Interpretable MRI-Based Machine Learning Model for Noninvasive Prediction of Axillary Lymph Node Metastasis After...

Area of Science:

• Oncology
• Radiology
• Artificial Intelligence

Background:

• Accurate prediction of axillary lymph node metastasis (ALNM) after neoadjuvant chemotherapy (NAC) is crucial for breast cancer treatment planning.
• Current methods for ALNM prediction post-NAC are often invasive and challenging.

Purpose of the Study:

• To develop and validate an interpretable machine learning model for noninvasive ALNM prediction after NAC in breast cancer.
• The model integrates MRI-based radiomics, deep learning features, and the Node-RADS score.

Main Methods:

• A multicenter retrospective study included 641 breast cancer patients.
• Preoperative MRI and clinicopathologic data were analyzed to extract radiomics and deep learning features.
• Combined clinical-deep learning-radiomics (CDLR) models were developed and validated across multiple cohorts.

Main Results:

• The CDLR model demonstrated superior predictive performance (AUCs ranging from 0.737 to 0.879) across training and validation cohorts.
• The CDLR model outperformed standalone clinical and deep learning-radiomics models.
• SHAP analysis identified key predictors including Node-RADS and specific radiomic/deep learning features.

Conclusions:

• An interpretable CDLR model provides accurate, noninvasive prediction of ALNM after NAC in breast cancer.
• This model can aid in individualized clinical decision-making for breast cancer patients.