AI 'Merlin' Predicts Chronic Disease From CT Scans

Stanford University researchers, supported by the National Institutes of Health, have introduced a machine-learning model called "Merlin" ^[1]. This foundation model analyzes 3D abdominal CT scans to predict the development of chronic conditions, including diabetes and heart disease ^[1][2]. The tool can forecast these health risks up to five years before clinical onset ^[1]. According to researchers, Merlin identifies subtle imaging biomarkers that are typically invisible to human radiologists ^[2][3]. It also demonstrates superior performance compared to existing specialized medical models ^[1]. By automating the analysis of routine scans, Merlin could significantly accelerate clinical assessments ^[2]. This technology offers a versatile approach to preventive medicine by utilizing data already captured during standard diagnostic procedures ^[3].