US-based researchers have created an AI-based test that can identify heart problems, even when regular examinations report normalities.
The analyses for the heart condition, atrial fibrillation, last just 10 seconds, compared to the current procedures which can take weeks or even years.
Atrial Fibrillation is a common heart disorder that if left untreated, can lead to strokes and heart failure. Identifying the condition is hard because of the unpredictable rhythms of a heart. Usually, electrocardiograph (EGC) machines use monitors that are placed on patients to detect abnormal rhythms. But these often have to be kept on for a long time.
“Atrial fibrillation is an arrhythmia where the atrium, or top chamber of the heart, loses its coordinated contractual activity and instead quivers, because the electrical impulse is changed in the way it courses through the atrium,” explains one of the researchers, Peter Noseworthy of the Mayo Clinic.
Therefore, the AI has been designed to look at the rhythm of the heart and identify abnormalities that people find hard to spot.
Using AI to find problems
As a way to speed up the testing process, Noseworthy and collaborators developed an AI model to detect issues in heart rhythm. The researchers used 650,000 ECG’s from 181,000 patients over a 24-year period. The investigation was carried out with a range of people. Some were known not to have the condition, whilst others had already been diagnosed with the disorder.
The findings showed that the artificial intelligence used was able to detect the differences in patients. There was a 79% accuracy reported for single tests with that figure rising to 83% for multiple tests on the same patient.
“When we look at the ECGs, they don’t have a single pattern to them, but the AI is able to identify many different, more subtle patterns that we may as cardiologists just recognise as mild abnormalities and do not really draw any attention to,” Noseworthy explains.
How this will benefit the population
Noseworthy believes that using AI in this sector will have 2 values. The first being to scan the general population and detect potential heart abnormalities. And the second being for it to be used on patients who have had a stroke and may benefit from longer-term tests.