Second Tumors of the Central Nervous System
Second neoplasms (SNs) were one of the first late effects recognized in survivors of childhood cancer.1 These events occur after essentially all childhood cancers and have been clearly linked to the characteristics of the treatment and of the childhood cancer survivor. As such, it is not surprising that SNs occur in survivors of childhood central nervous system (CNS) tumors and that SNs of the CNS are a well-documented occurrence in the survivors of childhood cancers that arise outside of the CNS. This chapter reviews both of these clinical settings and provides details on the incidence and risk factors (both therapy and host) that are associated with these events.
Interpretation of the data discussed here requires recognition of the limitations of the studies from which they were derived. CNS tumors that arise in a survivor of a primary CNS tumor can be difficult to differentiate from the primary neoplasm in number of settings. Primary CNS tumors may recur very late, often a decade or more after the original diagnosis. Distinguishing a “new” tumor from a recurrence can be particularly problematic if the second event is of a similar (i.e., a different grade of a glioma) histology and near the primary site. The occurrence of multiple CNS neoplasms in an individual may also be attributed to therapy, but in the setting where a primary genetic disorder (i.e., neurofibromatosis) is unrecognized, such attribution may be incorrect, at least in part. These situations can skew the interpretation of research results and may add to uncertainty about risk estimates.
One of the earliest descriptions of CNS second cancers came from the Late Effects Study Group in 1982.2 This consortium of 10 institutions constructed a cohort of almost 15,000 childhood cancer patients treated from 1950 to 1970 and documented the occurrence of 113 new primary malignancies. Within this cohort, the 14 of these new primary malignancies that occurred among the survivors of primary CNS tumors were as follows: sarcomas (three); subsequent CNS, bone, lymphoma, and skin cancers (two each); and thyroid, leukemia, and “other” (one each). Across this entire cohort nine subsequent cancers arose in the CNS with a wide range of primary cancer diagnoses. Another cohort, constructed from the British National Health Service, assessed the risk of second cancers among 10,000 three-year survivors of childhood cancer.3 Ninety of these patients developed a second primary tumor, and 24 of them were survivors of primary CNS tumors. Four of the second tumors were meningiomas. The greatest excess risks among CNS survivors were for subsequent thyroid cancers (standardized incidence ratio [SIR] = 48.7) and connective tissue cancers (SIR = 38.8), and were associated with radiotherapy treatment. Now, 30 years after this early publication, the nature of these risks has been refined and the contribution of treatment has become clearer.
Second Neoplasms in Survivors of Central Nervous System Tumors
As with any group of childhood cancer survivors, the survivors of childhood CNS tumors are at an increased risk of new malignant and nonmalignant neoplasms. This risk results from the characteristics of the individual and the nature of the therapy for the first cancer.
When considered among all childhood cancer survivors, the overall risk of a second malignant neoplasm among survivors of CNS tumors does not stand out as substantially different from the risk for the entire group of survivors. The Childhood Cancer Survivor Study (CCSS) is a retrospective cohort of 5-year or more survivors of selected childhood cancers in the United States and Canada.4 The study is a National Cancer Institute–supported resource for outcome and intervention research; it reported on the occurrence of subsequent neoplasms among 14,359 five-year-plus survivors, with a mean of 22.7 years of follow-up from the original diagnosis.5 The risk of a new, invasive cancer was increased sixfold over the expected risk for the entire cohort. The survivors of CNS tumors had a very similar risk, ranging from a SIR of 4.3 among survivors of astrocytoma to 7.3 among survivors of medulloblastoma. An analysis of data from the U.S. National Cancer Institute′s Surveillance, Epidemiology and End Results (SEER) program that included over 25,000 persons diagnosed before age 18 with cancer between 1973 and 20026 found the risk of a new malignancy was increased 5.9-fold for the entire cohort as compared with 6.3-fold for survivors of primary CNS cancers. Survivors of CNS tumors in Great Britain were also found to be at similar risk to these other cohorts and, in aggregate had a risk of a second invasive cancer that was similar to that of other large groups in the British cohort.7
Although the overall risk of a new invasive cancer among CNS survivors is similar to that of all childhood cancer survivors, the nature of the invasive and noninvasive neoplasms is clearly different and modified by therapy and host factors ( Table 49.1 ). Across the cohort studies referenced above, subsequent CNS malignancies, thyroid cancers, meningiomas, and melanoma and nonmelanoma skin cancers all occur at an increased frequency in CNS survivors. In the SEER analysis of new primary cancers the highest SIRs for subsequent brain and CNS cancers were found among astrocytoma (SIR = 12.6) and primitive neuroectodermal tumor (PNET; SIR = 44.6) survivors.6 The most recent CCSS report identified a total of 59 subsequent malignant CNS tumors among the entire cohort; 14 of them (13 glial tumors and one PNET) occurred in the CNS survivors.5 In a more detailed analysis of 1,877 CNS survivors within the CCSS, 20 subsequent malignant CNS neoplasms were observed versus less than one expected (SIR = 25.3). The median time to occurrence of the new CNS malignancy was 14 years.8 Meningiomas also occur in excess in the CNS survivor population. Of 170 meningiomas observed in the CCSS cohort, 68 occurred in CNS survivors, resulting in a 30-year cumulative incidence of meningiomas that ranged from 7.1% among astrocytoma survivors to 16.4% in survivors of medulloblastoma.5 More detailed analysis of the risk factors for the development of subsequent meningiomas has been completed for all childhood cancer survivors and is discussed below.
