Brain tumors represent a unique challenge in that they affect the organ that is the essence of the "self." Furthermore, because each area of the brain serves a different but vital function, the therapy that is most effective for other cancers--surgical removal of either the entire organ or the tumor with a generous surround of normal tissue--cannot be used to cure brain tumors. Unfortunately, most brain tumors are relatively insensitive to other cancer treatment, including radiation and chemotherapy.
Coupled with the difficulty in treating brain tumors is the unique biology of the brain:
• Many brain tumors extensively invade normally functioning brain, making complete surgical removal impossible.
• In their early stages, brain tumors are protected behind a blood-brain barrier; even when this barrier is disrupted in the bulk of the tumor, infiltrating tumor cells at the growing edge remain protected.
• Disruption of the blood-brain barrier leads to edema, which the brain tolerates poorly because of the limited intracranial space and the lack of lymphatics to rid itself of the products of edema and other debris.
• The brain itself is rich in expressed genes and therefore is a fertile field for the growth of both primary tumors and metastases.
• The brain and brain tumors appear to be less susceptible to attack by the immune system than are tumors in other organs. Even the term brain tumor, which suggests a single type of tumor, can be misleading. There are a bewildering variety of central nervous system tumors; the World Health Organization lists 126. Many of these tumors are not, strictly speaking, in the brain but arise from structures intimately associated with that organ, such as tumors of the covering membranes (meningiomas) and adjacent cranial and paraspinal nerves (schwannomas). Brain tumors range from benign (most meningiomas) to highly aggressive (glioblastomas). They affect both adults and children (although the distribution of tumors varies) and are often highly resistant to treatment.
The term brain cancer is also misleading. Most cancers that arise elsewhere in the body cause damage by metastasizing to other organs (including the brain). Primary brain tumors, however, rarely metastasize, although they may widely infiltrate the nervous system. Conversely, many cancers metastasize to the brain, making metastatic brain tumors much more common than primary brain tumors.
Throughout this document, the term brain tumor is used to refer to all tumors that grow inside the skull. The issues discussed in this document, however, also extend to tumors growing within the spinal canal.
STRUCTURE AND PROCESS OF THE PRG MEETING
For the reasons described in the introduction to this report, the Brain Tumor Progress Review Group (BT-PRG) required input
from participants with much more diverse expertise than was needed in previous PRGs. In addition to experts on cancer biology
and genetics, the BT-PRG required expertise in neurobiology, including areas such as progenitor cells, cellular migration,
and blood-brain barrier function. Clinically, expertise was required from both oncology and the clinical neurosciences, including
neurosurgery and neurology. In addition, to ensure inclusion of the wide diversity of brain tumors, breakout sessions were
held not only for those topics that apply to all solid tumors, including brain tumors, but also (different from other PRGs)
for specific types of brain tumors (i.e., intraaxial tumors, extraaxial tumors, pediatric tumors, and metastases). A total
of 16 breakout groups
were therefore convened (see box). Each participant attended three breakout sessions.
• Cancer Biology and Etiology
• Neurobiology: Progenitor Cells
• Neurobiology: Migration and Trafficking
• Cancer Genetics
• Tumor Immunology
• Epidemiology, Prevention, and Outcomes
• Radiation Biology
• Therapeutic Targeting: Blood-Brain Barrier, Gene Therapy, and Vascular Biology
• Intraaxial Tumors
• Pediatric Tumors
The participants in each of these 16 breakout sessions were asked to identify three important research priorities in their assigned areas. It was recognized that it might not be possible to place all of the research priorities formulated by the groups into an overall hierarchy. Although all of the priorities included in the appendices are important and meritorious, some arose in multiple breakout sessions and therefore appear to be of overarching importance. This report delineates those priorities considered by the BT-PRG to be overarching. The appendix contains the full reports of the individual breakout sessions and their priorities.
This report is divided into two sections. Section I, "Scientific Priorities," describes the overarching priorities in both the basic and the clinical sciences. These scientific research priorities are hypothesis driven. To meet them will require the scientific resources described in Section II. The resource priorities in Section II can be considered as hypothesis generating in that their development will generate hypotheses for further research.
