Multi-institutional external validation of a novel glioblastoma prognostic nomogram incorporating MGMT methylation

A recent nomogram for glioblastoma (GBM) was designed to incorporate methylguanine-DNA methyltransferase (MGMT) methylation status in trial patients receiving temozolomide. Since clinical trial patients are strictly selected, compared to the general population, we performed a multi-institutional, external, independent assessment of the nomogram. Consecutive adult patients with supratentorial GBM diagnosed between June 2007 and December 2014 who initiated TMZ-based concurrent chemoradiotherapy (CRT) and were not enrolled on RTOG 0525 or 0825 were eligible. We collected age, gender, MGMT status, performance status, resection extent, race, and tumor site and Cox regression analysis of overall survival (OS) was conducted with the 1-year nomogram-predicted survival (NPS). The predictive accuracy was quantified by the concordance index (c-index) as well as by separating patients into quintile-groups of the population distribution of NPS and comparing mean NPS and observed OS. Of 514 patients with GBM, 309 had all nomogram factors. Median OS was 18.7 months. NPS and observed OS demonstrated a c-index of 0.695. On univariate analysis, the NPS and all included factors except gender were significant. On multivariable analysis (MVA) the only significant factor for worse survival was lower NPS. When separated into quintile-groups of NPS, the observed survival was slightly better than the predicted survival for all but the worst prognostic group. Our multi-institutional cohort provides independent external validation of a novel GBM nomogram incorporating MGMT methylation status. No individual factor included in the nomogram retained significance on MVA after adjusting for NPS.