Serum and Tissue Metabolite-based Prediction of Sentinel Lymph Node Metastasis in Breast Cancer

Overview

Breast cancer is a malignant tumor with the highest morbidity and mortality among women worldwide. Accurate staging of axillary lymph nodes is critical for metastatic assessment and decisions regarding treatment modalities in breast cancer patient. Among patients who underwent sentinel lymph node biopsy, about 70 % of the patients had negative pathological results and in other words, these 70 % of the patients received unnecessary surgery. At present, imaging and pathological diagnosis is the main measure of lymph node metastasis in breast cancer. However, limitations remained. Artificial intelligence, including deep learning and machine learning algorithms, has emerged as a possible technique, which can make a more accuracy prediction through machine-based collection, learning and processing of previous information, especially in radiology and pathology-based diagnosis. With the intensification of the concept of precision medicine and the development of non-invasive technology, the investigators intend to use the artificial intelligence technology to develop a serum and tissue-based predictive model for sentinel lymph node metastasis diagnosis combined with imaging and pathological information, providing specific, efficient and non-invasive biological indicators for the monitoring and early intervention of lymph node metastasis in patient with breast cancer. Therefore, the investigators retrospectively include serum samples from early breast cancer patients undergoing sentinel lymph node biopsy, including a discovery cohort and a modeling cohort. Metabolites were detected and screened in the discovery cohort and then as the target metabolites for targeted detection in the modeling cohort. Combined with preoperative imaging and pathological information, a prediction model of breast cancer sentinel lymph node metastasis based on serum metabolites would be established. Subsequently, multi-center breast cancer patients will prospectively be included to verify the accuracy and stability of the model.

Conditions