SECTION I: SCIENTIFIC PRIORITIES
Three separate sets of breakout sessions addressed the scientific priorities. One set was devoted to fundamental biology and included sessions on models, neurobiology of progenitor cells and of cellular migration and dispersal, cancer biology, immunobiology, and cancer genetics. Another set of sessions was related to clinical issues, ranging from detection and diagnosis to treatment and outcomes. A third set was devoted to specific tumors. Several overarching scientific priorities emerged from all of these sessions and are described here.
Brain tumors are phenotypically and genotypically heterogeneous. Significant gaps exist in current understanding of the molecular pathways involved in the genesis, progression, and biological and clinical behavior of brain tumors. Brain tumors are unique among human cancers because of their complex interaction with the brain itself, which greatly complicates the use of existing therapies as well as the development of novel ones.
A cardinal feature of the most common malignant brain tumors--their diffuse infiltration into the surrounding brain--presents substantial barriers to the effective delivery of therapeutic agents and increases the possibility of therapeutic toxicity to a vital organ whose function greatly affects the patient's quality of life. Other obstacles to effective therapy include the blood-brain barrier and the difficulties it creates for therapeutic delivery, as well as the relative lack of information on the unique immunological aspects of brain tumors and the cerebral environment.
The biology of brain tumors is distinct from that of many other human tumors. Although tumors are named as though their lineage were understood (e.g., astrocytoma from astrocytes), the cells of origin for most human brain tumors remain enigmatic, complicating the interpretation of data that require a comparison between brain tumor cells and their "normal" counterparts. Highlighting these issues are childhood brain tumors, especially primitive neuroectodermal tumors that arise during brain development. Insights into the normal and aberrant regulation of neurodevelopmental genes may be significant in understanding the etiology of both childhood and adult brain tumors. Likewise, elucidating the genetic alterations in brain tumors may yield new insights into brain development. Achieving significant advances in the diagnosis, prognosis, therapy, and prevention of brain tumors requires unraveling and understanding many aspects of the cellular and molecular biology of brain tumors and their interactions with normal brain elements. These advances must proceed along a number of different fronts and will require the interaction of several disciplines in order to achieve the greatest chance of success (see "Communication" in Section II of this report).
Many of the priorities generated by the breakout sessions of the BT-PRG overlapped, particularly those concerning needed resources (see Section II). The highest scientific priorities in basic biology identified by such overlap are as follows:
-- Characterize the interactions of brain tumor cells with the normal brain. • Provide a detailed molecular classification of the cells of origin for distinct tumor types and define their lineage associations, as well as the signal transduction pathways that regulate cell fate and the mechanisms by which the local environment of the brain influences cell migration and differentiation.
• Understand genotypic influences on phenotypic behavior, tumor type, age at onset, anatomical position, cell of origin, and cellular biology.
• Isolate the genes that predispose to human brain tumors and understand their relationship to the genes that regulate normal development.
• Identify the genes that regulate patients' responses to chemotherapy and radiotherapy and those that mediate tumor chemoresistance and radioresistance.
• Characterize both central nervous system and systemic immune responses in patients with brain tumors.
• Understand the blood-brain barrier and its regulation.
• Understand the mechanisms underlying the spread and establishment of metastases in the central nervous system.Epidemiology
Little is known about the epidemiology of brain tumors. Germ line mutations (familial brain tumor syndromes) account for no more than 7% of patients. The only unequivocally established risk factors for nonfamilial brain tumors--therapeutic irradiation to the brain and chronic immunosuppression (e.g., AIDS)--are also infrequent causes of brain tumors. Other suggested etiologies, such as nonionizing radiation (e.g., from cellular telephones or high-tension wires), viral agents, household chemicals, or foods, have not been established as causal. In addition, little is known about the interaction of genetic factors and environmental toxins in the genesis of brain tumors.
