Population-based analysis of radiation-induced gliomas after cranial radiotherapy for childhood cancers

BACKGROUND: Cranial radiotherapy (RT) used for pediatric CNS cancers and leukemias carries a risk of secondary CNS malignancies, including radiation-induced gliomas (RIG). Our aim was to characterize the epidemiology of RIG. METHODS: This retrospective study used SEER data (1975–2016). Cohort 1 included patients diagnosed with glioma as a second malignancy ≥2 years after receiving treatment for a first malignancy diagnosed at 0–19 years, either a primary CNS tumor (1a, n = 57) or leukemia (1b, n = 20). Cohort 2 included patients who received RT for a pediatric CNS tumor and died of presumed progressive disease >7 years after diagnosis, since previous studies have documented many missed RIGs in this group (n = 296). Controls (n = 10 687) included all other patients ages 0–19 years who received RT for a first CNS tumor or leukemia. RESULTS: For Cohort 1, 0.77% of patients receiving cranial RT developed RIG. 3.39% of patients receiving cranial RT for primary CNS tumors fell in cohort 2. Median latency to RIG diagnosis was 11.1 years and was significantly shorter for cohort 1b than 1a. Median OS for cohort 1 was 9.0 months. Receiving surgery, radiation, or chemotherapy were all associated with a nonstatistically significant improvement in OS (P .1–.2). A total of 1.8% of all brain tumor deaths fell in cohort 1, with 7.9% in cohort 2. CONCLUSION: A total of 1%–4% of patients undergoing cranial RT for pediatric cancers later developed RIG, which can occur 3–35 years after RT. Given the substantial and likely underestimated impact on overall CNS tumor mortality, RIG is deserving of increased attention in preclinical and clinical studies.