The timing of chemoradiotherapy after surgical resection and its impact on overall survival in glioblastoma.

Abstract

2051 Background: Prior studies examining time to initiate chemoradiotherapy (CRT) after surgical resection (S) in glioblastoma (GBM) have not provided clear consensus on its clinical impact. We sought to evaluate the effect that differential timing of adjuvant therapy may have on overall survival (OS). Methods: With the National Cancer Database (NCDB), patients (pts) with GBM who underwent S and adjuvant CRT from 2004-2013 were analyzed. Analysis was performed for the entire cohort as well as by Radiation Therapy Oncology Group (RTOG) recursive partitioning analysis (RPA) classes (i.e. I, II, and III). Time from S to CRT was grouped weekly (i.e. 0-1, 1-2, 2-3, 3-4, 4-5, 5-6, 6-7, 7-8, and > 8 weeks). Pts were excluded if they died within the first 8 weeks to account for immortal time bias. Kaplan-Meier analysis, log-rank testing, and multivariate (MVA) Cox proportional hazards regression were performed with OS as the primary outcome. Results: A total of 30,414 pts were included for analysis. RPA class I, II, and III contained 903, 4,347, and 25,164 pts, respectively. The most common time to initiate CRT was week 4-5 (n = 7389), and this group served as reference for survival analysis. On MVA, weeks 0-1 (hazard ratio [HR] 1.18, 95% confidence interval [CI] 1.02-1.35), 1-2 (HR 1.24, CI 1.17-1.32), and 2-3 (HR 1.11, CI 1.07-1.15) demonstrated worse OS (all p < 0.03). For RPA class I pts, week 1-2 (HR 2.07, CI 1.08-3.95) was associated with worse OS (p = 0.028). For RPA class II pts, weeks 1-2 (HR 1.34, CI 1.14-1.57), 2-3 (HR 1.18, CI 1.07-1.31), and 3-4 (HR 1.10, CI 1.0-1.21) were associated with worse OS (all p < 0.05). For RPA class III pts, weeks 0-1 (HR 1.18, CI 1.02-1.38), 1-2 (HR 1.22, CI 1.14-1.3), and 2-3 (HR 1.09, CI 1.05-1.14) were associated with worse OS (all p < 0.03). No time point after week 5 was associated with change in OS for the overall cohort or any RPA class subgroup. Conclusions: These data provide insight into the optimal timing of CRT in GBM and describe RPA-class specific outcomes. In general, OS was negatively impacted if CRT started less than 3 weeks from S. Waiting up to 8 weeks, however, was not detrimental to OS and suggests delaying CRT beyond week 4-5 should be considered if clinically indicated without undue concern.