Molecular and Treatment-Associated Correlates of Survival in Glioblastoma: Complementary Surveillance, Epidemiology, and End Results (SEER) and The Cancer Genome Atlas-Glioblastoma (TCGA-GBM) Cohort Analyses.

Background Glioblastoma (GBM) carries a poor prognosis. We evaluated population-level and tumor-level determinants of survival in parallel Surveillance, Epidemiology, and End Results (SEER) and The Cancer Genome Atlas-Glioblastoma (TCGA-GBM) cohorts. Methods SEER cases (histology 9440/3, 2000-2022) were stratified by diagnosis era and registry-recorded trimodality therapy. TCGA-GBM IDH-wildtype cases were analyzed by MGMT promoter methylation status. Kaplan-Meier, log-rank, and multivariable Cox models were applied. Results SEER included 53,432 cases (48,715 deaths), and TCGA IDH-wildtype included 512 cases. Median overall survival (OS) rose modestly across eras (7.0, 9.0, 10.0 months) and remained independently associated with survival after adjustment. Recorded trimodality therapy was associated with longer median OS (14.0 vs 4.0 months). In TCGA, median OS was 15.7 vs 12.5 months for MGMT-methylated vs unmethylated GBM (p=0.012), but MGMT methylation was not significant after multivariable adjustment (HR 0.83, 95% CI 0.63-1.08, p=0.162). Conclusions Era-associated survival gains persisted after adjustment, suggesting that multimodality therapy uptake explains only part of population-level improvements. MGMT methylation predicted unadjusted but not adjusted survival in IDH-wildtype GBM, consistent with known age confounding.