Machine learning study shows AI’s potential for predicting kidney transplant outcomes

Editor's Note:

Artificial intelligence (AI) leveraging machine learning (ML) and natural language processing (NLP, a subset of machine learning) models can help identify donors with kidneys unsuitable for organ transplant, according to a study published November 1 in Jama Surgery.

Despite the unmet need, many deceased-donor kidneys are discarded or not recovered, researchers report. To evaluate the feasibility of AI for alleviating this challenge, they conducted a retrospective study consisting of two cohorts: training/validation, with 9,555 donors, and testing, with 2,481 donors. Collected data consisted of structured donor characteristics and unstructured donor narratives from the United Network for Organ Sharing (UNOS). Donor offers to a single transplant center between January 2015 and December 2020 were used to train and validate ML models to predict donors who had at least 1 kidney transplanted (at our center or another center). The donor data from 2021 were used to test each model.

Conventional ML models generated by structured data and NLP models with unstructured data both predicted donors with high-risk kidneys. Although he use of unstructured data is likely to expand the possibilities, researchers emphasize that further improvement will require exploring new approaches.