MRI-based radiomics model for predicting the nature of nodules in cirrhotic liver

Motivation: We developed and validated a machine learning model combining semantic and radiomic features from MRI to differentiate regenerative nodules, dysplastic nodules, and hepatocellular carcinoma in cirrhotic liver. Goal(s): To differentiate regenerative nodules, dysplastic nodules, and hepatocellular carcinoma in cirrhotic liver. Approach: We developed and validated a machine learning model combining semantic and radiomic features from MRI to differentiate regenerative nodules, dysplastic nodules, and hepatocellular carcinoma in cirrhotic liver. Results: The combined model achieved superior performance (AUC=0.936) compared to single-feature models in a multicenter study of 266 patients. Impact: It provides a promising tool for non-invasive characterization of cirrhotic nodules.