (b) (c)

Figure 22.2. MRI scans of a 72-year-old woman with ovarian cancer metastatic to the brain. (a) and (b) were obtained before 37.5 Gy in 15 fractions whole-brain radiation therapy (WBRT) was administered. A subcentimeter left posterior frontal metastasis is seen on the T1-weighted image with contrast (a). There are minimal surrounding T2 changes in the white matter (b), and no other white matter abnormalities present. Two years following WBRT, the patient has diffuse white matter demyelination and atrophy, as seen on the T2-weighted image (c). The patient died 1 year later of a progressive Alzheimer’s-like dementia

brain radionecrosis suggested benefit in the use mon with larger fraction sizes (>2Gy) (DeAngelis of pentoxifylline 400 mg tid and vitamin E 1000 et al., 1989). Symptoms range from cognitive slow- IU daily (Rogers, 2006). In medically unresponsive ing, poor attention and concentration, difficulty patients, surgical resection of the necrotic lesion, multi-tasking, decreased short-term (and eventu- providing it can be safely performed, will often allow ally long-term) memory, word finding problems, the dexamethasone dose to be reduced and also and decreased IQ (in children), to a progressive provide relief from the symptoms and signs of mass Alzheimer’s-like dementia, which is also character- effect associated with the cerebral edema (Rogers, ized by urinary incontinence and gait disturbance 2006). While the various interventions cited have (Figure 22.2). Changes in energy and mood, particu- no proven preventive role in the development of larly fatigue, anxiety, and depression, often accom- brain radionecrosis, very little research has been pany the cognitive changes that occur in long-term conducted in this area, pre-clinically or clinically. survivors of brain radiation (Shaw et al., 2006). Like One animal model of radiation-induced injury to brain radionecrosis, there are no proven preven- the optic nerves even suggested an increased inci- tive interventions for radiation-induced cognitive dence of optic neuropathy with “preventive” hyper- dysfunction. However, there are several therapies baric oxygen (Kim et al., 2004), in addition to or that have been reported as beneficial, one of which, instead of dexamethasone. methylphenidate (RitalinR ), is described in the sec- Cognitive dysfunction, a late delayed effect of tion on acute reactions. Methylphenidate is particu- whole-brain and large-field partial-brain radiation, larly useful when fatigue is one of the symptoms, or can occur with total doses as low as 20 Gy in the prominent symptom experienced by a particu- adults and 24 Gy in children given with conven- lar patient. tional fractions of 1.8–2 Gy (Armstrong et al., 2002; The Wake Forest University School of Medicine Crosson et al., 1994; Mulhern et al., 1992; Ochs Comprehensive Cancer Center, through its Commu- et al., 1991; Ris & Noll, 1994), and is more com- nity Clinical Oncology Program (CCOP) Research Chapter 22. Pharmacologic interventions for the treatment of radiation-induced brain injury 317

Base, has recently completed two sequential Anderson Cancer Center and its CCOP Research open-label Phase II clinical trials utilizing novel Base. Other neuroprotective and neurotherapeutic interventions for the treatment of symptomatic late strategies, particularly single drugs or combinations radiation-induced brain injury in adults. Eligibility of agents, are being studied in this patient popula- criteria included partial- or whole-brain radiation tion, including therapies that target the presumed to a dose of >25 Gy for a primary or metastatic pathophysiology of radiation-induced injury (Fike brain tumor completed 6 months or more prior to et al., 1994; Hopewell et al., 1993; Hornsey et al., study entry and no radiographic evidence of tumor 1990; Lynch et al., 2002; Nakagawa et al., 1996; progression in the 3 months prior to study entry. Nieder et al., 2000; Sminia et al., 2003; Spence et al., Endpoints included quality of life (measured by 1986) and newer approaches using neural stem cells the brain subscale of the FACT) (Weitzner et al., (Rezvani et al., 2002) or new agents that stimulate 1995a), mood (measured by the Profile of Mood neurogenesis (Monje et al., 2002). States, POMS) (McNair et al., 1992), and cognitive function (measured by a neurocognitive test battery including assessment of attention and concentra- ACKNOWLEDGMENT tion, verbal and visual memory, verbal fluency, and executive function) (Benton & Hamsher, 1983; Delis Supported by NCI grant CA81851. et al., 1987; Fastenau et al., 1999; Kaplan et al., 1978; Reitan, 1958; Wechsler, 1981). Both studies utilized a 24-week intervention with quality of life, mood, REFERENCES and cognitive function evaluations occurring at baselineandatweeks6,12,and24oftreatment. Armstrong CL, Hunter JV, Ledakis GE et al. (2002). Late cog- The first study utilized donepezil (Aricept R )5mg nitive and radiographic changes related to radiotherapy: daily for 6 weeks followed by 10 mg daily for 18 initial prospective findings. Neurology 59: 40–48. weeks, whereas the second study utilized Ginkgo Benton AL, Hamsher K (1983). Multilingual Aphasia Exam- biloba 40 mg three times a day. Both interventions ination. Iowa City, IA: AJA Associates. were based on data from previously reported ran- Bryson HM, Benfield P (1997). Donepezil. Drug Aging 10(3): 234–243. domized trials in dementia (Bryson & Benfield, Butler J, Case D, Atkins J et al. (2005). A phase III double 1997; Le Bars et al., 1997; Rogers et al., 1998a, blind placebo controlled prospective randomized clini- 1998b). Of patients who completed the 24-week cal trial of the effect of d-threo-methylphenidate HCL (d- study, significant improvements were seen in qual- MPH) on quality of life in brain tumor patients receiv- ity of life (increase in FACT brain subscale score), ing radiation therapy. Int J Radiat Oncol Biol Physics 63 mood (decrease in POMS score), and cognitive [Supp1.]: 80. function (improved attention and concentration Crossen JR, Garwood D, Glatstein E, Neuwelt EA (1994). and memory) with both donepezil and Ginkgo Neurobehavioral sequelae of cranial irradiation in biloba. In addition, Ginkgo biloba patients had sig- adults: a review of radiation-induced encephalopathy. nificant improvement in executive function. Patient J Clin Oncol 12: 627–642. dropout for lack of efficacy or toxicity was greater DeAngelis LM, Delattre J, Posner JB (1989). Radiation- induced dementia in patients cured of metastases. Neu- in those taking Ginkgo biloba than donepezil (Shaw rology 39: 789–796. et al., 2006). An open-label Pilot study of donepezil Delis DC, Kramer JH, Kaplan E, Ober BA (1987). California is currently being conducted in children at Wake Verbal Learning Test-Research Edition. San Antonio, TX: Forest. A Phase III randomized double-blind The Psychological Corp. placebo controlled trial of donepezil is being Dodd MJ, Miaskowski C, Paul SM (2001). Symptom clusters conducted by Wake Forest Cancer Center with its and their effect on the functional status of patients with CCOP Research Base in conjunction with the M. D. cancer. Oncol Nurs Forum 28(3): 465–470. 318 Section 3. Interventions and implications for clinical trials

Dodd MJ, Miaskowski C, Lee KA (2004). Occurrence of mone and insulin-like growth factor 1 [Abstract]. Neu- symptom clusters. J Natl Cancer Inst Monographs 32: 76– rooncology 4: 354. 78. McNair DM, Lorr M, Droppleman LF (1992). Profile of Mood Fastenau PS, Denburg NL, Hufford BJ (1999). Adult norms States Manual.SanDiego,CA:EducationalandIndustrial for the Rey-Osterrieth Complex Figure Test and for sup- Testing Service. plemental recognition and matching trials from the Monje ML, Mizumatsu S, Fike JR, Palmer TD (2002). Irra- Extended Complex Figure Test. Clin Neuropsychol 13: 30– diation induces neural precursor-cell dysfunction. Nat 47. Med 8: 955–962. Fike JR, Goebbel GT, Martob LJ, Seilhan TM (1994). Radi- Mulhern R, Hancock J, Fairclough D, Kun L (1992). Neu- ation brain injury is reduced by the polyamine inhibitor ropsychological status of children treated for brain alpha-difluoromethylornithine. Radiat Res 138: 99–106. tumors: a critical review and integrative analysis. Med Forsyth PA, Kelly PJ, Casano TL et al. (1995). Radiation Pediat Oncol 20: 181–191. necrosis or glioma recurrence: is computer assisted Mulhern RK, Khan RB, Kaplan S et al. (2004). Short-term stereotactic biopsy useful? J Neurosurg 82(4): 36–44. efficacy of methylphenidate: a randomized, double- Gleason J, Case D, Rapp S et al. (2006). Symptom clusters blind, placebo-controlled trial among survivors of child- in newly-diagnosed brain tumor patients [Abstract]. In hood cancer. J Clin Oncol 22: 4795–4803. 2006 ASCO Annual Meeting Proceedings Part I. J Clin Nakagawa M, Bellinzona M, Seilhan TM et al. (1996). Oncol 24(18S) [June 20 Suppl.]: 8587. Microglial responses after focal radiation-induced injury Halperin EC, Constine LS, Tarbell NJ, Kun LE (eds.) (1994). are affected by alpha-difluoromethylornithine. Int J Pediatric Radiation Oncology (2nd edn.). New York: Radiat Oncol Biol Phys 36: 113–123. Raven Press. Nieder C, Price RE, Rivera B, Ang KK (2000). Both early Hopewell JW, Van Den Aardweg GJMJ, Morris GM et al. and delayed treatment with growth factors can modulate (1993). Unsaturated lipids as modulators of radiation the development of radiation myelopathy (RM) in rats. damage in normal tissues. In Horrobin DF (ed.) New Radiother Oncol 56 [Suppl. 1]: S15. Approaches to Cancer Treatment (pp. 88–106). London: Nieder C, Zimmerman FB, Adam M, Molls M (2005). The Churchill Communications Europe. role of pentoxifylline as a modifier of radiation therapy. Hornsey S, Myers R, Jenkinson T (1990). The reduction of Cancer Treat Rev 31: 448–455. radiation damage to the spinal cord by postirradiation Ochs J, Mulhern R, Fairclough D et al. (1991). Compari- administration of vasoactive drugs. Int J Radiat Oncol son of neuropsychologic functioning and clinical indi- Biol Phys 18: 1437–1442. cators of neurotoxicity in long-term survivors of child- Kaplan EF, Goodglass H, Weintraub S (1978). The Boston hood leukemia given cranial radiation or parenteral Naming Test. Boston, MA: E. Kaplan & H. Goodglass. methotrexate: a prospective study. J Clin Oncol 9: 145– Kim JH, Brown SL, Kolozsvary A et al. (2004). Modifica- 151. tion of radiation injury by Ramipril, inhibitor of the Reitan RM (1958). Validity of the Trail Making Test as an angiotensin-converting enzyme, on optic neuropathy in indicator of organic brain damage. Percept Motor Skills the rat. Radiat Res 161: 137–142. 8: 271–276. Le Bars PL, Katz MM, Berman N, Itil TM, Freedman AM, Rezvani M, Birds DA, Hodges H, Hopewell JW, Milledew K, Schatzberg AF (1997). A placebo-controlled, double- Wilkinson JH (2002). Modification of radiation myelopa- blind, randomized trial of an extract of ginkgo biloba for thy by the transplantation of neural stem cells in the rat. dementia. J Am Med Assoc 278: 1327–1332. Radiat Res 156: 408–412. Leber KA, Eder HG, Kovac H, Anegg U, Pendl G (1998). Ris M, Noll R (1994). Long-term neurobehavioral outcome Treatment of cerebral radionecrosis by hyperbaric oxy- in pediatric brain-tumor patients: review and method- gen therapy. Stereotact Funct Neurosurg 70 [Suppl. 1]: ological critique. J Clin Exp Neuropsychol 16: 21–42. 229–236. Rogers LR (2006). Natural history and results of ther- Liu CY, Yim BY, Wozniak AJ (2001). Anticoagulation therapy apy in 50 patients with histologically confirmed cere- for radiation-induced myelopathy. Ann Pharmacother bral radiation necrosis [Abstract]. Neurooncology 8(4): 35: 188–191. 489. Lynch CD, Sonntag WE, Wheeler KT (2002). Radiation- Rogers SL, Doody RS, Mohs RC, Friedhoff LT, Donepezil induced dementia in aged rats: effects of growth hor- Study Group (1998a). Donepezil improves cognition and Chapter 22. Pharmacologic interventions for the treatment of radiation-induced brain injury 319

