To construct a gene prediction model that can effectively predict the response to neoadjuvant chemotherapy combined with targeted therapy in human epidermal growth factor receptor 2 (HER2)-positive breast cancer patients.

The datasets from patients with HER2-positive breast cancer undergoing neoadjuvant chemotherapy combined with targeted therapy were integrated from the GEO database. Differential gene analysis, WGCNA analysis, and LASSO regression were used to screen key genes and construct a predictive model. Potential associated pathways and mechanisms were explored using GSEA, immune cell infiltration levels, immune scoring, and drug sensitivity assessments.

Nine key genes (CD8A, CST7, CXCL13, IL2RB, LDHB, RTN3, SEMA4D, EPB41L1, MYO5C) were identified, and the constructed predictive model demonstrated good predictive performance in both the training set (AUC=0.836) and the validation set (AUC=0.827). The low-risk group (sensitive to treatment) exhibited characteristics of immune activation, while the high-risk group (resistant to treatment) was relatively in an immunosuppressive state.

The gene model constructed in this study can effectively predict the response to neoadjuvant therapy in HER2-positive breast cancer, with the activation status of the immune microenvironment closely related to treatment sensitivity. The key genes provide potential biomarkers and intervention targets for personalized treatment.