To investigate the expression characteristics of myeloid-derived suppressor cell (MDSC)-related genes in breast cancer, construct a risk prediction model, and evaluate its risk stratification and predictive performance.

RNA-seq data and clinical information of 1216 breast cancer patients were downloaded from the TCGA database, and randomly divided into a training set (n=851) and a validation set (n=365) at a ratio of 7∶3. A total of 494 MDSC-related genes were retrieved from the GeneCards database. In the training set, univariate Cox regression (P<0.05) was used to preliminarily screen prognosis-related genes, followed by LASSO-Cox regression for variable compression, and a multi-gene risk score (Riskscore) model was finally established. Kaplan-Meier curves and time-dependent ROC curves were adopted to assess the risk stratification ability and predictive efficacy of the model. Independent prognostic factors were integrated to construct a nomogram for predicting 2-, 3- and 5-year survival rates, which was comprehensively validated via calibration curves, decision curve analysis (DCA) and the C-index.

Forty-nine hub genes were screened out by LASSO regression to establish the Riskscore model. The model exhibited favorable predictive power in both the training and validation sets (3-year and 5-year AUC values of 0.785 and 0.787 in the training set, versus 0.691 and 0.765 in the validation set). Multivariate Cox regression analysis verified that Riskscore was an independent prognostic factor beyond conventional clinical indicators including age and tumor stage (HR=3.552, 95% CI: 2.933~4.301, P<0.001). The nomogram built on Riskscore, age and tumor stage achieved excellent predictive accuracy (5-year AUC=0.814 in the training set and 0.810 in the validation set) and satisfactory calibration. Decision curve analysis further confirmed its prominent clinical application value.

A prognostic prediction model based on MDSC-related genes was constructed and validated. This model can effectively stratify survival risks among breast cancer patients. As a visual assessment tool, the established nomogram enables clinicians to deliver individualized prognostic evaluation and facilitate clinical treatment decision-making.