Introduction
Metabolic dysfunction-associated steatohepatitis (MASH) is a severe liver condition linked to metabolism disorders that significantly increases risks of liver-related complications and overall morbidity and mortality. In individuals with obesity, the prevalence and impact of MASH are even more pronounced, necessitating accurate and accessible diagnostic tools. Traditional diagnosis involves liver biopsy, which, while the gold standard, is invasive, costly, and not feasible for large-scale screening. This study aimed to develop and validate a noninvasive scoring system that identifies at-risk MASH in obese populations, particularly among those undergoing bariatric surgery.
Background and Rationale
Obesity is a well-known risk factor for the development of nonalcoholic fatty liver disease (NAFLD) and its progressive form, MASH. MASH predisposes patients to serious liver-related outcomes such as cirrhosis, hepatocellular carcinoma, and liver failure, which contribute to substantial health burdens. Early identification of at-risk individuals is critical to initiating timely interventions, potentially preventing progression and associated complications.
However, liver biopsy presents limitations due to its invasiveness, sampling variability, and patient reluctance. Thus, there is a pressing need for simple, reliable, and noninvasive diagnostic methods utilising routine clinical and laboratory data for broader application in clinical practice.
Study Design and Methods
The research utilized data collected from 1,961 individuals across five independent bariatric surgery cohorts with biopsy-confirmed liver assessments. One cohort (n=1,095) served as the derivation set to develop the predictive scoring model, termed FMO (Fibrotic/at-risk MASH in Obesity). Internal validation was performed through bootstrapping techniques, while external validation utilized the remaining four cohorts (n=866).
Further validation extended internationally to overweight and obese populations within the UK Biobank (n=15,745) and the US NHANES database (n=1,573). Additionally, the model’s ability to predict severe liver-related outcomes (SLROs), such as cirrhosis and hepatocellular carcinoma, was tested in a UK Biobank subcohort (n=1,955) with a median 13.7-year follow-up period.
Development of the FMO Score
The FMO score integrates four routinely measured clinical indicators: aspartate aminotransferase (AST), alanine aminotransferase (ALT), triglycerides, and high-density lipoprotein cholesterol (HDL-C). Each component reflects important aspects of liver function and metabolic health:
- AST and ALT: Enzymes indicative of liver cell injury and inflammation.
- Triglycerides: Elevated levels correlate with metabolic dysregulation and fatty liver accumulation.
- HDL cholesterol: Lower levels are often associated with increased cardiovascular and metabolic risks.
Using statistical modeling approaches, these variables were combined to produce a score that stratifies patients by their risk of having MASH with fibrosis.
Validation Results
The FMO model demonstrated high discriminative accuracy with an area under the receiver operating characteristic curve (AUROC) of 0.874 (95% CI: 0.844–0.905) in the derivation cohort. External validation cohorts showed AUROCs ranging from 0.803 to 0.874, confirming robustness across diverse populations.
When tested in the global cohorts—the UK Biobank and NHANES datasets—the FMO maintained good predictive performance (AUROCs of 0.753 and 0.866, respectively). Longitudinal data analysis in the UK Biobank subcohort established the model’s practical value in predicting serious liver-related outcomes over more than a decade, with a Harrell C-index of 0.703.
Clinical Performance and Thresholds
In clinical application, the FMO score offers distinct cutoff values for ruling out or ruling in at-risk MASH:
- Rule-out cutoff (0.05): Sensitivity ≥90% and negative predictive value (NPV) of 97.6%, ensuring patients below this threshold are unlikely to have at-risk MASH.
- Rule-in cutoff (0.22): Specificity ≥90% and positive predictive value (PPV) of 48.1%, identifying patients at high risk who may need further evaluation or intervention.
External validations also showed strong NPVs (0.907–1.00) and PPVs (0.333–0.630), demonstrating the model’s utility in various clinical settings.
Comparison with Existing Models
The FMO score outperformed several previously established indices in diagnostic accuracy. Its simplicity, based on standard laboratory tests, and strong validation across independent and international cohorts underscore its superiority and potential for widespread adoption.
Implications for Practice
The adoption of the FMO scoring system can significantly enhance early identification of individuals with obesity at risk for MASH, enabling timely lifestyle, pharmacological, or surgical interventions such as bariatric surgery. It provides clinicians with a cost-effective, noninvasive tool to stratify patients and prioritize those needing detailed assessment or intensive management.
Moreover, its validation in large populations supports its role in public health and epidemiological surveillance, potentially impacting guidelines for metabolic and liver disease screening in obese individuals.
Limitations and Future Directions
While promising, the FMO score requires further prospective studies to evaluate its predictive capacity in routine clinical workflows and diverse ethnic groups. Integration with imaging modalities and novel biomarkers may further enhance accuracy. Continuous updates incorporating emerging knowledge about MASH pathophysiology may improve its diagnostic power.
Conclusion
The FMO score is a validated, accurate, and cost-effective noninvasive diagnostic tool for identifying at-risk metabolic dysfunction-associated steatohepatitis in individuals with obesity undergoing bariatric surgery. Its widespread implementation could transform current clinical practice by facilitating early diagnosis and enabling interventions that reduce progression to severe liver disease and associated mortality.