Because identification of the risk factors for brain tumors may aid prevention and suggest effective treatments, high-quality epidemiological studies are extremely important. Factors that inhibit epidemiological studies include the relatively small number of patients affected by brain tumors and the large number of histopathological types of these tumors. These factors complicate the design of research protocols and limit the statistical power of the data collected. In addition, existing tumor registries are neither linked nor structured to facilitate the collection of large numbers of samples for meaningful epidemiological research. Important epidemiological scientific priorities, therefore, include the following:
• Expand and enhance databases to include all primary brain and spinal tumors--malignant and nonmalignant, adult and pediatric--and to have the flexibility to accommodate new histological and molecular classifications of tumors.
• Develop epidemiological studies of patients' susceptibility to the toxic effects of current treatment modalities and investigate risk and protective factors with study designs that incorporate biological measures.
• Use validated animal models (see "Models," Section II) to study the potential causal factors of brain tumors and of treatment-induced neurotoxicity.Detection and Diagnosis
Because brain tumors are an extraordinarily heterogeneous group of lesions, accurate diagnosis is essential to proper management. Current imaging techniques provide a sensitive means for delineating the anatomical features of brain tumors but have not provided an effective means for early detection. Early detection could also be complicated by the ethical problems created by presymptomatic diagnosis of tumors for which there may not be effective treatment, and in an organ whose proper function is essential to quality of life. Nonetheless, early detection of brain neoplasms, particularly in the pediatric population, where these lesions are often treatable, could be facilitated by appropriate education of pediatricians, parents, school officials, and other caregivers.
The diagnosis of brain tumors is currently based on histological examination of brain tumor tissues after radiological characterization and surgical biopsy. These approaches are successful in classifying and grading most cases, but in many situations they do not allow accurate prediction of therapeutic responses or of prognosis. The situation may be further complicated by the small size of some diagnostic biopsy samples. There is therefore a critical need to improve the diagnosis of brain tumors in order both to improve current therapeutic management strategies and to form a basis for the evaluation of novel approaches.
The ability to characterize tumors comprehensively at the molecular level raises the possibility that diagnosis could be based on molecular profiling, either alone or with histological examination, rather than on histological phenotype alone. Once such techniques become possible and practical, molecular profiling could be accomplished by tissue analysis or imaging. In the future, molecular markers could also form the basis for screening at-risk individuals or populations. In light of such possibilities, the following priorities in the detection and diagnosis of brain tumors were identified:
• Develop techniques that can reliably detect brain injury related to tumor or treatment and use such techniques to assess the efficacy of neuroprotective interventions.Treatment
Treatment options for patients with brain tumors have been limited and, for most types of tumors, have provided only modest benefits. Some of the likely reasons for these limitations (see "Introduction") include the unique structural and physiological aspects of the central nervous system, especially its vulnerability to damage from many therapies as well as from neoplastic processes themselves. Research in the treatment of brain tumors has been hampered by the lack of clinically predictive model systems; by a minimal understanding, until quite recently, of fundamental tumor biology; and by a narrow range of available therapeutic agents for testing that have had little expected specificity for brain tumors. The major challenge for the future is to develop more effective techniques to treat brain tumors without damaging the brain.
Marked progress is currently being made in dissecting the molecular mechanisms of neoplasia in the brain and elsewhere. These advances are enabling the rapid identification of relevant molecular targets, and the result is a vast array of potential therapeutic approaches and agents in the development pipeline. At the same time, advances in neuroimaging are raising the tantalizing possibility of clinically assessing the capacity of an agent to alter its intended target. It therefore seems reasonable to expect an improved rate of success in research on the treatment of brain tumors. Because the special characteristics of these tumors will continue to present problems and challenges, however, the following priorities were identified:
• Increase knowledge about the mechanisms of existing therapies for both adult and pediatric brain tumors.
• Improve the therapeutic index of new agents that are specifically relevant to the central nervous system.
• Enhance the therapeutic ratio for radiation therapy for brain tumors. (Overcome radioresistance of primary brain tumors; overcome normal tissue toxicity such as necrosis/edema and functional deficits.)
• Develop novel drug targeting systems that enhance the uptake by brain tumors of small- and large-molecule diagnostic and therapeutic agents.
• Develop clinical consortia for immunotherapy that are similar to those for radiation and chemotherapy.
