Integrating Machine Learning with Serum Multi-Omic Biomarkers for the Early Detection and Prognostication of Non-Alcoholic Fatty Liver Disease (NAFLD)
- May 25
- 2 min read
Updated: May 26
DOI: https://doi.org/10.66715/jsccr/2026v3.i3.2937 | Original Research | 2026 | Volume 3 | Issue 3 | Page 29-37
Dr. Shahan Layek, Independent Researcher, West Bengal, India, Email: layekcallmeshahan@gmail.com
ABSTRACT
BACKGROUND: Non-Alcoholic Fatty Liver Disease is a leading global cause of chronic hepatopathy. Early detection is often hindered by the asymptomatic nature of the initial stages and the limitations of conventional diagnostic modalities. This study investigates the integration of serum multi-omic biomarkers with advanced machine learning algorithms to develop a highly sensitive, non-invasive predictive model for the early detection and prognostication of Non-Alcoholic Fatty Liver Disease.
METHODS: A comprehensive cross-sectional analysis was conducted on a cohort of three hundred participants, including two hundred clinically confirmed Non-Alcoholic Fatty Liver Disease cases and one hundred healthy controls. Serum samples were subjected to untargeted metabolomic and proteomic profiling. The resulting high-dimensional multi-omic datasets were processed using state-of-the-art machine learning classifiers, specifically Random Forest, Support Vector Machines, and Deep Neural Networks, to identify distinct biomarker signatures capable of differentiating disease stages.
RESULTS: The Deep Neural Network model demonstrated superior diagnostic performance, achieving an accuracy rate of ninety-four percent and an area under the curve of zero point nine six for early disease detection. A distinct multi-omic panel, comprising eleven specific lipid metabolites and four inflammatory proteomic markers, was identified as the primary driver of predictive accuracy. Furthermore, the machine learning algorithms successfully stratified patients by prognostic risk, accurately predicting the potential progression to non-alcoholic steatohepatitis.
CONCLUSION: The synergistic integration of serum multi-omic profiling with machine learning algorithms provides a robust, non-invasive framework for the early diagnosis and prognostic risk stratification of Non-Alcoholic Fatty Liver Disease. This data-driven approach holds significant potential for optimizing predictive healthcare and guiding personalized therapeutic interventions in clinical hepatology.
KEYWORDS: Non-Alcoholic Fatty Liver Disease, Machine Learning, Multi-Omics, Predictive Healthcare, Biomarkers, Data Analytics.