<h2>Summary</h2><h3>Background</h3> Preferred neoadjuvant regimens for early-stage triple-negative breast cancer (TNBC) include anthracycline-cyclophosphamide and taxane-based chemotherapy. IMpassion031 compared efficacy and safety of atezolizumab versus placebo combined with nab-paclitaxel followed by doxorubicin plus cyclophosphamide as neoadjuvant treatment for early-stage TNBC. <h3>Methods</h3> This double-blind, randomised, phase 3 study enrolled patients in 75 academic and community sites in 13 countries. Patients aged 18 years or older with previously untreated stage II–III histologically documented TNBC were randomly assigned (1:1) to receive chemotherapy plus intravenous atezolizumab at 840 mg or placebo every 2 weeks. Chemotherapy comprised of nab-paclitaxel at 125 mg/m<sup>2</sup> every week for 12 weeks followed by doxorubicin at 60 mg/m<sup>2</sup> and cyclophosphamide at 600 mg/m<sup>2</sup> every 2 weeks for 8 weeks, which was then followed by surgery. Stratification was by clinical breast cancer stage and programmed cell death ligand 1 (PD-L1) status. Co-primary endpoints were pathological complete response in all-randomised (ie, all randomly assigned patients in the intention-to-treat population) and PD-L1-positive (ie, patients with PD-L1-expressing tumour infiltrating immune cells covering ≥1% of tumour area) populations. This study is registered with ClinicalTrials.gov (NCT03197935), Eudra (CT2016-004734-22), and the Japan Pharmaceutical Information Center (JapicCTI-173630), and is ongoing. <h3>Findings</h3> Between July 7, 2017, and Sept 24, 2019, 455 patients were recruited and assessed for eligibility. Of the 333 eligible patients, 165 were randomly assigned to receive atezolizumab plus chemotherapy and 168 to placebo plus chemotherapy. At data cutoff (April 3, 2020), median follow-up was 20·6 months (IQR 8·7–24·9) in the atezolizumab plus chemotherapy group and 19·8 months (8·1–24·5) in the placebo plus chemotherapy group. Pathological complete response was documented in 95 (58%, 95% CI 50–65) patients in the atezolizumab plus chemotherapy group and 69 (41%, 34–49) patients in the placebo plus chemotherapy group (rate difference 17%, 95% CI 6–27; one-sided p=0·0044 [significance boundary 0·0184]). In the PD-L1-positive population, pathological complete response was documented in 53 (69%, 95% CI 57–79) of 77 patients in the atezolizumab plus chemotherapy group versus 37 (49%, 38–61) of 75 patients in the placebo plus chemotherapy group (rate difference 20%, 95% CI 4–35; one-sided p=0·021 [significance boundary 0·0184]). In the neoadjuvant phase, grade 3–4 adverse events were balanced and treatment-related serious adverse events occurred in 37 (23%) and 26 (16%) patients, with one patient per group experiencing an unrelated grade 5 adverse event (traffic accident in the atezolizumab plus chemotherapy group and pneumonia in the placebo plus chemotherapy group). <h3>Interpretation</h3> In patients with early-stage TNBC, neoadjuvant treatment with atezolizumab in combination with nab-paclitaxel and anthracycline-based chemotherapy significantly improved pathological complete response rates with an acceptable safety profile. <h3>Funding</h3> F Hoffmann-La Roche/Genentech.

Neoadjuvant atezolizumab in combination with sequential nab-paclitaxel and anthracycline-based chemotherapy versus placebo and chemotherapy in patients with early-stage triple-negative breast cancer (IMpassion031): a randomised, double-blind, phase 3 trial