• Develop therapies that are less toxic than existing therapies to both the mature and the immature nervous system.Outcomes
Traditional outcome measurements used in brain tumor studies have included overall and recurrence-free patient survival and, in some instances, radiological response to therapy. Such measurements, however, largely ignore crucial issues relating to quality of life and biological endpoints of response. These issues are of particular importance in tumors for which effective therapies may not exist and in pediatric tumors, for which effective tumor control may be associated with significant long-term morbidity. For these reasons, there is an immediate and crucial need for better measurement tools and surrogate markers to assess patient quality of life and tumor response to therapy. Such outcome markers would facilitate the assessment of neurotoxicity, thereby providing an opportunity to discard potentially neurotoxic therapies sooner. They would also facilitate more accurate assessment of therapeutic response, thereby allowing effective therapies to be continued while ineffective therapies are discontinued. The following priorities were therefore identified:
• Refine the ability to detect response to existing therapies, such as radiation, and to novel treatments, using surrogate markers measured either by imaging or in biological fluids (e.g., serum or cerebrospinal fluid).
• Establish clinical and imaging markers of neurotoxicity from existing therapies, such as radiation, and from novel treatments.
• Extend the use of such markers to preclinical evaluations in animal models.Specific Tumors
Recognizing the remarkable diversity of human brain tumors and the distinct clinical questions associated with different tumor types, the PRG members were concerned that most of the general scientific sessions would concentrate on the more common tumors, such as malignant gliomas and medulloblastomas, to the exclusion of other brain tumor types. To address the possibility that research priorities might relate to different types of brain tumors, the PRG convened four special breakout sessions to focus on particular groups of brain tumors: pediatric brain tumors, intraaxial brain tumors (excluding malignant gliomas and medulloblastomas), extraaxial brain tumors, and metastases to the brain. These four special breakout sessions met after the 12 general scientific sessions had adjourned. The special sessions included attendees from the earlier, general discussions, thereby allowing important issues from the general sessions to be applied to discussions of the specific tumor groups.
Remarkably, the research priorities and needed resources identified by these special groups echoed those of the general sessions, although some different emphases were placed according to tumor type:
• The group addressing intraaxial brain tumors highlighted issues relating to low-grade gliomas, primary central nervous system lymphomas, and germ cell tumors.
• The session on extraaxial brain tumors emphasized the need for studies that incorporate careful long-term follow-up for these often slowly growing lesions.
• The group discussing metastatic tumors of the brain made the unique recommendation to convene a PRG devoted to the biology of metastasis. The specific priorities from these sessions are detailed in the individual reports in Appendix A.
SECTION II: RESOURCE PRIORITIES
Although the scientific priorities set forth by the BT-PRG varied considerably across the different areas of scientific and clinical investigation, the resources required to accomplish those priorities were remarkably concordant. Indeed, nearly all of the five resource priorities listed below were deemed essential by most of the participants:
2. Tissue banks and databases
3. Genomics and high-throughput screening
These resources can be viewed as hypothesis generating because they will provide the information and abilities to accomplish the scientific and clinical priorities listed in Section I. The creation of these resources is deemed essential in order to develop new, effective therapies for brain tumors.
Models are central to making the transition from developing scientific concepts to understanding human tumors within the context of the tissues that they affect. Models may be used for therapeutic screens, in preclinical trials, or to study the basic biology of tumors. However, because currently available cellular, tissue, and animal models do not accurately represent the biology of human brain tumors, it is vital to:
• Create genetically and behaviorally accurate models for brain tumors in mice and other animals.
• Generate tissue-based, imaging, and genomic methods to validate and compare animal models with their human counterparts.
• Improve the availability of the reagents needed to create new animal models of brain tumors, the sophisticated technologies used to evaluate and validate those models, and the animal models themselves. To accomplish these priorities, a mechanism must be created to support the development and validation of model systems that more accurately reflect the biology of brain neoplasms. Although the National Cancer Institute (NCI) Mouse Models for Human Cancer Consortium (MMHCC) has been established to fund the development of mouse cancer models, additional mouse models of the various brain tumors that are not addressed through the MMHCC, as well as models in other animals, remain high priorities.
