Early tumour shrinkage as a survival predictor in patients with recurrent glioblastoma treated with bevacizumab in the AVAREG randomized phase II study

BACKGROUND Disease assessment for recurrent glioblastoma (GBM) represents a challenge, especially with the use of antiangiogenic agents. Moreover, validated neuroradiological predictors of outcome are lacking. Recently, the concept of early tumor shrinkage (ETS) has been developed to better assess the ability of treatments in determining a rapid and remarkable tumor response. The aim of the study was to evaluate the role of ETS in predicting survival of GBM patients treated with BEV. METHODS We examined the radiological data of patients with recurrent GBM treated with bevacizumab (BEV) or fotemustine (FTM) in the randomized phase II AVAREG trial (EudraCT: 2011-001363-46). Radiologic assessments at first disease assessment (day 46) were used to calculate the relative change in the sum of the products of perpendicular diameters of all measurable lesions determined by either T1 contrast and T2/FLAIR. RESULTS In patients treated with BEV, the best ETS cut-off was reduction of 15% with T1 contrast and of 40% with T2/FLAIR. Adopting this cut-off for T1 contrast radiological changes, ETS was a significant predictor of OS for patients treated with BEV (HR = 0.511, 95%CI:0.269-0.971, p = 0.040). The cut-off obtained for T2/FLAIR was not significantly correlated with OS (p = 0.102), but we found a trend for correlation with survival when considering the variable as continuous (p = 0.058). CONCLUSIONS ETS evaluating T1 contrast reduction is a helpful predictor of survival in patients with recurrent GBM treated with BEV, and if validated in a larger prospective trial could be a helpful surrogate endpoint.