Using machine learning to examine pre-transplant factors influencing De novo HLA-specific antibody development post-kidney transplant.

INTRODUCTION

The development of de novo donor-specific antibodies (DSAs) against HLA is associated with premature graft failure in kidney transplantation. However, reported rates and contributing factors vary widely. We aimed to identify pre-transplant factors influencing de novo HLA-specific antibody development using machine learning (ML).

METHODS

Data from 460 kidney transplant recipients at a single centre (2009-2014) was analysed. Pre-transplant clinical and immunological variables were collected, and post-transplant sera were screened for HLA antibodies. Positive samples underwent Single Antigen Bead (SAB) testing. ML models (CART, RF, XGBoost, CatBoost) were trained on a set of pre-transplant data to predict dnDSA formation, with and without SMOTE oversampling. Model performance was evaluated using F1 scores, and feature importance was assessed using SHAP.

RESULTS

In the full cohort, 115 patients (25 %) developed dnHLA-specific antibodies, including 36 (31 %) with dnDSAs. XGBoost achieved the best performance (F1 0.54-0.59 without SMOTE; 0.72-0.79 with SMOTE). Univariate analysis identified significant predictors: pre-transplant HLA-specific antibodies (p < 0.001), prior transplantation (p < 0.001), cold ischaemia time (CIT) (p = 0.02), female gender (p = 0.01), younger age (p = 0.03), HLA mismatch (p = 0.01), aminoacid mismatch (p = 0.01), and depleting induction (p = 0.01). SHAP plots confirmed the importance of pre-existing antibodies and re-transplantation. Extremes of CIT and age ≥ 65 was associated were associated with reduced predicted risk. Model performance in the unsensitised subgroup was limited (F1 < 0.2).

CONCLUSION

ML models can be used to identify pre-transplant risk factors for de novo HLA-specific antibody development. Monitoring and risk-stratification based on these factors may inform immunological strategies and recipient selection to improve long-term allograft outcomes.

TRANSLATIONAL STATEMENT: This study identified pre-transplant risk factors for the development of de novo DSA in kidney transplantation. Monitoring and risk-stratifying patients based on these factors may help guide preventive immunological strategies and recipient selection to improve long-term allograft outcomes.