Safety interim analysis (SIA) of atractib: A phase 2 trial of first-line (1L) atezolizumab (A) in combination with paclitaxel (P) and bevacizumab (B) in metastatic triple-negative breast cancer (mTNBC).

1084 Background: A substantial benefit from adding an immune checkpoint inhibitor to chemotherapy (CT) was reported in mTNBC patients (pts) with PD-L1+ tumors. However, many pts still have a poor outcome. ATRACTIB is exploring the synergism between A (anti-PD-L1 antibody) and B (a VEGF-targeted antibody) with P in mTNBC irrespective of PD-L1 status. We report results from protocol-specified SIA. Methods: ATRACTIB is an open-label, single-arm, phase 2 trial (NCT04408118). Pts aged ≥18 years, with unresectable locally advanced or mTNBC, ECOG performance status of 0–1, who had received no prior systemic therapy or ≥12 months since (neo)adjuvant taxane-based CT are eligible. Pts receive A (840 mg IV, days 1, 15) with P (90 mg/m 2 IV, days 1, 8, 15), and B (10 mg/kg IV, days 1, 15) on each 28-day cycle until disease progression, unacceptable toxicity, or patient withdrawal. Primary endpoint is investigator-assessed progression-free survival (PFS) as per RECIST v.1.1. Secondary endpoints include objective response and clinical benefit rates, overall survival, and safety. The trial was designed to detect a treatment effect in terms of median PFS (H 0 : ≤7 months; H 1 : ≥9.5 months) and 100 pts are needed to attain 80% power at a nominal one-sided α level of 5%. One SIA was planned for evaluating safety as per CTCAE v.5.0 on the first 20 pts who had completed a 3-month follow-up or reached the end of study. Results: From Oct 5, 2020, through Nov 21, 2021, 34 pts were enrolled at 13 sites in Spain and Germany and received at least 1 dose of study treatment. Median age was 57.5 (range 40–84) years, 23 (67.6%) pts had received prior CT for early disease, and 19 (56.0%) had visceral disease. At data cutoff (Sep 30, 2021), 25 (71.4%) pts were still receiving the drug regimen. Adverse events (AEs) led to drug discontinuation in 3 (8.8%) pts. Mean relative dose intensity was 90.2% for A, 96.5% for P, and 95.7% for B. P dose reduction was reported in 7 (20.6%) pts. Five (14.7) pts required a dose delay due to AEs (11.8% for A, 11.8% for P, and 8.8% for B). The most common AEs of any grade (G) were fatigue (47.1%; 8.8% G≥3), diarrhea (38.2%; 0% G≥3), and neurotoxicity (35.3%; 8.8% G≥3). Anemia (20.6%; 0% G≥3) and neutropenia (17.6%; 8.8% G≥3) were the most frequent hematological AEs. AEs of clinical interest (AECI) for B were hypertension (17.6%; 5.9% G≥3) and pulmonary embolism (2.9%; 0% G≥3). AECI for A were pneumonitis (2.9%; 0% G≥3), autoimmune hepatitis (2.9%; 2.9% G≥3), and alanine aminotransferase increased (2.9%; 2.9% G≥3). No treatment-related deaths were reported. Conclusions: The addition of A to P and B as 1L therapy for mTNBC shows a tolerable safety profile which is consistent with known safety profile of each agent without a significant synergistic toxicity. Based on the independent data monitoring committee recommendation, patient recruitment is ongoing. Clinical trial information: NCT04408118.