# Post‐Transplant DSA Monitoring After HLAi Kidney Transplantation: Key Time Points

## TL;DR

This study identifies optimal timing for monitoring antibodies after kidney transplants to improve patient outcomes and reduce unnecessary testing.

## Contribution

A machine learning approach identifies key DSA monitoring time points with high accuracy using fewer measurements.

## Key findings

- Measuring DSAs on Days 1, 10, and 26 or 34 achieves 93.7% accuracy in classifying antibody patterns.
- Including both Days 26 and 34 increases classification accuracy to 96.9%.
- Testing beyond Day 34 does not improve classification accuracy.

## Abstract

In HLA‐incompatible transplantation, monitoring donor‐specific antibodies (DSA) after transplantation informs clinical decision‐making and the management of antibody‐mediated rejection. Persistent post‐transplant DSAs are associated with poorer graft survival compared to cases where DSAs resolve. Previously, four distinct early post‐transplant DSA dynamic patterns were identified using a unique dataset with frequent DSA measurements and shown to be associated with AMR and graft outcomes. However, frequent testing is costly, and predicting outcomes using fewer DSA measurements could improve monitoring efficiency. A machine learning classifier based on dynamic time warping was used to identify early post‐transplant DSA patterns from a reduced set of measurements. Individual days and day combinations within the first 50 post‐transplant days were systematically evaluated. The combinations with the highest classification accuracy compared to using all available time points were selected as optimal. This study determined the minimum number of DSA screenings required to accurately classify early DSA dynamics. Measuring DSAs on Days 1, 10, and either 26 or 34 identified the four dynamic patterns with 93.7% accuracy. Including both Days 26 and 34 increased the accuracy to 96.9%. Testing beyond Day 34 did not yield further improvement in classification. Several monitoring regimens commonly used in clinical practice were evaluated, and the proposed approach outperformed all of them. This study identifies an efficient DSA monitoring strategy that can predict antibody behaviour early post‐transplant, enabling timely intervention. It also addresses inconsistencies in current monitoring practices, offering evidence for standardised, data‐driven protocols for more precise and efficient post‐transplant care.