EMA Approves AI Tool for MASH Clinical Trials

The European Medicines Agency (EMA) offered a positive opinion last week on a new artificial intelligence (AI)–based tool that supports pathologists in diagnosing and assessing the severity of metabolic dysfunction–associated steatohepatitis (MASH) during clinical trials. This is the first “qualification opinion” by the EMA’s human medicines committee (Committee for Medicinal Products for Human Use [CHMP]) on an innovative development methodology based on AI. It means that the committee can now accept evidence generated by the tool as scientifically valid in future applications and assessments for new MASH treatments.

MASH is a condition linked with obesity, type 2 diabetes, high blood pressure, abnormal cholesterol, and belly fat. It causes fat buildup in the liver, leading to inflammation, irritation, and scarring, even without significant alcohol use or other reasons for liver injury. If untreated, MASH, previously called non–alcoholic steatohepatitis (NASH), can lead to advanced liver disease.

Addressing Screening Variability in Trials

Liver biopsy is the gold standard for testing new investigational medicines for MASH. However, specialists reviewing samples may not always agree on the severity of inflammation or scarring they see, resulting in high variability of interpretations. This creates a challenge for clinical trials.

Juan M. Pericàs, MD, MPH, PhD, team leader of the Liver, Metabolism and Infection group at Vall d’Hebron University Hospital, Barcelona, Spain, told Medscape Medical News that inter- and intraobserver disagreement on liver biopsy readings (especially for liver fibrosis stage and the presence and severity of hepatocyte ballooning) is the main cause of screening failure in MASH clinical trials, hindering interpretation of trial outcomes.

The role of the expert pathologist will continue to be critical, he added, but the homogenous use of this type of tool “will likely lead to a more efficient and patient-centric scenario where investigators would be able to make the most of [a] liver biopsy.” Pericàs was not involved in the tool’s development.

How It Works

The new tool, called AI-based measurement (AIM)-NASH, uses a machine learning model that was trained on more than 100,000 annotations from 59 pathologists who assessed over 5000 liver biopsies in nine large clinical trials.

An EMA spokesperson explained to Medscape Medical News that AIM-NASH can analyze whole image scans of liver biopsy specimens used in clinical trials investigating MASH treatments. The tool can help determine disease activity by evaluating steatosis, lobular inflammation, and hepatocellular ballooning, in addition to the fibrosis stage. It is used by a single central pathologist for enrolling patients into clinical phase 2 and phase 3 trials and for evaluation of study outcomes.

Evidence submitted to CHMP showed that AIM-NASH biopsy readings could reliably determine MASH disease activity as verified by a pathologist. It offers less variability than the methodology currently applied in MASH clinical trials, which relies on consensus by panels of three independent human pathologists. This, in turn, increases reproducibility and repeatability of histologic assessments, the EMA told Medscape Medical News.

The tool will help researchers “obtain clearer evidence on the benefits of new treatments in clinical trials that include fewer patients. Ultimately, this may bring effective treatments to patients faster,” the EMA added.

Pericàs said the development could boost the validation of noninvasive tests and biomarkers in a field that is steadily moving away from histology as a primary endpoint, particularly in routine clinical practice and clinical trials of MASH cirrhosis.

As a result of the CHMP’s opinion, the tool is now “locked,” meaning it cannot be modified or replaced without requalification by the committee.

Pericàs reported no relevant financial relationships.

Sheena Meredith is an established medical writer, editor, and consultant in healthcare communications, with extensive experience writing for medical professionals and the general public. She is qualified in medicine and in law and medical ethics.