Association of VEGF single-nucleotide polymorphisms (SNPs) with response and toxicity in patients treated with bevacizumab.

Abstract

e13560 Background: Functional polymorphisms modulate the regulation of gene expression or influence the function of the coded protein, and polymorphisms of genes associated with angiogenesis may influence the likelihood of benefit or toxicity from antiangiogenic treatment. Although the ECOG 2100 study demonstrated the predictive value of VEGF polymorphisms in the narrow context of breast cancer patients receiving combination bevacizumab and paclitaxel, it is unknown whether this predictive value of VEGF polymorphisms is more broadly applicable across diverse malignancies or across different bevacizumab treatment combinations. Methods: Archived tissue and PMBCs were collected from 400 patients who were treated on bevacizumab-based clinical trials at our institution. DNA extraction, genotyping, and SNP assay evaluation were performed according to standard protocols. Presence or absence of VEGF SNPs were correlated with response to treatment, defined as partial response, complete response, or prolonged stable disease lasting > 4 months. Presence or absence of VEGF SNPs were also correlated with incidence of bevacizumab-related grade 3 or 4 toxicities. Results: Preliminary analysis of 75 patients, including 21 different tumor types and 6 different bevacizumab combinations, showed a trend of correlation of VEGF polymorphism status with response and toxicity, compared to alternate genotypes. Patients with VEGF-2578 AA demonstrated a trend of improved response rate and prolonged stable disease (38% vs. 17%, p=0.068). Patients with VEGF-1498 TT demonstrated a trend of less bevacizumab-related grade 3 or 4 toxicity (5% vs. 21%, p=0.086). Patients with VEGF-634CC demonstrated a trend of less bevacizumab-related grade 3 or 4 toxicity (9% vs. 16%, p=0.714). Conclusions: These clinical data support the concept that VEGF polymorphisms may be potential biomarkers of efficacy and toxicity for patients treated with bevacizumab combinations, independent of tumor site or histology and independent of bevacizumab combination. Updated results of the 400-patient data set will be presented, including a subanalysis of colorectal cancer patients.