Validation of 3 high-risk resectable hepatocellular carcinoma models.

Background: Breakthroughs in immunotherapy, targeted therapy, and locoregional therapy are transforming the treatment of hepatocellular carcinoma (HCC). Studies are investigating whether therapies can be used to downstage tumors to resection or escalate therapies to improve prognostic outcomes. However, there are no standard criteria for "high-risk resectable" HCC. This study aims to validate and compare the prognostic performance of three distinct models for defining "high-risk resectable" HCC in predicting recurrence and survival following upfront resection. Methods: We used a prospectively collected database of 1,779 HCC patients from a single institution to validate 3 models for defining "high-risk resectable" HCC. Clinical parameters included tumor number, size, vascular invasion, alpha-fetoprotein (AFP), and technical resectability. Patients who underwent upfront liver resection were classified as "high-risk" or readily resectable according to each model. Logistic regression was performed to predict recurrence, 1-, and 3-year survival adjusting for age, sex, ethnicity, hepatitis B, C, metabolic-associated steatohepatitis (MASH), body mass index (BMI), history of diabetes, hyperlipidemia, hypertension, and each model as predictors. C-statistics were used for comparison. Results: Of the 290 patients who underwent upfront liver resection, 109 experienced recurrences. Patients were classified as high-risk resectable by Technical Risk, Integrated Risk, and Simplified Integrated Risk models (60, 148, and 40 patients, respectively). The Integrated Risk model demonstrated the highest sensitivity for predicting recurrence and survival (63.3%, 79.6%, and 70.1% for recurrence, 1-, and 3-year survival, respectively). Although all models showed significant predictive value based on area under the receiver operating curve (AUROC), the Technical Risk and Simplified Integrated Risk models exhibited lower sensitivity but higher specificity. Conclusions: The three models demonstrated strong predictive performance across a diverse cohort. The Integrated Risk model, incorporating both technical and prognostic parameters was the most sensitive for identifying patients at high risk who may benefit from escalated therapy. Future studies should incorporate these models into treatment escalation strategies for HCC.