Refinements in Histological Categorization of Brain Tumors and Associated Developments

The International Classification of Diseases for Oncology, Third Edition (ICD-O-3) is the standard classification system for the registration of cancers in the United States (and most areas of the world) and contains widely-accepted histologic categories of brain tumors. Brain tumors are classified into the following major histologic groupings: tumors of neuroepithelial tissue (hereafter referred to as glioma, including astrocytoma [grade II], anaplastic astrocytoma [grade III], glioblastoma [grade IV], oligodendroglioma, and ependymoma), tumors of meninges (including meningioma and hemangioblastoma), germ cell tumors, and tumors of the sellar region (including pituitary tumors and craniopharyngioma). Glioma is the most common histological category, followed by meningioma, the category that includes the highest proportion of benign brain tumors (approximately 96% of meningiomas are considered benign).¹ Criteria for classifying brain tumors have varied substantially across time and geographic region. Prior to the past decade, most epidemiologic studies presented results for crude categories such as “all central nervous system (CNS) tumors” or “brain tumors.” It is now well established that both the descriptive and analytic epidemiology of brain tumors varies considerably according to histologic grouping. A meaningful contribution of the Central Brain Tumor Registry of the United States (CBTRUS), which collects information on brain tumors occurring among residents of 19 U.S. states, has been the presentation of descriptive statistics according to categories detailed in the ICD-O-3.¹

Progress in Understanding the Descriptive Epidemiology of Brain Tumors

Glioma incidence rates increased during the 1970s and 1980s (probably reflecting the use of new diagnostic imaging technologies²) and have remained relatively stable since the 1980s. Incidence rates for high grade gliomas among older age groups have increased over time from the late 1970s to the early 1990s, as have those for oligodendroglioma and mixed tumor histologies at the expense of less specific histologies.² A similar description of American meningioma incidence rates over time is not possible, because benign brain tumors were only recently required to be reported to American central cancer registries. However, Klaeboe et al.³ report an increase in European meningioma incidence rates, which is also probably explained by new imaging technology introduced in the 1970s.

For all brain and CNS tumors combined, benign and malignant (brain tumors accounting for approximately 88%), the age-adjusted average annual (2000 to 2004) incidence rate for females (17.2 per 100,000 females) is greater than that for males (15.8 per 100,000 males).¹ Table 2-1 shows age-adjusted average annual (2000 to 2004) incidence rates and median ages at diagnosis for the major histologic groupings and their selected common histologic subtypes of brain tumors.¹ As shown in Table 2-1, men have higher incidence rates of gliomas, germ cell tumors and cysts, while women have a higher incidence rate of meningiomas. As described in a following section, results from analytic studies concerning reproductive and hormonal factors have suggested compelling possible explanations for the sex difference in glioma and meningioma risks.

Newly diagnosed glioma is approximately two times more common among whites than among blacks, as are germ cell tumors. In contrast, meningioma incidence rates are similar between whites and blacks.¹ Although there is no well-accepted mechanism for the race differences in glioma risk, differences in the distribution of human leukocyte antigens (HLA) among races, described later in the chapter, may explain it. In the United States, for diagnoses made between 2000 and 2004, the median age at diagnosis of a primary brain tumor was 57 years.¹ For all major histologic groups except germ cell tumors and cysts, incidence rates increase with advancing age.¹ Rates of cancers of nearly every anatomic site increase with advancing age; still, the reason for this increase is not known.

Brain tumor incidence rates vary moderately by geographic region in the United States and throughout the world.^1,4 However, differences in diagnostic practices, completeness of reporting, access to and quality of health care, make geographic and especially international comparisons difficult to interpret.^5,6

