MASH/NASH: Why Therapeutic Breakthroughs Demand a New Era of Patient Identification

December 5, 2025

Clinical DevelopmentCommercial ExcellenceDiseases Earlier Detection Insights Liver Disease Medical AffairsPredicting patient outcomes Stratifying patient cohorts Value & Access HEOR

MASH/NASH: Why Therapeutic Breakthroughs Demand a New Era of Patient Identification

By Léon Van Wouwe, Clinical Innovation Director, Volv Global

Introduction: A Silent Disease Becoming a Public Health Emergency

Metabolic dysfunction–associated steatohepatitis (MASH), historically known as non-alcoholic steatohepatitis (NASH), has rapidly shifted from an obscure hepatology concern to one of the fastest-growing causes of advanced liver disease worldwide. Around one-third of adults globally are estimated to have metabolic dysfunction–associated steatotic liver disease (MASLD/NAFLD), with a meaningful subset progressing to inflammatory, fibrotic MASH. [1,2]

Prevalence is even higher in at-risk groups such as people with obesity or type 2 diabetes, where NAFLD prevalence exceeds 50% in many cohorts. [3] MASH has become a leading contributor to liver transplantation and liver-related morbidity, particularly among women, and is projected to increase further over the next decade. [4]

Yet despite its clinical importance, MASH remains chronically underdiagnosed. Most patients are asymptomatic or present with non-specific features such as fatigue or mildly abnormal liver enzymes. [5–7] Diagnostic pathways frequently depend on opportunistic discovery (e.g., incidental imaging findings) and invasive liver biopsy in selected cases. Typical patterns include:

Age at onset of underlying steatosis: often in the 30s–40s
Age at recognition of clinically significant disease: commonly late 40s–50s
Diagnostic delay: many patients are only diagnosed when fibrosis is already advanced or complications have emerged [5–7]
Common misattribution: findings labelled simply as “fatty liver,” or overshadowed by diabetes, obesity, or cardiovascular disease

These delays are not trivial. Fibrosis stage is the strongest predictor of liver-related outcomes in MASH; once advanced fibrosis or cirrhosis is established, the opportunity for disease modification narrows. [1,3,5]

Patient Journey and Disease Course: Heterogeneity and Missed Inflection Points

The typical MASH patient begins their journey unknowingly. The spectrum of metabolic dysfunction–associated steatotic liver disease runs from simple steatosis through MASH to progressive fibrosis, cirrhosis, and hepatocellular carcinoma (HCC). [3] Most patients traverse this trajectory without clear clinical signals until late stages.

Importantly, disease progression is heterogeneous: some individuals remain stable for years, while others progress rapidly to decompensated cirrhosis or HCC within a relatively short timeframe. [3,8] Real-world analyses show that a substantial minority of patients with MASH develop features of end-stage liver disease over just a few years of follow-up, underscoring the unpredictability and systemic burden of the condition. [8]

From a health-system perspective, the patient journey is misaligned with therapeutic opportunity:

Early, largely silent phase
Steatosis and low-grade inflammation often coexist with obesity, diabetes, dyslipidaemia, or hypertension. Without proactive risk-based screening, these patients are rarely recognized as having a liver disease requiring active management. [5–7]
Intermediate phase with emerging fibrosis
This is the critical inflection point: fibrosis is accumulating, but cirrhosis has not yet developed. Therapeutic interventions—including lifestyle modification and pharmacotherapy—have the greatest potential to alter long-term outcomes here. [3,5]
Advanced fibrosis and cirrhosis
At this stage, patients finally present more clearly—portal hypertension, decompensation, or HCC. Interventions become more complex and less effective, and costs rise sharply. [3,4,8]

Within this journey, outcome prediction becomes as important as diagnosis itself. Clinicians are being asked to decide:

Which patients are fast progressors with a high risk of fibrosis progression and complications?
Who is unlikely to respond adequately to lifestyle measures alone?
How can we distinguish patients who need immediate pharmacotherapy from those who can be safely monitored?

Emerging work using data-driven clustering and artificial intelligence has shown distinct clinical phenotypes associated with rapid fibrosis progression in MASH. [9] This underscores a central message for this series: we must evolve from simple detection to trajectory-aware, prognosis-informed care.

