UCLA Health researchers have developed a new machine learning model to predict which donor organs from donors after circulatory death (DCD) are most likely to be suitable for transplantation. This research Heart and Lung Transplant Journaladdress a key challenge in thoracic transplantation: determining whether a potential DCD donor dies within the critical time frame required for successful organ recovery.
Traditionally, donor organs have come from individuals declared brain dead. However, a severe shortage of these organs has led to increased reliance on DCD donors, those with no chance of neurological recovery but whose hearts continue to beat. These donors currently account for about half of all organ donations this year.
A major challenge with DCD donation is that many potential donors die within the required time period after life support is discontinued, rendering their organs unsuitable for transplantation.
If you donate your heart, it usually needs to be stopped within about 30 minutes after life support is removed. For the lungs, this period is usually less than 2 hours. Exceeding these time limits often renders the organ unusable.
“This addresses the ‘Achilles’ heel’ challenge of thoracic transplantation, which is determining which donor cases are worth pursuing for procurement,” said Dr. Abbas Al-Dehari, director of the UCLA Heart, Lung, and Cardiopulmonary Transplant Program, professor of surgery in the UCLA David Geffen School of Medicine, and corresponding author of the study.
Currently, there is no reliable formula to predict whether a donor will expire within these critical periods. As a result, transplant teams travel to procure organs, often returning nearly half empty-handed, placing significant emotional stress on recipients and their families and increasing medical costs.
The study analyzed data on more than 4,400 potential donors from 2014 to 2025 from three major organ donation organizations in the United States. These donors were people who had their life support removed for the purpose of organ donation. Researchers collected important clinical information, including medical history, neurological function, respiratory status, and laboratory values, and combined it into one dataset.
Using machine learning, the research team was able to more accurately predict whether a donor would die within the critical 30 minutes and two hours required for heart and lung transplants.
Previous predictive models were built using smaller patient groups, fewer medical indicators, and more limited datasets, resulting in lower accuracy. Researchers say machine learning offers significant improvements by better reflecting the complex physiological changes that occur in the final moments of life.
“With machine learning, we can get a much more accurate picture of what’s actually happening inside the body in its final moments,” Aldehari said. “While more tests are still needed, this approach avoids unnecessary organ recovery attempts, uses resources more efficiently, and could ultimately help more patients receive life-saving transplants.”
The findings were presented as a “highlighted summary” during the plenary session of the world’s largest international conference on heart and lung transplantation.
Researchers believe this innovation has the potential to transform the field of DCD organ transplantation by improving donor selection, reducing wasted procurement trips, minimizing emotional distress for recipients and families, and reducing the use of unnecessary medical resources.
