Comparative Analysis of Nomogram and Machine Learning Models for Predicting Axillary Lymph Node Metastasis in Early-Stage Breast Cancer: A Study on Clinically and Ultrasound-Negative Axillary Cases Across Two Centers.

OBJECTIVE

Early and accurate prediction of axillary lymph node metastasis (ALNM) is crucial in determining appropriate treatment strategies for patients with early-stage breast cancer. The aim of this study was to evaluate the efficacy of radiomic features extracted from ultrasound (US) images combined with machine learning (ML) methods in predicting ALNM to improve diagnostic accuracy and patient prognosis.

METHODS

In this retrospective study, data of 282 early-stage breast cancer patients from two centers were analyzed. We considered clinicopathological characteristics, conventional US features, contrast-enhanced ultrasound (CEUS) characteristics, and radiomics features. Radiomics features were extracted from US images, and using least absolute shrinkage and selection operator (LASSO) regression, 12 key features were selected to compute a Radiomics score (Rad-score). A nomogram was developed based on these features, alongside five ML models: Logistic Regression (LR), Naive Bayes (NB), Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Extreme Gradient Boosting (XGBoost). Model performance was evaluated using metrics such as the area under the curve (AUC), accuracy (ACC), sensitivity (SEN), specificity (SPE), negative predictive value (NPV), and positive predictive value (PPV).

RESULTS

Both the nomogram and ML models, including the Rad-score combined with histologic type, significantly predicted ALNM. Among all models, the XGBoost model showed the best performance with an AUC of 0.810 and an accuracy of 84.1% in the external test set, surpassing the nomogram and other ML models. SHapley Additive exPlanations (SHAP) analysis further provided insights into the influence of individual radiomics features on ALNM prediction.

CONCLUSIONS

While the nomogram provides a useful traditional statistical approach, integrating radiomics features with ML, particularly the XGBoost model enhanced by SHAP interpretability, offers superior predictive accuracy for ALNM in early-stage breast cancer patients.