Predicting Kidney Transplant Complications Using Infrared Spectroscopy

A team of researchers recently demonstrated how Fourier transform infrared (FT-IR) spectroscopy can be used to identify biochemical signals associated with delayed graft function (DGF). Published in the Journal of Clinical Medicine, the findings are a proof-of-concept for a technique that could one day supplement existing clinical risk tools.¹

What is delayed graft function?

Delayed graft function, or DGF, is when a patient needs dialysis a week after a kidney transplant.^2,3 This is a common affliction, but the risk of delayed graft function dramatically decreases when the kidney comes from a living instead of a deceased donor.² Having this condition results in longer hospital stays, increased organ rejection risk, and worse long-term outcomes. As a result, accurate pre-implant prediction remains an unmet clinical need, as standard donor and recipient variables offer only modest predictive power.¹

What was the experimental procedure?

In their study, the research team analyzed preservation fluid samples from 56 kidney transplants derived from 49 deceased donors at a single Portuguese center.¹ Rather than using costly machine perfusion systems, the team examined fluid from conventional static cold storage. Dried-film FT-IR spectra were captured using a plate-based, high-throughput accessory and analyzed across the biochemical fingerprint region.¹

The study compared three predictive models in their study. The first model used clinical variables only. The second model used one using FT-IR data only, and the third model combined both. Under stringent donor-blinded cross-validation, which was designed to prevent data leakage between paired kidneys from the same donor, the FT-IR-only model performed the best with a receiver operating characteristic–area under the curve (ROC-AUC) of 0.814, modestly outperforming the clinical-only model at 0.775.¹ The combined model reached 0.796.¹

What were the limitations of the study?

The authors acknowledged several limitations that temper immediate clinical application. One limitation is the sample size. The cohort studied was small, with DGF occurring in just 14 of 56 cases.¹ Calibration analysis revealed the models tend to produce overly extreme risk estimates, and the researchers explicitly recommend recalibration before any deployment scenario is considered.¹

References

1. Ramalhete, L.; Araujo, R.; Vieira, M. B.; et al. Machine Learning-Assisted FTIR Spectroscopy Analysis of Kidney Preservation Fluids for Delayed Graft Function Risk Stratification. J. Clin. Med. 2026, 15 (7), 2762. DOI: 10.3390/jcm15072762
2. NHS Blood and Transplant, Delayed graft function. NHS.uk. Available at: https://www.nhsbt.nhs.uk/organ-transplantation/kidney/benefits-and-risks-of-a-kidney-transplant/risks-of-a-kidney-transplant/delayed-graft-function/ (accessed 2026-05-19).
3. Ponticelli, C.; Reggiani, F.; Moroni, G. Delayed Graft Function in Kidney Transplant: Risk Factors, Consequences and Prevention Strategies. J. Pers. Med. 2022, 12 (10), 1557. DOI: 10.3390/jpm12101557