global function in Alzheimer’s disease. Arch Int Med 158: Stupp R, Dietrich PY, Ostermann Kraljevic S et al. 1021–1031. (2002). Promising survival for patients with newly Rogers SL, Farlow MR, Doody RS, Mohs R, Friedhoff LT, diagnosed glioblastoma multiforme treated with Donepezil Study Group (1998b). A 24-week, double- concomitant radiation plus temozolomide fol- blind, placebo-controlled trial of donepezil in patients lowed by adjuvant temozolomide. J Clin Oncol 20: with Alzheimer’s disease. Neurology 50(1): 136–145. 1375–1382. Saconn PA, Ip E, Rapp S, D’Agostino RB Jr, Naughton MJ, Wechsler D (1981). Wechsler Adult Intelligence Scale – Shaw EG (2006). Symptom clusters in irradiated brain Revised Manual. New York: Psychological Corp. tumor survivors [Abstract]. In 2006 ASCO Annual Meet- Weitzner MA, Meyers CA, Gelke CK, Byrne KS, Cella DF, ing Proceedings Part I. J Clin Oncol 24 (18S) [June 20 Levin VA (1995a). The Functional Assessment of Can- Suppl.]: 8581. cer Therapy (FACT) scale: development of a brain sub- Shaw EG (2000). Central nervous system overview. In Gun- scale and revalidation of the general version (FACT-G) derson LL, Tepper JE (eds.) Clinical Radiation Oncology. in patients with primary brain tumors. Cancer 75: 1151– Philadelphia, PA: Churchill-Livingstone. 1161. Shaw EG, Rosdhal R, D’Agostino RB Jr et al. (2006). A phase Weitzner MA, Meyers CA, Valentine AD (1995b). II study of donepezil in irradiated brain tumor patients: Methylphenidate in the treatment of neurobehavioral effect on cognitive function, mood, and quality of life. slowing associated with cancer and cancer treatment. J Clin Oncol 24: 1415–1420. J Neuropsychiatry Clin Neurosci 7: 347–350. Sminia P, van der Kleij AJ, Carl UM, Feldmeier JJ, Hart- Wen PY, Marks PW (2002). Medical management of mann KA (2003). Prophylactic hyperbaric oxygen treat- patients with brain tumors. Curr Opin Oncol 14: ment and rat spinal cord irradiation. Cancer Lett 191: 59– 299–307. 65. Yellen SB, Cella DF, Webster K, Blendowski C, Kaplan Spence AM, Krohn KA, Edmonson SW, Steele JE, Rasey JS E (1997). Measuring fatigue and other anemia-related (1986). Radioprotection in rat spinal cord with WR-2721 symptoms with the Functional Assessment of Cancer following cerebral lateral intraventricular injection. Int J Therapy (FACT) measurement system. J Pain Symptom Radiat Oncol Biol Phys 12: 1479–1482. Manage 13: 63–67. 23

Neurocognitive testing in clinical trials

Jennifer A. Smith and Jeffrey S. Wefel

Importance of formal neurocognitive testing of-phase-II meeting regarding a novel radiation sensitizing agent, October 21, 1998). Since the early 1990s it has been recognized that Impaired neurocognitive functioning occurs in the “net clinical beneﬁt” of a therapy includes not the majority of patients with central nervous system only traditional survival endpoints but also beneﬁts (CNS) tumors and has been shown to be impaired in terms of symptoms and quality-of-life end- by cancer therapies for tumors arising outside the points (O’Shaughnessy et al., 1991). With increasing brain (Meyers et al., 1995; Wefel et al., 2004b). Mag- awareness that it is often inadequate to measure netic resonance imaging (MRI) abnormalities may survival without consideration of the “quality” not always be clinically symptomatic (Fleissbach et of that survival, there has been a call to develop al., 2003) and patients with deteriorating neurocog- and include neurocognitive and patient-reported nitive function may not always show concomitant outcome (PRO) measures into modern trial design. structural brain changes on imaging. In fact, Meyers Members of the Food and Drug Administration and Hess (2003) demonstrated that, in patients with (FDA), National Cancer Institute (NCI), American primary brain tumors, neurocognitive dysfunction Association for Cancer Research (AACR), and Amer- occurred in advance of MRI evidence of tumor pro- ican Society of Clinical Oncology (ASCO) met in gression. The addition of measures of neurocog- 2006 to discuss endpoints for drug registration trials nitive function also predicts survival better than in primary brain cancer. The recommendations clinical prognostic factors alone in patients with generated from this meeting were provided for primary brain tumors, leptomeningeal disease, and the Oncology Drug Advisory Committee’s (ODAC) parenchymal brain metastases (Meyers et al., 2000, consideration and included a composite progression 2004; Sherman et al., 2002; Taphoorn & Klein, 2004). endpoint in which radiographical, neurocogni- Neurocognitive measures can be utilized in tive, neurological, and PRO are jointly considered clinical trials to monitor patient function and (http://www.fda.gov/cder/drug/cancer endpoints/ to determine if a therapeutic strategy results in: brain summary.pdf; accessed 10 April, 2008). The (1) improvement or stability in neurocognitive func- FDA has recently opined that a therapeutic agent tion associated with better tumor control; (2) less may be approvable if preservation of neurocogni- rapid decline in neurocognitive function associated tive function can be demonstrated even if survival with the disease; or (3) increased or decreased acute endpoints are equivalent (minutes of an end- or late neurotoxicity (Meyers & Brown, 2006). In the

Cognition and Cancer, eds. Christina A. Meyers and James R. Perry. Published by Cambridge University Press. C Cambridge University Press 2008.

320 Neurocognitive testing in clinical trials 321

case of CNS disease, serial monitoring may demon- and handedness, when appropriate, against which a strate improved neurocognitive function secondary patient’s performance can be referenced. Decisions to better tumor control or a reduced rate of expected regarding which tests to include in a trial must con- neurocognitive decline. Similar outcomes may be sider a myriad of issues including the psychomet- measured in disease outside the brain (Wefel ric properties of the measures, the population and et al., 2004b) and can separate adverse disease treatment under study, and their expected effect on effects from treatment-related neurotoxicity neurocognitive function, frequency of testing, and (Meyers et al., 1995; Wefel et al., 2004a). the study design. It is important to choose psychometrically sound, objective, and standardized measures of neurocognitive function. It has been routinely demonstrated Psychometric properties that mental status screening measures such as the When selecting tests for a clinical trial it is impor- Mini-Mental State Examination (MMSE), intended tant to establish that the tests are psychometrically to detect extreme changes in neurocognitive func- sound. Preferably the tests will have been validated tion such as delirium or significant dementia, in the population of interest and have adequate are inadequate (Meyers & Wefel, 2003). Patient- test-retest reliability to permit serial monitoring and reported outcomes (PROs) and subjective com- meaningful determination of change. Tests should plaints of neurocognitive dysfunction have been also be chosen to avoid ceiling and floor effects, demonstrated to have little relationship with for- such that the majority of patients are able to com- mal, objective assessment (Cull et al., 1996; van Dam plete the testing and the tests have sufficient sen- et al., 1998). Moreover, serial monitoring of PRO may sitivity and range to detect serial improvement or be susceptible to biases such as patient response deterioration in patient performance. shift, such that these measures offer insights into patient preferences and adjustment rather than changes in the actual level of function. Clinician- Focal versus diffuse function determined performance status, measures of symptoms (i.e., fatigue, pain, etc.), and activities of daily Measures of neurocognitive function can mea- living (ADL) have been thought to better cap- sure discrete processes such as expressive language ture important aspects of patient well-being. How- function or right upper-extremity strength. For each ever, they are also inappropriate as proxy mea- of these tests, regional damage located outside the sures of cognitive function and frequently appear area specifically assessed by these tests is likely to insensitive to meaningful changes in patient func- exert relatively little influence on test performance. tion during longitudinal trials (Meyers & Hess, In contrast, some tests (e.g., information process- 2003). ing speed) will reveal performance deficits following damage to any one of a number of areas in the brain due to the more distributed nature of these cogni- Choice of neurocognitive tests tive processes. If the goal is to identify where the damage is occurring then these tests are less than As discussed previously, neurocognitive testing ideal. They will identify that something is wrong but provides direct, objective evidence of patient cog- not where it is wrong. nitive function that often corresponds to the struc- There are tradeoffs associated with the test tural and functional integrity of the brain. An advan- choice. If the desire is to detect any damage then a tage of neurocognitive testing is the standardized single test sensitive to activity in a number of brain administration and scoring procedure coupled with regions may be preferable. Benefits of such a choice normative data adjusted for age, education, gender, include decreased time for both test administration 322 Section 3. Interventions and implications for clinical trials

