New AI can diagnose life-threatening heart arrhythmia in 10 seconds by spotting patterns that are ‘invisible’ on standard EKGs, study suggests
- Researchers trained AI to recognize atrial fibrillation (AFib), an irregular and rapid heart rate that can cause fatal heart attacks and strokes
- EKGs don’t always detect AFib and the results can often come back as negative
- The AI model detected AFib with about 79% accuracy on single tests and 83% accuracy with multiple tests
- Doctors said the technology could one day be implemented in smartphones
Artificial intelligence could detect a life-threatening heart arrhythmia in just 10 seconds, a new study claims.
Researchers from the Mayo Clinic say AI was able to detect atrial fibrillation (AFib) during an electrocardiogram (EKG), even if the heart was beating normally during the time of the test.
This means that a patient with the condition may not be diagnosed for years unless they are consistently monitored.
The team says the technology will increase both the speed and accuracy of diagnosis so physicians can focus on making sure their patients receive the proper care.
It could even one day be implemented in smartphones and smartwatches, making the technology more accessible.
A new study from the Mayo Clinic has found that an AI-run EKG can detect atrial fibrillation with about 79% accuracy on single tests and 83% accuracy with multiple tests (file image)
Normally, the heart contracts and relaxes to a regular beat so blood can flow to other organs.
However, in AFib, the upper chambers of the heart beat out of coordination with the lower chambers, which weakens heart muscles.
Clots can form and, if they enter the blood stream, they can become lodged in arteries and cause fatal heart attacks or strokes.
Some people have no symptoms. Others might experience symptoms including fatigue, a fluttering feeling in the chest, dizziness and shortness of breath.
According to the American Heart Association, at least 2.7 million Americans have this condition.
When undiagnosed, AFib can often interfere with other treatments, such as stroke treatments.
‘When people come in with a stroke, we really want to know if they had [AFib] in the days before the stroke, because it guides the treatment,’ said Dr Paul Friedman, chair of the department of cardiovascular medicine at Mayo Clinic.
‘Blood thinners are very effective for preventing another stroke in people with [AFib]. But for those without [AFib], using blood thinners increases the risk of bleeding without substantial benefit. That’s important knowledge. We want to know if a patient has [AFib].’
For the study, published in The Lancet, the team created and trained an AI model to recognize AFib from about 650,000 EKGs.
The results were from more than 181,000 patients at the Mayo Clinic from December 1993 to July 2017.
By comparing EKGs from ‘normal’ patients with AFib patients, the AI was able to recognize the subtle differences that are hard to spot by even the most trained cardiologists.
Researchers then tested the AI on 10-second EKGs that had normal readings from more than 36,000 patients – about 10 percent of whom had been diagnosed with AFib.
The AI-run EKGs were able to correctly identify people with potentially undetected AFib with about 79 percent accuracy on single tests and 83 percent accuracy with multiple tests.
This is important because AFib is often asymptomatic and only diagnosed when it causes serious complications such as strokes.
Dr Friedman admitted he was surprised by the findings but said that, one day, the technology could even be implemented on a smartphone or a smartwatch.
‘An EKG will always show the heart’s electrical activity at the time of the test, but this is like looking at the ocean now and being able to tell that there were big waves yesterday,’ said Dr Friedman.
‘AI can provide powerful information about the invisible electrical signals that our bodies give off with each heartbeat – signals that have been hidden in plain sight.’
Source: Read Full Article