This article is more than 1 year old
FDA clears way for an AI stethoscope to detect heart disease
Algorithm proves better than primary care doctors at recognizing signs of heart trouble
The US Food and Drug Administration has approved the first artificial intelligence algorithm powering a digital stethoscope for doctors to detect valvular heart disease more accurately.
Valvular heart disease arises when the heart's valves don't function correctly, impacting the pumping of blood to the body and the intake within the heart's chambers as the valves do not open or close properly.
Around 2.5 percent of the US population has valvular heart disease, and tens of thousands die each year from complications like heart failure or cardiac arrest, according to the Centers for Disease Control and Prevention.
Detecting valvular heart disease can be tricky. It requires doctors listening to a patient's heartbeat and being able to recognize an unusual sound or pattern, and figure out which valve is impaired and what issue it's causing. Recognizing whether the sound and pattern of a heartbeat is irregular can be subjective, and that means the condition is frequently misdiagnosed or missed entirely.
Here's where AI algorithms could come in handy. Eko, a digital health startup based in Oakland, California, has developed software to analyze a patient's pulse and help health care professionals detect heart murmurs. Its Eko Murmur Analysis Software (EMAS) is the first of its kind to receive FDA approval.
Heartbeat data collected by doctors using Eko's smart stethoscopes is analyzed by EMAS. EMAS characterizes the heart murmurs to detect and better understand what type of valvular heart disease a patient might have in seconds, we're told.
"EMAS is a cloud-based service that allows users to upload heart sounds and optional electrocardiogram data via an application programming interface for analysis," a company spokesperson told The Register. "The software uses signal processing, such as waveform filtering, as well as algorithms derived from machine learning, to analyze the acquired data and generate clinical decision support output for clinicians.
"The EMAS algorithm analyzes the heart sound data and outputs a JSON file with the algorithm results, which is passed down to the requesting application and displayed by the requesting application to the user in the human readable format."
- Health trusts swapped patient data for shares in an AI firm. They may have lost millions
- Tencent Cloud slaps googly eyes on a monitor, says it can care for oldies
- UK govt promises to sink billions into electronic health records for England
- Behold: The first images snapped by the James Webb Space Telescope
Eko claims its EMAS tool has an overall sensitivity and specificity – two measures of how accurately it can identify the disease – of 85.6 percent and 84.4 percent respectively. For comparison, similar tests performed with general practitioners using traditional stethoscopes to detect valvular heart disease reportedly had a sensitivity and specificity of 44 percent and 69 percent respectively.
"It takes expert clinicians many years to master the art of hearing and interpreting heart murmurs, and there is still a lot of variability. Experts such as cardiologists tend to detect murmurs with higher accuracy than primary care physicians, but most patients are seen in primary care practices. Only patients who are suspected of or known to have a cardiac condition are referred to a cardiologist," the Eko spokesperson told us.
FDA approval means Eko can market its EMAS algorithm in the US and begin selling its technology to the healthcare industry. "FDA clearance is an important step towards commercialization. We are looking to partner with healthcare institutions in the US that want to be among the early adopters of this new solution for increased detection of valvular heart disease." ®