Background: The randomized phase III TURANDOT trial compared first-line BEV plus paclitaxel (PAC) vs BEV plus capecitabine (CAP) in HER2-negative metastatic BC (mBC). BEV-based regimens are often favored in TNBC [Dawood 2012] because of efficacy in subgroup analyses and a lack of effective treatments. We performed an exploratory subgroup analysis of TURANDOT to provide more data on BEV-based therapy in TNBC. Methods: Patients (pts) with HER2-negative mBC who had received no prior chemotherapy for mBC were randomized to either BEV–PAC (BEV 10 mg/kg d1 and 15 + PAC 90 mg/m² d1, 8, and 15 q4w) or BEV–CAP (BEV 15 mg/kg d1 + CAP 1000 mg/m²bid d1–14 q3w). The primary endpoint was overall survival (OS); secondary endpoints included objective response rate (ORR), progression-free survival (PFS), and safety. Results: Of 561 pts treated, 130 had TNBC. Baseline characteristics were typical of a poor-prognosis population and generally balanced between treatment arms, although fewer pts receiving BEV–PAC than BEV–CAP had ECOG PS 1/2 (25% vs 40%, respectively), positive lymph nodes (56% vs 72%), metastatic disease at first diagnosis (19% vs 30%), and liver metastases (27% vs 43%). Median age was 54 vs 56 years, respectively. At data cut-off, median follow-up was 21.4 vs 19.2 mo for BEV–PAC and BEV–CAP, respectively. The safety profiles in the TNBC subgroup were similar to the overall population. The predominant grade ≥3 AEs were hematologic AEs and neuropathy with BEV–PAC and hand-foot syndrome and diarrhea with BEV–CAP. Conclusions: One-year OS rates up to 78% in TURANDOT are among the highest seen in TNBC. BEV-based therapy is a valid option in a setting with limited active treatments. BEV–PAC may be favored based on 1-year OS, PFS, and ORR. Clinical trial information: NCT00600340.

^a Univariate Cox proportional hazard model (CPHM). ^b Multivariate CPHM, treatment effect adjusted for ECOG PS, bone metastases, lymph nodes, and menopausal status.