and data processing. If the desire is to narrow down These considerations are illustrated in two recent the damage to a particular locus of functioning then trials conducted by Pharmacyclics, Inc. assessing it would often be necessary to employ multiple tests the effect of motexafin gadolinium on the neurocog- that specifically target discrete brain areas or func- nitive functioning in patients with brain metastases. tional domains. However, even in this case, a single As brain metastases arise in a number of regions and test could suffice if damage to a specific brain area as the effect of any one metastasis could be quite is suspected and the neural network underlying the subtle, the study design group chose a number of neurocognitive function of interest is not widely dis- tests that measured neurocognitive domains sensi- tributed. If, however, the study is more exploratory, tive to the functioning of specific and distinct brain a greater number of tests would be needed. As dis- regions. The desire was to capture the effect of a sin- cussed below, selection of tests that can localize dis- gle lesion but at the same time allow for sensitivity to crete versus distributed functions will be predicated functioning throughout the brain. Additionally, the on the expected treatment or disease effect on the trial was designed to describe the functional deficits brain. due to brain metastases, not simply to catalogue the presence or absence of any deficit.

Impact of disease and treatment

In selecting a test battery it is critical to consider Frequency of testing the expected impact of the disease or treatment on neurocognitive function. In some cases, the mecha- The frequency of testing depends, in part, on the nisms of action of a therapy may be well understood rate of functional change and the goal of the follow- and hypothesized to potentially disrupt a specific up assessments. If the rate of change is slow or the neurocognitive domain (e.g., memory dysfunction). goal is to characterize performance months or years For example, Meyers et al. (1997) studied CI-980 in after initiation of treatment then infrequent or a Phase II trial as a potential therapy for individu- delayed follow-up would suffice. Similarly, if mem- als with ovarian and colorectal cancer. This agent is ory for particular test information plays a large role a synthetic mitotic inhibitor that shares structural in subsequent test performance (practice effect) and functional similarities with colchicine, binds to and alternative validated versions of the test are not tubulin at the colchicine-binding site, and crosses available, then the scores may be biased if testing is the blood–brain barrier. Cognitive testing, includ- too frequent. ing monitoring of memory function, was a com- Additional factors may dictate test frequency, ponent of this protocol because of the ability of such as the monetary burden of testing or patient colchicine to selectively damage cholinergic neu- compliance. Patients having regular study visits are rons in and around the hippocampus and basal often willing to submit to additional neurocognitive forebrain, structures critical to learning and mem- testing (Herman et al., 2003; Mehta et al., 2002). In ory functions. Serial testing demonstrated declines a large multisite Phase III trial in patients with brain in memory function using standardized neuropsy- metastases (Meyers et al., 2004) the protocol speci- chological measures. fied monthly visits for 6 months and then every third However, often either the mechanism of the month thereafter. Frequent testing was necessary potential neurocognitive toxicity of a therapeutic due to the sharp decline in functioning observed strategy is not well understood or the expected in these patients as well as their short overall life brain lesions are not homogeneous enough (i.e., expectancy; 98% of patients completed testing at right parietal, left frontal, etc.) to permit selection of baseline and 87% completed the 6 months follow- tests based on known brain–behavior relationships. up testing. Neurocognitive testing in clinical trials 323

Study design: baseline versus repeated testing measures of abstract reasoning or problem solving, this procedure may inadvertently alter the nature of Aspects of the study design can greatly influence the test and thus invalidate the findings. The effects test selection. Baseline, pre-treatment assessments of specific learning on future performance may be are sometimes performed to serve as a predictor of reduced by choosing tests that have minimal prac- later functioning (e.g., Meyers et al., 2000). Again, tice effects or multiple versions that can be alter- the tradeoffs are apparent. There can be less empha- nated over time. These alternative versions should sis on the feasibility of the number of tests as they be standardized and externally validated. will be given only at baseline and not at repeated visits. A variety of tests may be included to pro- Implementation in clinical trials vide a comprehensive picture of functioning. Per- formance on these tests can be used to predict later There are several challenges to the successful imple- functioning or can aid the development of prognos- mentation of neuropsychological testing. In addi- tic categories (Meyers et al., 2000, 2004; Sherman tion to the issues surrounding the choice of tests et al., 2002; Taphoorn & Klein, 2004). Baseline per- and frequency of testing as summarized above, one formance can also be used to divide people into must consider the personnel required for testing, groups at baseline for stratification purposes in the the training of such personnel, the data collection analysis of related endpoints. Correlations between procedures and the particulars of data analysis. performance on neurocognitive tests and other standard tests can be computed. Required personnel and training Repeated measurement of neurocognitive performance is required if it is to be used as an end- During study planning, a neuropsychologist should point; measures of both baseline and follow-up per- be consulted to discuss potential treatment bene- formance are needed. The baseline measurement fits and toxicities in terms of cognitive function, the ensures balance across treatment arms or describes frequency of testing, the choice of appropriate tests the incoming level of functioning of patients on the (i.e., tests that assess cognitive processes of interest trial. Performance is then assessed over time for and that have adequate psychometric properties for change, especially due to treatment. the purpose of the study), and inclusion or exclusion Measurement of patient performance over time criteria for the study population given properties of may be confounded by the effects of repeated prac- the tests to be used (e.g., certain tests are not appro- tice. Test performance often improves despite the priate for patients over/under a certain age or those absence of underlying change or even a decline in not fluent in English). Additionally, a neuropsychol- neurocognitive function. This improvement can be ogist can train the study personnel and ensure qual- divided into two broad categories: general improve- ity control of the data collection process across the ment as the patient learns the task demands or entire trial. relaxes in the presence of the examiner, and more Neuropsychological testing offers the ability to specific improvement that results from learning of objectively assess neurocognitive functions using the unique information presented in a test, such as standardized procedures that limit the variability the particular words presented in a memory test. between examiners. However, this requires care- Procedures can be introduced to reduce the effects ful training and monitoring when personnel with- of both general and specific learning. Practice tests out formal education in neuropsychology are con- can be given prior to formal testing to reduce the ducting the tests. The feasibility of multinational effects of the general testing situation on later test repeated administration of testing by trained study performance. However, for measures that assess personnel was demonstrated in a recent Phase III performance in response to novel situations, such as trial in brain metastases (Meyers et al., 2004). Our 324 Section 3. Interventions and implications for clinical trials

experience suggests that the process of training and time at which the delayed recall portion of the test certifying study personnel to administer standard- was begun, to ensure that the standardized delay ized neuropsychological tests is most effective when interval is adhered to for each patient. Similarly, it is the following components are included in the study important to include indicator statements on each procedures: form that the certified test administrator completes, 1. All individuals administering the tests undergo which specify whether the test was completed or training including review of a training manual not completed due to issues such as patient non- designed by the neuropsychologist that includes compliance, neurological deficit (e.g., hemiparesis) standardized instruction sets and administration or altered mental status (e.g., confusion). This infor- procedures. mation greatly assists the determination of whether 2. Review of a training video demonstrating and the missing data represent informative (i.e., pro- explaining each test, accompanied by a post- gression of neurocognitive dysfunction) or non- test that is reviewed by the neuropsychologist to informative (i.e., patient refusal/non-compliance, ensure trainee comprehension of key procedural examiner error) trial information, which has an elements. impact on the data analyses. 3. Administration of a practice test to a non-patient Early and regular feedback regarding test admin- colleague by the trainee that is reviewed for accu- istration and data collection is recommended. racy by the consulting neuropsychologist and Unforeseen practices in the administration, timing, followed up to discuss administration errors or or scoring of tests may arise, especially in interna- concerns. tional, multicenter trials. For instance, if the study The successful completion of all of these com- endpoint is a measure of change from baseline, then ponents was required before a trainee was deemed incorrect baseline test administration for a number certified to administer the cognitive test battery to of patients would invalidate both their baseline and any protocol patient. Ideally, this training and cer- their subsequent change-from-baseline measure- tification will occur close in time to when the first ments; this could compromise the integrity of the protocol patient is seen for their baseline evalua- endpoint results. Two of the more common admin- tion to minimize the chance of examiner drift (i.e., istration and scoring issues include failing to stop the tendency to forget prior training or deviate from at the specified time for timed tests and failing to standardized procedures). From a pragmatic stand- distinguish between a score of zero (for instance point as well, a neuropsychologist may need to remembering zero words on a memory test) and be consulted in trials involving cognitive testing as an incomplete or missing test score; the two should test publishers often will not sell or permit use of be handled quite differently in the statistical analy- these instruments by individuals without appropri- sis. Ideally, the neuropsychologist will review all test ate training. results to ensure standardized procedures were followed and patient response errors were appropriately marked. Similarly, the neuropsychologist may Data collection calculate the final scores for each test to provide Modifications to the test forms such as including a blinded, central review that eliminates potential the standardized instruction set on each form can bias in assessing outcomes. facilitate consistent, accurate test administration. Additional formatting may also be undertaken to Statistical and interpretive considerations ensure proper test administration and data capture. For example, we have generated a space on Confounding variables and missing data our modified memory test forms where the certified test administrator is required to record the time The interpretation of the neurocognitive data at which the learning trials were completed and the depends in part on the context of the overall trial Neurocognitive testing in clinical trials 325

