The potential for artificial intelligence (AI) to revolutionize how medicine is practiced presents as a double-edged sword—that was one of the take-aways from back-to-back sessions presented by cardiologists Geoff Tison, MD, assistant professor of medicine in the division of cardiology from the University of California San Francisco and Partho Sengupta, MD, professor of medicine and director of interventional echocardiography and cardiac ultrasound research and core lab at the Mount Sinai Hospital in New York at the Health Tech Summit during the American Heart Association Scientific Sessions in Anaheim last week.
On one hand, machine learning could help re-engineer the workflows that now render so many physicians and other healthcare professionals critically overworked and overextended. On the other hand the algorithms that mimic human cognitive functions like learning and problem solving, are data hungry, have labeling problems, and can leave us wondering how exactly a particular solution was derived in that “black box” to which we have no access.
Drs Tison and Sengupta also reminded their audiences that scientists still control the quality of the information that serves as the substrate for automated neural processing and that the “garbage in, garbage out” rule of data processing still holds true.
With those caveats on board both physicians discussed real-world applications of AI, including Dr Sengupta’s presentation on AI used to automate morphologic and functional assessments in echocardiology. A specific example was use of AI to help better understand hypertrophic cardiomyopathy in athletes.
Dr Seth Martin, assistant professor at the Ciccarone Center for the Prevention of Heart Disease at the Johns Hopkins University School of Medicine, highlights the two presentations in this short podcast.