Milan criteria, University of California, San Francisco, criteria, and model for end-stage liver disease score as predictors of salvage liver transplantation.

The selection criteria for salvage liver transplantation (SLT) candidates have not been previously established. A global analysis for the association between the criteria and prognosis is required. All of the adult patients who underwent liver transplantation with a diagnosis of hepatocellular carcinoma (HCC) from January 1, 2000, to December 31, 2011, were retrospectively analyzed. A total of 1,554 cases were involved, including 1,392 primary liver transplantation (PLT) and 162 SLT cases. All the cases were classified into 3 groups according to the Milan criteria combined with the University of California, San Francisco (UCSF), criteria, and significant differences were found between the 2 groups. The overall graft survival rate was lower in all cases of SLT than in PLT (P = .030). Within the Milan criteria, no significant difference in the graft survival rate was found between PLT and SLT. In a Cox regression analysis, the Model for End-Stage Liver Disease (MELD) score and tumor levels graded according to the Milan/UCSF criteria were found to be independent risk factors for the graft survival rate. Receiver operating characteristic (ROC) curves were generated by the fatality risk values calculated by means of the Cox model and the 1-year graft survival rates of all the patients and of the SLT patients. The areas under the ROC curves were 0.922 and 0.935, respectively. Compared with PLT, the global graft survival rate of SLT was compromised. The MELD score and Milan/UCSF criteria were effective in predicting the prognosis of PLT and SLT. Therefore, when the recurrent lesions of HCC are within the Milan criteria, SLT can be performed with a good prognosis.