Shadows Behind Using Simple Risk Models in Selection of Hepatocellular Carcinoma Patients for Liver Transplantation

Objective: To assess the potential influence of replacing Milan criteria with simple risk scores on outcomes of hepatocellular carcinoma (HCC) patients undergoing liver transplantation. Summary Background Data: Several risk scores combining morphological and biological features were recently proposed for precise selection of HCC patients for transplantation. Methods: This retrospective study included 282 HCC liver transplant recipients. Recurrence-free survival (RFS), the primary outcome measure, was evaluated according to Metroticket 2.0 model and French AFP model with Milan criteria serving as benchmark. Results: Patients were well stratified with respect to RFS by Milan criteria, Metroticket 2.0 criteria, and AFP model cut-off <= 2 points (all P < 0.001) with c-statistics of 0.680, 0.695, and 0.681, respectively. Neither Metroticket 2.0 criteria (0.014, Z = 0.023; P = 0.509) nor AFP model (-0.014, Z = -0.021; P = 0.492) provided significant net reclassification improvement. Both patients within the Metroticket 2.0 criteria and AFP model <= 2 points exhibited heterogeneous recurrence risk, dependent upon alpha-fetoprotein (P = 0.026) and tumor number (P = 0.024), respectively. RFS of patients beyond Milan but within Metroticket 2.0 criteria (75.3%) or with AFP model <= 2 points (74.1%) was inferior to that observed for patients within Milan criteria (87.1%; P = 0.067 and P = 0.045, respectively). Corresponding microvascular invasion rates were 37.2% and 50.0%, compared with 13.6% in patients within Milan criteria (both P < 0.001). Moreover, Milan-out status was associated with significantly higher recurrence risk in subgroups within Metroticket 2.0 criteria (P = 0.021) or AFP model <= 2 points (P = 0.014). Conclusion: Utilization of simple risk scores for liver transplant eligibility assessment leads to selection of patients at higher risk of posttransplant HCC recurrence.