Pretransplant recurrent liver cancer prediction score (p-RELAPSE) to risk stratify hepatocellular carcinoma patients undergoing liver transplantation: validation in an international multicenter retrospective cohort

Background: Selection of patients with hepatocellular carcinoma (HCC) for liver transplantation (LT) increasingly integrates dynamic tumor biology alongside morphometric criteria. Methods: We developed and externally validated the pre-transplant REcurrent Liver cAncer Prediction ScorE (p-RELAPSE) to predict post-LT recurrence using two retrospective cohorts. Post-LT outcomes of 4420 HCC patients undergoing LT in the US Multicenter HCC Transplant Consortium (UMHTC) were analyzed to derive p-RELAPSE using a competing risks multivariable model, with external validation in 1736 HCC LT recipients from the European Hepatocellular Cancer Liver Transplant (EurHeCaLT) group. Results: The UMHTC training ( n = 2893), internal validation ( n = 1527), and EurHeCaLT ( n = 1736) cohorts showed comparable 5-year overall survival (69.4–71.0%), recurrence-free survival (66.2–66.3%), and recurrence incidence (12.6–14.4%; all P > 0.05). p-RELAPSE incorporated Milan criteria (MC) status, downstaging, maximum tumor diameter, dynamic alpha-fetoprotein (AFP) response, and waitlist time. The model stratified recurrence risk even within MC patients. In EurHeCaLT, p-RELAPSE achieved a C-statistic of 0.69, outperforming MC, French AFP, Up-to-7, pre-MORAL, and New York/California models ( P < 0.05). Conclusions: By integrating AFP dynamics and waitlist time, p-RELAPSE refines post-LT recurrence risk stratification, identifying MC patients unlikely to benefit from LT and non-MC patients with favorable outcomes. This externally validated model provides a robust framework for individualized pre-LT decision-making in HCC candidates.