A new way of identifying a common condition that causes the heart to beat irregularly may have been discovered by artificial intelligence.
Atrial fibrillation affects one million people in the UK and increases the risk of stroke and long-term heart problems.
It is relatively simple to diagnose when the heart is beating irregularly, but not when it returns to normal.
Computer modelling at the Mayo Clinic in the US may have identified signs that indicate previous abnormalities.
Researchers said it was still early days, but believe the system could lead to earlier and easier detection of the problem and, therefore, ensure patients get the right treatment, saving lives.
Atrial fibrillation symptoms
- Noticeable heart palpitations, when the heart feels like it is pounding, fluttering or beating irregularly
- Your heart may also beat very fast (often considerably higher than 100 beats per minute)
- You can work out your heart rate by checking the pulse in your neck or wrist.
- Other symptoms may include tiredness and being less able to exercise, breathlessness, feeling faint or light-headed and chest pain
- The way the heart beats in atrial fibrillation reduces the heart’s performance and efficiency
- This can lead to low blood pressure (hypotension) and heart failure
- You should see your GP immediately if you notice a sudden change in your heartbeat and experience chest pain
- Sometimes atrial fibrillation does not cause any symptoms and a person who has it is completely unaware that their heart rate is irregular.
Source: NHS England
Currently where these tests – known as electrocardiograms – do not find abnormal rhythms, doctors can ask the patient to undergo longer-term heart monitoring.
But instead the computer modelling was asked to look out for what doctors believe are subtle signs of past irregular rhythms, including scarring of the heart, that are unable to be spotted by the human eye from test results.
The computer modelling analysed tests carried out on nearly 181,000 patients between 1993 and 2017.
They were all patients who had had normal test results at first.
The modelling correctly identified the subsequent diagnosis from the normal test results in 83% of cases.
Dr Paul Friedman, from the Mayo Clinic, said it showed real potential: “It is like looking at the ocean now and being able to tell that there were big waves yesterday.”
But the team said the modelling now needed to be tested further to see if it could be deployed on the frontline.
Prof Tim Chico, an expert in cardiology at the University of Sheffield, described the findings as “very important”.
“This AI-based approach could provide a revolutionary advance, although it’s important to note that this research is still in the early stages and we need to see replicated results, and how the algorithm responds when tested on the general population.”