Work

Combining ultrasound and core needle histology imaging Radiopathomics model in predicting axillary lymph node metastasis in triple-negative breast cancer

Year: 2026

Type: article

Abstract: OBJECTIVE: This study aims to utilize preoperative ultrasound (US) and core needle biopsy (CNB) histology image images to construct a multimodal Radiopathomics model to predict the risk of axillary lymph node metastasis (ALNM) in patients with triple-negative breast cancer (TNBC). MATERIALS AND METHODS: This multicenter retrospective study included TNBC patients from two medical centers between December 2021 and January 2025. Patients were randomly assigned to a training group and a validation group in a 7:3 ratio, with TNBC patients from a third medical center serving as an external testing group. Radiomic features were derived from preoperative ultrasound (US) images, while Pathomics features were obtained from CNB histology images. Key predictive features were selected through univariate analysis, correlation analysis, and the LASSO algorithm. The Radiopathomics model was then built using the XGBoost algorithm, and its performance was assessed through metrics such as the area under the curve (AUC), decision curve analysis (DCA), accuracy, specificity, sensitivity, positive predictive value (PPV), and negative predictive value (NPV). RESULTS: A total of 243 TNBC patients were included in this study, with an average age of 51.9 ± 10.5 years, among whom 102 had ALNM and 141 did not. The AUCs for the training group, validation group, and testing group of the multimodal Radiopathomics model were 0.892, 0.848, and 0.831, respectively; the diagnostic efficiency of the multimodal Radiopathomics model was superior to clinical, Radiomics, and Pathomics models. CONCLUSION: Preoperative ALNM in TNBC patients can be accurately predicted by the multimodal Radiopathomics model based on preoperative US and CNB histology image, assisting clinicians in formulating treatment plans.

Source: World Journal of Surgical Oncology

DOI: 10.1186/s12957-026-04316-4

Authors: Li Yanhua, Tang Yan, Liu Xiaoling, He Fanding, Qin Xiachuan