Related Experiment Video
Updated: Sep 26, 2026
Mouse Model of Metabolic Dysfunction-Associated Steatotic Liver Disease with Fibrosis
Published on: July 18, 2025
Optimisation of metabolic dysfunction-associated steatotic liver disease (MASLD) screening algorithm for
Chamila Mettananda1, Kaveesha Sivasumithran2, Lakmali Ranaweera3
1Faculty of Medicine, University of Kelaniya, Ragama, Sri Lanka.
Background:
The European Association for the Study of the Liver (EASL) metabolic dysfunction-associated steatotic liver disease (MASLD) screening algorithm involves two steps: initial screening with FIB-4, followed by referral for vibration-controlled transient elastography (VCTE) in patients likely to have significant fibrosis (SF). However, VCTE is not widely available in resource-limited settings.
Aim:
To optimise the EASL MASLD screening algorithm for resource-poor settings using machine learning (ML).
Methods:
We analysed data from 964 adults aged ≥35 years who underwent VCTE at a tertiary referral centre in Sri Lanka between November 2024 and 2025. Multiple ML models using different methods and variable combinations were trained on 80% of the dataset and tested on the remaining 20%. The best models were selected based on performance and externally validated on data from 430 patients who underwent VCTE before November 2024. Model performance was compared with that of the FIB-4 score using confusion matrices.
Results:
A Random Forest model incorporating age, AST, ALT, and platelet count separately, rather than using the FIB-4 score, outperformed in predicting SF. The model using all variables showed the best predictive performance for SF, with an AUC-ROC of 0.808. The variables used in the model, in descending order of feature importance, were AST, platelet count, BMI, diabetes mellitus, ALT, age, hypertension, dyslipidaemia, sex, family history, diabetes complication, hypothyroidism, and smoking. External validation of the all-variable model demonstrated an AUC of 0.795 and predicted SF in 9.0% more patients without increasing negative VCTE referrals compared with using the FIB-4 score as the screening tool in the first step of the MASLD screening algorithm.
Conclusions:
ML-based models were more effective than the FIB-4 score as the first-line screening tool for VCTE referrals, substantially improving the identification of patients with significant fibrosis in this South Asian cohort.
