PATH-28. ANGIOTENSINOGEN PROMOTER METHYLATION TO PREDICT BEVACIZUMAB RESPONSE IN RECURRENT GLIOBLASTOMA PATIENTS

Abstract

Abstract Recurrent glioblastoma patients achieving response to bevacizumab combination therapy have clinical improvement and prolonged survival. High gene-expression of angiotensinogen (AGT) is associated with a poor response to bevacizumab combination therapy. Because AGT gene-expression is epigenetically regulated, we investigate if lower AGT promoter methylation in tumor tissue predicts a poor response to bevacizumab combination therapy in recurrent glioblastoma patients. Methods: Patients were assessed for eligibility using our clinical database comprising all recurrent glioblastoma patients consecutively treated with bevacizumab combination therapy at our center. The study included 159 response and biomarker evaluable patients: A training cohort of 77 patients and a validation cohort of 82 patients treated in the period between year 2005–2011 and 2012–2015. DNA methylation of 4 CpG sites in the AGT promoter was measured using pyrosequencing. Using logistic regression analysis, a predictive model for non-response was established. Results: In the training cohort, lower methylation of each of the four CpG sites was significantly associated with non-response (p < 0.05). Lower mean methylation of the AGT promoter was significantly associated with non-response (2-fold decrease: OR &eq; 3.01; 95% CI:1.41–6.44; p &eq; 0.004). A predictive model able to predict bevacizumab non-response in clinical practice was established. This predictor was significantly associated with non-response in the validation cohort (p &eq; 0.037). Conclusion: AGT promoter methylation is lower in tumor tissue from non-responsive recurrent glioblastoma patients treated with bevacizumab combination therapy. A predictive model for non-response was established and successfully validated. This model can be used to identify patients who will not benefit from bevacizumab combination therapy.