results. Information regarding possible confounders score of less than −1.65 × SD from this distribution of neurocognitive function (e.g., concomitant med- is said to have deteriorated (assuming higher scores ications such as steroids, medical complications indicate improved performance) as it is unlikely to such as seizures, and depression or other mood have occurred by chance (under the null hypothesis disorders) should be collected throughout the trial. only 5% of patients decline by that much or more). Performance of neurocognitive testing at the same Note that calculating the SD of the change scores time as other physical examinations or staging eval- is mathematically equivalent to calculating the RCI uations (e.g., neurological exams, MRI scans) pro- according to the formula presented by Jacobsen and vides contextual information for analyzing the neu- Truax, using the correlation between Time 1 and rocognitive outcomes over time. Time 2 scores and the SD of the Time 1 scores. Missing data, often not at random, can occur Since its introduction, the RCI has been sub- when patients who are experiencing diminished ject to improvements and some criticisms (Chelune neurocognitive capacity become unable to com- et al., 1993; Crawford & Garthwaite, 2006; Maassen, plete the neurocognitive testing. It is possible, as 2004; Tempkin, 2004; Tempkin et al., 1999). Despite described previously, to collect information on the this, it remains a common and frequently justifi- reason for such missing values and to then incorpo- able method for analysis. Probably the most helpful rate this information into the analysis. For instance, improvement is an adjustment for practice effects inability to perform the test due to a new impair- introduced by Chelune et al. (1993). Practice effects ment of comprehension, such that the person can- are also estimated from the reference data and man- not understand test instructions as they could ifest as a positive shift in the reference distribution. before, might rightfully be considered an indication This can be conceptualized as a distribution that is of neurocognitive decline. not centered at zero but instead at some increased delta indicating a mean effect of practice. Analysis of patient data in the trial then incorporates this prac- Analytical approaches tice effect such that a smaller decrease is now considered indicative of decline (−1.65 × SD + delta). Reliable change index A criticism of the RCI method is that the parame- The reliable change index (RCI) is a method for ters from the reference data are treated as fixed val- determining if and when a patient has deterio- ues not subject to measurement error themselves. rated. According to the RCI method (Jacobsen & This topic is addressed in an article that promotes Truax, 1991), a distribution of change scores is cre- regression-based methods to predict change scores ated under the null hypothesis of no change. The (Crawford & Garthwaite, 2006). The authors per- distribution parameters can be generated from formed simulations that varied the size of the ref- published normative data or can be generated from erence dataset and found that the regression-based previous study data. Consider a hypothetical distri- method outperformed the RCI at small reference bution of follow-up (Time 2) minus baseline (Time dataset sizes (e.g., n values of 5, 10, or 20), but 1) scores for 100 normal individuals or control for moderate (n = 50) to large (n = 100 or more) patients tested soon after baseline. Barring prac- reference datasets the RCI performed well, in that tice effects, the scores are not predicted to change the predicted number of declines was within 1% of over time and the differences in the follow-up minus the actual number. In fact, for reference datasets as baseline scores reflect measurement errors. Thus small as 20 the predicted number of declines was the distribution is centered approximately at zero within 2% of the actual number. and 90% of the scores are within 1.65 standard devi- As mentioned above the parameters from the ref- ations (SD) of the center. Taking this distribution erence distribution can be obtained from actual pre- as the reference distribution for future trials, an vious study data (calculate the SD of the change observed patient with a follow-up minus baseline scores directly or use the correlation of Time 1 and 326 Section 3. Interventions and implications for clinical trials

Time 2 data, and Time 1 SD to calculate the SD of deteriorated; the best predictor of Time 2 scores was the change scores) or from published test-retest reli- the Time 1 score. The prediction intervals based on ability (correlation) coefficients and baseline SDs. the regression techniques were on a par with the It’s important in the latter case that the reliability intervals from the practice-corrected RCI; relatively coefficients and the Time 1 SDs are calculated from similar misclassification rates were observed. Before the same study population. It is not recommended, choosing between these methods it would be advis- for instance, to use a published correlation coeffi- able to read the relevant literature as many authors cient and the baseline SD from another source. This have weighed in on the pros and cons of these meth- will not result in an accurate estimate of the SD of ods (Chelune et al., 1993; Crawford & Garthwaite, the change score distribution because the published 2006; Maassen, 2004; Tempkin, 2004; Tempkin et al., correlation coefficient is dependent on the baseline 1999). spread of data from which it was derived, a phe- The above discussions center on the statistical nomenon known as “range restriction.” properties of a single follow-up minus baseline score. In practice there are often repeated follow-up measurements. Without any correction for multiple Regression-based methods for predicting comparisons it is likely that more than the nomi- change scores nal level of patients will be misclassified as declined. These methods are similar to the RCI as they pre- Researchers have thus tried to control the effect of dict follow-up scores under the null hypothesis of multiplicity by requiring confirmation of the sus- no treatment effect; if a patient’s score is beyond pected decline, by either requiring decline in more the limits of the prediction interval the patient is than one measure (often from a similar domain of determined to have deteriorated (Crawford & Garth- functioning), or maintenance of that decline on a waite, 2006). For large reference datasets the RCI subsequent follow-up test. can be considered a subset of this method where the predictor (X) is the Time 1 data and the outcome value (Y) is the Time 2 score in the reference Time-to-event analyses dataset. The spread of the residuals around Y forms the reference distribution to which the observed The RCI and regression-based methods result in follow-up Y values are compared. A strength of these a categorization of decline, yes or no, at a given methods is the ability to incorporate additional pre- time. This leads nicely to a time-to-event analysis dictor variables such as baseline demographics (e.g., in which the endpoint is transformed from one age and education) in addition to the Time 1 data. of repeated continuous measurements to a single In addition, these methods do not treat the param- time variable representing time to neurocognitive eters from the reference sample as fixed values but decline. One advantage of this analysis is that instead treat them as having random variation. They sporadically missing values that are followed by test incorporate potential regression to the mean effects scores indicating preserved functioning no longer and allow for greater variation in outcome variables require compensation in the analysis. In addition, for patients with extreme values on the predictor patients who exit the study early or have differential variables. follow-up can contribute to the estimate of time A comparison of RCI and regression-based meth- to progression in an unbiased manner because of ods was conducted by Tempkin et al. (1999) using the ability to include censoring in time-to-event simulations. They concluded that although the analyses. Lastly, the time-to-event analysis confers additional demographic variables were related to an advantage over a repeated measures analysis of performance, they had relatively little impact on the the continuous outcome: the endpoint offers an number of patients incorrectly classified as having interpretation of the proportion of patients with Neurocognitive testing in clinical trials 327

clinically meaningful neurocognitive decline exploratory analyses and studies rather than conﬁr- (events) or change in the timing of such events. matory Phase III studies where it is necessary to pre- specify effects.

Repeated measures analyses Q-TWiST methods The models discussed above turn a continuous outcome measure into a dichotomous variable indi- Statistical methodologies have been developed to cating decline (yes/no). It is also possible to ana- analyze a combined endpoint of more traditional lyze the outcome measures directly using general- measures, such as survival, with neurocognitive and ized linear mixed effects modeling. These models PRO measures. Techniques include quality-adjusted estimate and incorporate the correlations induced survival (Glasziou et al., 1990) and quality-adjusted by multiple testing over time and avoid the mul- time without symptoms of disease or toxicity of tiplicity issues present in the RCI and regression- treatment (Q-TWiST) methodologies (Gelber et al., based methods. They are powerful and flexible and 1993). allow for random and fixed effects, varying functions of decline, and a variety of distributions for Conclusions the outcome measures. For example a skewed outcome could be modeled with a gamma distribu- In recent years, regulatory agencies have empha- tion. Although they can incorporate missing val- sized the importance of clinical trial endpoints that ues, there are most often assumptions about the directly measure clinical benefit, in contrast to sur- random nature of the missing variables that would rogate measures of benefit such as tumor response. need to be justified. More generally, the strength of They have also emphasized the importance of “net (and the liability of) these models lies in the abil- clinical benefit” that captures not just survival but ity to (and need to) specify the form of the relation- the quality of that survival. Neurocognitive end- ship between treatment, time, and outcome. These points consist of objective, standardized tests that models generally require parametric assumptions directly measure a patient’s cognitive function. The about the data. information obtained with these measures is not captured through patient self-report (e.g., quality of life questionnaires), mental status screening instru- Growth mixture modeling ments, or a general neurological exam. Thus, they Growth mixture modeling has garnered some atten- provide a unique and direct assessment of patient tion recently for the ability to estimate decline benefit. (growth) curves separately for latent classes of Several theoretical and logistical issues must be patients. The models can be modified to allow considered to successfully incorporate neurocogni- for separate control arm curves but similar treat- tive testing into clinical trials. The expected impact ment effects across the latent classes, or for differ- of disease and treatment on neurocognitive func- ential treatment effects across the classes as well. tioning, choice of test, and frequency of testing must These models have been successfully used to iden- be determined. In addition, successful implemen- tify subgroups of patients that have unique trajec- tation includes personnel training, management of tories across time (Muthen et al., 2002). Although practice effects, early and frequent data review, and these models allow for discovery of subgroups of appropriate statistical methodologies. Despite these patients whose performance over time is distin- challenges, the feasibility of neurocognitive test- guished from the rest, the same issues of multiplicity ing and analysis has been demonstrated in recent arise as might be expected in the case of post-hoc large randomized multinational Phase III trials in subgroups. These models may be better tools for oncology. 328 Section 3. Interventions and implications for clinical trials

