PD-L1 Expression as a Predictor of Response to Neoadjuvant Immunotherapy in Triple-Negative Breast Cancer

Resumen

Background: Immune checkpoint blockade has changed the treatment of high-risk early-stage triple-negative breast cancer (TNBC), yet the clinical role of programmed death-ligand 1 (PD-L1) expression in the neoadjuvant setting remains unsettled. PD-L1 may identify an immune-enriched tumor microenvironment and therefore correlate with pathologic complete response (pCR), but a predictive biomarker must also demonstrate differential treatment benefit according to biomarker status.

Methods: A structured systematic search of PubMed/MEDLINE and trial-specific publisher records was performed through 31 July 2026, supplemented by backward citation searching of major randomized trials, regulatory summaries, and contemporary meta-analyses. Prospective neoadjuvant PD-1/PD-L1 inhibitor studies in early TNBC were included when pretreatment PD-L1 was assessed and response data were reported. Owing to heterogeneity in PD-L1 assays, cutoffs, chemotherapy backbones, and study designs, findings were synthesized narratively rather than pooled quantitatively.

Results: Nine prospective primary studies were identified, including four pivotal randomized chemoimmunotherapy trials and five biomarker-rich early-phase or single-arm studies. Across KEYNOTE-173, KEYNOTE-522, GeparNuevo, NeoPACT, and camrelizumab studies, higher PD-L1 expression was generally associated with higher absolute pCR probability. However, randomized evidence did not show a reproducible PD-L1-by-treatment interaction. In KEYNOTE-522, pCR improved with pembrolizumab in both PD-L1-positive (68.9% vs 54.9%) and PD-L1-negative tumors (45.3% vs 30.3%). IMpassion031 similarly showed higher pCR with atezolizumab in PD-L1- positive (69% vs 49%) and PD-L1-negative disease (48% vs 34%). NeoTRIP identified PD-L1 as a strong correlate of pCR, yet atezolizumab did not significantly improve pCR overall. Assay discordance between 22C3, SP142, and SP263 and variable use of CPS versus immune-cell scoring further constrain interchangeability.

Conclusions: Pretreatment PD-L1 expression is best regarded as a response-enrichment and immune-context biomarker in early TNBC, not a validated stand-alone selector for neoadjuvant checkpoint blockade. Current evidence does not support withholding standard perioperative pembrolizumab solely because a tumor is PD-L1 negative. Composite biomarkers integrating PD-L1 with tumor-infiltrating lymphocytes, immune gene-expression signatures, spatial features, and early on-treatment dynamics are more promising for treatment personalization.