•The phase III IMpassion131 trial evaluated atezolizumab combined with paclitaxel as first-line therapy for aTNBC.•The primary endpoint was investigator-assessed PFS, tested hierarchically in the PD-L1+ and then ITT populations.•Neither PFS nor OS was improved with the combination of atezolizumab plus paclitaxel in either population.•These findings may result from imbalances in prognostic features or chance findings in a relatively small trial.•IMpassion131 results highlight the need for further research into immunotherapy for TNBC. BackgroundIn the phase III IMpassion130 trial, combining atezolizumab with first-line nanoparticle albumin-bound-paclitaxel for advanced triple-negative breast cancer (aTNBC) showed a statistically significant progression-free survival (PFS) benefit in the intention-to-treat (ITT) and programmed death-ligand 1 (PD-L1)-positive populations, and a clinically meaningful overall survival (OS) effect in PD-L1-positive aTNBC. The phase III KEYNOTE-355 trial adding pembrolizumab to chemotherapy for aTNBC showed similar PFS effects. IMpassion131 evaluated first-line atezolizumab–paclitaxel in aTNBC.Patients and methodsEligible patients [no prior systemic therapy or ≥12 months since (neo)adjuvant chemotherapy] were randomised 2:1 to atezolizumab 840 mg or placebo (days 1, 15), both with paclitaxel 90 mg/m2 (days 1, 8, 15), every 28 days until disease progression or unacceptable toxicity. Stratification factors were tumour PD-L1 status, prior taxane, liver metastases and geographical region. The primary endpoint was investigator-assessed PFS, tested hierarchically first in the PD-L1-positive [immune cell expression ≥1%, VENTANA PD-L1 (SP142) assay] population, and then in the ITT population. OS was a secondary endpoint.ResultsOf 651 randomised patients, 45% had PD-L1-positive aTNBC. At the primary PFS analysis, adding atezolizumab to paclitaxel did not improve investigator-assessed PFS in the PD-L1-positive population [hazard ratio (HR) 0.82, 95% confidence interval (CI) 0.60-1.12; P = 0.20; median PFS 6.0 months with atezolizumab–paclitaxel versus 5.7 months with placebo–paclitaxel]. In the PD-L1-positive population, atezolizumab–paclitaxel was associated with more favourable unconfirmed best overall response rate (63% versus 55% with placebo–paclitaxel) and median duration of response (7.2 versus 5.5 months, respectively). Final OS results showed no difference between arms (HR 1.11, 95% CI 0.76-1.64; median 22.1 months with atezolizumab–paclitaxel versus 28.3 months with placebo–paclitaxel in the PD-L1-positive population). Results in the ITT population were consistent with the PD-L1-positive population. The safety profile was consistent with known effects of each study drug.ConclusionCombining atezolizumab with paclitaxel did not improve PFS or OS versus paclitaxel alone.ClinicalTrials.govNCT03125902. In the phase III IMpassion130 trial, combining atezolizumab with first-line nanoparticle albumin-bound-paclitaxel for advanced triple-negative breast cancer (aTNBC) showed a statistically significant progression-free survival (PFS) benefit in the intention-to-treat (ITT) and programmed death-ligand 1 (PD-L1)-positive populations, and a clinically meaningful overall survival (OS) effect in PD-L1-positive aTNBC. The phase III KEYNOTE-355 trial adding pembrolizumab to chemotherapy for aTNBC showed similar PFS effects. IMpassion131 evaluated first-line atezolizumab–paclitaxel in aTNBC. Eligible patients [no prior systemic therapy or ≥12 months since (neo)adjuvant chemotherapy] were randomised 2:1 to atezolizumab 840 mg or placebo (days 1, 15), both with paclitaxel 90 mg/m2 (days 1, 8, 15), every 28 days until disease progression or unacceptable toxicity. Stratification factors were tumour PD-L1 status, prior taxane, liver metastases and geographical region. The primary endpoint was investigator-assessed PFS, tested hierarchically first in the PD-L1-positive [immune cell expression ≥1%, VENTANA PD-L1 (SP142) assay] population, and then in the ITT population. OS was a secondary endpoint. Of 651 randomised patients, 45% had PD-L1-positive aTNBC. At the primary PFS analysis, adding atezolizumab to paclitaxel did not improve investigator-assessed PFS in the PD-L1-positive population [hazard ratio (HR) 0.82, 95% confidence interval (CI) 0.60-1.12; P = 0.20; median PFS 6.0 months with atezolizumab–paclitaxel versus 5.7 months with placebo–paclitaxel]. In the PD-L1-positive population, atezolizumab–paclitaxel was associated with more favourable unconfirmed best overall response rate (63% versus 55% with placebo–paclitaxel) and median duration of response (7.2 versus 5.5 months, respectively). Final OS results showed no difference between arms (HR 1.11, 95% CI 0.76-1.64; median 22.1 months with atezolizumab–paclitaxel versus 28.3 months with placebo–paclitaxel in the PD-L1-positive population). Results in the ITT population were consistent with the PD-L1-positive population. The safety profile was consistent with known effects of each study drug. Combining atezolizumab with paclitaxel did not improve PFS or OS versus paclitaxel alone.
