Immunogenic cell death-related classification reveals prognosis and effectiveness of immunotherapy in breast cancer.

Lack of specific biomarkers and effective drug targets constrains therapeutic research in breast cancer (BC). In this regard, therapeutic modulation of damage-associated molecular patterns (DAMPs)-induced immunogenic cell death (ICD) may help improve the effect of immunotherapy in individuals with BC. The aim of this investigation was to develop biomarkers for ICD and to construct ICD-related risk estimation models to predict prognosis and immunotherapy outcomes of BC. RNA-seq transcriptome information and medical data from individuals with BC (n = 943) were obtained from TCGA. Expression data from a separate BC cohort (GEO: GSE20685) were used for validation. We identified subtypes of high and low ICD gene expression by consensus clustering and assessed the connection between ICD subtypes and tumor microenvironment (TME). In addition, different algorithms were used to construct ICD-based prognostic models of BC. BC samples were categorized into subtypes of high and low ICD expression depending on the expression of genes correlated with ICD. The subtype of ICD high-expression subtypes are correlated with poor prognosis in breast cancer, while ICD low-expression subtypes may predict better clinical outcomes. We also created and verified a predictive signature model depending on four ICD-related genes (ATG5, CD8A, CD8B, and HSP90AA1), which correlates with TME status and predicts clinical outcomes of BC patients. We highlight the connection of ICD subtypes with the dynamic evolution of TME in BC and present a novel ICD-based prognostic model of BC. In clinical practice, distinction of ICD subtype and assessment of ICD-related biomarkers should help guide treatment planning and improve the effectiveness of tumor immunotherapy.