Development of a novel risk signature revealing prognostic and tumor microenvironmental features in breast cancer.

This study aimed to devise a breast cancer (BC) risk signature for based on pyrimidine metabolism-related genes (PMRGs) to evaluate its prognostic value and association with drug sensitivity. Transcriptomic and clinical data were retrieved from The Cancer Genome Atlas database and Gene Expression Omnibus repository. Pyrimidine metabolism-associated genes were identified from the Molecular Signatures Database collection. A risk signature was constructed through Cox regression and Lasso regression methods. Further, the relationship between the PMRG-derived risk feature and clinicopathological characteristics, gene expression patterns, somatic mutations, drug susceptibility, and tumor immune microenvironment was thoroughly investigated, culminating in the development of a nomogram. PMRGs displayed differential expression and diverse somatic mutations in BC. Univariate Cox analysis identified 36 genes significantly associated with BC prognosis, leading to the categorization of 2 BC molecular subtypes with discernible differences in prognosis. Using Lasso Cox regression, a risk signature composed of 16 PMRGs was established, wherein high-risk scores were indicative of poor prognosis. The PMRG-derived risk feature was also related to chemotherapy regimens and showed significant correlations with sensitivity to multiple drugs. Furthermore, distinct tumor immune microenvironment properties, gene expression profiles, and somatic mutation patterns were evident across varying risk scores. Ultimately, a nomogram was constructed incorporating the PMRGs-based risk signature alongside stage, and chemotherapy status, demonstrating excellent performance in prognosis prediction. We successfully developed a PMRG-based BC risk signature that effectively combines with clinicopathological attributes for accurate prognosis assessment in BC.