Machine Learning Accurately Predicts Recurrence and Death in Liver Cancer Patients

A novel random survival forest model effectively stratifies hepatocellular carcinoma patients into high- and low-risk groups for recurrence and mortality post-surgery. Utilizing critical clinicopathological features and molecular markers, the model demonstrates significant differences in 5-year recurrence and death probabilities. For instance, in the training cohort, high-risk patients faced an 87.3% recurrence rate compared to 51.5% for low-risk patients. This stratification aids surgeons in devising targeted post-surgical management plans, potentially improving patient outcomes.