Brain tumor survival time varies greatly by histologic type and age at diagnosis. In every adult age group, the lowest relative 2-year survival is found for patients with glioblastoma multiforme (GBM), ranging from 30.4% among those aged 20 to 44 years to 1.3% among those aged 75 years and older.¹ In general, within histologic types, survival time decreases with advancing age. The mechanisms for the strong and consistent inverse association between age and survival are poorly understood and deserve further exploration. Among adult (ages 20 years and older) patients diagnosed with primary malignant brain tumors between 2000 and 2005, only 32.5% and 23.7% survived 2 and 5 years, respectively, from the time of diagnosis.⁴ Although there has been little change in the poor survival rates for patients diagnosed with GBM, for adult (ages 20 years and older) patients diagnosed with all malignant brain tumors combined, 2-year survival has increased from 21.7% in 1975 to 31.9% in 2004.⁴ Among patients diagnosed with glioma, the largest improvements in survival occurred among patients with earlier stage glioma and those who were younger than 65 years of age. Population-based data from Norway and Finland suggest that survival for patients diagnosed with meningioma also improved between the 1950s and 1990s.^7,8 McCarthy et al.⁹ estimate 5-year survival of 69% for meningioma, with those younger at diagnosis and those with benign meningioma having a more favorable prognosis.⁹ It is likely that improvements in imaging technology, which allow the earlier identification of tumors, explain the progress for both glioma and meningioma. GBM probably has only a brief preclinical period; therefore, there may be little opportunity for technology-associated improvement in GBM survival. In the past decade, there has been very little improvement in 2-year survival among adults diagnosed with a primary malignant brain tumor (1.4% difference between the 2-year survival among patients diagnosed in 1994 and those diagnosed in 2004). The only relatively important difference in length of survival following GBM diagnosis, in the past decade, has resulted from concomitant addition of, and maintenance with, the chemotherapeutic agent, Temodar, which has improved median survival time for GBM patients by only 2.5 months.¹⁰

Advances in Identifying Prognostic Factors for Brain Tumors

Previous research has shown that GBM prognosis is associated with the following factors: age, marital status, tumor size, Karnofsky Performance Scale (KPS) score, patient condition before radiation therapy, degree of necrosis, enhancement on preoperative magnetic resonance imaging studies, therapeutic approach including extent of resection and capacity for complete resection, volume of residual disease, location of tumor, patient deterioration, presurgical serum albumin level, and persistent hyperglycemia 1 to 3 months following surgical resection.^11–16

Most recent advances in the search for prognostic factors related to GBM and other high-grade gliomas have focused on molecular or serologic factors or on inherited genetic variation. Loss of heterozygosity (LOH) on chromosome 10q has been associated with shorter duration of survival from GBM,^17,18 and the combined LOH on 1p and 19q may afford a more favorable prognosis to GBM patients.¹⁸ For oligodendroglioma, it is now well established that the combined tumor loss of 1p and 19q confers a more favorable prognosis.¹⁹ Results submitted by Yang et al. showed that two genotypes associated with the 19q deletion region, GLTSCR1-exon-1 and ARCC2-exon-22, are independent predictors of glioma survival.²⁰ Results reported in the past decade provide evidence for the following as prognostic indicators for GBM and other glioma subtypes: p53 mutation and expression,^18,21–29 overexpression or amplification of epidermal growth factor receptor (EGFR),^{22,24,25,27–29} CDKN2A alterations and deletions,^22,24,27 and MDM2 amplifications.^{21,24,27,29,30} For example, EGFR expression is associated with nearly three-fold poorer survival among anaplastic astrocytoma patients.²⁹ In addition, Wrensch et al.²⁹ reported that glutathione-S-transferase (GST) theta (T)1 deletion afforded a less favorable prognosis for glioma patients, while higher survival is afforded to glioma patients with the ERCC1 (a DNA excision repair gene) C8092A polymorphism.

Recently, we have learned that p53 protein expression probably decreases with advancing age,^22,31 and investigators have reported interactions between age and molecular prognostic factors. For example, Simmons et al.³¹ showed that there was a shorter survival time among younger patients whose tumors overexpressed EFGR but had normal p53 immunohistochemistry.³¹ Age-dependent associations between GBM survival and 1p and CDKN2A have also been identified.²²

GBM survival time varies with change in the MNS16A human telomerase polymorphism. Median survival time is 24.7 months for the SS-genotype, compared to 14.0 months and 13.1 months for the SL- and LL-genotypes respectively.³² These results suggest that MNS16A may be used as a biomarker of treatment success.

In order to develop and progress, brain tumors must evade anti-tumor immunity. Recently, several immunological factors have been implicated in glioma prognosis. GBM patients with elevated immunoglobulin (Ig) E live an average of 9 months longer than do patients with lower or normal IgE levels.²⁹ Furthermore, amplification of interleukin (IL)-6, a cytokine which stimulates immune response, is significantly associated with decreased GBM survival time.³³

Prognostic factors for meningioma patients have not been as thoroughly evaluated as those for glioma, perhaps because meningioma patients have more favorable prognoses. The recent requirement of central cancer registries to report benign brain tumors may increase our knowledge of demographic and geographic factors related to meningioma prognosis. Generally, younger meningioma patients survive longer than do older patients. A large study revealed that age, tumor size, and surgical and radiation treatments were associated with benign meningioma survival; however, only age and surgical and radiation treatments were associated with malignant meningioma survival.⁹ Loss of tumor suppressor in lung cancer-1 (TSLC1) protein and abnormalities of chromosome 14 have been associated with meningioma prognosis.^34,35 We know little about the molecular and genetic factors that determine meningioma prognosis, but studies of such characteristics are being conducted at the present time.