Bridging Innovation and Identification: The Case of Madrigal’s Resmetirom

The approval of resmetirom (Rezdiffra™) marked a watershed in the management of this disease. In March 2024, the US Food and Drug Administration (FDA) approved resmetirom as the first pharmacologic treatment for adults with noncirrhotic NASH/MASH with moderate to advanced liver scarring (F2–F3), to be used alongside diet and exercise. [10,11]

Resmetirom, a liver-directed thyroid hormone receptor-β agonist, has shown clinically meaningful improvements in steatohepatitis and fibrosis markers in pivotal trials, and regulators in Europe have since recommended conditional approval, signalling global momentum. [10,12,13]

At the same time, GLP-1–based and other investigational therapies are entering or expanding within the MASH space, including assets from Novo Nordisk, Eli Lilly, 89bio, Akero Therapeutics, Boehringer Ingelheim, and Boston Pharmaceuticals, among others. Wegovy (semaglutide) has recently received an additional indication in the US for MASH with fibrosis, becoming the second approved therapy in this setting. [14]

This is a striking moment: the therapeutic bottleneck is beginning to lift. But these innovations immediately expose the fragility of current identification practices:

Resmetirom is indicated for noncirrhotic disease with F2–F3 fibrosis – precisely the stage at which many patients are not yet recognised or accurately assessed for fibrosis severity. [10,11]
Non-invasive tests (NITs) such as FIB-4, VCTE, and ELF are valuable but imperfect, with misclassification risks that complicate treatment decisions. [4,5]
Claims and real-world data show that patients with MASH are frequently diagnosed late or coded inconsistently, limiting systematic identification of eligible populations. [7,15]

To realise the promise of resmetirom and the broader pipeline, diagnostic prediction and outcome prediction must evolve in lockstep with therapeutics.

1. Earlier detection using multimodal algorithms

Machine-learning approaches have demonstrated that routinely collected data – demographics, comorbidities, laboratory values, medication patterns – can be used to identify patients with likely undiagnosed MASH/NASH from large at-risk populations. [15–17] Models trained on claims and electronic health records have successfully flagged patients for further diagnostic work-up, reducing reliance on chance discovery.

2. Probabilistic fibrosis staging

Advanced imaging and serum biomarkers are now being combined with AI-based tools to refine fibrosis staging. Regulatory bodies have even endorsed AI-supported histologic assessment (e.g., AIM-NASH) to reduce variability in biopsy interpretation and improve the reliability of treatment effect assessment in trials, a paradigm that can ultimately inform clinical practice. [18]

3. Outcome prediction for treatment optimisation

Beyond “does this patient have MASH?”, clinicians increasingly need tools that answer “how will this patient’s disease behave?” AI-driven models that classify patients into phenotypes with different risks of rapid fibrosis progression and adverse outcomes are beginning to emerge. [8,9,13] These models can underpin decisions such as:

Earlier initiation of resmetirom or GLP-1–based therapies in high-risk phenotypes
Prioritisation of patients for hepatology referral or more intensive NIT-based monitoring
Selection and enrichment strategies for clinical trials of combination regimens

4. Precision-based patient selection in real-world practice

For health systems, the long-term value of resmetirom and pipeline agents will depend on a precision hepatology infrastructure capable of:

Systematically identifying at-risk patients from primary care and metabolic clinics
Applying non-invasive, prediction-enhanced staging algorithms to define fibrosis and risk
Matching specific mechanisms of action (e.g., thyroid hormone receptor-β agonists, GLP-1 agonists, FGF21 analogues) to the right subgroups at the right time

In other words, therapeutic innovation and predictive innovation must be co-designed. Without earlier detection and robust risk stratification, even the most effective MASH therapies will be deployed too late, in too few patients, to meaningfully alter the population-level trajectory of liver disease.

A Convergence Point for the Future of MASH Care

MASH is poised to become one of the defining metabolic liver diseases of this century, with substantial clinical and economic burden. [1,2,4] The arrival of resmetirom and subsequent therapies has shifted the narrative from “we have nothing to offer” to “we must decide whom to treat, when, and how.”

If diagnosis continues to occur late and opportunistically, these therapies will be reserved for advanced disease that might otherwise have been prevented or attenuated. But if we build a prediction-enabled ecosystem – integrating AI-based early detection, non-invasive fibrosis staging, and outcome prediction into routine care – then resmetirom and its peers can fundamentally change the natural history of MASH.

The central theme of this series [LC1] is clear in MASH more than almost any other disease area:

Novel therapies only achieve their full value when paired with better diagnostic and prognostic prediction.

Reflective Questions

Where in the MASH patient journey could predictive modelling create the largest near-term impact: early detection in at-risk populations, fibrosis staging, or treatment response prediction?

What policy, reimbursement, and workflow barriers currently prevent integration of non-invasive diagnostics and AI-enabled outcome prediction into routine metabolic and primary care pathways?

About the author

Léon van Wouwe has 20+ years’ global experience in clinical development and operations, uniting data science with pharma and research. He drives cross-functional collaboration to advance innovative treatments.