Hepatic Steatosis Analysis in Metabolic Dysfunction-Associated Steatotic Liver Disease Based on Artificial Intelligence

Background: Metabolic dysfunction-associated steatotic liver disease (MASLD) is characterized by the accumulation of fat in the liver, excluding excessive alcohol consumption and other known causes of liver injury. Animal models are often used to explore different pathogenic mechanisms and therapeutic targets of MASLD. The aim of this study is to apply an artificial intelligence (AI) system based on second-harmonic generation (SHG)/two-photon-excited fluorescence (TPEF) technology to automatically assess the dynamic patterns of hepatic steatosis in MASLD mouse models.

Methods: We evaluated the characteristics of hepatic steatosis in mouse models of MASLD using AI analysis based on SHG/TPEF images. Six different models of MASLD were induced in C57BL/6 mice by feeding with a western or high-fat diet, with or without fructose in their drinking water, and/or by weekly injections of carbon tetrachloride.

Results: Body weight, serum lipids, and liver enzyme markers increased at 8 and 16 weeks in each model compared to baseline. Steatosis grade showed a steady upward trend. However, the non-alcoholic steatohepatitis (NASH) Clinical Research Network (CRN) histological scoring method detected no significant difference between 8 and 16 weeks. In contrast, AI analysis was able to quantify dynamic changes in the area, number, and size of hepatic steatosis automatically and objectively, making it more suitable for preclinical MASLD animal experiments.

Conclusions: AI recognition technology may be a new tool for the accurate diagnosis of steatosis in MASLD, providing a more precise and objective method for evaluating steatosis in preclinical murine MASLD models under various experimental and treatment conditions.