Automatic Recognition and Prognostic Prediction of Colorectal Liver Metastases Using a Multi-Scale Deep Learning Framework: Model Development and Validation Study

Key Points

• The study aims to develop and validate a deep learning framework for better recognition and prognostic prediction of colorectal liver metastases.
• Developed the CLM-Net framework integrating various deep learning architectures.
• Used multi-scale mechanisms for enhanced model performance.
• Focused on clinical translation and generalization potential.
• Significantly improved recognition and prognostic prediction of colorectal liver metastases.
• Demonstrated excellent generalization capabilities.
• Showed high clinical translation potential for future applications.

Abstract