Chaoni Cai, Jialing Huang, Jinfeng Li, Saixiang Zheng, Dijin Lin, Chaohui Duan, Ying Xu
Peripheral T-cell activation profiles were associated with NAT response and clinical outcomes in this cohort, suggesting their potential as dynamic biomarkers for recurrence/metastasis risk stratification. However, given the retrospective, single-center design and lack of external validation, these internally validated models warrant independent prospective validation prior to clinical application.
PURPOSE: Peripheral blood T-cell monitoring may predict treatment response and recurrence in breast cancer patients receiving neoadjuvant therapy (NAT).
METHODS: The retrospective, single-center study enrolled 210 primary breast cancer patients (2020-2024). Peripheral T-cell subsets (CD69+, HLA-DR+, etc.) and tumor markers (CEA, CA125, etc.) were assessed at baseline, post-NAT, and post-surgery. Receiver operating characteristic (ROC) curves, Logistic and Cox regression, decision curve analysis, longitudinal trajectory analysis, and Gene Set Enrichment Analysis (GSEA) were performed.
RESULTS: Baseline CD3+CD69+T /CD3+T and CD3+HLA-DR+T /CD3+T ratios independently predicted pathological complete response (pCR) with a combined AUC of 0.87. Post-NAT Treg /CD4+T and CD3+HLA-DR+T /CD3+T predicted recurrence/metastasis (AUC = 0.85). Post-surgery, the combination of CD3+CD8+HLA-DR+T /CD3+T and CA125 achieved an AUC of 0.88. Post-treatment T-cell activation correlated with improved survival. Three-year trajectory analysis revealed distinct T-cell dynamics between no recurrence/metastasis and recurrence/metastasis patients. GSEA showed CD69-high CD8+T were enriched in inflammatory pathways, whereas HLA-DR-high CD8+T were enriched in cell-cycle, metabolism, and effector pathways.
CONCLUSIONS: Peripheral T-cell activation profiles were associated with NAT response and clinical outcomes in this cohort, suggesting their potential as dynamic biomarkers for recurrence/metastasis risk stratification. However, given the retrospective, single-center design and lack of external validation, these internally validated models warrant independent prospective validation prior to clinical application.