Role of Artificial Intelligence in Revolutionizing the Diagnosis of Fatty Liver Disease
DOI:
https://doi.org/10.61919/jhrr.v5i10.1887Keywords:
Steatotic liver disease, MASLD, NASH, artificial intelligence, deep learning, liver fibrosis, non-invasive diagnostics, digital pathology, EHR models, risk prediction.Abstract
Background: Steatotic liver disease (SLD), including metabolic dysfunction-associated steatotic liver disease (MASLD) and non-alcoholic steatohepatitis (NASH), affects nearly 30% of the global population and poses a major public health challenge. Diagnosis still depends heavily on liver biopsy, which is invasive, prone to sampling error, and unsuitable for widespread screening. These limitations have driven growing interest in artificial intelligence (AI) as a non-invasive diagnostic solution. Objective: To evaluate the current state and performance of AI technologies used for diagnosing, staging, and predicting progression in steatotic liver disease across imaging, electronic health record (EHR) analysis, and digital pathology. Methods: A comprehensive review was conducted, examining recent AI applications in CT, MRI, and ultrasound imaging, EHR-based machine learning models, and digital pathology platforms. Reported diagnostic accuracy, predictive performance metrics, technical strengths, and clinical limitations were analyzed. Results: Deep learning models applied to CT imaging demonstrated high accuracy in staging fibrosis, with AUC values of 0.97 for advanced fibrosis (≥F3) and 0.95 for cirrhosis (F4). AI-assisted ultrasound achieved an AUC of 0.98 for NAFLD detection. EHR-based tools, such as NASHmap, showed moderate predictive ability (AUC 0.76). Progression-prediction models reached AUROC values of 0.87 for forecasting fibrosis within four years. In digital pathology, AI systems like qFibrosis® provided superior reproducibility (89–93%) and identified treatment effects missed by conventional histology. Conclusion: AI offers accurate, scalable, and objective alternatives to invasive diagnostics in liver disease, particularly for ruling out advanced fibrosis and predicting progression. However, challenges persist, including algorithmic bias, limited generalizability, opacity of deep learning models, regulatory constraints, and slow clinical translation. Future advancement requires multimodal data integration, robust external validation, improved transparency, and clear governance frameworks.
Downloads
References
Younossi ZM, et al. (2023). The global epidemiology of NAFLD and NASH in 2023: A systematic review and meta-analysis. Hepatology, 77(4), 1335–1347.
Rinella ME, et al. (2023). A multisociety Delphi consensus statement on new fatty liver disease nomenclature. Hepatology, 78(6), 1966–1986.
Powell EE, et al. (2021). Non-alcoholic fatty liver disease. The Lancet, 397(10290), 2212–2224.
Chalasani N, et al. (2018). The diagnosis and management of nonalcoholic fatty liver disease: Practice guidance from the American Association for the Study of Liver Diseases. Hepatology, 67(1), 328–357.
Bedossa P, et al. (2021). Histological assessment of NAFLD: Current standards and future perspectives. Journal of Hepatology, 74(5), 1219–1229.
Sato K, et al. (2024). Revolutionizing MASLD: How artificial intelligence is shaping the future of liver care. Diagnostics, 14(3), 256.
Kim BH, et al. (2023). Artificial intelligence in NAFLD: Current status and future directions. Clinical and Molecular Hepatology, 29(2), 235–254.
LeCun Y, Bengio Y, Hinton G. (2015). Deep learning. Nature, 521(7553), 436–444.
Caussy C, et al. (2020). MRI proton density fat fraction for assessment of hepatic steatosis. Hepatology, 71(4), 1371–1383.
Park HJ, et al. (2020). Deep learning regression model for fat fraction estimation using clinical MRI sequences. Journal of Magnetic Resonance Imaging, 51(5), 1387–1396.
Wu CC, et al. (2021). Deep learning system for staging liver fibrosis on computed tomography. Radiology, 299(2), 338–347.
Byra M, et al. (2022). Deep learning models for grading hepatic steatosis: Comparing performance across disease stages. Radiology, 303(2), 445–454.
Byra M, et al. (2020). Artificial intelligence-based ultrasound imaging for detection and quantification of hepatic steatosis. IEEE Transactions on Medical Imaging, 39(8), 2422–2431.
Han A, et al. (2020). Deep learning for hepatic steatosis quantification in ultrasound imaging. European Radiology, 30(5), 2889–2898.
Sonic Incytes. (2023). FDA 510(k) clearance of AI-powered features for Velacur liver ultrasound. Company Press Release.
