<p>Immune infiltration in triple-negative breast cancer (TNBC) is associated with patient prognosis. However, the precise phenotypic classifications of immune infiltration in TNBC remain unclear, and the key regulatory mechanisms underlying this process require further investigation. In this study, we employed transcriptomic and single-cell data analysis, utilizing unsupervised clustering to classify the immune microenvironment of TNBC. We constructed a prognostic model based on immune phenotypes and further validated the regulatory role of key genes in immune cell modulation. Through unsupervised clustering analysis of the GEO and METABRIC databases, we identified two distinct immune infiltration phenotypes in TNBC tumor tissues. Notably, patients with high immune infiltration exhibited significantly prolonged overall survival (OS). Subsequent analysis of tumor tissues from high/low immune infiltration groups revealed five differentially expressed genes (DEGs), which were significantly correlated with patient prognosis: lymphotoxin β (LTB), interferon regulatory factor 8 (IRF8), indoleamine 2,3-dioxygenase 1 (IDO1), integral membrane protein 2A (ITM2A), and lymphocyte cytosolic protein 1 (LCP1). A prognostic model for TNBC was developed based on the five key genes. Furthermore, we demonstrated a significant association between LTB expression and enhanced immune effects in TNBC tissues. Single-cell sequencing data revealed that LTB expression was predominantly localized to immune-infiltrating T cells, which tended to differentiate toward cytotoxic T cells. We also confirmed that LTB expression improved T cell proliferation and enhanced the tumoricidal effect by modulating T cell differentiation pathways. In conclusion, immune infiltration phenotype is a crucial prognostic biomarker for TNBC patients. LTB contributes to enhanced immune infiltration, making as a promising target for immune therapy in TNBC.</p> Graphical abstract