Treatments associated with all-cause mortality among children with primary brain and central nervous system tumors: a retrospective cohort study from the SEER database.

Background: Evidence on treatment modalities and survival in childhood primary brain and central nervous system (CNS) tumors remains contradictory, with previous studies often lacking sufficient patient cohort sizes to assess the differences in histological subtypes. This cohort study based on a large population investigated the effects of various treatments on the mortality of patients with different histological types of primary brain and CNS tumors from the Surveillance, Epidemiology, and End Results (SEER) database. Methods: Data of demography, primary tumor site, histology, tumor grade and treatments from all pediatric patients with primary brain and CNS tumors were extracted in this retrospective cohort. The outcomes were overall, 1-, 5-, and 10-year all-cause mortality. Multivariate Cox proportional hazards models were to explore the associations of treatment with overall, 1-, 5-, and 10-year all-cause mortality, with hazard ratios (HRs) and 95% confidence intervals (CIs). Results: Totally 10,994 children were included, with the mean age at diagnosis of 7.3 years, and the median follow-up time of 5.0 years. Of which, 2,003 (18.2%) were diffuse astrocytoma, 3,188 (29.0%) were embryonal tumors, and 3,691 (33.5%) were malignant glioma. Then 4,333 (39.41%) children died during the follow-up. For diffuse astrocytomas and malignant gliomas, patients who received all other treatments were associated with overall, 1-, 5- and 10-year all-cause mortality compared to those only received resection. Embryonal tumors patients receiving resection with radiation only and those receiving resection with both chemotherapy and radiation were associated with lower odds of overall, 1-, 5- and 10-year all-cause mortality compared to patients who only received resection. For ependymal tumors, no surgery/only biopsy with chemotherapy, resection with chemotherapy only, resection with both chemotherapy and radiation, and other treatments had increased risks of overall all-cause mortality compared with resection. The risk of 1-year all-cause mortality increased in ependymal tumors with treatment involving resection and radiation. However, resection with both chemotherapy and radiation was not significantly associated with the 1- nor 5-year all-cause mortality. Conclusions: Resection may be recommended for children with diffuse astrocytoma, ependymal tumors, and malignant glioma, while resection with radiotherapy or chemoradiation may be recommended for children with embryonal tumors.