Machine Learning Prediction of Liver and Lung Metastasis in Colorectal Cancer

Using a machine learning model, researchers accurately predicted liver and/or lung metastasis in colorectal cancer patients, assisting clinical decision-making. A total of 51,265 patients were analyzed, with 15.3% having distant metastasis. The random forest algorithm showed the best predictive ability with high accuracy, AUC, and AUPR values, validated in an external set, providing a valuable tool for assessing metastasis risk.