Abstract PD10-03: BEGONIA: Phase 1b/2 study of durvalumab (D) combinations in locally advanced/metastatic triple-negative breast cancer (TNBC): Results from Arm 1 D + paclitaxel (P), Arm 2 D+P + capivasertib (C), and Arm 5 D+P + oleclumab (O)

Abstract Background: Chemotherapy, together with immune checkpoint inhibitors, improves outcomes vs chemotherapy alone for patients (pts) with metastatic (m)TNBC PD-L1+ disease. Most of these pts progress within a year. In a previous study, D (anti-PD-L1) combined with chemotherapy enhanced antitumor immune responses in early TNBC (Loibl. Ann Oncol 2019). In TNBC, activation of the PI3K/AKT/PTEN pathway and high CD73 expression are common. BEGONIA is an ongoing 2-part, multicenter, multi-arm, open-label platform study, evaluating safety and efficacy of D or D+P combined with novel therapies as first-line treatment for mTNBC (NCT03742102). Preliminary results from 2 arms were presented at ASCO 2021 (Abstract #1023). Here, we report results from Arm 1 D+P, Arm 2 D+P+C, and Arm 5 D+P+O. C is an oral, selective, ATP-competitive catalytic inhibitor of all 3 AKT isoforms, and O is a mAb targeting CD73. Methods: Eligible pts had untreated, unresectable, locally advanced or metastatic TNBC. In Arms 1 and 5, pts received D 1500 mg IV Q4W + P 90 mg/m2 IV day (d)1, d8, d15 of every cycle. Pts in Arm 5 also received O 3000 mg IV on d1 and d15 for the first 2 cycles, then Q4W. In Arm 2, pts received D 1500 mg IV Q4W + P 80/90 mg/m2 IV in 4-week cycles (d1, d8, d15, 1 week off) + C 400 mg BID in 4-week cycles (d2-5 × 3 weeks, 1 week off). Primary objectives were safety and tolerability. Secondary endpoints included objective response rate (ORR) and duration of response. Tumors were assessed Q8W per RECIST v1.1. The first 6 pts treated in Arms 2 and 5 were evaluated for dose-limiting toxicities (DLTs), with additional pts enrolled if treatment was tolerated. PD-L1 expression was assessed retrospectively. Previously presented data from Arm 1 D+P, are included for reference (Schmid. ASCO 2021, #1023). Results: In Arm 2 (data cutoff Mar 2021), 30 pts received D+P+C (15 P[80], 15 P[90]; total 13 ongoing); 2 pts (6.7%) discontinued all treatment due to AEs. The rates of dose delays were 13 pts (43%) for D and 15 (50%) for P; dose interruptions were 1 (3%) for D, 12 (40%) for P, 15 (50%) for C; dose reductions were 12 (40%) for P and 14 (47%) for C. Treatment-related (tr)SAEs and G3/4 trAEs were experienced by 7 (23%) and 22 (73%) pts. In Arm 5 (data cutoff Sep 2020), 33 pts received D+P+O (14 ongoing); no pts discontinued due to AEs. The rates of dose delays were 13 pts (39%) for D, 10 (30%) for P, 10 (30%) for O; dose interruptions were 2 (6%) for D, 10 (30%) for P, 3 (9%) for O; and dose reductions were 12 (36%) for P. trSAEs and G3/4 trAEs were experienced by 1 (3%) and 5 (15%) pts. In both arms, there were no DLTs or deaths due to AEs. The Table presents follow-up time and efficacy outcomes for Arms 1, 2, and 5. Responses were observed regardless of PD-L1 expression. The potential value of mutations in the PI3K pathway and CD73 expression as predictive biomarkers will be discussed for Arms 2 and 5, respectively. Updated data for Arm 1 will be presented. Conclusions: The safety profiles of triplet combinations in Arms 2 and 5 were consistent with the individual agents; however, in Arm 2, there was a relatively high rate of G3/4 trAEs but a low discontinuation rate for AEs. Although BEGONIA was not designed to compare activity across arms and numbers were small, the ORR of each triplet therapy was numerically similar to D+P. Biomarker analysis may elucidate pts that benefit from the combination of C or O with D+P. Funding: AstraZeneca Table. Efficacy outcomes in Arms 1, 2, and 5 of BEGONIAArm 1Arm 2Arm 5D+P N=23D+P(80)+C n=15D+P(90)+C n=15All D+P+C N=30D+P+O N=33Duration of follow-up at data cutoff, months, median (range)16.6 (8.5-19.8)6.7 (2-9)16.8 (6-21)8.2 (2-21)8.6 (4.1-14.6)Confirmed ORR, n (%)13 (56.5)8 (53.3)8 (53.3)16 (53.3)15 (45.5)95% CI34.5-76.826.6-78.726.6-78.7NC28.1-63.3CR, n10111PR12871514SD (Unconfirmed PR)7 (3)6 (2)4 (2)10 (4)13 (4)PD31345Percentage with ongoing response at data cutoff53.8%75.0%25.0%50.0%66.7%Arm 1 data cutoff was Sep 2020. C, capivasertib; CI, confidence interval; CR, complete response; D, durvalumab; NC, not calculable; O, oleclumab; ORR, objective response rate; P, paclitaxel; PD, progressive disease; PR, partial response; SD, stable disease. Citation Format: Peter Schmid, Zbigniew Nowecki, Seock-Ah Im, Wei-Pang Chung, Simon Lord, Anne Armstrong, Cynthia X Ma, Robert Huisden, Ross Stewart, Rakesh Kumar, Gaia Schiavon, Hannah Dry, Ana Nunes, Kyung Hae Jung, Yeon Hee Park. BEGONIA: Phase 1b/2 study of durvalumab (D) combinations in locally advanced/metastatic triple-negative breast cancer (TNBC): Results from Arm 1 D + paclitaxel (P), Arm 2 D+P + capivasertib (C), and Arm 5 D+P + oleclumab (O) [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr PD10-03.