Applying Artificial Intelligence techniques to cardiac ultrasound data may make it easier to identify patients with advanced heart failure. Investigators from Weill Cornell Medicine, Cornell Tech, Cornell Ann S. Bowers College of Computing and Information Science, Columbia University Vagelos College of Physicians and Surgeons and NewYork-Presbyterian reported a method designed to improve detection of a condition that is often difficult to diagnose and therefore undertreated.
Advanced heart failure is currently detected through cardiopulmonary exercise testing, which requires specialized equipment and trained staff and is typically only available at large medical centers. Due in part to this diagnostic bottleneck, only a few of the estimated 200,000 people in the United States with advanced heart failure get appropriate care each year. In the study, published March 3 in npj Digital Medicine, the researchers tested a novel Artificial Intelligence-powered method that predicts the most important cardiopulmonary exercise testing measure, peak oxygen consumption, using more easily obtainable ultrasound images of the heart along with electronic health records. The work emerged from the Cardiovascular Artificial Intelligence Initiative, a collaboration among Cornell, Columbia and NewYork-Presbyterian focused on improving heart failure diagnosis and management.
The research team developed a multi-modal, multi-instance machine learning model that can process several data types, including moving ultrasound images of the heart, waveform imagery showing heart valve dynamics and blood flow, and information from electronic health records. The model was trained on deidentified data from 1,000 patients with heart failure seen at NewYork-Presbyterian/Columbia University Irving Medical Center. Once trained, the model was then tasked with predicting peak VO2 for a new set of 127 patients with heart failure from three other NewYork-Presbyterian campuses.
The results were better than any reported before for Artificial Intelligence-based peak VO2 prediction. That figure in this case indicated an overall accuracy of roughly 85%, which suggests it will be useful in clinical settings. Researchers said the close collaboration between clinicians and Artificial Intelligence specialists shaped the development of new techniques, with heart failure experts helping identify where the technology could have the most clinical value.
The team has already begun planning clinical studies of the approach, which would be needed for U.S. Food and Drug Administration approval and routine clinical adoption. Researchers said wider use of this method could help identify advanced heart failure patients who would not otherwise be recognized, with the potential to change clinical practice and improve patient outcomes and quality of life.
