Researchers use AI to spot fatty liver disease earlier in routine tests

In short: Researchers are exploring AI tools that can flag fatty liver disease using data many patients already have, like routine blood tests and common X-rays.

What's going on

Fatty liver disease is becoming common worldwide. More than a billion people are thought to have too much fat in their liver, and studies estimate it affects about 30 percent of adults globally. It often causes no clear symptoms, so many people find out only when the liver is badly scarred.

Because early stages can be reversed, doctors want better ways to catch it sooner. Lifestyle changes like losing weight, cutting back on alcohol, exercising, and even drinking more coffee can help in some cases. For more advanced cases, newer medicines such as semaglutide and resmetirom have shown promise.

Several groups think AI could help by working in the background, like a spellcheck tool that reviews what is already on your screen. One idea is to have AI automatically calculate a Fib-4 score, which estimates the risk of serious liver scarring using age and common blood test results. Another approach uses imaging, including a study from Osaka Metropolitan University where an AI system reviewed routine chest X-rays and identified fatty liver disease with 82 percent accuracy, even though chest X-rays are mainly used to look at the heart and lungs.

What to watch

Many of these tools are still mostly in research, but some are moving toward real-world use. For example, Danish startup Evido has developed an AI tool called LiverPRO that uses nine routine blood markers and is being commercialized with Roche. The next step to watch is whether hospitals and primary care clinics actually adopt these systems, and how well they reduce missed cases without flooding specialists with unnecessary referrals.

Source: Wired