Artificial Intelligence shows strong early predictive capabilities for LV dysfunction

February 26, 2019

A CNBC interview with Dr. Paul Friedman, the Chair of Cardiovascular Medicine at the Mayo Clinic, describes his work with Artificial Intelligence as an early predictor of Asymptomatic Left Ventricular (LV) Dysfunction.  Using only EKG readings, the computer system has been “trained” to read the signals and identify abnormalities. Dr. Friedman tested the systems’ capability to evaluate unknown EKG’s and asked it to identify if the patient had Asymptomatic LV Dysfunction or as Dr. Friedman describes a “weak heart pump”.  The AI system was a very accurate test of patients having this condition (who may benefit from early treatment) and is also a powerful predictor of patients who will actually develop the condition over the next few years. Dr. Friedman reports that since these AI neural networks are trained on hundreds of thousands of examples, they can see subtle patterns in a EKG that a human is unable to interpret.  

EHC NOTE: Although the application discussed on the video is not designed for perioperative care, the potential for adaptation is evident.  Similar technology may be used to scour preoperative data and test for high risk patients, or to signal early warning signs intraoperatively by constantly evaluating the anesthesia data feed.  It may be that over the next few years we may add an AI member to the Anesthesia and OR care teams.