AI detects liver diseases sooner

By combining blood samples with artificial intelligence, the research project LiverPRO at Odense University Hospital (OUH) detects scar tissue in the liver sooner and thereby prevents cirrhosis of the liver.

The LiverPRO project has received DKK 500.000 from BETA.Health – the Danish national innovation platform of future healthcare – to implement and market the solution. In short, LiverPRO combines a blood sample with AI to detect scar tissue in the liver – a symptom of advanced fatty liver disease, which is an initial stage of cirrhosis of the liver.

PhD student Katrine Prier Lindvig from the liver research centre FLASH at OUH is driving the project together with fellow researchers and AI experts from OUH and the University of Southern Denmark.

“Today, the approach of general practitioners (GPs) to detect liver diseases is not accurate enough. it only finds 53% of the patients with advanced liver fibrosis. In comparison, LiverPRO finds 92% of the patients with liver fibrosis”, Katrine Prier Lindvig explains.

In addition to improving the patient course, the LiverPRO project also has the potential to optimise the use of resources at the hospital. Measurements from OUH show that 54% of the patients referred to OUH by GPs do not have liver diseases, which means that a certain amount of the resources at the hospital are allocated to testing healthy patients, while the patients who are actually ill are not identified soon enough.

“LiverPRO is a good example of how AI in the near future can help patients and save or prolong lives. The evaluation of the project will also show if the project is cost-effective when the healthy can be filtered out, and the resources are dedicated for those who need it the most”, explains Peter Børker Nielsen, Programme Manager at CAI-X.

CAI-X contributes to the project with a go-to-market strategy and assessment by use of the new MAS-AI model for assessment of AI in healthcare.

LiverPRO will be launched in autumn 2023.

Read more

Read more about the algorithm behind LiverPRO in the preceding project LiverTrail.