US FDA approves first AI tool for liver disease trials: How AIM-NASH transforms biopsy reading

The FDA has officially approved the first AI tool to help with liver disease drug trials, and it is designed to make reading liver biopsies for MASH faster, more consistent, and easier for pathologists to use in clinical research.

"The U.S. Food and Drug Administration (FDA) has qualified the first AI drug development tool, the AI-Based Histologic Measurement of NASH (AIM-NASH), to help pathologists assess metabolic dysfunction-associated steatohepatitis (MASH) disease activity in clinical trials. This cloud-based tool helps pathologists score liver biopsy components, including fat infiltration (steatosis), inflammation (hepatocellular ballooning and lobular inflammation), and scarring (fibrosis) stage," the US FDA has said in a press release.

MASH is a serious form of fatty liver disease where too much fat in the liver causes inflammation and scarring.

Over time, this can progress to severe scarring (cirrhosis), liver failure, liver cancer, the need for a liver transplant, or death.

The tool is called AIM-NASH and it uses artificial intelligence to look at digital images of liver biopsy samples taken from patients in clinical trials.

It helps score key features of liver damage: how much fat is in the liver (steatosis), how much inflammation and ballooning of liver cells is present, and how advanced the scarring (fibrosis) is.

"Currently in MASH clinical trials, multiple experts independently assess liver histology, a time-consuming process made complicated by variable scoring. AIM-NASH could help standardize histologic assessment and reduce the time and resources needed for MASH drug development," the US FDA has said. "The AIM-NASH system uses AI algorithms to analyze images of liver biopsies and provides scores according to the NASH Clinical Research Network scoring system. The process keeps humans involved, as pathologists are fully responsible for final interpretation, reviewing the whole slide image and AIM-NASH outputs before accepting or rejecting the AI-generated scores."