Machine Learning–Guided Echocardiogram Image Acquisition

Novel software, developed using deep-learning technology and recently authorized by the US Food and Drug Administration, can provide real-time prescriptive guidance to novice operators about how to obtain transthoracic echocardiographic (TTE) images that allow for limited diagnostic assessment of cardiac chambers. This video demonstrates the interaction of a user with the software, illustrating the turn-by-turn instructions; the quality meter, which dictates when a clip is automatically recorded (termed an auto-capture); and the ability to manually capture the best clip image if the autocapture is never achieved. A study of the technology concluded that untrained nurses could obtain images sufficient for trained echocardiographers to evaluate left ventricular (LV) size and function, right ventricular (RV) size, and presence vs absence of pericardial effusion. If the software utility and accuracy is confirmed across studies and users, it could make echocardiography more accessible outside of specialty offices and centers. Click the Related Article link for complete details.