Retrospective analysis of candidate predictive tumor biomarkers (BMs) for efficacy in the GOG-0218 trial evaluating front-line carboplatin–paclitaxel (CP) ± bevacizumab (BEV) for epithelial ovarian cancer (EOC).

Abstract

5505 Background: In EOC, the double-blind placebo (PLA)-controlled randomized phase 3 GOG-0218 trial showed significantly improved progression-free survival (PFS) (but not overall survival [OS]) with BEV added to front-line CP and continued alone. Plasma BM correlative analyses identified no predictive BMs for efficacy. Tumor BM evaluation was an exploratory objective. Methods: Patients (pts) with stage III (incompletely resected) or IV OC were randomized to receive 6 cycles of CP with: PLA (CPP); BEV 15 mg/kg q3w → PLA (CPB15); or BEV for 15 mo (CPB15+). Five tumor BMs (CD31, tVEGF-A, VEGFR-2, NRP-1, MET) with a biologic rationale for evaluation were assessed by immunohistochemistry (IHC). The BM-evaluable population (BEP; all pts with an evaluable BM sample and ≥ 1 post-baseline efficacy assessment) was analyzed using 1st, 2nd and 3rd quartile (Q) expression levels for each BM as the cutoff for high vs low BM subgroups. Correlations between tumor BM levels and PFS and OS were analyzed. Results: The BEP, comprising 1455 (78%) of 1873 pts in the ITT population, had very similar baseline characteristics and efficacy to the ITT population. No prognostic or predictive association was seen for VEGFR-2, NRP-1 or MET. However, when comparing CPB15+ vs CPP (control), higher microvascular density (MVD) measured by CD31 IHC showed prognostic (not shown) and potential predictive value for PFS ( > Q3 MVD HR 0.38 [95% CI 0.25–0.58]; ≤ Q3 MVD HR = 0.68 [95% CI 0.54–0.86]; interaction p = 0.018) and OS ( > Q3 MVD HR 0.57 [95% CI 0.39–0.83]; ≤ Q3 MVD HR 1.03 [95% CI 0.83–1.27]; interaction p = 0.0069). tVEGF-A showed potential predictive value for OS (and PFS) for CPB15+ vs CPP with a Q3 cutoff ( > Q3 tVEGF-A OS HR 0.62 [95% CI 0.43–0.91]; ≤ Q3 tVEGF-A OS HR 1.01 [95% CI 0.82–1.25]; interaction p = 0.023). Conclusions: These retrospective tumor BM analyses suggest a positive correlation between expression levels of molecular (tVEGF-A) and cellular (endothelial cell) targets of anti-VEGF and magnitude of PFS and OS improvement from BEV in EOC. The predictive value of these candidate efficacy BMs requires validation in other relevant datasets. Clinical trial information: NCT00262847.