Objective: This retrospective study analyzed whether the type of radiologic progression, classified according to contrast enhancement on MRI T1-weighted sequences and changes in T2-hyperintense signal, is relevant for outcome in patients with progressive glioblastoma (pGB) treated with bevacizumab.
Methods: MRI scans of 83 patients with pGB treated with bevacizumab were evaluated prior to and at disease progression. Based on initial decrease in and subsequent flare-up of contrast enhancement in T1 and 2 patterns of T2-hyperintense tumor progression, progression types (PTs) were categorized as cT1 flare-up, T2-diffuse, T2-circumscribed, or primary nonresponder. Overall survival (OS), survival from start of bevacizumab therapy (OS_Bev), survival after bevacizumab failure (OS_PostBev), time from initial diagnosis until initiation of bevacizumab therapy (StartBevT), and time to bevacizumab progression were evaluated using Kaplan-Meier curves, log-rank test, and Cox regression analyses.
Results: The time observed for development of a T2-diffuse (n = 15) or a cT1 flare-up (n = 35) progression was longer than for progression in primary nonresponders (n = 16) or T2-circumscribed progression (n = 17). The T2-diffuse PT showed longer OS, OS_Bev, OS_PostBev, and StartBevT compared to the other PTs. Postprogression therapy tended to be relevant only for patients with a T2-circumscribed PT.
Conclusions: Radiologic PTs following bevacizumab treatment failure show differences in time to development and are related to outcome. We therefore hypothesize that these PTs reflect a different glioma biology, including differential resistance mechanisms to bevacizumab, and may be associated with different responses to postprogression therapy.