Researchers have identified multiple causal biomarkers for metabolic dysfunction-associated steatotic liver disease (MASLD), using genetic tools and machine learning models. The findings, published in the Journal of Clinical and Translational Hepatology, offer a framework for future diagnostic and risk stratification strategies, though the clinical use of these biomarkers will require further investigation.

Genetic screening reveals causal associations

The study examined over 2,900 molecular biomarkers derived from proteomic data and 35 clinical biomarkers using Mendelian randomization, a method that assesses potential causal relationships using genetic variation. This approach was used to determine which biomarkers are likely to influence the development of MASLD, a condition formerly known as nonalcoholic fatty liver disease (NAFLD).

Six molecular biomarkers were found to be causally associated with MASLD. These included canopy FGF signaling regulator 4 (CNPY4), ectonucleoside triphosphate diphosphohydrolase 6 (ENTPD6), and the human leukocyte antigen gene HLA-A. In parallel, eight clinical biomarkers were identified, including serum total protein (STP).

Protein mediation and external validation

Mediation analysis showed that STP levels partially explained the relationship between HLA-A and MASLD, suggesting that STP may play a role in immune-metabolic pathways. The mediation accounted for approximately 24% of the total effect.

To assess the reproducibility of the findings, researchers conducted external validation using clinical data from a hospital-based cohort of 415 individuals. In this group, STP was positively associated with MASLD risk. Specifically, for each unit increase in STP, the odds of MASLD rose by 8%, with a confidence interval that excluded the null value.

Diagnostic model shows high accuracy

A random forest machine learning model trained on the molecular biomarker data demonstrated strong diagnostic performance. The model achieved an area under the receiver operating characteristic curve (AUC) of 0.941 in training and 0.875 in validation. These results suggest that molecular biomarkers could contribute to the development of noninvasive diagnostic tools for MASLD.

Links to liver cancer and overall survival

Two molecular biomarkers, CNPY4 and ENTPD6, were also linked to the development of hepatocellular carcinoma (HCC), a form of primary liver cancer. Elevated levels of these proteins were associated with poorer survival outcomes in patients with HCC, highlighting their potential prognostic value.

Low STP levels, defined as under 60 grams per liter, were also found to be predictive of all-cause mortality. Individuals with low STP had more than twice the risk of death compared to those with higher levels, based on hazard ratios calculated during follow-up.

A multi-omics approach to MASLD

The study supports a multi-omics approach that combines genetic, proteomic, and clinical data to better understand the biological mechanisms underlying MASLD. While the findings highlight potential biomarkers for diagnosis and prognosis, the authors note that further research is needed to assess how these markers perform across diverse populations and clinical settings.


Reference: Feng G, Targher G, Byrne CD, et al. Biomarker discovery for metabolic dysfunction-associated steatotic liver disease utilizing Mendelian randomization, machine learning, and external validation. J Clin Transl Hepatol. 2025. doi: 10.14218/JCTH.2025.00270

This content includes text that has been generated with the assistance of AI. Technology Networks’ AI policy can be found here.