A cell-free DNA methylation test identified treatment-eligible patients more accurately than current clinical guidelines in a 528-patient study.
Hepta, a liquid biopsy company, released clinical data showing its blood test identifies patients with metabolic dysfunction-associated steatohepatitis (MASH) eligible for treatment more accurately than the three non-invasive tests currently recommended by clinical guidelines.
The study compared the performance of Hepta’s test against FIB-4, ELF, and FibroScan. The results, involving a 528-patient multi-cohort study, were presented at the European Association for the Study of the Liver 2026 Congress by Jörn Schattenberg, MD, director of the department of gastroenterology, hepatology, endocrinology, and nutritional medicine at Saarland University Medical Center.
While the drugs Rezdiffra and Wegovy are approved for MASH with moderate-to-advanced fibrosis (stages F2 and F3), identifying these specific patients has remained a clinical challenge. Current non-invasive blood tests were originally designed to identify more advanced F3 and F4 stages and fail to detect nearly one-third of the patients now eligible for therapy, according to a press release from Hepta.
Using AI to Analyze DNA Methylation
The study utilized Hepta’s LiquidTransformer platform, an artificial intelligence model with more than 100 million parameters designed for cell-free DNA (cfDNA) methylation data. This technology analyzes epigenetic marks on DNA fragments shed from the liver and other organs into the bloodstream, providing a molecular readout of disease biology from a standard blood draw.
The platform was trained on 349 samples and validated on an independent cohort of 179 patients from a Phase 3 program. In head-to-head comparisons, the platform achieved an area under the curve of 0.83 in detecting significant fibrosis. It also outperformed FIB-4, ELF, and FibroScan in separating treatment-eligible F2/F3 patients from those with earlier-stage or cirrhotic disease.
“Existing non-invasive tests for MASH predate effective therapies and were never optimized for treatment selection,” says Hamed Amini, PhD, CEO and co-founder of Hepta, in a release. “We built a classifier for the clinically actionable question of identifying treatment-eligible patients. On hold-out validation, head-to-head on the same samples, it outperforms FIB-4, ELF, and FibroScan.”
Clinical Implications for Treatment Monitoring
The validation cohort was collected across more than 50 sites and remained independent of the training data. According to the company, this design rules out overfitting and supports the generalization of results to new patient populations.
The findings follow additional data presented at Digestive Disease Week 2026, which suggested the same cfDNA methylation approach could identify weight-loss responders before treatment and track biological changes over time.
“The ultimate blood test for MASH needs to do three things: detect treatment-eligible disease, guide therapy selection, and monitor patient response over time to signal whether a therapy is working,” says Soheil Damangir, PhD, chief technology officer and co-founder of Hepta, in a release. “This data establishes that we can identify the patients who should be on therapy more accurately than the tests in current guidelines.”
ID 53986442 © Seoterra | Dreamstime.com
