Development of models to predict early post-transplant recurrence of hepatocellular carcinoma that also integrate the quality and characteristics of the liver graft: A national registry study in China

Background: Donor characteristics and graft quality were recently reported to play an important role in the recurrence of hepatocellular carcinoma after liver transplantation. Our aim was to establish a prognostic model by using both donor and recipient variables. Methods: Data of 1,010 adult patients (training/validation: 2/1) undergoing primary liver transplantation for hepatocellular carcinoma were extracted from the China Liver Transplant Registry database and analyzed retrospectively. A multivariate competing risk regression model was developed and used to generate a nomogram predicting the likelihood of post-transplant hepatocellular carcinoma recurrence. Results: Of 673 patients in the training cohort, 70 (10.4%) had hepatocellular carcinoma recurrence with a median recurrence time of 6 months (interquartile range: 4-25 months). Cold ischemia time was the only independent donor prognostic factor for predicting hepatocellular carcinoma recurrence (hazard ratio = 2.234, P=.007). The optimal cutoff value was 12 hours when patients were grouped according to cold ischemia time at 2-hour intervals. Integrating cold ischemia time into the Milan criteria (liver transplantation candidate selection criteria) improved the accuracy for predicting hepatocellular carcinoma recurrence in both training and validation sets (P < .05). A nomogram composed of cold ischemia time, tumor burden, differentiation, and a-fetoprotein level proved to be accurate and reliable in predicting the likelihood of 1-year hepatocellular carcinoma recurrence after liver transplantation. Additionally, donor anti-hepatitis B core antibody positivity, prolonged cold ischemia time, and anhepatic time were linked to the intrahepatic recurrence, whereas older donor age, prolonged donor warm ischemia time, cold ischemia time, and ABO incompatibility were relevant to the extrahepatic recurrence. Conclusion: The graft quality integrated models exhibited considerable predictive accuracy in early hepatocellular carcinoma recurrence risk assessment. The identification of donor risks can further help understand the mechanism of different patterns of recurrence. (C) 2018 Elsevier Inc. All rights reserved.