Purpose: To identify the predictive and prognostic factors for a decrease or recurrence of brain edema in patients who developed radiation-induced brain necrosis (RN) after radiation therapy for nasopharyngeal carcinoma (NPC) and who received bevacizumab monotherapy.
Methods and materials: This was a retrospective study. The charts of 50 patients who were diagnosed with RN after radiation therapy for NPC, treated with bevacizumab, and followed up for 6 months were reviewed. Clinical data of these patients were collected, and their brain edema volume before bevacizumab administration, after bevacizumab administration, at 3-month follow-up, and at 6-month follow-up was evaluated on the basis of brain magnetic resonance imaging findings. The baseline serum vascular endothelial growth factor levels of 15 patients were measured by enzyme-linked immunosorbent assay. A random forests model was developed for statistical analysis.
Results: The median percentage of decrease in RN volume shown on T2-weighted fluid-attenuated inversion recovery images at the end of bevacizumab therapy was 72.6% (interquartile range, 34.5% to 89.5%; P < .001). Twelve of these 50 patients (24.0%) did not have an effective response, and 38 patients (76.0%) showed an effective response after bevacizumab administration. Fifteen of the 38 patients showed RN recurrence. According to the random forests model the maximum radiation dose of the temporal lobe (Dmax of the temporal lobe) was a highly ranked predictor for the therapeutic effect of bevacizumab. The duration between radiation therapy and bevacizumab treatment and the duration between radiation therapy and RN diagnosis were highly ranked predictors for RN recurrence after bevacizumab treatment.
Conclusions: Prediction models for the therapeutic effect of bevacizumab in RN patients were developed, using the random forests model. Bevacizumab might be more effective in patients with a lower maximum radiation dose to the temporal lobe.
Copyright © 2017 Elsevier Inc. All rights reserved.