METHODS
In this retrospective, population-based cohort study, we included children between the ages of 0-19 years newly diagnosed with primary CNS tumors between 2000 - 2015 in the Surveillance, Epidemiology, and End Results (SEER) Program sponsored by the National Cancer Institute (NCI). These population-based cancer registries include all incident cancer from 18 geographical locations in the US representing more than 28% of the US population.23 SEER collects demographic, clinical, treatment and survival information on all incident cancers (excluding non-melanoma skin cancer). Tumor histology is classified using International Classification of Diseases for Oncology (ICD-O-3) codes.24 We additionally received special permission to access the Specialized Census Tract-level SES and Rurality Database developed by SEER. In these data, socioeconomic status for patients represented by Yost quintile reflects the SES group for the patients’ census tract of residence at the time of disease (see below for measure definition). 25,26 The SES Yost quintile classification of census tracts is based on a time dependent composite score of the social and economic stratification of the area in which a patient resides at the time of diagnosis.26
Cancers diagnosed only by direct visualization or clinical impression without any imaging or laboratory confirmation were excluded from the study (n=73). Cases diagnosed by autopsy or death certificate only were also excluded (n=41), as were cases without information on their SES (n=151). Patients who had the optic nerve as the primary site of tumor were also excluded (n=723) from the data set due to its association with neurofibromatosis and complex natural history and therapy.27 Our final sample included 8,713 children aged 0-19 at diagnosis.