Doctors have developed an AI tool that could reduce wasted efforts to transplant organs by 60%.

Thousands of patients worldwide are waiting for a potentially life-saving donor, and more candidates are stuck on waiting lists than there are available organs.

Recently, in cases where people need a liver transplant, access has been expanded by using donors who die after cardiac arrest. However, in about half of these donations after circulatory death (DCD) cases, the transplant ends up being cancelled.

That is because the time between the removal of life support and death must not exceed 45 minutes. If the donor does not die within the timeframe needed to preserve organ quality, surgeons often reject the liver because of the increased risk of complications to the recipient.

Now doctors, scientists and researchers at Stanford University have developed a machine learning model that predicts whether a donor is likely to die within the timeframe during which their organs are viable for transplantation.

The AI tool outperformed the judgment of top surgeons and reduced the rate of futile procurements – which occur when transplant preparations have begun but the donor dies too late – by 60%.

“By identifying when an organ is likely to be useful before any preparations for surgery have started, this model could make the transplant process more efficient,” said Dr Kazunari Sasaki, a clinical professor of abdominal transplantation and senior author on the study.

“It also has the potential to allow more candidates who need an organ transplant to receive one.”

Details of the breakthrough were published in the Lancet Digital Health journal.

The advance could reduce the number of instances in which healthcare workers prepare organs for recovery, only to determine they are unsuitable for recovery and transplantation, putting financial and operational strain on transplant centres.

Hospitals primarily rely on surgeons’ judgment to estimate this critical timeframe, which can vary widely and lead to unnecessary costs and wasted resources.

The new AI tool was trained on data from more than 2,000 donors across several US transplant centres. It uses neurological, respiratory and circulatory data to predict a potential donor’s progression to death with greater accuracy than previous models and human experts.

The model was tested retrospectively and prospectively, achieving a 60% reduction in futile procurements compared with surgeons’ predictions. Importantly, it maintains accuracy even when some donor information is missing, researchers said.

A reliable, data-driven tool could help healthcare staff make better decisions, optimising organ use and reducing wasted efforts and costs.

The approach could be a significant step forward in transplantation, the research team said, highlighting “the potential for advanced AI techniques to optimise organ utilisation from DCD donors”.

Next, they plan to vary the AI tool to trial it with heart and lung transplants.