AI project
LiverTrail
Full title: The LiverTRAIL software for early detection of severe liver fibrosis in patients with alcoholic- and non-alcoholic fatty liver disease
Aim
This Ph.D. project will develop and validate a decision aid for use in primary health care to assess the risk of advanced fibrosis in patients with alcoholic- and non-alcoholic fatty liver disease. The decision aid – LiverTRAIL – works algorithm-based, by combining results from routine liver blood tests into multiple diagnostic algorithms, thus utilising computational power for large-scale data aggregation and pattern recognition.
Our aim is that LiverTRAIL can transform knowledge from original research data into a rational decision tool for the general practitioner.
LiverTRAIL will unite three main challenges within the study of fatty liver disease:
- Identify patients with severe fibrosis due to ALD and NAFLD (rule-in).
- Exclude patients with ALD and NAFLD at very low risk of advanced fibrosis from further diagnostic investigations (rule-out).
Monitor patients at high- and moderate risk of advanced fibrosis for disease progression and fibrosis improvement, e.g. during anti-fibrotic treatment.We hypothesize that through the implementation of LiverTRAIL in the primary sector, it is possible to achieve early detection of patients with asymptomatic, early-stage severe fibrosis and cirrhosis.
Participants
FLASH – Center for Liver Research, OUH
Professor Esmaeil S. Nadimi, Applied AI and Data Science, The Maersk Mc-Kinney Moller Institute, Faculty of Engineering, SDU
Funding
The project is funded by Innovation Fund Denmark.
Project period
Start: December 2018
End: November 2022
Read more
Read more about the first results from the project in the article Artificial intelligence outperforms standard blood-based scores in identifying liver fibrosis patients in primary care. The article was published in Scientific Reports, part of the Nature Portfolio.
Contact

Katrine Prier Lindvig
PhD student, MD
FLASH – Center for Liver Research
(+45) 28 83 87 54
[email protected]