Transplantology · Published 2025-03-03 · DOI 10.3390/transplantology6010006
Background/Objectives: Kidney transplantation (KT) is the preferred treatment for end-stage renal disease (ESRD), offering improved quality of life, superior survival rates and lower economic burden. However, improving long-term kidney allograft survival post transplantation remains a significant challenge. HLA eplet matching has emerged as a promising strategy to minimize immunological risk and enhance long-term graft survival. Still, our understanding of HLA immunogenicity remains limited. This study aims to evaluate if Electrostatic mismatch score (EMS) and eplet mismatch (EpMM) are significant for predicting KT outcomes and their optimal cut-off values associated with improved graft survival. Methods: Our study analyzed over 10,000 kidney transplant records from the Scientific Registry of Transplant Recipients (SRTR) dataset using traditional survival analysis and machine learning (ML) techniques. The immunogenicity scores EMS and EpMM were calculated based on donor-recipient HLA molecular mismatches. Kaplan–Meier plots, Cox proportional hazards (CPH), random survival forests (RSF), and survival decision trees (SDT) were utilized in assessing the significance of EpMM and EMS in improving KT outcomes and their optimal cut-offs. Results: EpMM and EMS were found to be significant predictors of kidney graft survival. The optimal cutoff values for improved outcomes for EMS and EpMM were 11 and 7 respectively, beyond which graft failure risk increased. The RSF model was the best-performing model in KT outcome prediction (C-index = 0.6945, Brier score = 0.1460). Conclusions: EMS and EpMM were significant in the prediction of kidney transplantation outcomes at cutoffs of 11 and 7, respectively. Incorporating these measures in KT organ allocation strategies could improve long-term survival outcomes.
Abstract from DOAJ. Public domain (CC0 1.0).
Read the article at the publisher →