Dear Editor,
A recently published study in
Clinical and Molecular Hepatology offers valuable insights into the application of the Korean Association for the Study of the Liver (KASL)’s two-step approach for risk stratification in patients with metabolic-associated steatotic liver disease (MASLD). Using a large real-world cohort of over 8,000 individuals with a median follow-up of nearly four years, the authors demonstrated that combining the fibrosis-4 (FIB-4) index with vibration-controlled transient elastography can effectively identify patients at varying risk for liver-related events (LREs), such as hepatic decompensation and hepatocellular carcinoma (HCC) [
1]. While the findings support the practicality and prognostic utility of this noninvasive strategy, deeper interrogation of the methodology and interpretation reveals areas in need of refinement and further exploration.
A central issue arises from the initial four-tier risk classification, which is defined by combinations of FIB-4 and liver stiffness (LS) thresholds: low, intermediate-low, intermediate-high, and high risk. While the risk of LREs increased significantly from the low-risk group to the high-risk group, the lack of clear differences in fibrotic burden between the low-risk category and the intermediate-low-risk category, in terms of both LS and composite indices such as Agile 3+ and FibroScan aspartate aminotransferase (FAST), calls into question the usefulness of this classification. The study appropriately explored a simplified three-group scheme that merged the low- and intermediate-low-risk categories. This scheme preserved similar prognostic accuracy while enhancing clinical usability. This finding echoes longstanding concerns in hepatology about balancing diagnostic precision with real-world feasibility. In particular, over-fragmentation of risk strata may lead to confusion in primary care settings, where decisions must be intuitive and actionable. However, while simplification may facilitate clinical translation, it is crucial to consider whether such streamlining diminishes the capacity to detect subtle yet clinically signifi-cant variations, particularly among patients with borderline risk.
Another important limitation of the current framework is its reliance on a static baseline assessment. Fibrosis progression in MASLD is nonlinear, and patients may experience shifts in fibrotic burden over time due to evolving metabolic status, comorbidities, and therapeutic interventions. However, despite acknowledging previous work showing that serial increases in these markers are strong predictors of future decompensation or HCC, the study did not incorporate longitudinal measurements of FIB-4 or LS. For example, Loomba et al. demonstrated that changes in LS, independent of baseline fibrosis stage, are associated with progression to cirrhosis and liver-related complications [
2]. Similarly, a persistent rise in FIB-4 has been linked to a greater than 50-fold increase in HCC risk. The absence of longitudinal tracking in the current study limits the algorithm’s ability to identify high-risk patients whose fibrosis trajectory accelerates despite an initially low-risk profile. In clinical practice, especially for patients in intermediate-risk categories, dynamic monitoring could play a pivotal role in identifying individuals who could benefit from early intervention.
Furthermore, the apparent discrepancy between steatotic burden, as measured by controlled attenuation parameter, and fibrotic indices among the various risk groups raises significant pathophysiological and diagnostic concerns. While the authors observed differences in controlled attenuation parameter values between groups, these differences were not consistently aligned with fibrosis progression. This is not unexpected, as fibrosis and steatosis often evolve independently in MASLD. For example, patients with advanced fibrosis may exhibit low levels of hepatic fat due to hepatocyte loss, and those with substantial steatosis may have minimal fibrotic involvement. This complex interplay highlights the limitations of using steatosis as a surrogate for disease severity and suggests that risk models focusing on fibrosis rather than combined fibrosis-steatosis indices may be more appropriate for predicting long-term outcomes. However, the cross-sectional nature of the current data restricts our understanding of how these processes interact over time, particularly under the influence of weight loss, glycemic control, or pharmacological interventions.
Despite the mathematical complexity of these composite indices, the study revealed only modest differences in predictive accuracy when comparing the KASL algorithm with other two-step approaches incorporating Agile 3+, Agile 4, or FAST scores. Notably, Agile 3+ had the highest area under the curve, suggesting its potential for refining high-risk identification. However, the increased computational burden and limited availability of the Agile calculator in routine settings may constrain its practical adoption. Integrating these scores into electronic health records or creating simplified look-up tables could help address this issue. Alternatively, the authors proposed a conditional use model wherein more complex indices, such as Agile, are reserved for patients with discordant FIB-4 and LS values. Further investigation of this selective deployment strategy is warranted, as it may strike a balance between diagnostic performance and workflow efficiency.
In addition to these two-step algorithms, alternative strategies could be considered to complement or refine the current risk stratification of MASLD. For example, incorporating serial measurements of FIB-4 or LS over time into a dynamic risk assessment may more accurately capture disease progression. Additionally, integrating biomarkers beyond fibrosis and steatosis, such as markers of inflammation, apoptosis, or metabolic stress, could provide a more comprehensive risk profile. Machine learning approaches that synthesize clinical, laboratory, and imaging data could enhance predictive performance, especially in intermediate-risk populations. Finally, applying simpler indices for broad screening and more complex scores for discordant cases in a tiered manner may optimize diagnostic precision and clinical feasibility.
