SAT-682 Innovative Artificial Intelligence (AI) SXR platform use to aid diagnosis of metabolic dysfunction associated steatohepatitis (MASH) and liver fibrosis

Abstract Disclosure: V.A. Preda: None. M.K. Wilson: co-inventor of software. T. Carney: co-inventor of software. M.V. Preda: None. Background: The prevalence of cardiometabolic complications of obesity and diabetes are rising and this includes hepatic disease. Patients are often complex and have a number of chronic health needs with polypharmacy and a burden of care provision. As presented at US Endo 2024, there are medications that can be used for both cardiovascular, diabetic and hepatic improvements, such as GLP-1RAs and SGLT-2i. The global prevalence of non-alcoholic fatty liver disease (NAFLD) continues to rise in patients with obesity and diabetes. Hepatic disease is often overlooked and may not be diagnosed until there is progression to cirrhosis. The gold standard of testing is burdensome liver biopsy. Non-invasive diagnostic modalities including ultrasonography and clinical scoring systems have been proposed as alternatives to liver biopsy but with limited performance and application. The previously used liver transaminases may not capture many patients at risk. Artificial intelligence (AI) is currently being integrated with conventional diagnostic methods in the hopes of performance improvements. We describe our novel platform for assessing NAFLD, non-alcoholic steatohepatitis (NASH), and liver fibrosis by use of artificial intelligence (AI) software using the SXR® platform. Methods: Patients undergoing endocrine assessment for obesity and cardiometabolic disease were able to participate in this study. Data was deidentified. The integrated fit for purpose AI software from SXR® was used to capture, store and analyse patient demographic data, clinical anthropometric data and combine this with liver transaminases and ultrasound imaging to produce a risk assessment score for screening. Changes in time were also able to be longitudinally assessed. Results: Rapid liver screening is possible in patients with cardiometabolic risk without significant additional work load burden to the clinician or invasive testing to the patient. Assessment of at risk patients can be made quickly using existing datasets and clinicians can quickly screen patients at higher risk for hepatic disease without the need for invasive testing. Conclusion: AI-supported systems provide promising performance improvements for diagnosing NAFLD, NASH, and identifying high risk liver fibrosis cohorts among NAFLD patients. Recommendations for care can then be tailored to the patient including options for therapy that treat both obesity/diabetes and MAFLD/ NAFLD. Presentation: Saturday, July 12, 2025