Johns Hopkins' heart-scanning AI predicts cardiac arrests up to 10 years ahead

Researchers at Johns Hopkins University have developed an artificial intelligence approach they say can help predict if and when a person could die of cardiac arrest based on imaging scans of the heart.

By analyzing the patterns of scar tissue that can develop in the cardiac muscle over time due to heart disease, the researchers found that the AI algorithms could offer accurate predictions of a person’s risk up to 10 years ahead.

That scar tissue can affect how the heart’s timekeeping electrical pulses travel to the contracting ventricles, resulting in abnormal heart rhythms that can become dangerous if left untreated. 

"Sudden cardiac death caused by arrhythmia accounts for as many as 20% of all deaths worldwide, and we know little about why it's happening or how to tell who's at risk," said Natalia Trayanova, a professor at Johns Hopkins and senior author of a paper published in Nature Cardiovascular Research.

The technology uses contrast-enhanced MRI scans of the heart combined with standard data points from the patient’s medical history that are typically used to generate cardiovascular risk factors such as age, weight, race and prescription drug use.

After gathering years of heart scans from hundreds of Johns Hopkins patients with cardiac scarring, the researchers said the AI was able to outperform physicians’ predictions that relied on simple visual estimates of scar features like their total volume and mass.

"The images carry critical information that doctors haven't been able to access," first author Dan Popescu, a former Johns Hopkins doctoral student, said in a statement. "This scarring can be distributed in different ways, and it says something about a patient's chance for survival. There is information hidden in it."

With predictions delivering about 74% accuracy when put up against an independent set of testing data gathered from 60 health centers across the U.S., the AI could help physicians personalize long-term treatments for heart disease.

“There are patients who may be at low risk of sudden cardiac death getting defibrillators that they might not need and then there are high-risk patients that aren't getting the treatment they need and could die in the prime of their life,” Trayanova said. “What our algorithm can do is determine who is at risk for cardiac death and when it will occur, allowing doctors to decide exactly what needs to be done.”

Other efforts to find helpful heart signals unviewable by the human eye have focused on spotting the signs of early heart disease in electrocardiogram data or ultrasound scans.

Researchers at the Mayo Clinic previously demonstrated that learning AI algorithms could be trained to find signs of low ejection fraction and poor heart performance based on a standard, 12-lead EKG reading—leading to 32% more diagnoses compared to standard diagnostic workups.

Meanwhile, a human-versus-AI study by the University of Oxford spinout Ultomics showed that the technology was more effective at predicting which COVID-19 patients would suffer severe heart complications, using data from ultrasound scans.

By viewing the contractions of the beating heart muscle and highlighting signs of damage caused by COVID-19, the AI could forecast potentially higher mortality rates both inside the hospital and in the months following treatment.