Triple-negative breast cancer (TNBC) is an aggressive subtype with a relatively immunogenic tumor microenvironment that may respond to immune checkpoint inhibitors (ICIs). The multicenter, multinational, double-blind NSABP B-59/GeparDouze phase 3 trial evaluated whether adding the anti–PD-L1 antibody atezolizumab to contemporary taxane–carboplatin–anthracycline neoadjuvant chemotherapy would improve outcomes for patients with stage II–III, centrally confirmed TNBC.
Prior neoadjuvant studies in TNBC reported increases in pathologic complete response (pCR) rates when ICIs were added to chemotherapy, and KEYNOTE-522 demonstrated EFS and OS benefits with pembrolizumab. Atezolizumab had previously shown activity in metastatic PD-L1–positive TNBC and improved pCR in early-stage studies, motivating this randomized trial.
NSABP B-59/GeparDouze (ClinicalTrials.gov NCT03281954) was an international, double-blind, placebo-controlled, academic phase 3 study. Between 29 December 2017 and 28 May 2021, 2,188 patients were screened and 1,550 patients with centrally confirmed TNBC were enrolled across 353 sites in four countries (United States, Germany, Canada and Spain). Of these, 773 patients were randomized to receive neoadjuvant atezolizumab and 777 to placebo, administered every 3 weeks concurrently with neoadjuvant chemotherapy.
The trial included prospectively planned translational sub-studies incorporating tumor mRNA analysis and TNBC molecular subtyping, as well as evaluation of tumor-infiltrating lymphocytes (TILs) to explore predictive and prognostic biomarkers.
Neoadjuvant chemotherapy comprised weekly paclitaxel (80 mg m−2 IV × 12 doses) combined with carboplatin (AUC 5 IV on day 1 every 3 weeks for four cycles), followed by anthracycline–cyclophosphamide (either doxorubicin 60 mg m−2 or epirubicin 90 mg m−2 with cyclophosphamide 600 mg m−2) administered every 2 or 3 weeks for four cycles per investigator discretion. Patients received study drug (atezolizumab 1,200 mg IV or matched placebo) every 3 weeks during neoadjuvant therapy and continued adjuvant study drug to complete 1 year of therapy.
The primary endpoint was event-free survival (EFS) with a planned 4-year analysis. The addition of atezolizumab did not produce a statistically significant improvement in EFS: hazard ratio (HR) 0.80 (95% CI, 0.62–1.03); stratified log-rank P = 0.083. The absolute difference in 4-year EFS rates between arms was 3.3% in favor of atezolizumab.
Overall survival (OS) showed an HR of 0.86 (95% CI, 0.62–1.19) with a 4-year absolute benefit of 0.7% reported in the primary analysis. Secondary endpoints including pCR and distant disease-free survival (DDFS) were included in the planned analyses; the source article reports these were analyzed but detailed numeric results beyond the primary EFS and OS HRs were not provided in the available text.
Prespecified subgroup analyses indicated heterogeneity of treatment effect for EFS. Notably, a statistically significant interaction was observed for patients presenting with clinical lymph node involvement (P interaction = 0.039), suggesting that node-positive patients may have experienced greater EFS benefit from adding atezolizumab.
Treatment-emergent adverse events (AEs) of grade ≥3 occurred in 75.3% of patients receiving atezolizumab compared with 73.4% receiving placebo. Immune-related adverse events were reported in 27.6% of patients on atezolizumab versus 11.4% on placebo. Discontinuation during the neoadjuvant phase was more frequent with atezolizumab: 196 patients (25.5%) discontinued atezolizumab compared with 143 patients (18.8%) who discontinued placebo.
These safety findings underscore an increased frequency of immune-mediated toxicity with atezolizumab when added to an intensive multi-agent neoadjuvant chemotherapy backbone.
A prospectively planned translational sub-study integrated mRNA analysis and molecular TNBC subtyping of primary tumor samples. In an exploratory mRNA-based subset analysis, patients with basal-like immune-activated tumors appeared to derive benefit from atezolizumab. Similarly, quantification of tumor-infiltrating lymphocytes (TILs) identified tumors with high TIL counts that may be more likely to benefit from checkpoint inhibition.
The authors propose that combining TNBC molecular subtyping with TIL quantification could be a promising approach to select patients who will benefit from adding ICIs to neoadjuvant chemotherapy, although these findings were exploratory and require validation.
In the overall study population of stage II–III TNBC, adding atezolizumab to taxane–carboplatin–anthracycline neoadjuvant chemotherapy did not achieve a statistically significant improvement in the trial’s primary endpoint of 4-year EFS. Subgroup signals — notably in clinically node-positive patients — and exploratory biomarker analyses suggest there may be clinically relevant benefit in selected molecular and immune-activated subgroups.
These results emphasize the importance of identifying robust predictive biomarkers, such as TILs and mRNA-defined TNBC subtypes, to guide use of ICIs in early TNBC. Further validation of the exploratory findings is needed before routine selection of patients for atezolizumab based on these markers.
ClinicalTrials.gov registration: NCT03281954.