Brain Tumor Registry Reporting Training Materials
Other Materials

Surveillance of Primary Intracranial and Central Nervous System Tumors: Recommendations from the Brain Tumor Working Group
Background

Surveillance of Primary  Intracranial and Central  Nervous System Tumors Preface Executive Summary Introduction Background Methods Results Discussion Recommendations References Tables Appendix A Appendix B Appendix C Appendix D

Classification
Historically, cancer registries have used the International Classification of Diseases (ICD) system, which describes tumors by their location (topography) and behavior (benign versus malignant),¹ to code neoplasms. Tumor nomenclature describing histology was developed by pathologists and incorporated into the International Classification of Diseases for Oncology (ICDO) system as a morphology code in 1976² and revised in a second edition (ICD-O-2).³ However, not all of the morphology codes in the current ICD-O classification used by cancer registries are consistent with the recent World Health Organization (WHO) brain tumor codes^.3,4 The recent WHO brain tumor classification has new terms that are not included in ICD-O-2; however, a new third edition to ICD-O may include these terms. ICD-O-2 also includes terms which are not categorized by WHO.

The CNS environment also contributes to the difficulty in characterizing tumors that involve this site. In contrast to tumors arising in other organ systems, the terms "benign" and "malignant" are only relative distinctions for CNS neoplasms. Depending on location, histologically benign CNS tumors can result in similar or worse outcomes compared with malignant tumors. For these and other reasons, classic oncologic concepts predicated on histologic grade, nodal status, and staging strategies are not entirely applicable to tumors of the CNS.⁵

When the BTWG discussed whether to consider anatomic sites other than the brain for this report, the unique features of the CNS and the morbidity and mortality associated with non-CNS intracranial tumors were closely examined. For the sake of simplicity in reporting and analyzing data for all of these sites, the designation of "primary intracranial and CNS tumors" was suggested. This designation, proposed by Walker et al,⁶ is used by the Central Brain Tumor Registry of the United States (CBTRUS) and includes the brain, cerebral meninges, cranial nerves, pituitary gland, craniopharyngeal duct, pineal gland, spinal cord, cauda equina, and spinal meninges. The intracranial sites are the brain, cerebral meninges, cranial nerves and other intracranial parts of the CNS; the craniopharyngeal duct; and the pineal and pituitary glands. The extracranial sites are the spinal cord, cauda equina, and spinal meninges. The BTWG also included lymphomas in the histology groupings for these sites. The topography codes for the sites are in Appendix B, and the histology groupings are listed in Appendix C. In this report, no distinctions are made between different age groups for intracranial and CNS tumors.

Incidence
The age-adjusted incidence rate for malignant brain and other CNS tumors in the U.S. for 1991–1995 was 6.1 per 100,000 person years.⁷ Approximately 17,400 new cases will be diagnosed in 1998.⁸ The incidence of nonmalignant brain and other CNS tumors is difficult to ascertain because few cancer registries collect or report these data. However, CBTRUS reported an incidence rate for all primary brain and CNS tumors, including the pituitary and pineal glands, of 11.8 per 100,000 for 1990–1993;⁹ the organization also estimated that 28,600 new cases of primary malignant and benign brain tumors were diagnosed nationwide in 1995.¹⁰ For both men and women, rates for malignant and nonmalignant brain tumors decline after a peak in childhood (younger than 10 years), increase after age 25, and stabilize after age 75; overall rates are higher for males.¹¹

Although the Surveillance, Epidemiology, and End Results (SEER) Program reported a 2.8 percent decrease in the incidence of invasive brain and other nervous system cancer from 1991 to 1995,⁷ within the last decade there have been many reports of dramatic increases in the incidence of brain tumors,^12-21 particularly among children^22,23 and the elderly.^12,16,17 Some consider the increased rates as histology-specific¹⁴ and indicative of a true shift in incidence.¹³ Other studies have concluded that many of the new cases are an artifact of changing diagnostic procedures.^16,17 In the U.S. the discussion regarding the increased incidence of brain tumors has been limited to rates for malignant tumors. Standard reporting practices would aid in the interpretation of time trends.

Molecular studies^24-26 have demonstrated that some low grade or benign intracranial tumor subtypes transform to malignant tumors. To understand the factors that might contribute to this transformation and whether incidence rates for both malignant and nonmalignant intracranial tumors are affected, the full spectrum of the disease needs to be observed.