Advances in Identifying Causal Factors for Brain Tumors

Only a few causal factors have been identified for brain tumors. Diseases or syndromes associated with rare mutations in highly penetrant genes (including neurofibromatosis types 1 and 2, tuberous sclerosis, retinoblastoma, Li-Fraumeni cancer family syndrome, and Turcot syndrome) and ionizing radiation are known to increase both glioma and meningioma risks. Analytic studies conducted over the past decade have sought to refine these established risk factors or to identify new factors.

Established Environmental Causal Factors for Brain Tumors

There are consistent and strong results from prospective studies of people exposed to ionizing radiation, providing unquestionable evidence of a linear dose-response association between ionizing radiation exposure and glioma risk.³⁶ Exposure to ionizing radiation comes from therapeutic and diagnostic medical procedures, occupation, atmospheric testing of nuclear weapons, natural sources, industrial accidents, and atomic bomb explosions. Therapeutic doses of ionizing radiation probably contribute to the development of only a small proportion of brain tumors, as exposure to therapeutic levels of ionizing radiation is rare.³⁷ Atomic bomb survivors have higher incidence rates of glioma and meningioma, as well as of schwannoma and pituitary tumors,³⁸ and meningioma risk increases with estimated radiation dose to the brain.³⁹ Ionizing radiation used to treat conditions such as tinea capitis in infants and children is associated with elevated relative risks for both glioma and meningioma, along with nerve sheath tumors and pituitary adenoma.^38,40,41 There are mixed results pertaining to diagnostic and therapeutic x-rays of the head and neck.^42–45 Second primary brain tumors occur more frequently than expected among patients previously treated for brain tumors with radiation therapy⁴⁶; however, these results are confounded by the fact that people with higher grade tumors are more likely both to receive radiation and to have a recurrence. Future studies should consider the potential for underreporting of ionizing radiation exposure and imprecise estimates of age at first exposure.³⁶ There may be interaction between ionizing radiation and both age at exposure and genetic variation that may mediate exposure; future studies should consider these possible interactions.

In addition to rare mutations in penetrant genes and ionizing radiation, exogenous hormone use among women is now an established risk factor for meningioma. In part because females have a greater meningioma risk than do males, investigators have examined factors associated with estrogen (menopausal status, ages at menarche and menopause, parity, and uses of oral contraceptives and hormone replacement therapy [HRT]). The ratio of the female-to-male meningioma incidence rate has increased in European countries over time with the increased use of hormone replacement therapy by perimenopausal and postmenopausal women.³ Several results suggest that meningioma risk is greatest among women during their reproductive years^7,47,48; however, SEER statistics reveal that ratios of female-to-male meningioma incidence rates are greatest (greater than 2.5) during the ages 30 to 54 years.⁴ There is no consistent or convincing evidence that parity and oral contraceptive use are associated with meningioma risk.^47–50 However, results from a population-based case-control study conducted by Wigertz et al.⁴⁸ revealed elevated meningioma risks among women who had used long-acting hormonal contraceptives (odds ratio [OR] for at least 10 years of use = 2.7; 95% CI: 0.9-7.5) and postmenopausal women who had ever used HRT (OR = 1.7; 95% CI: 1.0-2.8). While exogenous female hormones probably play a role in the development of some meningiomas, our understanding of mechanisms governing their role is limited, perhaps in part because the menstrual and reproductive factors that have been examined are insufficient to accurately characterize lifetime estrogen or other hormonal exposure.

Probable Causal Factors for Brain Tumors

Previous repeated null findings have allowed the dismissal of some risk factors (such as head injury/trauma and residential exposure to low-level electromagnetic fields), and more focused lines of inquiry have emerged to elucidate the complex roles of family history and allergic conditions and associated immunological factors, including varicella-zoster virus (VZV) infection and associated IgG levels. Over the past decade, the most compelling results from analytic epidemiologic studies have been those concerning immunological factors.