Artificial Intelligence for Detecting and Quantifying Fatty Liver in Ultrasound Images: A Systematic Review

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Hepatology

Background:

• Non-alcoholic Fatty Liver Disease (NAFLD) prevalence is increasing globally.
• Conventional non-invasive diagnostic methods for NAFLD lack sufficient efficacy.
• Artificial Intelligence (AI) integration aims to enhance non-invasive diagnostic performance.

Purpose of the Study:

• To evaluate the diagnostic and quantitative performance of various AI methods on ultrasound (US) images for NAFLD.
• To assess AI's role in diagnosing and staging NAFLD, including Non-Alcoholic Steatohepatitis (NASH) and liver fibrosis.

Main Methods:

• A systematic review of five major scientific databases was performed.
• Included studies were peer-reviewed English articles, conferences, theses, and book chapters.
• Data synthesis followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

Main Results:

• Forty-nine studies were included in the systematic review.
• AI significantly improved the diagnosis of NAFLD, NASH, and liver fibrosis.
• Performance metrics like accuracy, sensitivity, and specificity were analyzed across different AI approaches.

Conclusions:

• AI-supported systems demonstrate potential for enhanced detection and quantification of NAFLD, NASH, and fibrosis.
• Prospective studies comparing AI-assisted and conventional techniques are required for clinical implementation.