Integrative Transcriptomic Analysis Identifies a Novel 22-Gene Signature Driving Breast Cancer Progression via the Proteasome-Chemokine Axis

Abstract Breast cancer heterogeneity often limits traditional clinicopathological predictions. Using transcriptomic data from the SCAN-B cohort (N = 1,185), we developed a 22-gene prognostic signature via LASSO Cox regression. This signature successfully stratified patients into distinct risk groups, with the high-risk group showing significantly decreased overall survival ( p < 0.0001). Multivariate analysis confirmed the risk score as a robust independent predictor (HR = 3.44, 95% CI: 2.10–5.65, p < 0.001), surpassing traditional markers like age and ER/PR status. Mechanistically, the signature defines a "Proliferation-Immune Axis": risk correlates with proteasome hyperactivity and proliferation ( TK1, MARCO ), while survival advantage links to chemokine-driven immune recruitment ( CXCL13, TCRVB ). The model demonstrated high accuracy in predicting early recurrence (AUC = 0.847). When integrated into a calibrated nomogram, this signature provides a precise tool for capturing the proteasome-chemokine interplay, enhancing individualized risk assessment and precision oncology decisions in breast cancer.