Int J Mol Sci. 2026 Mar 06. pii: 2431. [Epub ahead of print]27(5):
Tumor-infiltrating lymphocytes (TILs) are an established predictor of pathological complete response (pCR) after neoadjuvant chemotherapy (NACT) in breast cancer. However, TILs primarily reflect the extent of immune infiltration and provide limited insight into immune functional state. We investigated whether integrating measures of immune infiltration (TILs), effector T-cell presence (CD8), and functional context (immune checkpoint components) may improve prediction of pCR beyond TILs alone. Pretreatment tumor biopsies from 166 patients with early breast cancer treated with standard NACT were assessed for stromal TILs and mRNA expression of CD8, PD-1, LAG-3, and TIM-3. Associations with pCR were evaluated using univariate and multivariable logistic regression, and composite immune phenotypes were constructed to capture functional immune states. In univariate analyses, higher TILs, CD8, PD-1, and LAG-3 were associated with pCR (all p < 0.05), whereas TIM-3 was not (p = 0.801). In multivariable models, TILs remained independently associated with pCR when adjusted for checkpoint markers, but this association was attenuated when CD8 was included, consistent with the strong biological correlation between TILs and CD8, and neither CD8 nor checkpoint markers retained independent significance. PD-1 and LAG-3 expression strongly correlated with CD8 and moderately correlated with TILs, indicating that checkpoint expression predominantly reflects an immune effector-engaged tumor microenvironment. Composite immune phenotypes based on CD8/PD-1 co-expression identified distinct immune functional states, with CD8-high/PD-1-high tumors demonstrating the highest pCR rates. Hierarchical modeling showed modest improvements in discrimination with sequential addition of immune variables to clinical predictors, with the integrative CD8/PD-1 model achieving the highest discrimination within the cohort (AUC = 0.849), although the magnitude of improvement beyond TIL assessment alone was limited. In conclusion, immune infiltration, effector T-cell presence, and functional immune context represent complementary dimensions for pCR prediction following NACT in breast cancer. However, TILs remain the most robust and clinically feasible immune biomarker.
Keywords: CD8-positive T cells; PD-1 immune checkpoint; breast cancer; immune microenvironment; neoadjuvant chemotherapy; pathological complete response; predictive biomarkers; tumor-infiltrating lymphocytes