Presented at the American Heart Association Scientific Sessions 2025 (AHA25) by researchers from Juntendo University (Tokyo, Japan), the AI-SCREEN-CA study evaluated the real-world performance of artificial intelligence for detecting hidden cases of cardiac amyloidosis (CA) from routine echocardiograms.
The study leveraged Us2.ai's automated echocardiography analysis technology to assess its ability to surface previously unrecognized CA in an unselected clinical population.
Using Us2.ai for cardiac amyloidosis screening has been a transformative experience. The parameter-based algorithm centered on the IWT score provides exceptional specificity, reliably identifying patients with a high likelihood of amyloidosis and minimizing false positives. Meanwhile, the deep-learning model captures myocardial texture and subtle motion abnormalities, enabling sensitive detection even in early or atypical stages. In our study, Us2.ai successfully identified a large number of “hidden” cardiac amyloidosis cases in real-world clinical practice — a remarkable breakthrough for this field. Importantly, by integrating parameter-based and deep-learning approaches, Us2.ai has achieved a balance of high specificity and high sensitivity. This capability represents a major advance in the diagnostic pathway for cardiac amyloidosis and paves the way for future clinical workflows that combine AI-enhanced echo analysis with other testing modalities and emerging treatments. – Dr. Nobuyuki Kagiyama Associate Professor, Department of Cardiovascular Biology and Medicine / AI Incubation Farm, Juntendo University Graduate School of Medicine
Insights from Dr. Kagiyama