Artificial intelligence for detecting and quantifying steatotic liver disease

The prevalence of hepatic steatosis is increasing globally. While non-invasive diagnostic methods like ultrasonography and clinical scoring systems have been suggested as alternatives to liver biopsy, their effectiveness has been questioned. Integrating Artificial Intelligence (AI) with traditional diagnostic methods is being explored to enhance the accuracy of non-invasive approaches. The research utilized science bibliographic databases for data retrieval, namely PubMed, Scopus, and Google Scholar. The search terms utilized were “fatty liver,” “hepatic steatosis” “artificial intelligent”, “machine learning”, “deep learning”, “convolutional neural network”, “artificial neural network” and “ultrasound” etc. The systematic review encompassed studies, which collectively demonstrated that AI had a notable impact on improving the diagnosis of various liver conditions including liver steatosis, steatohepatitis, liver fibrosis, and liver cirrhosis. Through qualitative analysis, it was found that AI was particularly effective in enhancing diagnostic accuracy for these conditions. The integration of AI-supported systems has shown promising advancements in the detection and quantification of steatosis, NASH, and liver fibrosis in patients with liver steatosis. These systems have demonstrated the ability to improve performance in accurately diagnosing and assessing the severity of liver diseases, providing healthcare professionals with valuable tools for more effective clinical management.