Multiparametric MRI‐based model for prediction of local progression of hepatocellular carcinoma after thermal ablation

Findings: In this multicenter study that included 417 patients, a deep learning radiomics model was proposed, which achieved high prediction accuracy with AUC of 0.864‐0.858 across one training and two external testing cohorts, and realized successful patient stratification of high and low local recurrence risk. Meaning: The proposed prediction system may help physicians in therapeutic decision making and surveillance strategy selection for hepatocellular carcinoma patients after thermal ablation in clinical practice.