Related Experiment Video
Updated: May 14, 2025
07:03
Author Spotlight: Establishing MASLD Cell Models for Investigating Disease Mechanisms and the Lipid-Lowering Effects of Koumiss
Published on: July 19, 2024
696
Predicting metabolic dysfunction associated steatotic liver disease using explainable machine learning methods
Yihao Yu1, Yuqi Yang2,3, Qian Li2,3
1Master of Finance, Australian National University, Canberra, Australia.
Scientific Reports
|April 11, 2025
Summary
Developing an accurate machine learning model aids in early identification of metabolic dysfunction-associated steatotic liver disease (MASLD). This explainable 10-feature random forest model demonstrates high predictive performance for MASLD risk.
Area of Science:
- Hepatology
- Medical Informatics
- Machine Learning
Background:
- Metabolic dysfunction-associated steatotic liver disease (MASLD) requires early identification for effective management.
- Accurate risk prediction is crucial for preventing disease progression and improving patient outcomes.
Purpose of the Study:
- To develop and validate an explainable machine learning (ML) model for predicting MASLD in adults.
- To identify key clinical features for MASLD risk assessment.
Main Methods:
- A national cross-sectional study utilized data from 13,436 participants (2017-2020 NHANES).
- Six ML algorithms were trained and validated using 50 easily obtainable medical characteristics.
- Recursive feature elimination and Shapley Additive exPlanations were employed for feature selection and model interpretability.
Main Results:
- The random forest (RF) model achieved the highest predictive performance.
- An optimal 10-feature RF model demonstrated excellent discrimination in internal and external validation cohorts (AUC: 0.928, 0.918).
- The developed model outperformed traditional MASLD risk indicators.
Conclusions:
- An explainable, 10-feature RF model was successfully developed and validated for MASLD prediction.
- The model offers high accuracy and interpretability using readily available clinical data.
- This tool can aid in early MASLD risk stratification and clinical decision-making.
