artificial intelligence

AI Technology Can Predict Life-Threatening Heart Trouble, Researchers Say

"We can say the patient will have an episode in five years and we are certain 80% that that's going to happen"

NBC Universal, Inc.

Researchers at Johns Hopkins University developed artificial intelligence technology that may be able to assess a patient's risk of sudden cardiac death, which is when the heart abruptly stops beating.

Sometimes, modern medicine isn’t enough to help keep us healthy. The Johns Hopkins University researchers said artificial intelligence can help accurately predict if and when someone’s heart will stop beating years in advance.

"It uses deep learning on images in combination with deep learning also on clinical data to predict the patient's risk of sudden cardiac death over a period of 10 years," said Dr. Natalia Trayanova, a professor of biomedical engineering and medicine.

Trayanova's team developed the AI technology and published their work in a medical journal.

She said this survival predictor is the first of its kind and works by analyzing scar tissue in a patient’s heart, which detects patterns invisible to the naked eye.

"We can say the patient will have an episode in five years and we are certain 80% that that's going to happen," Trayanova said. "For other patients, we can say that they'd be such that we can predict the patient and will have it in six years, but our certainty is only 40% for it."

The artificial intelligence works off an algorithm created from MRI scans and the medical history of more than 200 real patients with heart disease at Johns Hopkins Hospital, all uploaded into a network to create this high-tech tool.

"We trained the algorithm to look at the images, learn from the images," Trayanova said.

Dr. Trayanova said the technology could revolutionize the way clinical decisions are made, protecting those who may otherwise be sent home, not knowing they’re at high risk.

"Some of them might die in the prime of their life because they are not protected. So what these do is provides a much more accurate prediction of what is the chance of a patient to have an event," Trayanova said.

According to researchers, the algorithm’s predictions were more accurate than those of doctors.

And although more clinical trials are needed to confirm the findings, the hope is this AI technology could one day help physicians tailor treatments for patients based on their own unique risk.

"This is our goal at Hopkins, is to take what is developed by engineers and computer scientists and bring it in the clinic directly," Trayanova said.

Researchers are now working to build algorithms to test other types of heart disease.

Heart disease remains the No. 1 killer in the United States, and sudden cardiac death caused by arrhythmia accounts for as many as 20% of all deaths worldwide.

Contact Us