Dianne Daniels
Our group has previously introduced delayed-contrast MR imaging for calculating treatment-response-assessment maps (TRAMs) in patients with brain tumors undergoing standard treatments, providing a clear separation between an active tumor and treatment effects.1,2 Following this application of the TRAMs methodology to aid patient management and decision-making in patients with newly diagnosed high-grade glioma (HGG) undergoing standard first-line treatment, we decided to focus on seeking a better means for assessing response to bevacizumab treatment in patients with recurrent HGG. Although bevacizumab has become a leading treatment option for recurrent HGGs in the United States, the interpretation of the radiologic response of these patients still represents a unique challenge.
Furthermore, in our studied cohort, 62.5% of patients showed a positive radiologic response to bevacizumab on standard MR imaging 1 month after initiation of treatment. However, only 29.2% demonstrated a long-term response. These numbers suggest that a more reliable tool for the early prediction of long-term response to bevacizumab is required.
Our study identified radiologic response and progression patterns from the mean change in lesion volumes as a function of time, calculated from conventional T1-weighted MR imaging, TRAMs, and perfusion MR imaging. Thresholds for early (1 month) prediction of response were calculated using these sequences. The predictor calculated from TRAMs demonstrated higher sensitivity, specificity, and positive/negative predictive values. The benefit of delayed-contrast MR imaging in separating responders from nonresponders was further confirmed by using log-rank and receiver operating characteristic analyses.
In this study, response was determined by overall survival (OS), a reliable measure of clinical outcome by all means, and predictors of response were confirmed to provide significant separation between short- and long-term responders (in terms of OS) by log-rank analysis.