Development and External Validation of a Prognostic Nomogram for Overall Survival in Male Breast Cancer Using SEER and Multicenter Data

Male breast cancer (MBC) represents an uncommonly diagnosed condition. The infrequency of its occurrence has led to a scarcity of investigations examining prognostic outcomes in this population. The present work aimed to develop a nomogram to estimate overall survival in MBC patients, followed by external validation in a Chinese patient sample. Drawing upon records within the Surveillance, Epidemiology, and End Results (SEER) database, all male subjects who first received a breast cancer diagnosis within the window spanning January 2010 to December 2015 were eligible for inclusion. Through random allocation following a 7:3 split, these individuals formed a training subset (n = 1610) and an internal validation subset (n = 713). An external corroboration cohort was assembled from 22 MBC patients whose diagnoses occurred at the First Affiliated Hospital of Guangxi Medical University across the period from January 2013 through June 2021, applying June 10, 2023, as the conclusion of follow-up observation. Application of the Cox regression methodology permitted identification of statistically meaningful risk determinants, which subsequently informed construction of a nomogram projecting overall survival for MBC patients. The resultant model was tested using data from the test subset. Evaluations of discriminative capacity and dependability incorporated the concordance index (C-index), receiver operating characteristic (ROC) curves, decision curve analysis (DCA), and Kaplan-Meier survival plotting. The analytical sample comprised 2301 MBC cases extracted from the SEER database, along with 22 MBC cases contributed by the participating hospital facility. Seven independent predictors were retained within the finalized model: patient age (hazard ratio [HR] = 1.89, 95% CI: 1.50‐2.38), receipt of surgical intervention (HR = 0.38, 95% CI: 0.29‐0.51), marital circumstance (HR = 0.75, 95% CI: 0.63‐0.89), tumor classification (HR = 1.17, 95% CI: 1.05‐1.29), clinical disease stage (HR = 1.41, 95% CI: 1.15‐1.74), administration of chemotherapy (HR = 0.62, 95% CI: 0.50‐0.75), and HER2 receptor expression (HR = 2.68, 95% CI: 1.20‐5.98). For the training, internal validation, and external validation subsets, C-index measurements reached 0.72, 0.747, and 0.981, respectively. Satisfactory calibration properties were evident in the nomogram, and ROC analysis demonstrated the model’s superior clinical validity. Results from DCA suggested favorable applicability within clinical settings. Additionally, the risk classification scheme derived from the nomogram permitted sharper demarcation among prognostic subgroups, with individuals stratified into the low-risk tier experiencing considerably more favorable survival trajectories than their medium- and high-risk counterparts (P < .001). Through this investigation, a prognostic nomogram incorporating 7 variables was successfully developed to predict survival in MBC patients. The resultant predictive instrument can estimate survival endpoints for affected individuals and provides an evidence-based reference point for diagnostic and therapeutic clinical judgment.