Episode 22: Neoadjuvant and adjuvant pembrolizumab in early-stage high-risk ER+/HER2- breast cancer: KEYNOTE-756

In this episode of WND Breast, Dr Kevin Punie from GZA Hospitals and Prof Evandro de Azambuja from Institut Jules Bordet discuss the KEYNOTE-756 trial.

This global phase 3 trial, employing a double-blind, placebo-controlled design, investigated the efficacy of adding pembrolizumab to neoadjuvant therapy, continuing in the adjuvant setting, in high-risk breast cancer patients.

The study enrolled patients with invasive ductal breast carcinoma, T1c to T2 disease with one to two positive lymph nodes, or T3 to T4 disease with zero to two positive lymph nodes. All patients had centrally confirmed ER-positive, HER2-negative, grade 3 disease and were treatment-naive. Dose-dense chemotherapy was permitted, allowing anthracycline chemotherapy administration every 2 or 3 weeks. Pembrolizumab or placebo was administered every 3 weeks.

The study showed an overall 8% increase in pCR, particularly pronounced in patients with lymph node-positive disease. However, the benefit was minimal in patients without lymph node involvement, albeit from a small subgroup, cautioning against overemphasis on this result.

Biomarker analysis revealed a positive correlation between PD-L1 expression and CPS on pCR outcome with the addition of pembrolizumab, while ER expression showed an inverse association. Patients with PD-L1-negative disease experienced minimal benefit from pembrolizumab in terms of pCR. This study supports the approach of treating patients with grade 3 tumours and ER-low (<10%) PD-L1 positive expression as triple-negative breast cancer.

While these findings are significant, they are preliminary and do not yet correlate with improved cure. Further efficacy data are needed to inform the decision to add a checkpoint inhibitor.

The KEYNOTE-756 study is powered for EFS, so assessing whether the 8.5% pCR difference translates into improved survival, particularly across different subgroups, is crucial. Biomarker analysis will aid in optimizing patient selection, given the heterogeneity of this disease, once robust long-term data are available.