Predicting Coronary Bypass Failure with Artificial Intelligence.

Coronary artery bypass grafts fail with a decade of surgery and there is no reliable way to predict graft failure prior to the event. Using artificial intelligence in collaboration with the Stanford Artificial Intelligence Group and with the use of computational flow dynamics with super-computers, we can estimate the blood flow stress and other factors that may lead to bypass graft failure. We have shown in preliminary work that these measures are superior to simply relying on graft appearance. Once confirmed, this methodology can be used to predict graft failure as well as determine the need for coronary stenting well before symptoms or closure. Studies are planned retrospectively and prospectively.

(O. Kahn, PhD; R. Packard, MD, PhD; R. Karlsberg, MD)