Brivanib (BMS-582664) in advanced soft-tissue sarcoma (STS): Biomarker and subset results of a phase II randomized discontinuation trial.

Abstract

10000 Background: Pts with sarcomas, a family of >70 types of cancer, have limited options once cytotoxic agents fail. Brivanib (B) is an oral once daily selective dual inhibitor of FGF and VEGF signaling currently in phase III. Data from a phase II trial were analyzed to evaluate the potential utility of biomarkers and STS subtypes in predicting response to B. Methods: Pts aged ≥18 y with unresectable STS and no other treatment options received open-label B 800 mg qd for a 12-wk lead-in period and were assessed by CT/MRI. Pts with RECIST response 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), stratified by FGF2 expression (IHC + or -), until PD or unacceptable toxicity. Pts with PD on P could crossover to resume open-label B. The primary endpoint was PFS for B vs P in FGF2+ pts; secondary endpoints included ORR, DCR (ORR+SD), and exploratory biomarkers. Results: Of 251 pts (52% with ≥2 prior systemic regimens; subtypes: leiomyosarcoma, n=60; liposarcoma, n=61; angiosarcoma, n=20; other, n=110) enrolled from 21 centers, 76 were randomized. PFS from wk 12 was significantly (prespecified 10% alpha level) prolonged for B vs P in randomized FGF2+ pts (n=53): median 2.8 vs 1.4 m (HR 0.58, 90% CI 0.34–0.97; P=.08). HR for the smaller FGF2- subset (n=23) was 0.80 (90% CI 0.36–1.79, P=.69). At wk 12, ORR was 2.8% (7 PR) and DCR 30% overall, with 3 PR in angiosarcomas (15% ORR for subtype). Exploratory analyses (eg, mPFS, HR, PFS rate) showed no apparent differences between subtypes. Of 38 patients randomized to P, 30 crossed over to B at PD; median PFS for these pts from crossover was 4.1 m (95% CI, 2.8–6.2). Most common AEs were as expected; grade 3-4 AEs ≥5% were hypertension (15%), fatigue (10%), dyspnea (5%), and abdominal pain (5%). Conclusions: Single-agent B significantly improved PFS vs P in pts with heavily pretreated advanced FGF2+ STS (benefit in FGF2- not excluded), with most common AEs as expected. All STS subtypes appeared to potentially benefit, with possibly increased activity in angiosarcoma. Further investigation is warranted, potentially in combination with other agents.