The study, published in the journal Metabolism, Clinical and Experimental, describes the team’s use of five artificial intelligence techniques across two platforms to build predictive models that could diagnose liver health status based on the levels of fats, hormones and carbohydrate and sugar molecules in a patient’s blood, without requiring the traditional invasive liver biopsy.
“We measured as many circulating molecules as reasonably possible and then let machine learning and artificial intelligence pick the best sets of molecules that would most accurately predict outcomes,” said senior author Christos Mantzoros, MD, DSc, director of the Human Nutrition Unit at BIDMC and a professor of medicine at Harvard Medical School in Boston.
Dr. Mantzoros continued, “Although the number of subjects appears small given conventional study designs, employing powerful and novel artificial intelligence models allowed us to derive accurate results, as high as 98 percent in some cases. These models may serve as low-risk, cost-effective alternative method to liver biopsy for diagnosis and monitoring [non-alcoholic fatty liver disease].”
More articles on AI:
DeepMind co-founder transitions to Google’s applied AI team
Facebook VP: AI is useful and rapidly advancing — ‘but don’t overhype it’
AHA report: 4 keys to successful implementation of clinical AI