Zhang YN, et al. (2022). Artificial intelligence for liver disease diagnosis: Strengths and limitations. Frontiers in Medicine, 9, 850656.
Cao W, et al. (2021). Deep learning performance in staging steatosis severity: A comparative analysis. Medical Image Analysis, 69, 101965.
Loomba R, et al. (2021). Noninvasive diagnosis of nonalcoholic steatohepatitis using machine learning: Development of the NASHmap model. JMIR Medical Informatics, 9(5), e26565.
Ioannou GN, et al. (2020). Machine learning to predict progression of non-alcoholic fatty liver to steatohepatitis or fibrosis. Alimentary Pharmacology & Therapeutics, 51(10), 1038–1049.
Konerman MA, et al. (2021). Artificial intelligence in identifying patients with undiagnosed nonalcoholic steatohepatitis. Journal of Health Economics and Outcomes Research, 8(1), 70–78.
Kleiner DE, et al. (2005). Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology, 41(6), 1313–1321.
Ratziu V, et al. (2011). Sampling variability of liver biopsy in nonalcoholic fatty liver disease. Gastroenterology, 128(7), 1898–1906.
Goodman ZD, et al. (2021). Second harmonic generation microscopy with artificial intelligence for quantitative fibrosis assessment in NASH. Hepatology Communications, 5(7), 1209–1221.
Sun Y, et al. (2021). Repeatability and reproducibility of artificial intelligence-based digital pathology for liver fibrosis evaluation. Journal of Pathology Informatics, 12, 28.
Tsochatzis EA, et al. (2022). qFibrosis-assisted reads improve interpathologist agreement in NASH clinical trials. Hepatology Communications, 6(9), 2341–2351.
Loomba R, et al. (2023). qFibrosis detects treatment effects missed by conventional NASH-CRN scoring in clinical trials. Hepatology, 77(3), 712–724.
PathAI. (2022). AIM-NASH: AI-powered digital pathology for reproducible NASH assessment. PathAI Technical Report.
PharmaNest. (2023). FibroNest: Sub-staging fibrosis with AI-based quantitative pathology. PharmaNest White Paper.
FDA. (2023). Use of biomarkers for diagnosing and assessing treatment response in noncirrhotic NASH trials: Guidance for industry. U.S. Food and Drug Administration.
Topol EJ. (2019). High-performance medicine: The convergence of human and artificial intelligence. Nature Medicine, 25(1), 44–56.
FDA. (2021). Artificial intelligence and machine learning in software as a medical device. U.S. Food and Drug Administration.
Ronneberger O, Fischer P, Brox T. (2015). U-Net: Convolutional networks for biomedical image segmentation. In: MICCAI. 234–241.
Zhou T, et al. (2021). The evolution of artificial intelligence technology in non-alcoholic fatty liver disease research. Journal of Clinical Medicine, 10(16), 3528.
Char DS, et al. (2020). Implementing machine learning in healthcare: Addressing ethical challenges. New England Journal of Medicine, 378(11), 981–983.
He J, et al. (2020). The practical implementation of artificial intelligence technologies in medicine. Nature Medicine, 25(1), 30–36.
Obermeyer Z, et al. (2019). Dissecting racial bias in an algorithm used to manage the health of populations. Science, 366(6464), 447–453.
Rajkomar A, et al. (2018). Ensuring fairness in machine learning to advance health equity. Annals of Internal Medicine, 169(12), 866–872.
Gianfrancesco MA, et al. (2018). Potential biases in machine learning algorithms using electronic health record data. JAMA Internal Medicine, 178(11), 1544–1547.
Rudin C. (2019). Stop explaining black box machine learning models and use interpretable models instead. Nature Machine Intelligence, 1(5), 206–215.
Vayena E, et al. (2018). Machine learning in medicine: Addressing ethical challenges. PLOS Medicine, 15(11), e1002689.
Selvaraju RR, et al. (2017). Grad-CAM: Visual explanations from deep networks via gradient-based localization. In: ICCV. 618–626.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2025 Muhammad Ahmad, Shanawar Hayat
This work is licensed under a Creative Commons Attribution 4.0 International License.
Public Licensing Terms
This work is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0). Under this license:
- You are free to share (copy and redistribute the material in any medium or format) and adapt (remix, transform, and build upon the material) for any purpose, including commercial use.
- Attribution must be given to the original author(s) and source in a manner that is reasonable and does not imply endorsement.
- No additional restrictions may be applied that conflict with the terms of this license.
For more details, visit: https://creativecommons.org/licenses/by/4.0/.
