Nomogram combining clinical and ultrasonographic features for differentiating granulomatous lobular mastitis from invasive lobular carcinoma.

OBJECTIVE: This study aims to develop and validate a nomogram model based on clinical and ultrasonographic features for differentiating granulomatous lobular mastitis (GLM) from invasive lobular carcinoma (ILC) of the breast in clinical practice and evaluate its diagnostic performance and clinical utility. METHODS: A total of 205 patients with pathologically confirmed GLM and 166 patients with ILC were retrospectively enrolled and randomly allocated to training and validation cohorts in a 7:3 ratio. Clinical data and ultrasonographic characteristics were collected. Univariate analysis was performed to compare variables between GLM and ILC groups. Variables with statistical significance were subsequently entered into multivariate logistic regression analysis to identify independent predictors. A visual nomogram model was constructed based on these predictors. Model performance was evaluated using receiver operating characteristic curves and the area under the curve. Calibration curves were plotted to evaluate the agreement between predicted probabilities and observed outcomes. Decision curve analysis (DCA) was conducted to assess the clinical net benefit of the model. RESULTS: Univariate analysis revealed significant differences across several clinical and ultrasonographic features between GLM and ILC (P < 0.05). Within the training cohort, multivariate logistic regression identified age category, menopausal status, lesion classification, shape, margin characteristics, posterior echo, and axillary lymph nodes enlarged as independent predictors for distinguishing GLM from ILC. The nomogram derived from these variables exhibited strong predictive capability, with AUCs of 0.97 and 0.96 for the training and validation cohorts, respectively. Calibration curves exhibited good agreement between predicted probabilities and observed outcomes. Furthermore, DCA demonstrated a substantial net clinical benefit associated with the model. CONCLUSION: A nomogram integrating clinical and ultrasonographic features was developed and validated, demonstrating robust discrimination, satisfactory calibration, and favorable clinical utility in differentiating GLM from ILC. As a noninvasive and visually intuitive tool, this model requires external validation, it still may facilitate more precise diagnostic and therapeutic decision-making.