Objective We aimed to develop a nomogram in corporating multidetector computed tomography (MDCT) imaging features and clinicopathological indicators for the preoperative prediction of axillary lymph node metastasis (ALNM) in patients with triple-negative breast cancer (TNBC).

Methods We retrospectively analyzed data from 265 female patients with pathologically confirmed TNBC treated at Harbin Medical University Cancer Hospital between November 2020 and October 2024. Patients were randomly assigned into a training cohort (n = 161) and a validation cohort (n = 104) in a 6:4 ratio. Feature selection was performed using least absolute shrinkage and selection operator (LASSO) regression with 10-fold cross-validation. Independent predictors of ALNM were identified by multivariate Logistic regression analysis, and a nomogram was constructed accordingly. Model performance was assessed using receiver operating characteristic (ROC) curves, calibration plots, and decision curve analysis (DCA).

Results Three independent predictors of ALNM were identified: clinical N-stage (odds ratio OR = 6.789; 95% confidence interval CI: 2.203-22.20; P = 0.001), short-axis diameter of lymph nodes on CT (OR = 1.686; 95% CI: 1.349-2.257; P< 0.001), and cortical thickness (OR=6.296; 95% CI: 2.170-19.310; P=0.001). The nomogram showed strong discrimination, with areas under the ROC curve (AUC) of 0.918 (95% CI: 0.860-0.977) in the training cohort and 0.885 (95% CI: 0.809-0.962) in the validation cohort. Calibration was confirmed by Hosmer–Lemeshow tests (P=0.609 and P=0.694 for training and validation cohorts, respectively). DCA demonstrated clinical utility across probability thresholds of 0.02-0.96 and 0.03-0.87 in the training and validation cohorts, respectively.

Conclusions This nomogram, integrating MDCT imaging features and clinical indicators, provides a practical tool for individualized preoperative risk assessment and may aid clinical decision-making in patients with TNBC.