Global Variability in Non-Invasive Testing Accuracy for MASLD: Key Findings from the G-MASLD Study

Highlight

The G-MASLD study, involving 17,792 MASLD patients across 41 countries, assessed the global performance of non-invasive tests (NITs) to diagnose liver fibrosis stages. Key highlights include: 1) Significant regional variation in the diagnostic accuracy of common NITs such as FIB-4, ELF, and LSM for detecting advanced liver fibrosis and cirrhosis; 2) Agile-3+ and Agile-4 composite scores demonstrated superior and consistent diagnostic performance worldwide; 3) Non-invasive tests still show limitations, particularly in North America and Latin America, indicating the need for region-specific adaptation or further refinement of NIT algorithms.

Study Background

Metabolic dysfunction-associated steatotic liver disease (MASLD), previously termed non-alcoholic fatty liver disease (NAFLD), represents a spectrum of liver pathology characterized by fat accumulation associated with metabolic dysfunction. It is a leading cause of chronic liver disease globally and a major health burden due to its progression potential toward advanced fibrosis, cirrhosis, and hepatocellular carcinoma. Detecting liver fibrosis non-invasively is critical for disease management and risk stratification, as liver biopsy, the diagnostic gold standard, is invasive with inherent risks and limited accessibility.

Various non-invasive tests (NITs), including serum biomarkers (FIB-4, ELF) and imaging-based methods like liver stiffness measurement (LSM) using transient elastography (FibroScan), have been developed to estimate fibrosis stage. Composite scores such as FibroScan-AST (FAST) and Agile-3+/Agile-4 combine clinical, laboratory, and imaging parameters to enhance accuracy. However, their global diagnostic performance and regional variability have not been thoroughly evaluated on a large scale before the G-MASLD study.

Study Design

The G-MASLD study is a large-scale, multinational observational analysis coordinated through the Global NASH Council. It enrolled 17,792 MASLD patients who underwent liver biopsy alongside concomitant non-invasive fibrosis assessments. Patient data were collected from 41 countries spanning multiple continents, encompassing diverse ethnicities and healthcare settings.

Non-invasive tests examined included the fibrosis-4 index (FIB-4), enhanced liver fibrosis (ELF) score, liver stiffness measurement (LSM) via FibroScan, and derived composite scores such as FibroScan-AST (FAST) and Agile-3+ and Agile-4 algorithms. The main endpoints were the diagnostic accuracy of these NITs for identifying clinically relevant fibrosis stages: significant fibrosis (≥F2), advanced fibrosis (≥F3), and cirrhosis (F4) based on histological classification.

Key Findings

The patient cohort had a fibrosis distribution: 14% F0 (no fibrosis), 32% F1, 18% F2 (significant fibrosis), 22% F3 (advanced fibrosis), and 13% F4 (cirrhosis). Almost half (48%) had a NAS (NAFLD Activity Score) ≥5, indicating high disease activity. The study found marked differences in diagnostic accuracy of NITs across geographic regions:

• FIB-4 Index: The pooled area under the receiver operating characteristic curve (AUC) for advanced fibrosis (≥F3) was 0.80 (95% CI, 0.79–0.81). Regionally, Latin America had the lowest AUC of 0.75 (0.71–0.79), while the Middle East and North Africa (MENA) region had the highest at 0.84 (0.82–0.87).
• ELF Score: Overall pooled AUC was 0.77 (0.76–0.79), with Europe showing the lowest performance (AUC 0.72; 0.69–0.76) and North America the highest (AUC 0.80; 0.78–0.82).
• Liver Stiffness Measurement (LSM): LSM showed a pooled AUC of 0.84 (0.83–0.85) for detecting advanced fibrosis with relatively consistent performance worldwide, except for a lower AUC in North America (0.78; 0.76–0.81).
• FAST and Agile Scores: FAST score had an AUC of 0.75 (0.74–0.76), while Agile-3+ achieved an AUC of 0.87 (0.86–0.88) with stable performance across regions. For cirrhosis (F4), Agile-4 demonstrated excellent accuracy (AUC 0.90; 0.89–0.91), though performance was lower in North America (AUC 0.85; 0.83–0.87) and highest in MENA (AUC 0.96; 0.94–0.98).

These data highlight that composite scores combining multiple biomarkers and imaging (Agile-3+ and Agile-4) outperform single biomarker or elastography-based tests and that regional factors such as genetic diversity, comorbidities, and healthcare practices may influence test accuracy.

Expert Commentary

The G-MASLD study provides the most comprehensive global evaluation of NITs in MASLD to date. The observed regional variability calls for cautious interpretation of test results and suggests the need for contextual calibration or regional validation cohorts to optimize clinical decision-making.

While liver biopsy remains the reference standard, its impracticality in routine care necessitates reliable NITs. Agile-3+ and Agile-4 scores represent promising tools for fibrosis staging, integrating clinical and biochemical variables with LSM, achieving higher diagnostic accuracies especially for critical fibrosis stages impacting prognosis and treatment decisions.

Limitations of the study include potential heterogeneity in biopsy assessment, retrospective data collection, and the influence of underlying liver disease etiology that may vary by region. The diminished performance in North America could reflect differences in patient demographics, prevalence of comorbidities such as diabetes or obesity, or biopsy quality and scoring practices.

Conclusion

The G-MASLD study underscores significant global heterogeneity in the diagnostic performance of NITs for fibrosis in MASLD. Agile-3+ and Agile-4 composite scores demonstrated superior accuracy and cross-regional consistency, advocating their wider clinical adoption for fibrosis risk stratification. However, regional variations in test performance highlight the necessity for localized validation and potential modification of diagnostic thresholds.

Future research should focus on refining NITs with machine learning approaches incorporating diverse populations, improving standardization of fibrosis assessment, and integrating novel biomarkers to enhance early detection and personalized management of MASLD worldwide.

Funding and Registrations

This large-scale collaborative research was facilitated by the Global NASH/MASLD Council (GNC) and supported by multiple international institutions. Specific funding sources were not detailed in the study abstract.

References

Younossi ZM, de Avila L, Petta S, et al. Global performance of non-invasive tests in MASLD: Insights from the G-MASLD study. Hepatology. 2025;84(1):161-174. PMID: 41100867. https://pubmed.ncbi.nlm.nih.gov/41100867/