REFERENCES Meyers CA, Byrne KS, Komaki R (1995). Cognitive deficits in patients with small cell lung cancer before and after Chelune GJ, Naugle RI, Luders H et al. (1993). Individual chemotherapy. Lung Cancer 12: 231–235. change after epilepsy surgery: practice effects and base- Meyers CA, Kudelka AP, Conrad CA et al. (1997). Neurotox- rate information. Neuropsychology 7: 41–52. icity of CI-980, a novel mitotic inhibitor. Clin Cancer Res Crawford JR, Garthwaite PH (2006). Detecting dissociations 3: 419–422. in single-case studies: type I errors, statistical power and Meyers CA, Hess KR, Yung WK et al. (2000). Cognitive func- the classical versus strong distinction. Neuropsychologia tion as a predictor of survival in patients with recurrent 44: 2249–2258. malignant glioma. J Clin Oncol 18: 646–650. Cull A, Hay C, Love SB et al. (1996). What do cancer patients MeyersCA,SmithJA,BezjakAet al. (2004). Neurocognitive mean when they complain of concentration and mem- function and progression in patients with brain metas- ory problems? Br J Cancer 74: 1674–1679. tases treated with whole-brain radiation and motexafin Fliessbach K, Urbach H, Helmstaedter C et al. (2003). Cog- gadolinium: results of a randomized phase III trial. JClin nitive performance and magnetic resonance imaging Oncol 22: 157–165. findings after high-dose systemic and intraventricular MuthenB,BrownCH,MasynKet al. (2002). General growth chemotherapy for primary central nervous system lym- mixture modeling for randomized preventive interven- phoma. Arch Neurol 60: 563–568. tions. Biostatistics 3: 459–475. Gelber RD, Goldhirsch A, Cole BF (1993). Evaluation of O’Shaughnessy JA, Wittes RE, Burke G et al. (1991). Com- effectiveness: Q-TWiST. Cancer Treat Rev 19 [Suppl. A]: mentary concerning demonstration of safety and effi- 73–84. cacy of investigational anticancer agents in clinical trials. Glasziou PP, Simes RJ, Gelber RD (1990). Quality-adjusted J Clin Oncol 9: 2225–2232. survival analysis. Stat Med 9: 1259–1276. Sherman AM, Jaeckle K, Meyers CA (2002). Pretreatment Herman MA, Tremont-Lukats I, Meyers CA et al. (2003). cognitive performance predicts survival in patients with Neurocognitive and functional assessment of patients leptomeningeal disease. Cancer 15: 1311–1316. with brain metastases: a pilot study. Am J Clin Oncol 26: Taphoorn MJ, Klein M (2004). Cognitive deficits in 273–279. adult patients with brain tumours. Lancet Neurol 3: Jacobson NS, Truax P (1991). Clinical significance: a sta- 159–168. tistical approach to defining meaningful change in psy- Tempkin NR (2004). Standard error in the Jacobson and chotherapy research. J Consult Clin Psychol 59: 12–19. Truax reliable change index: the “classical approach” Maassen GH (2004). What do Temkin’s simulations of reli- leads to poor estimates. J Int Neuropsychol Soc 10: 899– able change tell us? J Int Neuropsychol Soc 10: 902–903. 901. Mehta MP, Shapiro WR, Glantz MJ et al. (2002). Lead-in Tempkin NR, Heaton RK, Grant I et al. (1999). Detecting sig- phase to randomized trial of motexafin gadolinium and nificant change in neuropsychological test performance: whole-brain radiation for patients with brain metastases: a comparison of four models. J Int Neuropsychol Soc 5: centralized assessment of magnetic resonance imaging, 357–369. neurocognitive and neurologic end points. J Clin Oncol van Dam FS, Schagen SB, Muller MJ et al. (1998). Impair- 20: 3445–3453. ment of cognitive function in women receiving adjuvant Meyers CA, Brown PD (2006). Role and relevance of neu- treatment for high-risk breast cancer: high-dose versus rocognitive assessment in clinical trials of patients with standard-dose chemotherapy. J Natl Cancer Inst 90: 210– CNS tumors. J Clin Oncol 24: 1305–1309. 218. Meyers CA, Hess KR (2003). Multifaceted end points in Wefel JS, Lenzi R, Theriault R et al. (2004a). “Chemobrain” brain tumor clinical trials: cognitive deterioration pre- in breast cancer? A prologue. Cancer 101: 466–475. cedes MRI progression. Neurooncology 5: 89–95. Wefel JS, Lenzi R, Theriault R et al. (2004b). The cognitive Meyers CA, Wefel JS (2003). The use of the Mini-Mental sequelae of standard dose adjuvant chemotherapy State Examination to assess cognitive functioning in can- in women with breast cancer: results of a prospec- cer trials: no ifs, ands, buts, or sensitivity. J Clin Oncol 21: tive, randomized, longitudinal trial. Cancer 100: 3557–3558. 2292–2299. Index

abstract thinking, assessment of 15 academic achievement assessment of 71 children with NF-I 221, 223–24 diagnosis relating to 8 HSCT (hematopoietic stem cell transplantation) children 232 importance of early intervention 74–75 reduced following radiation therapy 204 tests measuring 62 acute lymphoblastic leukemia (ALL) incidence of 57–58 neuropsychological deﬁcits 58–60 survival rates 57 treatment 58 effect on child’s development 58 MRI studies 23–24 adaptive functioning 67, 71–73, 205 adrenocorticotropic hormone (ACTH) 251, 273 affective distress and cognitive performance 51–52 Alzheimer’s disease cognitive rehabilitation 285 external memory aids 286 versus mental stimulation 286 donepezil, effectiveness of 177 and low testosterone levels 117–18 memantine 178–79 susceptibility to 117 androgens androgen deprivation therapy 126–35 androgen substitution therapy 118 anti-androgens 120 effect on spatial abilities 116–18, 125 intermittent androgen suppression (IAS) 120

329 330 Index

anemia and cognition Beck Depression Inventory II (BDI-II) 296 in cancer patients 51, 102 behavioral intervention strategies 281–82 hematological cancer patients 232–33 degenerative conditions 285–87 in non-cancer patients 233 traumatic brain injury (TBI) 281, 283 children and infants 233 attention deficits, strategy training for 283–84 animal models 270, 276 memory and learning deficits, strategy training for associated with cancer cells 271–72 284–85 attention deficit hyperactivity disorder (ADHD) 271 problem solving for executive function deficits cognitive tests 274–75 285 criteria for 271 benzodiazepines (BZPs) 262–63 distinction from animal tests 270 biological response modifiers (BRMs) 106 fatigue tests 275–76 see also interferon alpha (IFN-α) laboratory studies, radiation-induced CNS injury 87–88 brain edema models of depression and stress 272–73 executive function deficits 174 and NF-I gene 215, 217, 224 pharmacological interventions 35, 314–16 olfactory bulbectomy model of depression 271 symptoms of 175 rationale for using and problems with 270–71 brain metastases 170 tests of depression 273–74 clinical presentation 171 anti-epileptic drugs, negative effect on cognitive functions cognitive sequelae of cancer therapy 175 51, 151, 177 adjuvant medications 177 anti-Hu antibodies 244–46 radiotherapy 175 anti-VGKC antibodies 246–47 surgery 176 antidepressant treatments 266, 274, 298 systemic anticancer therapy 176 antineoplastic therapies and cognitive impairment 264 diagnosis and treatment in children 58 antineuronal antibodies, paraneoplastic limbic effect of motexafin gadolinium 322 encephalitis (PLE) patients 240–41, 244–47 epidemiology 170–71 antipsychotic drugs for delirium 262 neurocognitive impairment 173 anxiety, effect on cognitive performance 51–52, 263 due to cancer 174–75 apoptosis, effects of abnormal signaling 214–15 etiology and pathogenesis 174 aqueductal stenosis 215 incidence of 173–74 aromatase inhibitors, breast cancer 119, 126 management of 177;donepezil 177–78;hyperbaric assessment see neuropsychological assessment oxygen 178; methylphenidate 177; NMDA receptor astrocytes 84 178–79; transplantation of purified astrocytomas 143, 216 oligodendrocytes 178 attention prevention of; avoidance of hippocampus in assessment of 12–13, 62, 68–69, 74 whole-brain radiation therapy 180–81; effect of anemia 233 erythropoietin (EPO) 180; radiation dose and effect of radiation therapy on sustained 204–5 fractionation 179; radiosurgery alone 179–80 remediation programs for 73–74, 287, 307 prognosis 171 strategy training for deficits in 283–84 risk of developing 34 and working memory 69 treatment 171–73 attention deficit hyperactivity disorder (ADHD) 215, 271 brain structure and function Attention Process Training (APT) technique 73–74, 287 impact of disease, genetics and therapies autoimmunity 104–6 children with paraneoplastic opsoclonus-myoclonus nervous system development 56–57 250–51 brain tumors patients with paraneoplastic limbic encephalitis 244–47 growth rate 35 theory of 239–40 location and size of see also immunotherapy and degree of deficits 34–35, 174, 203 Index 331

and presenting symptoms 34 osteosarcoma patients treated with high-dose and psychiatric symptoms 37 methotrexate 24–25 neuroﬁbromatosis NF-I 215–17 childhood brain tumors 58, 198 symptoms/symptom clusters 312, 314 lifelong annual medical follow-up 206 see also childhood brain tumors; high-grade gliomas; medical late effects 198–99 low-grade gliomas cardiac complications 199–200 breast cancer endocrine complications 200–1 and chemotherapy 36 gastrointestinal and hepatic complications 201–2 MRI studies of cognitive changes 25–28, 104 hearing and sight disorders 201 and hormonal therapy 119 neurological complications 202 impact on cognitive function 120–26 pulmonary complications 200 renal complications 201 cardiac complications of childhood cancer survivors secondary malignancies 199 199–200 neuropsychological late effects 202 central nervous system (CNS) neurocognitive late effects 202–5 disease, assessment of patients with 48–49 psychosocial late effects 205–6 radiation-induced toxicities 91 treatment and rehabilitation 206 see also nervous system; radiation-induced CNS injury childhood cancer cerebral vascular accident (CVA), behavioral intervention assessment see neuropsychological assessment, strategies for 283–85 children “chemobrain”/“chemofog” 49–50 behavioral interventions 73–74 in children 98 common forms of 57–58 patient-reported 98–99 acute lymphoblastic leukemia (ALL) 23–24, 58 retrospective studies 98 brain tumors see childhood brain tumors symptoms of 97–98 increase in survival rates 57 chemotherapy neuropsychological deﬁcits 58–59 for brain metastases 173 children treated for ALL 59–60 chemotherapeutic agents children with brain tumors 202–6 neurotoxicities associated with 50, 99; risk factors school and educational support 301–2 underlying 99–103 support for parents 303 and cognitive dysfunction 49–50, 97, see also paraneoplastic opsoclonus-myoclonus (POM) 176 childhood leukemia see acute lymphoblastic leukemia etiological mechanisms underlying 176 (ALL) hematological malignancies 230–31 Children’s Oncology Group (COG) incidence and nature of 97–99 neuropsychological test battery 72–73 neuroimaging and EEG studies 103–4 screening guidelines 206 pharmacogenetic modulation of 104–6 ciclosporin 229 high-dose methotrexate for primary central nervous cisplatin system lymphoma 188–89 and hearing loss in children 70, 201 intrathecal methotrexate, effects on children 59 nephrotoxic effects of 201 for low-grade gliomas 147, 151 and osteosarcoma 25 neurocognitive changes following clinical neuropsychology 6 risk factors implicated in 20 assessment goals 8–10 studies demonstrating 19–20 assessment tests 12–16 neuroimaging of cognitive changes 19, 23 historical background 7–8, 56 breast cancer patients 25–28 standard evaluation procedures 10–12 childhood leukemia treated without radiation therapy see also neuropsychological assessment 23–24 clinical trial endpoints 4, 40, 45, 48, 162, 320, 327 leukoencephalopathy in adults 25 clinical trials see neurocognitive testing 332 Index