Subsequent central nervous system tumor (glioma/primitive neuroectodermal tumor)
Soft tissue sarcoma
Melanoma Nonmelanoma skin cancer
Neoplasms outside of the CNS also occur in excess among CNS survivors. In a large, population-based study of the occurrence of second malignancies following childhood CNS, malignant tumors data were obtained on 8,431 children under the age of 15 years.9 The cumulative incidence of a non-CNS malignancy was 3.3% at 45 years after the original cancer diagnosis. Excess risk was evident for subsequent thyroid cancer (SIR = 10.6), leukemia (SIR = 2.8), and lymphoma (SIR = 2.5). The authors estimated that approximately one in 180 CNS survivors will develop a non-CNS second malignant neoplasm within 15 years after the diagnosis of the primary disease. Survivors of CNS tumors who participated in the CCSS cohort also were at increased risk for later development of thyroid cancers (SIR = 11.2), soft tissue sarcoma (SIR = 8.4), and bone cancers (SIR = 15.1).8
Central nervous system tumor survivors are at risk for both melanoma and nonmelanoma skin cancers (NMSCs). The occurrence of NMSC in childhood cancer survivors is increasingly being recognized as a substantial contributor to the long-term morbidity of this population.10
Nonmelanoma skin cancer in survivors of childhood and adolescent cancer (report from the CCSS) has a cumulative incidence in CNS survivors approaching 8% at 30 years after diagnosis.5 In many cases a single individual may have numerous NMSCs that occur over many years. The occurrence of multiple NMSCs in patient with medulloblastoma who also carry the mutation for Gorlin syndrome is well recognized.11 Malignant melanoma may also occur and, although at marginal excess, it should be considered in the care of these survivors.8
The cumulative risk of a second neoplasm for survivors of childhood CNS tumors who participated in the CCSS is presented in Table 49.2. 8 By definition of the cohort, second neoplasms occurring before 5 years were excluded; thus, a small number are likely missed. Nevertheless, there is a steady rise in the occurrence of both malignant and nonmalignant neoplasms evident at 10 years after diagnosis and continuing to the limits of follow-up so that at 25 years after the original diagnosis over 10% of survivors had developed an SN. The highest cumulative incidence was for meningiomas (3.3%), followed by NMSCs (2.9%). Among the non-NMSC and non-meningioma SNs in this study (n = 76), the most frequent cancers were malignant CNS tumors (n = 20) and thyroid cancers (n = 12).
The risk of any second neoplasm is modified by host and treatment characteristics. Overall, there has been an increase in risk associated with the use of more aggressive chemotherapy and radiation therapy. Specific considerations for CNS tumor survivors include the radiation-associated risk of thyroid cancer, skin cancers, sarcomas, bone tumors, and other neoplasms in or near the radiation field. Children with an underlying genetic predisposition to cancer (e.g., neurofibromatosis types 1 or 2 [NF-1 or NF-2], nevoid basal cell carcinoma syndrome, and others) are at increased risk as well.