Related Experiment Video
Updated: Jul 9, 2026
07:15
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Predicting sentinel lymph node metastasis in breast cancer using an interpretable machine learning approach based on
Ruoyan Wang1, Xiongwu Li2, Hongni Zhu1
1Department of Bioinformatics, College of Artificial Intelligence Medicine, Chongqing Medical University, Chongqing, 400016, China.
Breast Cancer Research : BCR
|July 7, 2026
Summary
A new machine learning model accurately predicts sentinel lymph node metastasis (SLNM) in breast cancer patients. This tool can help spare eligible node-negative patients from unnecessary sentinel lymph node biopsy (SLNB).
Area of Science:
- Oncology
- Medical Informatics
- Surgical Oncology
Background:
- Accurate sentinel lymph node (SLN) status is crucial for breast cancer treatment decisions.
- Sentinel lymph node biopsy (SLNB) carries surgical risks for node-negative patients.
- Existing prediction models lack comprehensiveness and interpretability.
Purpose of the Study:
- To develop a comprehensive and explainable machine learning (ML) model for predicting sentinel lymph node metastasis (SLNM).
- To evaluate the feasibility of omitting SLNB in breast cancer patients.
- To improve surgical and adjuvant therapy decisions.
Main Methods:
- Trained nine ML algorithms on clinical, imaging, and pathological features from 1,485 patients.
- Validated the best model (SVM) on a prospective cohort of 103 patients.
- Utilized SHapley Additive exPlanations (SHAP) for model interpretability.
Main Results:
- The optimized SVM model achieved an AUC of 0.892 internally and 0.775 prospectively.
- The model outperformed existing clinical and imaging-based models.
- It could spare 58.1% of node-negative patients from SLNB, with key predictors including tumor margin smoothness and AR expression.
Conclusions:
- Developed and validated an SLNM prediction model using routine diagnostic features.
- The model demonstrates clinical applicability and near-surgical accuracy in assessing axillary status.
- Provides evidence for potential de-escalation of surgical approaches in breast cancer management.
