EPIDEMIOLOGY, PREVENTION, AND OUTCOMES

Co-Chairs: Christina Meyers, Ph.D., and Susan Weiner, Ph.D.

Participants: James Boyette Pim Brouwers Greta Bunin Peter Burger Faith G. Davis Lisa DeAngelis

Fred Hochberg Peter Inskip Carol Krutchko Minesh Mehta Raymond Mulhern Scott Pomeroy

David Ramsay Lynn A.Ries Lucy B. Rorke Libby Stevenson

EPIDEMIOLOGY

Statement of the Problem

Gaps in the fundamental understanding of brain tumors, from basic biological interactions to epidemiology, hinder systematic approaches to prevention. Although research exploiting new molecular and genetic technologies is attracting scientists and funds, epidemiological information about primary brain tumors remains scarce. The absence of high-quality epidemiological studies is an impediment to understanding who and how brain tumors arise in both children and adults.

Challenges and Barriers

The histopathological variability in brain tumors and the relatively small numbers of persons affected complicate the design of research protocols and historically have limited the statistical power of studies. Existing tumor registries and surveillance systems (including the Surveillance, Epidemiology, and End Results [SEER] Program and the North American Association of Central Cancer Registries [NAACCR]) are not linked or structured to allow rapid case ascertainment or case sample sizes adequate for meaningful epidemiological research of all brain tumor types. Although promising epidemiological clues have been uncovered in recent years, few are being pursued through funded studies.

Many previous epidemiological studies on primary brain tumors have significant methodological flaws. Methodological issues themselves warrant investigation to advance the field. Questions that need to be addressed include how to assess recall bias among study subjects who may have cognitive impairment; what types of exposures for which interview data and the potential for recall bias may be strong enough to result in spurious associations; how the interval between exposure and interview may affect recall bias; and whether there are feasible sources of control groups in the United States other than the current approach of random-digit dialing and interviews with hospital patients.

Most registries collect data only on malignant tumors, as defined by the International Classification of Diseases, 9th Edition (ICD-9). That registries are limited to malignant tumors artificially constrains epidemiological understanding of brain tumors. For example, the just-revised classification system used by all cancer registries now codes juvenile pilocytic astrocytomas as benign, and therefore registries--including SEER--are likely to stop collecting data on these tumors. Brain tumor specialists include and treat all neoplasia of the central nervous system as cancer, and the rationale for including all brain tumors in population-based surveillance systems has been well documented.

Better surveillance tools are critical to facilitate epidemiological research. Two avenues of epidemiological investigation are especially compelling. In pediatrics, mechanisms are in development for a surveillance system that routinely collects biological samples of blood and tumor. A similar system is needed for adult tumors.

Research and Scientific Priorities

Priority 1: Enhance and expand the existing infrastructure for a surveillance database that does the following:

• Includes all primary brain tumors (malignant and nonmalignant, central nervous system and extraaxial

• Is designed with enough flexibility to accommodate new histological and molecular classifications of tumors

• Gives individual investigators access to rapid case ascertainment for studies that require questionnaires, tissue samples, pathological review, and pooling of results to increase sample size. SEER and state registries capture 50% or more of neuroepithelial tumors, and extraaxial tumors can be captured only through a central brain tumor registry. Approximately 52% of "quality" registries cover the U.S. population. If definitions were changed by consensus, data from diverse sources could be integrated into the database of the National Cancer Institute and other databases.

Priority 2: Target funding to support basic science and population-based human studies to evaluate leads on animal neurocarcinogens, such as exposure to nitrosamines, viral agents, and polymorphisms. For known animal carcinogens, molecular exposure measures are needed to allow direct exposure measures for human studies.

Priority 3: Investigate the epidemiology of neurotoxicity and other toxic effects of treatment, such as what makes some patients susceptible and what exerts a protective effect.

