NITs in MASH: disease-activity vs response

A letter published on 5 July 2026 in Alimentary Pharmacology and Therapeutics (PMID 42402644) raises a methodological question about how non-invasive tests are being used in semaglutide MASH trials^[1]. The authors, Guofu Mao and Haizhen Lan from Danyang People's Hospital, China, argue that after semaglutide treatment, these tests should be read as disease-activity readouts rather than as response biomarkers that predict long-term histological benefit.

The distinction is a methodological one, not a claim that semaglutide fails in MASH. Understanding it matters for anyone reading trial results or wondering why investigators keep coming back to the biopsy question.

What non-invasive tests measure in MASH

MASH (metabolic dysfunction-associated steatohepatitis) is formally diagnosed by liver biopsy: a pathologist grades hepatocyte injury, lobular inflammation, steatosis, and fibrosis stage. Because repeat biopsies are invasive, expensive, and carry a small but real procedural risk, researchers and clinicians use non-invasive tests (NITs) as stand-ins. Common ones include the FIB-4 score (a formula combining age, platelet count, and two liver enzymes), liver stiffness measured by elastography, and serum alanine aminotransferase (ALT)^[2]. The SAMARA trial, published in Alimentary Pharmacology and Therapeutics in April 2026, went further and used NIT-based criteria both to screen patients and to track response, reporting that FAST scores, ALT, AST, and MRI-PDFF tracked changes in the semaglutide arm compared with placebo^[3].

The appeal is obvious: if a NIT moves in the right direction after treatment starts, you know something is changing without putting a patient through a second biopsy. The question the Mao and Lan letter poses is what that movement actually tells you.

The argument: disease-activity readout versus response biomarker

A response biomarker, in the strict sense used in clinical trial design, predicts a long-term clinically meaningful outcome. If a NIT falls after semaglutide treatment and that fall reliably predicts fibrosis regression at year two or year five, it is serving as a response biomarker^[1]. The Mao and Lan letter argues that the existing data do not yet demonstrate this predictive relationship.

What NITs appear to reflect after semaglutide treatment is metabolic inflammation and hepatic fat reduction, both of which are genuine effects of the drug^[1]. Semaglutide drives weight loss, reduces hepatic steatosis, and lowers systemic inflammatory markers. These changes show up in NITs. But showing up in a NIT is not the same as predicting that the underlying fibrosis architecture of the liver is improving in a way that will persist or that will reduce the long-term risk of cirrhosis.

Fibrosis is a structural process driven by hepatic stellate cell activation. Disease activity (the degree of inflammation and fat) and fibrosis stage are related but not identical. A drug could suppress inflammation strongly, produce impressive NIT improvements, and still leave underlying fibrosis largely unchanged. The letter's concern is that reading NIT improvements as treatment response biomarkers risks conflating those two things.

Why this matters for interpreting semaglutide MASH data

The ESSENCE Phase 3 trial, published in the New England Journal of Medicine in June 2025, used histological endpoints: resolution of steatohepatitis without fibrosis worsening (62.9% on semaglutide vs 34.3% on placebo) and fibrosis improvement without steatohepatitis worsening (36.8% vs 22.4%)^[4]. Those biopsy-based results are the foundation of the FDA accelerated approval for semaglutide in noncirrhotic MASH with moderate-to-advanced fibrosis. The histological endpoint is the gold standard precisely because it avoids the ambiguity the Mao and Lan letter describes.

The concern becomes sharper when trials move away from mandatory biopsies and rely on NITs for primary endpoints. Several ongoing or planned MASH trials are attempting to do exactly that, in part to reduce patient burden and trial cost. If NITs are validated only as disease-activity readouts and not as response biomarkers, trials built around NIT endpoints alone may not tell regulators or clinicians what they need to know about long-term histological outcomes.

The SAMARA trial^[3] showed that NITs moved in the expected direction with semaglutide, which is useful evidence. What the Mao and Lan letter flags is that showing movement is the starting point of validation, not the end of it. The field still needs prospective data showing that a given NIT threshold, or a given magnitude of NIT change, corresponds to histological response that holds up on follow-up biopsy.

What this does not mean for semaglutide's efficacy

This methodological argument does not undermine semaglutide's liver evidence base. The biopsy-confirmed ESSENCE results are substantial: roughly two in three patients on semaglutide 2.4 mg weekly achieved steatohepatitis resolution without fibrosis worsening at 72 weeks, versus about one in three on placebo. Those are histological endpoints, not NIT-derived ones, and they were the basis for regulatory approval.

What the letter addresses is a forward-looking question about trial design: as the field moves toward biopsy-free or biopsy-reduced trials for MASH, the NIT endpoints being proposed need to demonstrate predictive validity before they can stand on their own. The authors are not arguing that semaglutide does not improve MASH. They are arguing that the tools used to measure improvement in future trials need more rigorous validation. That is a reasonable methodological position, and one thebroader NIT validation literature is starting to address.

This article is educational and is not medical advice. If you have questions about liver disease, MASH, or whether semaglutide may be appropriate for your situation, speak with a qualified hepatologist or gastroenterologist.