Development and validation of a nomogram to predict allograft survival after pediatric liver transplantation.

BACKGROUND: Liver transplantation is the main treatment for cholestatic liver disease and some metabolic liver diseases in children. However, no accurate prediction model to determine the survival probability of grafts prior to surgery exists. This study aimed to develop an effective prognostic model for allograft survival after pediatric liver transplantation.

METHODS: This retrospective cohort study included 2032 patients who underwent pediatric liver transplantation between January 1, 2006, and January 1, 2020. A nomogram was developed using Cox regression and validated based on bootstrap sampling. Predictive and discriminatory accuracies were determined using the concordance index and visualized using calibration curves; net benefits were calculated for model comparison. An online Shiny application was developed for easy access to the model.

RESULTS: Multivariable analysis demonstrated that preoperative diagnosis, recipient age, body weight, graft type, preoperative total bilirubin, interleukin-1β, portal venous blood flow direction, spleen thickness, and the presence of heart disease and cholangitis were independent factors for survival, all of which were selected in the nomogram. Calibration of the nomogram indicated that the 1-, 3-, and 5-year predicted survival rates agreed with the actual survival rate. The concordance indices for graft survival at 1, 3, and 5 years were 0.776, 0.757, and 0.753, respectively, which were significantly higher than those of the Pediatric End-Stage Liver Disease and Child-Pugh scoring systems. The allograft dysfunction risk of a recipient could be easily predicted using the following URL: https://aspelt.shinyapps.io/ASPELT/ / CONCLUSION: The allograft survival after pediatric liver transplantation (ASPELT) score model can effectively predict the graft survival rate after liver transplantation in children, providing a simple and convenient evaluation method for clinicians and patients.