Resources Needed

• Database of household carcinogens. A database listing the major household sources of known carcinogens (e.g., dry-cleaned clothing, oven cleaner, specific types of paints, weed killer), including neurotoxins and neurocarcinogens, would be valuable for both brain tumor and cancer epidemiology in general.

• Training. Training in epidemiological methods is needed for researchers involved in etiological and outcome studies, including neurooncology and neurosurgery residents.

PREVENTION

Statement of the Problem

Although the literature and the current National Institutes of Health (NIH) research portfolio hold almost no information relevant to the prevention of primary brain tumors, factors that may reduce the risk of primary brain tumors are gradually being identified. Specifically, studies of pediatric cancers suggest that maternal diets high in fruit and vegetable intake or maternal use of vitamin supplementation, particularly folates, during pregnancy may reduce tumor development. Other studies suggest a protective role for allergic conditions and selected infectious diseases in adult tumors. If these isolated results can be replicated and clarified and the underlying biological mechanisms understood, intervention and prevention strategies may become feasible.

Challenges

At this juncture, not enough is known about the basic epidemiology of primary brain tumors to guide research initiatives. For this reason, the most pressing challenge is to acquire data that will drive scientific inquiries in prevention.

Research and Scientific Priorities

Priority 1: Because of the paucity of data and ideas about prevention, a request for proposals should be issued to stimulate new research approaches.

Priority 2: For a subset of brain tumors, some information is available to suggest fruitful research directions. For example, it may be feasible to explore the prevention of certain familial cancers, such as neurofibromatosis or von Hippel-Lindau syndrome. In addition, data from patients treated in pediatric cancer cooperative groups may yield insights about prevention for patients at risk of second primary tumors resulting from previous cancer treatment.

OUTCOMES

Statement of the Problem

The endpoints of survival and disease-free survival, which traditionally have been used to assess outcome in patients with cancer, fall painfully short as measures of success in treating brain tumors. From the perspective of patients and families, "outcome" is a multidimensional, daily reality, and quality of life can be at least as important as survival. So, too, assessment of quality of life is increasingly used to evaluate the risks and benefits of new treatments.

Researchers who are not directly involved in patient care may lose sight of the importance of the functional impact of brain tumors and their treatment on survivors and the families who care for them. Because the current treatment armamentarium has little to offer many patients with brain tumors, saving a life can be a considerable achievement. Saving a life without considering future constraints on how life can be lived, however, may offer an unacceptable outcome. Parents want to normalize life for an affected child, and adult patients want to weigh the functional risks they face with each treatment option.

To allow patients and parents to make more informed treatment decisions, clinicians need more information on the expected functional outcomes of disease and treatment. Such assessment data can also contribute to the drug approval process, when the survival benefits of two treatments are not substantially different. Finally, greater understanding of the impact of treatment can lead to interventions that will allow parents and adult patients to rehabilitate damaged functioning and normalize their lives.

Posttreatment follow-up of brain tumor patients and survivors does not routinely include functional assessment, nor is patient functioning routinely evaluated in clinical trials. This lack places a serious limitation on the ability to weigh the risks and benefits of new therapies and to develop strategies to improve functional outcomes. In this context, "function" includes neurocognitive status, symptoms, ability to perform activities of daily life, and psychosocial status. Another important but neglected aspect of assessment and intervention is the impact of brain tumor diagnosis and treatment on families. Because families are the primary caretakers of patients with brain tumors, families' functioning can have a profound and direct impact on patients' functioning and quality of life.

Challenges and Barriers

A major barrier to the conduct of functional assessments is a negative attitude among clinicians. Some feel that it should be sufficient for a treatment to avert a patient's death. Some clinicians interpret as presumptuous patients' wishes to have maximum levels of functioning and minimal damage to the central nervous system and loss of function. Further, clinicians may feel defensive when faced with a patient's serious or unexpected functional deficits. Finally, clinicians may be unfamiliar with the assessment process and may inappropriately feel that assessment is too costly and burdensome for patients.

