Treatment and survival of patients harboring histological variants of glioblastoma

It is unclear whether the survival difference observed between glioblastoma (GBM), giant cell glioblastoma (gcGBM), and gliosarcoma (GSM) patients is due to differences in tumor histology, patient demographics, and/or treatment regimens. The USA National Cancer Database was utilized to evaluate patients diagnosed with GBM, gcGBM, and GSM between 1998 and 2011. Kaplan-Meier survival estimates and Cox proportional hazards models were utilized to estimate overall survival. A cohort of 69,935 patients was analyzed; 67,509 (96.5%) of these patients had GBM, 592 (0.9%) gcGBM, and 1834 (2.6%) GSM. The median age for GBM and GSM patients was 61 versus 56 years for gcGBM (p < 0.0001). Higher extent of resection (p < 0.0001) and radiation (p = 0.001) were observed in gcGBM patients compared to other histologies. Multivariate analysis showed that gcGBM patients had a 20% reduction in the hazards of mortality (hazard ratio [HR] 0.80, 95% confidence interval [CI] 0.69-0.93) compared to GBM, while GSM patients trended towards higher hazards of mortality (HR 1.04, 95% CI 0.96-1.12) than the GBM cohort. Previous studies have suggested a disparity in the survival of patients with GBM tumors and their histological variants. Using a large cohort of patients treated at hospitals nationwide, this study found a 20% reduction in the hazards of mortality in gcGBM patients compared to GBM. Similarly, gcGBM patients had a 24% reduction in the hazards of mortality compared to the GSM cohort. GSM patients had a 3% increase in the hazards of mortality compared to GBM.