FDA Clears AI That Predicts Heart Failure During Routine Checkups

News Room

Heart disease remains the number one killer of Americans, with nearly 700,000 people dying each year. Over 6 million Americans are estimated to live with heart conditions–though they may not know it. But about half of them have a risk factor where their heart isn’t pumping blood as effectively as it could be — what doctors call “low ejection fraction.” Catching this early can mean prevention of more serious heart conditions, but it often goes undetected because there are few symptoms.

Eko Health, a startup that creates digital stethoscopes and their software, aims to change that. On Tuesday, the Food and Drug Administration cleared an algorithm co-developed by the Mayo Clinic and Eko which can help Eko’s AI stethoscopes accurately detect low ejection fraction. (A clinical study found it was able to correctly identify it about 85% of the time.) This could enable primary care physicians who use Eko’s stethoscopes to identify at-risk patients during routine checkups–possibly even before any symptoms of heart failure are present.

“Physicians get an early warning for heart disease,” Eko CEO Connor Landgraf told Forbes. “And we can actually help get the patients to the cardiologist or the pulmonologist or surgery or whatever that intervention is a lot sooner.”

Founded in 2013, Eko Health has sold over 500,000 of its digital stethoscopes. The company has raised $125 million in venture backing to date, and its valuation is estimated at $415 million, according to Pitchbook. The Mayo Clinic is an investor in the company through its venture arm, as are ARTIS Ventures, 3M’s venture arm and Highland Capital Partners, among others.

This isn’t Eko’s first FDA approval. In 2020, the agency also gave a green light to algorithms that can accurately detect atrial fibrillation and heart murmurs, both of which can be early warning signs of more serious conditions such as heart failure and strokes.

The algorithm was first developed at the Mayo Clinic, according to its cardiovascular chair Paul Friedman. The clinic developed a machine learning algorithm for electrocardiogram machines that was capable of detecting low ejection fraction when a person was hooked up to the machine. The algorithm was licensed to Eko, who adapted it for use with its stethoscope.

“These tools are incredibly powerful — they help us screen for conditions for which we have treatments,” Friedman said, adding that testing it in a retrospective study found that it was able to detect heart issues better than conventional methods like a treadmill test, which measures oxygen and other vitals.

In a peer-reviewed study published in The Lancet in January 2022, researchers at Imperial College London who weren’t affiliated with Eko found the algorithm detected low ejection volume about 85% of the time in patients. In November 2023, it received approval in the U.K., where doctors have been using the algorithm at multiple primary care facilities within the country’s National Health Service.

Now that the FDA has cleared the algorithm, the company will begin rolling it out as a pilot to some of its customers, with full commercialization in its software systems to follow in the U.S. In the meantime, Landgraf said the company will continue to develop new algorithms for its digital stethoscopes in order to provide an even broader array of diagnostics.

“We’ll release new algorithms on a frequent basis that unlock new functionality,” he said. “And make this tool even more useful for the clinicians who have it. We’ve got the devices in the field and now we need to unlock new value with new generations of AI tools.”

MORE FROM FORBES

Read the full article here

Share this Article
Leave a comment