cognitive dysfunction Cognitive Remediation Program (CRP) 73–74, 287 in adult cancer patients 33 cognitive set shifting, assessment of 15–16 due to tumor progression in high-grade glioma 160 community re-integration 301 paucity of research 33 complementary and alternative medicine 307–8 in brain metastases patients 173–74 computerized axial tomography (CAT) 20–21 deficits due to cancer 174–75 low-grade gliomas 144 correlation to caregiver burden 39 in osteosarcoma 24 due to antineoplastic therapies 264 patients with paraneoplastic limbic encephalitis 248 IFN-α causing 107–8 core cognitive processes late delayed effect of radiation therapy 316 assessment of 68 in pediatric brain tumors, late effects 202–5 attention 68–69 potential causes of 4 processing speed 69 predictors of 4 working memory 69 symptoms and effects of 3–4 deficits in children cognitive function treated for ALL 59–60 assessment of 4 treated for brain tumor 204–5 for descriptive reports 8–9 corticosteroids 25, 36, 160, 251 following surgery 9 counseling 304 goals of 8–10 cranial radiation therapy reasons for referral 9–10 childhood brain tumors serial assessments 9 age risk factor 203 see also neuropsychological assessment brain injury caused by 203 role of reproductive hormones 115–16 cognitive deficits caused 203 androgen substitution therapy 118 impact on attention, processing speed and working estrogen in women 103, 116 memory 204–5 hormone replacement therapy (HRT), studies of 116 impact on global IQ 203–4 testosterone in men 117–18 impact on non-dominant hemisphere processes 204 cognitive rehabilitation 281–82 neuropsychological outcome of 59 cancer patients 287, 307 treatment of optic glioma 216–17 adults with primary brain tumors 287–88, 307 see also radiation therapy (RT) clinical application 289; guided questions 290–91; cytokines selecting appropriate intervention 289–91; timing and immune response 233–34 of intervention 291 indirect effects on CNS function 176 Cognitive Remediation Program (CRP) 287 neurotoxic side-effects of 50 low referrals for rehabilitation 287 pro-inflammatory activity of 103, 105 problem-solving therapy 288, 291–93 see also interferon alpha (IFN-α) and randomized controlled trials 289 research questions 288–89 data collection, test administration 324 definition of 282–83 delirium degenerative conditions 285 assessment of 260 cognitive rehabilitation versus mental stimulation 286 EEG and neuroimaging 262 external memory aids 286 laboratory assessment 261–62 strategies for acquired brain injury rating scales 261 Cicerone recommendations 283 classification of 259 strategy training 282 common causes of in cancer patients 260 attention deficits 283–84 definition and classification of 258–59 executive function deficits 285 management of 262–63 external and internal procedures 282–83 in palliative care settings 264–65 memory and learning deficits 284–85 risk factors 259–60 Index 333

dementia following radiotherapy 148–50, 175 cognitive rehabilitation 286–87 hematological cancer patients 229 and complementary and alternative therapies 308 and high-dose methotrexate treatment 24 depression dementia/pseudodementia 263 methotrexate-induced 229 evaluation to rule out 46–47 see also leukoencephalopathy; paraneoplastic disorders and hormone replacement therapy (HRT) studies 116 endocrine complications in long-term survivors of screening tools 162 childhood brain tumors 200–1 and whole-brain radiation therapy 173, 175 epidemiology demyelination, radiation-induced 84, 148, 188 brain metastases 170–71 depression 37, 298 low-grade gliomas 142–43 animal models of 272–73 secondary malignancies 199 animal tests of 273–74 epileptic seizures see seizures Beck Depression Inventory II (BDI-II) 296 erythropoietin (EPO), brain metastases 180 differentiating from neurocognitive impairment 37 estrogen 115 effect on cognitive performance 51–52 aromatase inhibitors, impact of 119, 126 major depression in cancer survivors 263 effect on cognitive function 103, 116–17, 126–35, 176 side-effect of interferon treatment 107 estradiol therapy for prostate cancer 136 dexamethasone 87, 314–16 luteinizing-hormone-releasing hormone (LHRH) differential diagnosis 8, 40 suppressing production of 119 patients with suspected paraneoplastic limbic see also tamoxifen (TAM) encephalitis 247 European Organization for Research and Treatment of of seizures in cancer patients 161 Cancer QLQ-C30 (EORTC) 38, 48, 91, 146, 158, 163 diffusion tensor imaging (DTI) 22 executive function (EF) 69 Distress Thermometer and Problem List 296–97 assessment of 15–16 donepezil in children 66, 69–70 Alzheimer’s disease 177 sample tests for 62 for fatigue and sedation 266 compensation-oriented cognitive rehabilitation 289–91 late radiation-induced brain injury 317 impairment of lung cancer patients 177–78 following IFN-α treatment 232 trials for brain tumors/metastasis 164, 178 formal problem-solving strategies for 285 Drosophila NF-I protein 217 improved with Ginkgo biloba 317 dyslexia, MRI studies 215, 223–24 rehabilitation strategies 307 experimental neuropsychology 6 edema see brain edema external radiation therapy (XRT) 36–37 electroencephalography (EEG) 104, 244, 262 eye, formation of 215–16 embryogenesis, impact of NF-I gene 214–16 emotional functioning family caregivers assessment of in children 71–72 factors predicting improvement in children with brain demoralization and distress 298–99 tumors 205–6 psychiatric medications 234 quality of life (QOL) beneﬁts of neuropsychological support services 297–99 assessment 38–40 see also anxiety; depression; mood support for 302–3 employment farnesyltransferase inhibitors 213–14, 217 return to work 46, 299 fatigue 51, 295 support needed in work settings 302 animal models and tests 275–76 vocational status of children with cancer 205 associated with anemia 232–33 encephalopathy complementary and alternative medicine therapies for chemotherapeutic agents associated with 99–100, 307–8 102–3 pharmacological interventions 266, 312–13 334 Index

fatigue (cont.) neurocognitive deficits and return to normal functioning 299 causes of 158–62 side-effect of interferon treatment 107, 232 cognitive rehabilitation 165 Food and Drug Administration (FDA) 40, 45, 48, 320 prevention of 164–65 Functional Assessment of Cancer Therapy (FACT) 48, 313, treatment of 163–64 317 and survival 156–58, 163 functional MRI (fMRI) 22–23 hippocampus breast cancer survivors 26–27, 104 avoidance of during whole-brain radiation therapy high-grade gliomas (HGG) 160 180–81 neurofibromatosis (NF-I) studies 222–23 hippocampal neurogenesis 85–86, 103, 164 reading and visual-spatial processing 223 history of neuropsychology 7–8, 56 hormonal deficiencies in long-term survivors of childhood genetics brain tumors 200–1 animal studies and NF-I gene 215, 217, 224 hormonal therapy 115 low-grade gliomas 143–44 in cancer treatment 119, 176 polymorphisms of genes and cognitive function 104–7 breast cancer 119; influence on cognitive functioning gliomas see high-grade gliomas; low-grade gliomas 50, 120–26 glucocorticoids 35, 177, 273 factors to include in studies 138 gray matter (GM) prostate cancer 119–20; influence on cognitive volume increase in NF-I patients 222 functioning 50, 126–35 volume loss following chemotherapy 26–27, 103 hormone replacement therapy (HRT) growth mixture modeling 327 dementia studies 116 for growth hormone deficiency 201 hair loss following radiation therapy 313 methodological aspects 136 effects on cognition cross-sectional versus prospective studies 136 prospective studies 190–91 group means comparison versus individual test score retrospective studies 191–93 analysis 137 interactions with whole-brain radiation therapy 188–89 mood and psychosocial factors 137 and white matter abnormalities 193–94 sample size 136 hearing loss in childhood tumor survivors 201 selection of cognitive tests 137 hematological malignancies 228 self-reported cognitive problems 137 cognitive deficits 229 neuroimaging studies 118 due to anemia and iron deficiency 232–33 reproductive hormones and cognition 115–18 immune response mechanisms 233–34 hospice agencies and quality of life 300 treatment effects 229–30; biological response hyperactive delirium 259–60, 262 modifiers 232;chemotherapy 230–31; hyperbaric oxygen (HBO) hematopoetic stem cell transplantation (HSCT) myelopathy prevention, animal studies 87 231–32; radiation therapy 231 treatment for brain metastasis 178 neurological complications of 228–29 hypoactive delirium 259, 262–63, 266, 300 treatment for cognitive and emotional dysfunction 234 hematopoetic stem cell transplantation (HSCT) 231–32 IFN-α see interferon alpha effects on neuropsychological functioning 230, 232 immune response mechanisms 233–34 neurological complications 228–29 immunotherapy high-dose-methotrexate-based chemotherapy for primary biological response modifiers (BRMs) 106 central nervous system lymphoma 187–88 impact on cognitive function 50 high-grade gliomas (HGG) 156 for paraneoplastic disorders 241 nervous system problems 158 factors influencing response to 241–42 neurocognitive assessment 162 poor prognosis for neurological improvement importance of 162–63 242 Index 335

