Panel of serum biomarkers to predict benefit from bevacizumab (BEV) in advanced NSCLC patients.

Abstract

e21069 Background: BEV has produced modest benefits in patients (PTS) with advanced NSCLC. Identification of positive predictors for BEV would have important implications for individual PTS and health care costs. Methods: We performed a prospective exploratory analysis to identify serum biomarkers as predictors of improved outcomes with BEV. Pre treatment sera were collected from 93 pts prior to initiation of first line treatment for advanced NSCLC. Treatment drugs, including BEV, were prescribed according to treating physician’s discretion. Seventy two serum biomarkers, relevant to angiogenesis and tumor progression, were recorded using Luminex immunobead platform. Serum levels were correlated with progression free survival (PFS) and overall survival (OS) and compared between patient treated with or without BEV containing regimens, BEV + and BEV- groups respectively. Log-rank and interaction p value tests were used to identify markers associated with longer PFS and OS in the BEV+ group but not in BEV- group. Results: Characteristics for each group were: BEV+ (n=43, median age 65 y/o, 72% smokers, 60% females, 100% non-squamous). BEV– (n=50, median age 64 y/o, 84% smokers, 50% female, 70 % non-squamous). The BEV+ group had longer PFS (5.8 vs. 3.0 mos, log-rank p= 0.039) and OS (13.1 vs. 8.5 mos, log-rank p =0.11) when compared to the BEV- group. High serum levels of these markers resulted in a differential decreased hazard in the BEV+ group: PDGF-AB/BB (interaction p <0.01 for PFS, p=0.04 for OS), FGF (interaction p=0.15 for PFS, p<0.04 for OS), tenascin-c (interaction p=0.18 for PFS, p=0.04 for OS), RANTES (interaction p=0.04 for PFS, p=0.6 for OS), epiregulin (interaction p=0.31 for PFS, p=0.04 for OS) and anti-HGF (interaction p=0.18 for PFS, p=0.03 for OS). In the BEV+ group higher levels of PDGF-AB/BB were associated with a better outcome (log-rank p=0.05 and p=0.01 for PFS and OS respectively). We did not find significant correlations between serum levels of VEGF, anti-VEGF or VEGFR and benefit from BEV. Conclusions: This exploratory analysis suggests that these biomarkers may have predictive value for BEV in NSCLC PTS and should be considered for further studies.