Predicting preoperative axillary lymph node metastasis to guide surgical decisions in invasive breast cancer.

Background: In the clinical management of patients with invasive breast cancer (IBC), the precise identification of axillary lymph node metastasis (ALNM) is of paramount importance for guiding axillary surgery and formulating corresponding treatment strategies. Currently, the clinical need to accurately predict the risk of ALNM in patients with IBC, so as to determine whether axillary surgery (including sentinel lymph node biopsy) can be safely omitted, has not been fully met. Methods: The clinical data of 454 patients with IBC were retrospectively analyzed, and the patients were randomly divided into a training group and a validation group at a ratio of 3:1. Independent predictors related to ALNM were identified through univariate and multivariate Logistic regression analysis. Accordingly, a nomogram model integrating clinicopathological, ultrasound and serological indicators along with additional clinical parameters was constructed and validated by receiver operating characteristic (ROC) curve analysis, calibration plots, and decision curve analysis (DCA). The primary outcome measure was the incidence of ALNM. Results: The ALNM rates were 31.4% and 27.4% in the training and verification groups, respectively. Multivariate regression analysis indicated that tumor size, circumscribe margin, ultrasonic lymph node status, WBC, ER, and BI-RADS were independent risk factors for ALNM in patients with IBC. In both the training group and the validation group, the nomogram exhibited modest predictive performance (training group AUC = 0.741; validation group AUC = 0.705). The Ultrasonic lymph node status (normal vs. abnormal) alone yielded an AUC of 0.620 in the training cohort and 0.627 in the validation cohort for predicting ALNM. In comparison, our nomogram achieved significantly higher AUCs (training: 0.741, P = 0.0017; validation: 0.705, P = 0.1772; DeLong test). Additionally, the nomogram demonstrated satisfactory calibration and clinical utility as evidenced by the calibration curve and DCA. Conclusion: By integrating clinicopathological, ultrasonic and serological indicators, this nomogram can enhance the accuracy of predicting ALNM in patients with IBC. However, given its moderate discriminative ability, the model can support preoperative risk stratification but cannot replace standard axillary staging procedures (including sentinel lymph node biopsy). Our findings provide a basis for decision-making regarding individualized axillary lymph node surgery.