Tissue Banks and Databases
Addressing the complex biology of brain tumors requires innovative tumor banking and characterization facilities with relevant and appropriate clinical and radiological databases. Tissue banks linked to clinical databases are also vital for translating research discoveries into clinically relevant information. Because current tissue banks are typically institution based, they are limited in scope and amount of available specimens. These banks also process tissues in different ways, and their specimens are usually not sufficiently annotated with clinical and radiological information. Because of the rarity of many brain tumor types, including both adult and pediatric neoplasms, there is a great need for organized, interinstitutional approaches to banking and data management of both adult and pediatric neoplasms.
An effective tissue bank or database must do the following:
• Maintain a comprehensive database of relevant clinical and demographic, pathologic, biologic, and therapeutic information on all patients whose tissue is banked. Develop links to population databases to enhance potential etiological and other epidemiological studies.
• Involve multidisciplinary participation of surgeons, pathologists, scientists, and other professionals, including neurooncologists, to ensure reliable and consistent tissue processing.
• Provide mechanisms to ensure access, on a competitive and open basis, by researchers to the material and data in the bank.
• Employ approved and ethical methodologies to protect patient confidentiality and ensure appropriate patient consent.
• Feature local and regional facilities and facilitate effective communication and collaboration among centers.
• Be supported by ongoing funding, potentially for longer than 5-year periods, to facilitate study of tumors with long clinical courses, such as meningiomas.Genomics and High-Throughput Screening
The explosion of information in genomics, together with the promise of similar advances on the near horizon in proteomics, raise the need for technologies that allow high-throughput screens of brain tumors and related specimens (e.g., other tissues from patients with brain tumors). Such high-throughput screens would allow large amounts of information to be gleaned quickly and would facilitate further translational research toward more tailored therapeutic approaches. These screens can occur at the tissue level ex vivo or, in the future, at the molecular neuroimaging level in vivo. For such large-scale approaches to be functional, considerable emphasis will need to be placed on bioinformatics support. The need for high-throughput screening technologies was identified by a number of the different breakout sessions; the highest priorities were the following:
• Develop high-throughput laboratory approaches to identify the genes and genetic variations that underlie tumor resistance to chemotherapy and radiation therapy, as well as the allelic variations that influence responses to therapy in individual patients.
• Develop high-throughput laboratory approaches to identify antigens that may be used to further understanding of the immunological features of brain tumors and to develop novel immunological therapies.
• Develop high-throughput neuroimaging approaches for the in vivo characterization of the molecular features of tumors and the surrounding brain that could monitor and influence therapies.
• Develop the bioinformatics support necessary for rapid and accurate analysis of data generated via these high-throughput approaches.
• Establish a consortium of brain tumor modeling laboratories for the purpose of testing novel therapies.
• Allocate resources for the generation of cDNA microarrays based on the mouse equivalent of the human sequences identified through the Brain Tumor Genome Anatomy Project (BT-GAP).
• Create a mechanism to ensure affordable access to these reagents and models.Communication
The PRG Roundtable meeting provided a unique opportunity for scientists from different disciplines--including cancer biology and genetics, neurobiology, immunology, and radiation biology--to meet and discuss brain tumor biology. These stimulating interactions highlighted the potential for novel insights arising from such interdisciplinary interactions. A central goal that emerged from these discussions was the need for further communication among these various disciplines on the subject of brain tumor biology. Such enhanced communication would in turn lead to interdisciplinary collaborations that would approach problems in brain tumor research from a unified, and therefore novel, perspective.
It was recognized that one reason for the relative lack of such communication and collaboration among disciplines has been the historically different funding and oversight mechanisms that have supported such research. For example, neurobiology research is largely funded through the National Institute of Neurological Disorders and Stroke (NINDS), whereas cancer biology and immunology research is generally funded by NCI and other agencies. Recent attempts to bring together NCI and NINDS to address questions in brain tumor research--the BT-PRG, the BT-GAP, and the establishment of a combined NCI-NINDS Neurooncology Branch--have been widely applauded and further interinstitutional interactions strongly encouraged. The possible extension of such interactions to the grants review process was also deemed an important area for discussion.
