Impact of extent of resection for recurrent glioblastoma on overall survival

Abstract

Object Extent of resection (EOR) has been shown to be an important prognostic factor for survival in patients undergoing initial resection of glioblastoma (GBM), but the significance of EOR at repeat craniotomy for recurrence remains unclear. In this study the authors investigate the impact of EOR at initial and repeat resection of GBM on overall survival. Methods Medical records were reviewed for all patients undergoing craniotomy for GBM at the University of California San Francisco Medical Center from January 1, 2005, through August 15, 2009. Patients who had a second craniotomy for pathologically confirmed recurrence following radiation and chemotherapy were evaluated. Volumetric EOR was measured and classified as gross-total resection (GTR, > 95% by volume) or subtotal resection (STR, ≤ 95% by volume) after independent radiological review. Overall survival was compared between groups using univariate and multivariate analysis accounting for known prognostic factors, including age, eloquent location, Karnofsky Performance Status (KPS), and adjuvant therapies. Results Multiple resections were performed in 107 patients. Fifty-two patients had initial GTR, of whom 31 (60%) had GTR at recurrence, with a median survival of 20.4 months (standard error [SE] 1.0 months), and 21 (40%) had STR at recurrence, with a median survival of 18.4 months (SE 0.5 months) (difference not statistically significant). Initial STR was performed in 55 patients, of whom 26 (47%) had GTR at recurrence, with a median survival of 19.0 months (SE 1.2 months), and 29 (53%) had STR, with a median survival of 15.9 months (SE 1.2 months) (p = 0.004). A Cox proportional hazards model was constructed demonstrating that age (HR 1.03, p = 0.004), KPS score at recurrence (HR 2.4, p = 0.02), and EOR at repeat resection (HR 0.62, p = 0.02) were independent predictors of survival. Extent of initial resection was not a statistically significant factor (p = 0.13) when repeat EOR was included in the model, suggesting that GTR at second craniotomy could overcome the effect of an initial STR. Conclusions Extent of resection at recurrence is an important predictor of overall survival. If GTR is achieved at recurrence, overall survival is maximized regardless of initial EOR, suggesting that patients with initial STR may benefit from surgery with a GTR at recurrence.