DCE-MRI-based machine learning model for predicting axillary lymph node metastasis in breast cancer

Area of Science:

• Oncology
• Radiology
• Artificial Intelligence

Background:

• Accurate staging of breast cancer, specifically axillary lymph node (ALN) status, is crucial for treatment planning.
• Metastasis to ALNs significantly impacts patient prognosis and therapeutic strategies.

Purpose of the Study:

• To develop and validate an artificial intelligence (AI) model for predicting ALN metastasis in breast cancer.
• To leverage pre-treatment dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and radiomics features for metastasis prediction.

Main Methods:

• A dataset of 166 breast cancer patients' pre-treatment DCE-MRI scans was analyzed.
• Radiomics features were extracted from 2D and 3D images, including pre-enhancement and early post-enhancement phases.
• Machine learning algorithms, specifically the least absolute shrinkage selection operator (LASSO), were used to build predictive models.

Main Results:

• A 3D radiomics model achieved a mean area under the curve (AUC) of 82% and accuracy of 82% via 10-fold cross-validation.
• A combined model integrating radiomics and clinical features demonstrated a high C-index of 90% and AUC of 90%.

Conclusions:

• The developed AI model, integrating DCE-MRI radiomics and clinical data, accurately predicts axillary lymph node metastasis in breast cancer patients.
• This approach offers a promising non-invasive tool to improve the diagnostic accuracy of ALN staging.