Mortality
Brain tumors, regardless of behavior, are the second leading cause of death from neurological disease.²⁷ In 1995, malignant brain tumors accounted for 12,062 deaths, and 947 deaths were due to benign brain tumors. Another 131 deaths were reported for brain tumors of uncertain behavior, and 2,788 persons died of brain tumors for which the behavior was unspecified.²⁸ The estimated number of deaths in 1998 due to malignant brain and other CNS tumors is 13,300.⁸ Increased mortality rates for malignant brain tumors, particularly among the elderly, have been reported. One study concluded that the increases were largely related to better diagnostic technology and the introduction of support programs such as Medicare that facilitate diagnostic procedures in the elderly.¹⁷ Another study attributed the rising primary malignant brain tumor mortality to differential survival and its effect on the surviving gene pool in an aging population.²⁸

Survival
Prognosis for CNS tumors depends on at least four variables: tumor histopathology, anatomic location, patient age, and neurologic status.⁵ According to a recent analysis using SEER data, from 1973–1991 overall survival rates for malignant brain tumors and rates for patients with three specific histological types—astrocytoma, medulloblastoma, and oligodendroglioma—improved.²⁹ Few data are available for survival rates for nonmalignant primary intracranial and CNS tumors. However, Surawicz et al used data from the National Cancer Data Base (NCDB), a joint project of the Commission on Cancer (COC) of the American College of Surgeons (ACoS) and the American Cancer Society, to examine survival rates for patients diagnosed from 1985-1988 and from 1990-1992 with malignant or benign brain tumors;³⁰ NCDB collects data from hospital tumor registries. Based on the records of more than 60,000 patients, the authors found a 21.6 percent 5-year survival rate for patients with malignant tumors and a 72.4 percent rate for those with benign tumors. The most favorable prognosis was associated with neurilemmomas, pilocytic astrocytomas, and meningiomas. In contrast, microgliomas, lymphomas, malignant gliomas, and anaplastic astrocytomas were associated with a poor outcome. Surawicz et al also found variations in survival for some tumors depending on location. For example, survival rates for glioblastomas, which generally have a poor prognosis, improved if the tumors were located in the cerebellum; similarly, astrocytomas and anaplastic astrocytomas were associated with a better outcome if the tumors were in the ventricles or the cerebellum.

Risk Factors
Although malignant and nonmalignant intracranial and CNS tumors have undergone considerable study, etiologic and pathophysiologic details concerning their genesis remain obscure. Definite genetic predispositions for the development of brain tumors have been identified; however, population-based studies suggest that no more than 4 percent of thesetumors can be attributed to heredity. ³¹ Several environmental carcinogens may be associated with brain tumors, including ionizing radiation,¹¹ electromagnetic fields,³² and pesticides;³³ sustained exposure to vinyl chlorides, polycyclic hydrocarbons, and nitrosoureas has been implicated as well.⁵ Also, the presence of the Epstein-Barr virus contained in the DNA of primary lymphoma suggests that a viral etiology for human brain tumors cannot be entirely ignored.⁵ Accurate and complete data are necessary to develop hypotheses to identify the causes of intracranial and CNS tumors. The heterogeneity of brain tumors may mask the identification of causes when histology-specific studies are limited by the number of available cases.

Surveillance
Population-based cancer registries generally provide incidence rates and trends for cancer surveillance purposes, and support related clinical and epidemiological research. To receive accreditation from the COC, hospital cancer registries are required to report only in situ and primary malignant tumors for all sites, including the brain and other CNS sites.³⁴ Because data collection for nonmalignant intracranial and CNS tumors is not standardized at either the central tumors can be attributed to heredity or hospital registry level, collecting data for all primary intracranial tumors is a challenge.

In 1997, CBTRUS conducted a state survey (Appendix A) of benign brain and other CNS tumor collection.³⁵ This survey gathered information from 64 member registries of the North American Association of Central Cancer Registries (NAACCR) representing 48 states. All 64 of the registries collected data for primary malignant brain and other CNS tumors; however, only 15 state registries collected data for nonmalignant (i.e., benign) tumors at the time of the survey. Seven state registries had stopped collecting data for benign brain and other CNS tumors, three collected the data passively, one had started to collect the data in 1997, and one planned to start in 1998. The majority of the registries surveyed indicated that the requirements of their brain and other CNS tumor data collection had not been altered in the past 5 years.