Brivanib (B) in Advanced Ovarian Cancer (OC): Subset Results of a Phase 2 Randomized Discontinuation Trial (RDT)

Abstract

Background: Brivanib (B) is an oral once daily selective dual inhibitor of FGF and VEGF signaling with preclinical activity against various tumor types. An RDT assessed B in pts with advanced solid tumors; data showing activity in sarcoma pts were reported previously, and data for the OC pts are presented here. Methods: OC pts progressing after previous treatment received open-label B 800 mg QD for a 12-wk lead-in period and were assessed by CT/MRI. Pts with complete (CR) or partial response (PR) continued on open-label B, those with progressive disease (PD) went off study. Pts with stable disease (SD) were randomized 1:1 to B or placebo (P) until PD or intolerance. Pts with PD on P could crossover to open-label B. Endpoints included progression-free survival (PFS) in randomized pts, objective response rate (ORR), disease control rate (DCR = ORR + SD), and safety. Results: One hundred twenty-six pts (63% with >3 prior systemic regimens; 88% FGF2+ by IHC) entered 12-wk lead-in and received B. At wk 12, 12 pts had PR and 43 pts had SD (ORR 10%; DCR 44%). Three of 18 pts pretreated with bevacizumab (BEV) achieved PR at wk 12. After 12 wks, 49 continued on the study, 10 with PR remained on open-label B; and 39 were randomized to either B (n = 19) or P (n = 20). Two pts with PR were randomized in error. Among randomized pts (n = 39), median PFS was 4 mo for B vs 2 mo for P with HR of 0.54 (90% CI: 0.28-1.03; p = 0.11). Among randomized FGF2+ pts (n = 36), HR was 0.56 (90% CI: 0.29-1.07; p = 0.14). In the randomized phase, 3 additional pts achieved PR (overall population: 15 PRs; ORR 12%). The most common (≥5%) grade ≥3 AEs were increased ALT (20%) and AST (14%), fatigue (12%), hyponatremia (9%), asthenia (7%), diarrhea (7%), hypertension (7%), abdominal pain (6%), and decreased appetite (6%). Discontinuation due to treatment-related AE occurred in 13% of pts. Conclusions: The observed PFS prolongation with B vs P, ORR and DCR in this RDT indicate clinical activity of B in pts with heavily pretreated OC, including those with prior BEV. The high frequency of FGF2+ pts precluded the assessment of FGF2 as a predictive biomarker. The safety profile was acceptable. These results support further investigation of B in OC, potentially in pts with prior BEV.