The integration of artificial intelligence into cardiac ultrasound analysis offers a significant advancement in diagnosing advanced heart failure, according to recent research by Weill Cornell Medicine, Cornell Tech, and other leading institutions. The study, published in npj Digital Medicine, demonstrates the potential of AI to streamline the identification of high-risk patients, addressing a critical bottleneck in current diagnostic practices.
Currently, advanced heart failure is diagnosed through cardiopulmonary exercise testing (CPET), a resource-intensive process confined to major medical centres. As a result, many of the estimated 200,000 affected individuals in the United States are left without appropriate care. The new AI-driven method predicts peak oxygen consumption (peak VO2) using accessible ultrasound images and electronic health records, offering a more practical alternative.
“This opens up a promising pathway for more efficient assessment of patients with advanced heart failure,” stated Dr. Fei Wang, senior author and associate dean for AI and data science at Weill Cornell Medicine. The project involved collaboration across several institutions, including contributions from Dr. Deborah Estrin at Cornell Tech and Dr. Nir Uriel at NewYork-Presbyterian.
The study is part of the Cardiovascular AI Initiative, aiming to harness AI for improved heart failure management. Recent technological advancements have enabled AI applications beyond consumer use, with machine learning models now capable of detecting disease patterns in medical data. Dr. Uriel highlighted the potential of AI in cardiac ultrasound data as a promising application, leading to the development of a sophisticated machine learning model.
The model, developed by Dr. Wang’s team, processes various data types, including ultrasound images and health records. Trained on deidentified data from 1,000 heart failure patients, it demonstrated an impressive 85% accuracy in predicting peak VO2 in a new cohort of patients. This accuracy suggests its viability for clinical use.
Looking ahead, clinical trials are planned to validate this approach further, paving the way for potential FDA approval and routine clinical adoption. “If we can use this approach to identify many advanced heart failure patients who would not be identified otherwise, then this will change our clinical practice and significantly improve patient outcomes,” Dr. Uriel concluded.