response of patients with paraneoplastic limbic neurocognitive deficits, factors causing 147 encephalitis 248–49 medical therapy 151 individual psychotherapy 304 mood disorder 151 indomethacin, preventative trials 164 primary tumor 147–48 intellectual functioning radiation 148–50 children with NF-I 217–18 surgery 148 clinical tests of intelligence 12, 62 pathology 143 global IQ 66–67, 203–4 prognostic factors 144–45 impact of cranial radiation therapy on children 59, therapeutic management 202–5 chemotherapy 147, 151 impact of intrathecal methotrexate 59, 231 radiotherapy 146–47 see also academic achievement; core cognitive processes surgery 145–46 interferon alpha (IFN-α) 106–7 luteinizing-hormone-releasing hormone (LHRH) agonists animal studies of fatigue 276 119 and cognitive impairment 107–8, 232 impact on women 120 mechanisms underlying 108–9 prostate cancer treatment 50 side-effects of 107 impact on cognitive functioning 126, 135, interventions for 109 176 intrathecal methotrexate, effects of treating children with prevention of tumor flare with anti-androgens 59, 231 120 iron deficiency, effects of 232–33 irradiation see cranial radiation therapy; radiation macrocephaly in patients with NF-I 222 therapy (RT) magnetic resonance imaging (MRI) 20 and CNS tumors 320 Judgment of Line Orientation (JLO) test 219–20 dyslexia/reading-disabled patients 215 effects of tamoxifen on hippocampal atrophy 125 language leukoencephalopathy in adults 25 assessment of 14, 62, 70, 74 low-grade gliomas (LGG) 144, 146 critical period for development of 56 MR hyperintensities in children with NF-I 221–22 see also reading problems in NF-I children osteosarcoma patients 24–25 learned helplessness, animal stress models 273 paraneoplastic limbic encephalitis patients 244 learning disability (LD), children with NF-I 212, primary central nervous system lymphoma patients 217–19 189, 194 learning and memory see also functional MRI (fMRI); positron emission animal tests 274–75 tomography (PET); structural brain imaging assessment children 62, 67–68 methods strategies for Alzheimer’s disease 286 magnetic resonance spectroscopy (MRS) 22 see also memory comparison with MRI in breast cancer patients 28, leukoencephalopathy 34 in adults, chemotherapy-induced 25 impact of tamoxifen and hormone replacement therapy chemotherapeutic agents associated with 99, on brain function 125 189 medications following high-dose methotrexate treatment 24 antidepressants 266, 274, 298 limbic encephalitis see paraneoplastic limbic encephalitis antipsychotics for managing delirium 262–63 low-grade gliomas (LGG) 142 assessment results confounded by 51 clinical features 144 confounding tumor-related causes of cognitive deficits epidemiology and biology 142–43 160–61 genetics 143–44 delirium-inducing in cancer patients 260 imaging features 144 symptomatic pharmacotherapy 265–66 336 Index

medulloblastomas 57–58, 164 for osteosarcoma 24–25 effect of age at time of cranial radiation therapy 203 see also high-dose-methotrexate-based chemotherapy evaluation of 10-year survivors 299 for PCNSL psychosocial effects of treatment 205 methylphenidate (RitalinR ) 177 risk of secondary malignancies 199 for attention deﬁcit hyperactivity disorder (ADHD) treatment affecting attention, memory and new 271 learning 204–5 for fatigue 313, 316 memory high-grade glioma patients 164, 177 animal tests 274–75 for hypoactive delirium 266, 300 assessment of 14–15, 67–68, 74 microglia, impact of irradiation on 84–85 effects of androgen deprivation therapy 126, 135 Mini-Mental State Examination (MMSE) 38, 46, 117, 162, impairment 3 261, 321 in Alzheimer’s disease 285 modaﬁnil 164, 266 amnesia in paraneoplastic limbic encephalitis molecular imaging methods 23 patients 243 mood combined modality therapy in patients with primary assessment of 16 central nervous system lymphoma 192–93 cluster in newly diagnosed brain tumor patients 312 due to anxiety and depression 51–52 detrimental effect on test performance 137 following chemotherapy 27–28, 98, 230 disorders, link to cognitive dysfunction 51–52, 98–99, following hormonal therapy 135 151, 263 from radiation therapy 49 motor and sensory-perceptual function, assessment of 16, hematopoietic stem cell transplantation patients 62, 71 232 myelination 56–57 IFN-α treatment 232 demyelination, radiation-induced 84, 91, 148, 188 in patients with delirium 260 remyelination, transplantation of oligodendrocytes 178 and tamoxifen 120 myeloencephalopathy 103 processes 15 myoclonic encephalopathy of infancy 249 rehabilitation external memory aids 286 nervous system internal procedures, other regulated 286 development of 56–57 memory notebook 282, 284–85, 288 effects of early brain damage 57 using calendar system 291–92 plasticity of 57 role of hippocampus 180–81 treatment-related injury, effects on children 58 role of NMDA receptor 178–79 see also central nervous system (CNS) see also non-verbal memory; verbal memory; working neurocognitive impairment see cognitive dysfunction memory neurocognitive testing menopause Children’s Oncology Group (COG) test battery 72–74 chemotherapy and surgically induced 102, 116 implementation in clinical trials 323 effect of hormone levels during 116 data collection 324 hormone replacement therapy (HRT), studies of 116 personnel and training needs 323–24 mental stimulation tasks for Alzheimer patients 286 importance of formal testing 320–21 metastatic brain tumors see brain metastases Oregon Health Sciences University test battery 73 methotrexate (MTX) statistical and interpretive issues causing microvascular injury and myeloencephalopathy analytical approaches; growth mixture modeling 327; 103 Q-TWiST methods 327; regression-based methods for hematological malignancy 229 for predicting change scores 326; reliable change intrathecal methotrexate, effects of treating children index (RCI) 325–26; repeated measures analyses with 58–59 327; time to event analyses 326–27 leading to metabolic abnormalities 99–102 confounding variables and missing data 324–25 Index 337

test selection location 34–35; prognostic value of assessment factors influencing; focal versus diffuse function 37–38; side-effects of treatment 35–37 321–22; frequency of testing 322;impactofdisease cancer research benefits 40 and treatment 322; psychometric properties 321; clinician benefits 38 study design: baseline versus repeated testing 323 disease and treatment issues; adjunctive medications see also neuropsychological assessment 51; affective distress 51–52;anemia 51; assessment neurodegenerative conditions, cognitive rehabilitation of patients with CNS disease 48–49;chemotherapy 285–87, 291 49–50; fatigue 51;hormonaltherapies 50; neurofibromatosis (NF) immunotherapy 50; radiation therapy 49 brain structure and function correlates 221 high-grade gliomas 162–63 brain tumors 221 principles of assessment 45–46; assessment process congenital malformation 222 47; cognitive areas of interest 47–48; macrocephaly 222 patient-reported outcomes (PROs) 48;test MR hyperintensities 221–22 selection 46–47; timing of assessments 47 children with NF-I 217 quality of life benefits for patient and family academic achievement of; impact of visual spatial caregivers 38–40 abilities 219–21; and learning disability (LD) children 60 218–19 approaches to assessment 61–66 intellectual functioning 217–18 areas to evaluate; academic achievement 71; core diagnostic criteria 212, 219–21 cognitive processes 68–69; executive function (EF) functional imaging studies 222–23 69–70; language 70; learning and memory 67–68; genotype 212–13, 224 motor and sensory-perceptual function 71; history 211 non-verbal skills 70–71; overall intellectual phenotype 211–13, 223–24 functioning 66–67; psychosocial and adaptive Ras signaling system 213, 217 functioning 71–72 systemic impact future directions 75 benign tumors, focal growth dysregulation in 214 liaison with schools 61 brain malformations 214–15 serial assessments 60–61 brain tumors 215–17 test batteries; Children’s Oncology Group (COG) farnesyltransferase inhibitors, positive effect on 72–73; Oregon Health Sciences University 73 learning 217 see also cognitive function, assessment of on function 217 neuropsychology 6 role of NF-I gene, animal studies 217 ecological validity 75 neurogenesis evaluation attempts to rectify RT-induced decrease 87, 164 for descriptive reports 8–9 in hippocampus of rat brain 86 following surgery 9 impact of chemotherapies on 103 goals of 8 of neural stem cells, effects of radiation 85–86 reasons for referral 9–10 neuroimaging technologies 20–23 serial assessments 9 see also magnetic resonance imaging (MRI) standard procedures; clinical interview 10–11; neurons dissemination of findings 11–12; history taking 10; and neurogenesis 85–87 interpretation of results 11; test administration 11; rapid growth and pruning 56 test selection 10 sensitivity to radiation 85 future directions in 73–75 neuropsychological assessment 44–45 history of 7–8, 56 adult cancer patients 33–34, 44 terminology issues 71 benefits for patient’s clinical medical management neurosurgery see surgery 34; cancer surveillance 35; depression, effects on NF-I (neurofibromin-1) see neurofibromatosis (NF) cognition 37; presenting symptoms and tumor NMDA (N-Methyl-d-aspartate) receptor 178–79 338 Index