Another important consideration is the limited effectiveness of vibration-controlled transient elastography (VCTE) in patients with MASLD. While VCTE is an established tool for non-invasively assessing liver stiffness, its performance may be affected by common MASLD factors, such as obesity and high body mass index. These factors can result in unreliable or failed measurements, reducing diagnostic accuracy. Furthermore, VCTE sometimes overestimates liver stiffness compared to the actual fibrotic burden, particularly in the presence of hepatic inflammation or congestion. For example, studies have shown that severe hepatic steatosis, inflammation, and hepatic congestion can elevate liver stiffness values measured by VCTE independently of the actual fibrosis stage [
3,
4]. These findings underscore the importance of carefully interpreting VCTE results, especially in patients with metabolic comorbidities that can transiently increase liver stiffness. Operator variability remains a concern as well, underscoring the importance of standardized protocols and operator training. Despite these limitations, VCTE is a valuable tool when used with other noninvasive tests. However, its performance in specific subgroups and the potential for overestimation warrant further exploration.
One final consideration is the low overall incidence of LREs in this cohort, even among high-risk individuals, which raises the question of whether current non-invasive tools are sensitive enough to detect risk in MASLD. Compared to biopsy-based cohorts, which often include patients referred for suspected advanced disease, VCTEbased registries may capture a broader population with less severe disease. While this improves generalizability, it may also reduce event rates and decrease the algorithm’s ability to distinguish true progression. Indeed, even among high-risk patients, the seven-year incidence of HCC was 6.1%, and the incidence of decompensation was 8.0% [
5]. These rates are lower than those reported in cohorts with histologically confirmed advanced fibrosis. These findings suggest that, although non-invasive strategies offer significant clinical utility, they may underestimate risk in a subset of patients, particularly those with non-cirrhotic nonalcoholic steatohepatitis who remain biologically primed for progression. Future studies should consider combining non-invasive tests with biomarkers of inflammation, apoptosis, or fibrosis to improve sensitivity, especially in early disease stages.
In summary, this study confirms the real-world usefulness of the KASL two-step strategy as a scalable, reproducible, and clinically accessible tool for MASLD risk stratification. However, several limitations—notably, the absence of dynamic monitoring, limited discriminative power between adjacent risk groups, and modest event rates— highlight areas for refinement. Improving risk prediction will likely require integrating longitudinal data, incorporating biomarkers beyond fibrosis and steatosis, and tailoring model complexity to specific clinical contexts. As the global burden of MASLD continues to increase, it is essential to develop robust yet implementable risk stratification algorithms to prevent progression to cirrhosis and HCC.
FOOTNOTES
-
Authors’ contributions
Chunyan Wang and Jun Sun wrote the manuscript, Yinyan Li provided methodological and revised the manuscript.
-
Conflicts of Interest
The authors have no conflicts to disclose.
Abbreviations
FibroScan aspartate aminotransferase
metabolic-associated steatotic liver disease
vibration-controlled transient elastography
REFERENCES
- 1. Lee HW, Lee JS, Kim MN, Kim BK, Park JY, Kim DY, et al. Risk stratification by noninvasive tests in patients with metabolic dysfunction-associated steatotic liver disease. Clin Mol Hepatol 2025;31:1018-1031.
- 2. Loomba R, Adams LA. Advances in non-invasive assessment of hepatic fibrosis. Gut 2020;69:1343-1352.
- 3. Castéra L, Foucher J, Bernard PH, Carvalho F, Allaix D, Merrouche W, et al. Pitfalls of liver stiffness measurement: a 5-year prospective study of 13,369 examinations. Hepatology 2010;51:828-835.
- 4. Millonig G, Reimann FM, Friedrich S, Fonouni H, Mehrabi A, Büchler MW, et al. Extrahepatic cholestasis increases liver stiffness (FibroScan) irrespective of fibrosis. Hepatology 2008;48:1718-1723.
- 5. Angulo P, Kleiner DE, Dam-Larsen S, Adams LA, Bjornsson ES, Charatcharoenwitthaya P, et al. Liver fibrosis, but no other histologic features, is associated with long-term outcomes of patients with nonalcoholic fatty liver disease. Gastroenterology 2015;149:389-397.e10.
Citations
Citations to this article as recorded by
- Correspondence to letter to the editor on “Risk stratification by noninvasive tests in patients with metabolic dysfunction-associated steatotic liver disease”
Hye Won Lee, Seung Up Kim
Clinical and Molecular Hepatology.2026; 32(2): e244. CrossRef
