CINCINNATI — Artificial intelligence is revolutionizing how doctors detect deadly heart attacks at local hospitals.
As some celebrate American Heart Month, a new study shows AI can catch these life-threatening events much quicker than traditional methods.
"if you have a high cholesterol, or you smoke, high blood pressure, diabetes, you get blockages in your arteries. ... A heart attack comes because that blockage breaks open and ... then your body forms a blood clot there, and when the blood clot stops the flow of blood that leads to death of the heart muscle," said Dr. Timothy Henry, lead researcher and director of The Carl and Edyth Linder Research Center at Christ Hospital.
WATCH: New study shows artificial intelligence can detect severe heart attacks, saving more lives
Henry said the new study is critical to saving more lives. Within seconds, the tool can analyze an EKG and tell doctors if a patient is having a severe heart attack, also known as ST-segment elevation myocardial infarction (STEMI), and needs to go to a cardiac cath lab right away.
"The key part when you have a blockage in your artery, and the key part is to open that blockage as fast as possible. And the way to do that is to get rapidly to the hospital that has a cath lab, and you go down with a wire, and you open that up directly," Henry said.
The study looked at more than 1,000 patients at three hospitals over four years and shows AI software can now analyze heart rhythm tests better than current methods.
"More importantly, about half of patients who have a STEMI, the EKGs are complicated, and we call it STEMI mimics, or a lot of people wouldn't read it as a STEMI, and in that situation, AI had a 95% probability and read it the correct way," Henry said. "So what that allows is you diagnose it quicker, which means you then get to the cath lab quicker, which means people live longer and have better outcomes."
Henry said severe heart attacks must be treated within 90 minutes. AI detected 553 out of 601 confirmed heart attacks compared to just 427 detected by standard care.
"It aids the health care provider. So it's a way that you can quickly get information on EKGs, but it will never take the place of the health care provider," said Henry.
It also had a false positive rate of less than 8% compared to more than 40% with current methods.
"Hypertrophic cardiomyopathy is a specific kind of abnormality that is the highest risk for sudden death in athletes, and AI can detect that from the EKG," Henry said. "We will have an answer about FDA approval within the next few months. And once it's FDA approved, then I think there will be very rapid utilization."
Researchers say the technology could be especially helpful for rural hospitals that don't have specialized heart treatment centers. The study was published in a major cardiology journal, JACC: Cardiovascular Interventions.
