Overview

• EchoNext outperformed cardiologists with 77% accuracy versus 64% in identifying structural heart disease from ECGs in a head-to-head study published in Nature.
• Columbia University has filed a patent for EchoNext and launched an ongoing clinical trial across eight emergency departments to assess real-world performance.
• Imperial College London’s valve-disease predictor achieved 69%–79% accuracy in spotting regurgitant valvular heart diseases and has been validated on U.S. and Chinese patient cohorts.
• Both AI systems were trained on more than two million paired ECG-echocardiogram records from over 600,000 patients, demonstrating applicability across diverse populations.
• Integrating AI analysis into routine ECGs could leverage 400 million annual tests worldwide to screen early for hidden heart conditions and improve outcomes.