non-verbal learning disability (NVLD) 218–19 management of 248–49 non-verbal skills, assessment of in children 70–71 pathology 243 non-verbal memory 14, 59, 68 paraneoplastic opsoclonus-myoclonus (POM) autoimmunity 250–51 object perception, assessment of 13 clinical features 249–50 oligodendrocytes pathology 250 effect of radiation on 84, 91, 188 patient management 251–52 oligodendroglioma 144–45, 147 parenchymal hypothesis for radiation-induced CNS injury transplantation of promoting remyelination 178 84 onconeural antigens 239–41, 251 parents of childhood cancer survivors online support groups 306 information provision 75 opsoclonus-myoclonus see paraneoplastic support for 303 opsoclonus-myoclonus (POM) see also family caregivers optic tumors in neurofibromatosis (NF-I) patients 215–17 patient-reported outcomes (PROs) 48 Oregon Health Sciences University test battery 73 pediatric oncology see childhood cancer osteosarcoma, neuroimaging studies 24–25 peer support groups 304–6 oxidative stress, radiation-induced 86, 92 personality, assessment of 16 oxygen treatment for brain metastasis 178 phonological processing, functional MRI studies 223 pituitary tumors 200 palliative care settings planning ability, assessment of 15 hospice agencies 300 PLE see paraneoplastic limbic encephalitis management of delirium 264–65 polyunsaturated fatty acids, reducing paralysis, animal symptoms requiring pharmacotherapy 265–66 studies 87 see also supportive care POM see paraneoplastic opsoclonus-myoclonus (POM) paraneoplastic disorders 239 Porsolt Forced Swim Test, animal test of depression 273–74 antineuronal antibodies 240–41 positron emission tomography (PET) 23 and autoimmunity theory 239–40 effects of tamoxifen and estrogen on brain glucose clinical importance of 239 metabolism 125 clinical neurological outcome 241 hippocampal hypermetabolism in patients with factors causing neuronal injury 241 paraneoplastic limbic encephalitis 244 immunotherapy for 241–42 neural effects of chemotherapy in breast cancer patients limbic encephalitis (PLE) 242–43 27–28 autoimmunity 244–47 PPARγ agonists, role of pioglitazone in reducing clinical features 243–44 inflammation 87–88 diagnostic studies 244 practical needs of cancer patients 299 differential diagnosis 247 practice effects, test administration 9, 60, 136, 323, 325 pathology 243 primary CNS lymphoma (PCNSL) 187 patient management 248–49 cognitive dysfunction 189–90 opsoclonus-myoclonus (POM) prospective studies 190–91 autoimmunity 250–51 retrospective studies 191–93 clinical features 249–50 and treatment-related white matter abnormalities pathology 250 193–94 patient management 251–52 diagnosis 187 prognosis for neurological improvement 242 treatment 187–88 paraneoplastic limbic encephalitis (PLE) 242–43 delayed neurotoxicity 188–89 autoimmunity 244–47 problem-solving therapy 292–93 clinical features 243–44 for adults with cancer 288 diagnostic studies 244 for executive functioning deficits 285 differential diagnosis 247 at the Mayo Clinic 288, 291–93 Index 339

processing speed pathogenesis 83, 175 assessment of 13, 69 astrocytes 84 psychometric tests 62 classical model 83 deﬁcits following cranial radiation therapy 204–5, 230 contemporary view 86 effects of hormone therapy 126, 135 microglia 84–85 and IFN-α treatment 232 neural stem cells and neurogenesis 85–86 prophylactic interventions neurons 85 for children with cancer 73–75 parenchymal hypothesis 84 prophylactic cranial irradiation (PCI) 36 vascular hypothesis 83 prostate cancer, hormonal therapy 119–20 pharmacological interventions for 312 impact on cognitive function, studies 127 acute reactions 312–14 inﬂuence of luteinizing-hormone-releasing hormone early delayed reactions 314 agonists 50, 126–36 late delayed reactions 314–17 psycho-education 296–97, 306 preventative and therapeutic interventions 92 psychosocial functioning quantitative scoring of CNS toxicity 90–92 assessment of in children 62, 71–72 radiation necrosis 83, 91–92, 202 late effects, childhood brain tumor survivors 205–6 radiation tolerance of CNS tissues measures of 73, 75 factors associated with 88 psychosocial interventions 234, 304–6 importance of dose and fraction size 90 pulmonary complications in long-term survivors of mathematical models for tolerance doses 89 childhood brain tumors 200 neuret model of brain tolerance 90 see also central nervous system (CNS); radiation Q-TWiST methods 327 therapy (RT) quality of life (QOL) of cancer survivors 156 Radiation Therapy Oncology Group (RTOG) 91, 149, 171, assessment feedback helping to maximize 39–40 174, 179 clinical trial endpoint 40 radiation therapy (RT) 231 for primary brain tumors 320 acute reactions 175, 312–14 for radiation-induced brain injury 317 early delayed reactions 91, 175, 314 complementary and alternative medicine therapies endocrinological sequelae 200–1 improving 307–8 focal versus whole brain 149 hospice care 300 late delayed reactions 91, 175, 314–17 importance of multidisciplinary approach to for low-grade gliomas 146–47 interventions 312 negative effects on cognitive performance 23 interventions to improve, symptom-focused 281 childhood brain tumor patients 203–5 patient-reported outcomes (PROs) 48 CNS disease patients 49 patient’s appreciation of 163 low-grade glioma patients 148–51 in primary central nervous system lymphoma patients tolerance doses 192 for normal CNS tissues 88–89 and problem-solving therapy 288 for other normal CNS tissues 90 psychosocial support 234, 304–6 treatment volumes and doses for brain tumors 313 relationship to cognitive impairment 38–39 see also cranial radiation therapy; whole-brain research on health related 58 radiation therapy (WBRT) support services improving 296 radiosensitizers 173 tests measuring 73, 159, 296–97 radiosurgery 172–73, 179–80 raloxifene 125 radiation-induced CNS injury reactive oxygen species (ROS) 85–86 indomethacin reversing 86, 164 reading problems in NF-I children 218–19 laboratory studies of therapeutic interventions 87–88 functional MRI studies 215, 223 oxidative stress 86, 92 and malformation in brain regions 222 340 Index

Recklinghausen’s disease see neuroﬁbromatosis (NF) statistical methodologies 325–27 regression-based methods for predicting change scores steroids 326 for brain edema 314–15 rehabilitation side-effects 35, 51 brain tumor patients 165 for somnolence syndrome 314 children with brain tumors 206 see also corticosteroids; dexamethasone; glucocorticoids children with cancer 73–75 stress see also cognitive rehabilitation; support services animal and human models of 272–73 reliable change index (RCI) 325–26 and brain norepinephrine (NE) activity 273 renal complications, long-term survivors of childhood reduction of with complementary and alternative brain tumors 201 medicine therapies 307–8 repeated measures analyses 327 underlying physiology of 273 reports, neuropsychological evaluations 8–9, 12 structural brain imaging methods 20–22 reproductive hormones and cognition 119 analysis of data using voxel-based morphometry 21–22 in healthy men 117–18 in breast cancer patients 25–27 in healthy women 116–17 childhood leukemia studies 23–24 neuroimaging studies 118 support services 295–96 see also hormonal therapy community re-integration 301 response inhibition, assessment of 15 complementary and alternative medicine 307–8 during active treatment 298–99 secondary malignancies 199 end of life/palliative care 300 seizures 160 family settings 302–3 anti-epileptic drugs, negative effect on cognitive modalities of support 303–4 functions 51, 151, 177 needs assessment 296–97 chemotherapeutic agents associated with 100 across the disease continuum 297 differential diagnosis of patients with cancer 161 newly diagnosed patients 297 effects on cognitive function 151 post-treatment 299–300 in low-grade glioma 144 psychosocial support 304–6 selective estrogen receptor modulators (SERMs) see rehabilitation programs 306–7 raloxifene; tamoxifen school settings 301–2 self-instruction training 285 work settings 302 sensorimotor ability, assessment of 16, 62 supportive care 258 serial assessments 9, 60–61 drugs used 264 small cell lung carcinoma (SCLC) 239 future research antineuronal antibodies associated with detection and screening 266 and multifocal encephalomyelitis 243–44 treatment of delirium 266 onconeural antigens 241 see also delirium presence of anti-Hu antibodies 245–46, 248 surgery prognosis 242, 248 childhood brain tumors 58 prophylactic cranial irradiation (PCI), effects of 36, low-grade gliomas 145–46, 148 173–74 neuropsychological evaluation to clarify outcome of 9 whole-brain radiation therapy, response to 171 risks of damage to healthy tissues 35 socialization problems in childhood brain tumor survivors in treatment of brain metastases 172 205–6 effects on cognition 176 somnolence syndrome 91, 314 radiosurgery alone 172–73, 179–80 spatial perception radiosurgery with whole-brain radiation therapy 172 assessment of 13 survival role of androgens 116–18 acute lymphoblastic leukemia (ALL) 57, 228 special education services 301–2 childhood brain tumors 198–99 Index 341

high-grade gliomas 156–58 and children with non-verbal learning disability 218–19 low-grade gliomas 138, 142, 144–45 functional MRI studies 223 predictors of length of 37–38 and reading problems in children with NF-I 219 primary CNS lymphoma with methotrexate 188 tests measuring 74 versus “compression of morbidity” 156, 163 tests predicting NF-I diagnosis 219–21 switching, assessment of 15–16 see also spatial perception symptom management see supportive care voxel-based morphometry (VBM) 21–22 systemic anticancer therapies, effect on brain function breast cancer studies 25–27 176 Wake Forest University School of Medicine clinical trials tail suspension test (TST) 274 316–17 tamoxifen (TAM) Wechsler tests 12, 62, 67 effects on cerebral metabolism 28 white matter (WM) estrogenic and anti-estrogenic qualities 119 abnormalities in primary central nervous system impact on cognitive function 36, 50 lymphoma patients following treatment 193–94 studies of 120–26 changes in breast cancer patients after chemotherapy testosterone levels 25–26, 28 anti-androgens 120 demyelination 84, 148, 188 effect of aromatase inhibitors 119 investigation of using diffusion tensor imaging 22 effect on cognitive performance 116–18, 126–34 necrosis following radiation therapy 83, 91–92 and visual-spatial abilities 137 animal studies showing delayed 87 see also androgens; luteinizing-hormone-releasing neurological complications caused by 202 hormone (LHRH) agonists pharmacological interventions 315–16 thyroid gland abnormalities 108–9, 201 radiation dose influencing 90, 150 Time Pressure Management (TPM) 283–84 reduction in volume following chemotherapy and time to event analyses 326–27 radiation therapy 24, 103 traumatic brain injury (TBI), cognitive rehabilitation 281, whole-brain radiation therapy (WBRT) 175 283–85 brain metastases treatment 171–72 twin studies, breast cancer and chemotherapy 27 conflicting results of studies 175 regulation of dose and fractionation 179 vascular damage, radiation-induced 83, 202 response to 171 verbal memory primary CNS lymphoma (PCNSL) 187–88 assessment of 14 delayed neurotoxicity 188–89 influence of high levels of estrogens 116 Women’s Health Initiative Memory Study (WHIMS) influence of hormonal therapy 116–17 in men 126, 135 work settings in women 120, 126 providing support in 302 tests for measuring in children 62 returning to work 299 visual ability 70–71 working memory assessment of 13–14 assessment of 13, 69 psychometric tests 62 impact of chemotherapy on 27, 98 visual spatial abilities influence of hormonal therapy 135 animal studies 217 psychometric tests 62, 74