Apparent diffusion coefficient changes predict survival after intra-arterial bevacizumab treatment in recurrent glioblastoma

Abstract

Superselective intra-arterial cerebral infusion (SIACI) of bevacizumab (BV) has emerged as a novel therapy in the treatment of recurrent glioblastoma (GB). This study assessed the use of apparent diffusion coefficient (ADC) in predicting length of survival after SIACI BV and overall survival in patients with recurrent GB. Sixty-five patients from a cohort enrolled in a phase I/II trial of SIACI BV for treatment of recurrent GB were retrospectively included in this analysis. MR imaging with a diffusion-weighted (DWI) sequence was performed before and after treatment. ROIs were manually delineated on ADC maps corresponding to the enhancing and non-enhancing portions of the tumor. Cox and logistic regression analyses were performed to determine which ADC values best predicted survival. The change in minimum ADC in the enhancing portion of the tumor after SIACI BV therapy was associated with an increased risk of death (hazard ratio=2.0, 95% confidence interval(CI) [1.04-3.79], p=0.038), adjusting for age, tumor size, BV dose, and prior IV BV treatments. Similarly, the change in ADC after SIACI BV therapy was associated with greater likelihood of surviving less than 1year after therapy (odds ratio=7.0, 95% CI [1.08-45.7], p=0.04). Having previously received IV BV was associated with increased risk of death (OR 18, 95% CI [1.8-180.0], p=0.014). In patients with recurrent GB treated with SIACI BV, the change in ADC value after treatment is predictive of overall survival.