Because the Center for Scientific Review (CSR) reviews most unsolicited brain tumor grant applications, the brain tumor research community believes that better coordination among the institutes and CSR is needed. Improved communication could prevent brain tumor biology from "falling between the cracks" among the various review groups that may have relatively few brain tumor biologists. It is anticipated that coordinated efforts by NINDS, NCI, and CSR on the referral, review, and funding of brain tumor research applications would facilitate the implementation of the national plan for brain tumor research.
Goals for improved communication extend to clinical problems as well. There is clearly a need for increased dissemination of information to patients, as well as to clinicians outside of neurooncology centers, with regard to the variety of available treatment options. The relatively low percentage of patients with adult malignant gliomas who are enrolled in clinical trials may reflect an inadequate knowledge of treatment options on the part of both patients and physicians. This area of need represents an ideal opportunity for patient advocacy groups to collaborate with physicians to develop strategies to educate patients and clinicians about treatment options, including clinical trials, as well as about the specialized expertise that is available at neurooncology centers. For these reasons, the following priorities were identified:
• Facilitate collaborations among different disciplines by encouraging interdisciplinary grant applications in brain tumor biology and etiology.
• Continue to develop combined programs in brain tumor research from NCI and NINDS and explore the possibility of revisions in the grant review process for brain tumor research.
• Encourage coordinated activities by advocacy groups toward further education of patients and clinicians about available treatment options for brain tumors.Training
Achieving the goals for brain tumor research outlined in this report requires an adequately sized and well-trained scientific and clinical work force specializing in brain tumor research. Unfortunately, there is a dearth of basic scientists working in the field of brain tumors, which lacks sufficient numbers of clinicians who are cross-trained in brain tumor biology and scientists who are aware of the problems driving clinical neurooncology research. As is the case for biomedical science in general, there exists a true crisis caused by the small number of clinician-investigators now entering academic medicine. This issue has been discussed elsewhere and will not be recapitulated here, but its importance should not be underestimated. High priorities for brain tumor research are therefore as follows:
-- Recruit new talent and sustain proven talent in the field of brain tumor research. • Create innovative public and private programs to stimulate promising young investigators to choose a career in clinical or laboratory brain tumor research through, for example, tuition loan payback or forgiveness and fellowships.
• Develop a joint NCI-NINDS campaign to encourage students to pursue interdisciplinary careers in the field of brain tumor research.
• Develop at NIH a model for a joint NCI-NINDS interdisciplinary training program in neurooncology at both the basic science and the clinical level. This program might include not only training at NIH for 2-3 years, but also additional support for the first 3 years of the individual's career as an independent investigator.CONCLUSION
Although not among the most common of neoplasms, brain tumors are among the most devastating. Mental impairment, seizures, and paralysis afflict the very core of the person and have a demoralizing effect on loved ones. Added to these burdens is the knowledge that, for most brain tumors, adequate treatment is not available and the likelihood for long-term survival is poor. In children, even if they do survive, the devastating impact of disease and treatment often leaves permanent neurological damage.
Recent advances in the clinic, as well as in neuroscience and cancer biology, make the present an opportune time for a major attack on brain tumors. As indicated in this report, progress has been made in the basic understanding of many aspects of brain tumor biology. These advances promise to provide new targets for therapies and more rational ways of delivering these novel therapies. In the clinic, new techniques in surgery and radiation therapy are just beginning to be exploited in the treatment of brain tumors. Other innovative approaches, such as gene and immunological therapies, are still in their infancy but represent substantial hopes for the future. Preventive factors identified in recent epidemiological studies, if replicated and understood at the biological level, may lead to intervention strategies.
The priorities outlined in this report provide a framework to guide progress in the field of brain tumor research. A concerted, interdisciplinary, and timely approach to addressing these priorities will allow the development of new diagnostic and therapeutic techniques that may ameliorate and, it is hoped, eventually cure brain tumors.
Last updated February 9, 2005