To incorporate functional assessments into the design of clinical trials, two things are needed:

1. Criteria to determine which trials may be suitable (functional assessments are not an appropriate component of all clinical trials)

2. Identification of the appropriate time points during a clinical trial for administration of the assessments.

Research and Scientific Priorities

The goals in the area of outcomes are to be able to predict the impact of the tumor and treatment on a patient's functional status and to achieve the best possible outcome. Routine functional assessments and interventions are necessary to improve patients' functioning and quality of life and should become the standard of care for patients with primary brain tumors.

Priority 1: Identify the pathogenesis of the injury related to the tumor (e.g., rapid versus slow growth, site of the tumor), treatment, and comorbid processes (for example, those involving the neuroendocrine and vascular systems and cytokines). To do this, the following are necessary:

• Correlate patient demographic characteristics, imaging, and cognitive/quality of life assessments.

• Develop animal models in tandem with human studies.

• Identify markers for patients at risk of developing toxic effects (e.g., the apolipoprotein E genotype).

• Develop and assess drugs that might offer neuroprotection during primary cancer therapy.

Priority 2: Design and evaluate innovative interventions to ameliorate undesirable symptoms and functional deficits. Studies should include the following:

• Empirical investigations of what works--pharmacologic, behavioral, cognitive, or a combination of these approaches

• Mechanism-based interventions focused on cytokine antagonists, neurotransmitter agonists, and neuroprotective agents

Priority 3: Alter primary therapy to reduce toxic effects while maintaining efficacy, such as conformal radiotherapy, local therapy, and reduced doses of radiotherapy.

To satisfy third-party payers, in anticipation of eventual coverage of the cost of such interventions, cost-effectiveness should be built into the research design prospectively. The efficacy of these interventions should be correlated with imaging measures, including functional magnetic resonance imaging, and with physiological correlates, such as levels of pro-inflammatory cytokines and neuroendocrine markers.

Resources Needed

Intra-institutional cooperative research initiatives, both within and outside NIH, should be fostered to address rehabilitation, education, and medical issues. Several working groups or consensus panels should be convened under NIH auspices to accomplish the following:

• Standardize assessment tools. The tools now used to assess patient function are not standardized across studies, sites, or patient populations. Standardized assessment instruments will enhance the generalization of findings; if a core of standard content exists, institutions can tailor aspects to specific studies. Tools need to be valid, reliable, easy to use, and inexpensive to administer.

A panel of neuropsychologists, neurologists and patient advocates should select well established, user-friendly instruments from the literature to form a "practice guideline protocol" for use in evaluating patients in clinical trials. Such a protocol of assessment instruments can be constructed to allow the investigator to select the tools that will evaluate hypotheses related to, for example, memory, attention, language, and spatial deficits.

• Survey successful interventions. A review of cognitive-behavioral and psychopharmacological interventions used in other rehabilitation-related disciplines, such as special education and rehabilitative practices related to traumatic (acquired) brain injury, dementia and aging, and stroke, should be undertaken to determine whether successful strategies might be used successfully with brain tumor patients. This review should be comprehensive and include projects funded and/or conducted through the National Institute of Neurological Disorders and Stroke, the National Institute of Child Health and Human Development, the National Institute of Mental Health, the National Institute of Dental and Craniofacial Research, and other NIH and Department of Health and Human Services institutions, centers, and working groups. Similarly, a review of family support interventions developed for traumatized families should be conducted to cull potentially effective interventions for families with a member affected by a brain tumor.

• Develop assessment guidelines. Guidelines are needed on what and when to assess in different types of clinical trials. Currently, questions about symptoms such as headaches are asked to determine side effects of treatment, but patients may also be asked questions such as how they feel about the symptoms they report. These questions are important for patient care but vary dramatically from person to person and may not be informative in clinical trials. It would also be helpful for guidelines to be framed with an understanding of the World Health Organization's three-tiered level of analysis (impairment, disability, and handicap).