A new model based on gamma-glutamyl transpeptidase to lymphocyte ratio and systemic immune-inflammation index can effectively predict the recurrence of hepatocellular carcinoma after liver transplantation

Background: Liver transplantation (LT) is one of the most effective treatment modalities for hepatocellular carcinoma (HCC), but patients with HCC recurrence after LT always have poor prognosis. This study aimed to evaluate the predictive value of the gamma-glutamyl transpeptidase-to-lymphocyte ratio (GLR) and systemic immune-inflammation index (SII) in terms of HCC recurrence after LT, based on which we developed a more effective predictive model. Methods: The clinical data of 325 HCC patients who had undergone LT were collected and analyzed retrospectively. The patients were randomly divided into a development cohort (n = 215) and a validation cohort (n = 110). Cox regression analysis was used to screen the independent risk factors affecting postoperative recurrence in the development cohort, and a predictive model was established based on the results of the multivariate analysis. The predictive values of GLR, SII and the model were evaluated by receiver operating characteristic (ROC) curve analysis, which determined the cut-off value for indicating patients' risk levels. The Kaplan-Meier survival analysis and the competing-risk regression analysis were used to evaluate the predictive performance of the model, and the effectiveness of the model was verified further in the validation cohort. Results: The recurrence-free survival of HCC patients after LT with high GLR and SII was significantly worse than that of patients with low GLR and SII (P<0.001). Multivariate Cox regression analysis identified GLR (HR:3.405; 95%CI:1.954-5.936; P<0.001), SII (HR: 2.285; 95%CI: 1.304-4.003; P=0.004), tumor number (HR:2.368; 95%CI:1.305-4.298; P=0.005), maximum tumor diameter (HR:1.906; 95%CI:1.121-3.242; P=0.017), alpha-fetoprotein level (HR:2.492; 95%CI:1.418-4.380; P=0.002) as independent risk factors for HCC recurrence after LT. The predictive model based on these risk factors had a good predictive performance in both the development and validation cohorts (area under the ROC curve=0.800, 0.791, respectively), and the performance of the new model was significantly better than that of single GLR and SII calculations (P<0.001). Survival analysis and competing-risk regression analysis showed that the predictive model could distinguish patients with varying levels of recurrence risk in both the development and validation cohorts. Conclusions: The GLR and SII are effective indicators for evaluating HCC recurrence after LT. The predictive model based on these indicators can accurately predict HCC recurrence after LT and is expected to guide preoperative patient selection and postoperative follow-up.