LBA1 Final analysis of the placebo-controlled randomised phase III IMpassion031 trial evaluating neoadjuvant atezolizumab (atezo) plus chemotherapy (CT) followed by open-label adjuvant atezo in patients (pts) with early-stage triple-negative breast cancer (eTNBC)

IMpassion031 met its primary objective, significantly improving pathological complete response (pCR) rate with atezo added to neoadjuvant CT in pts with eTNBC (58% vs 41% with CT alone in the intent-to-treat [ITT] population; 17% difference, 1-sided p=0.0044) [Mittendorf 2020]. Benefit from atezo was seen regardless of PD-L1 status. Pts with untreated stage II/III eTNBC and a primary tumour >2 cm were randomised 1:1 to placebo or atezo 840 mg q2w + neoadjuvant CT (nab-paclitaxel 125 mg/m2 qw for 12 wk, then doxorubicin 60 mg/m2 + cyclophosphamide 600 mg/m2 q2w for 8 wk). After surgery, pts randomised to atezo received open-label atezo 1200 mg q3w for 11 cycles. Pts without pCR could have standard adjuvant systemic therapy (± atezo). Stratification factors were disease stage (II vs III) and PD-L1 status (IC <1% vs ≥1% by SP142). Secondary efficacy endpoints included event-free survival (EFS), disease-free survival (DFS) and overall survival (OS) in the ITT and PD-L1+ populations. After ∼39 mos' median follow-up, EFS, DFS and OS numerically favoured atezo consistently across key clinical subgroups (Table). Among pts without pCR, 14/70 (20%) in the atezo arm vs 33/99 (33%) in the placebo arm received additional adjuvant systemic therapy, most often capecitabine (4/70 [6%] vs 26/99 [26%]). In exploratory analyses, pCR was highly prognostic for long-term outcome. Exploratory ctDNA analyses showed clearance in most pts (including all ctDNA-evaluable pts with pCR) at surgery. Positive ctDNA at/after surgery was associated with poor prognosis. There were no new safety signals or treatment-related deaths with atezo.Table: LBA1EndpointITTPD-L1+aPD-L1-expressing tumour-infiltrating immune cells on ≥1% tumour area by SP142.PlaceboAtezoPlaceboAtezoEFSEvents, n/N (%)41/168 (24)31/165 (19)17/75 (23)11/77 (14)HRbStratified.0.76 (0.47–1.21)0.55 (0.26–1.18)1 y, %91 (86–95)95 (92–98)89 (82–96)96 (92–100)2 y, %80 (74–86)85 (79–90)80 (71–89)89 (82–96)DFScIn operated pts. HR = hazard ratio.Events, n/N (%)31/153 (20)24/155 (15)11/67 (16)8/73 (11)HRbStratified.0.76 (0.44–1.30)0.57 (0.23–1.43)2 y, %83 (77–89)87 (82–93)87 (79–95)91 (85–98)OSEvents, n/N (%)28/168 (17)16/165 (10)9/75 (12)7/77 (9)HRbStratified.0.56 (0.30–1.04)0.71 (0.26–1.91)2 y, %90 (85–95)95 (91–98)91 (84–98)95 (89–100)Ranges in brackets are 95% CIs.a PD-L1-expressing tumour-infiltrating immune cells on ≥1% tumour area by SP142.b Stratified.c In operated pts. HR = hazard ratio. Open table in a new tab Ranges in brackets are 95% CIs. The significant pCR benefit with the addition of atezo to CT for eTNBC translated into numerically improved EFS, DFS and OS. Long-term safety results are consistent with previous reports.