AI taught to look for subtle signs of heart failure on ECG

The authors of the new work have created an AI-based algorithm that can recognize subtle changes in

electrocardiograms indicating heart failure.

Heart failure or congestive heartfailure occurs when the heart pumps less blood than the body needs. Usually, doctors use echocardiography to assess if a patient has any symptoms. This is a rather laborious procedure, during which a lot of equipment is used, so scientists have tried to find another way to quickly track down heart problems.

Researchers have found that you can alsousing an electrocardiogram is a quick and readily available alternative. The results of the procedure must be analyzed by the AI. Deep learning can detect even minor weakness in the left ventricle of the heart, which pushes fresh oxygenated blood to the rest of the body.

Algorithm developers at the time of its creationtrained AI to read patients' electrocardiograms, as well as information from written reports. The authors used a pattern search engine to show AI hearts that are healthy and those that need treatment. Next, the algorithm compared the obtained data with those on which it was trained in order to learn how to identify pathologies.

As a result, the AI identified patients with healthy left ventricular function with 94% accuracy, and also correctly identified patients with a weak heart in 87% of cases.

Read more

Hawking was right, but sometimes wrong: the scientist's most daring ideas

Astronomers have figured out that the Earth and solar system are in a giant magnetic tunnel

A Chinese hypersonic missile orbited the entire globe. How is this possible and what is known about her