Predicting older-donor kidneys' post-transplant renal function using pre-transplant data.

Area of Science:

• Nephrology
• Transplantation
• Biomedical Engineering

Background:

• Predicting post-transplant kidney function is crucial for optimizing donor-recipient matching.
• Current methods may not fully leverage available pre-transplant data for accurate prognostication.

Purpose of the Study:

• To develop and validate a methodology for predicting 1-year post-transplant estimated Glomerular Filtration Rate (eGFR-1).
• To identify donor-candidate pairs with a high probability (≥80%) of achieving a successful eGFR-1 (≥30 ml/min).

Main Methods:

• Utilized customized machine-learning algorithms applied to pre-transplant donor and recipient data.
• Analyzed data from discarded older-donor kidneys to identify suitable candidates.

Main Results:

• The algorithm accurately predicted the probability of achieving eGFR-1 ≥30 ml/min for specific donor-candidate pairs.
• Identified 5282 candidates who could have received a discarded kidney with an ≥80% chance of success.
• Transplants with ≥80% predicted success showed improved 10-year death-censored graft survival (73.61% vs. 70.48%).

Conclusions:

• Machine learning offers a powerful tool for predicting kidney transplant outcomes.
• This methodology can help optimize the utilization of older-donor kidneys, potentially increasing transplant rates and improving patient survival.