[ad_1]
Preliminary data indicate that a 10-second non-invasive test can identify American patients with an abnormal intermittent heart rhythm.
Caroline White
Friday, August 2, 2019
Artificial intelligence (AI) can help detect undiagnosed atrial fibrillation in 10-second ECGs taken from patients at a normal rate, preliminary data * published in The lancet to suggest.
The study of nearly 181,000 patients is the first to use in-depth learning to detect patients with abnormally abnormal heart rhythm potentially undetected.
The technology detects signals in the ECG that could otherwise be invisible but contain important information about the presence of atrial fibrillation.
Atrial fibrillation is difficult to detect on a single ECG because the heart of patients can enter and exit at an abnormal rate.
"The application of an artificial intelligence model to the ECG can detect atrial fibrillation even if it is not present at the time of registration of the # 39; s ECG. It's like watching the ocean now and being able to say that there were big waves yesterday, "comments Dr. Paul Friedman, co-investigator, director of the cardiovascular department at the Mayo Clinic. , United States.
"Currently, IA has been trained with the help of ECG in people requiring clinical investigations, but not in people with unexplained stroke nor in the whole Population. We are not yet sure how to diagnose these groups.
"However, the ability to test quickly and inexpensively with a non-invasive and widely available test could one day help identify undiagnosed atrial fibrillation and guide an important treatment to prevent stroke and other serious diseases."
After an unexplained stroke, it is important to accurately detect atrial fibrillation in order to be able to administer anticoagulants to patients to avoid the risk of another stroke, while other patients who do not have not and who could be hurt by this treatment are not. .
But this currently requires monitoring for several weeks to a few years, sometimes with an implanted device.
Hearts with atrial fibrillation develop structural changes. Before these changes are visible for standard imaging techniques such as echocardiograms, the heart probably has scars. In addition, atrial fibrillation may temporarily alter the electrical properties of the heart muscle, even after its termination.
The researchers set out to form a neural network – an AI clbad of deep learning – in order to recognize the subtle differences in a standard ECG supposed to be due to these changes.
The authors used the heart rate ECGs of nearly 181,000 patients (approximately 650,000 ECG examinations) between December 1993 and July 2017, dividing the data between those with and without atrial fibrillation.
When testing the first cardiac ECG produced by each patient, the accuracy of the AI was 79% for a single scan and, with the use of multiple ECGs for the same patient, the accuracy was improved at 83. %.
The authors of the study believe that it may be possible to use this technology as a diagnostic test at the place of treatment in the doctor to identify high-risk groups. Current detection methods are expensive, cumbersome and imprecise.
"It is possible that our algorithm can be used on inexpensive and widely available technologies, including smartphones, but this will require more research before widespread application," notes Dr. Xiaoxi Yao, co-researcher.
In a related comment **, Dr. Jeroen Hendriks of the University of Adelaide and Royal Adelaide Hospital of Adelaide, Australia, said: "Since AI algorithms have recently reached the level of diagnostic AI-ECG interpretation is revolutionary an algorithm to reveal the risk of atrial fibrillation in ECGs with sinus rhythm. "
But Dr. Malcolm Finlay, consulting cardiologist at Barts Heart Center, warned: "This study used patients already under investigation for atrial fibrillation for the test and algorithm dataset. This really limits the applicability of the results to real world situations, where we would like to use electrocardiograms to determine which patients will benefit from certain treatments for [atrial fibrillation] decrease their risk of having a stroke.
He added, "One of the problems with the AI approach is that detection can be good, but it can not tell you why it detects." Therefore, the detection can be based on an already known element of the ECG. This means that simpler methods than an artificial intelligence algorithm could be used to detect [atrial fibrillation] using this marker. It is also unclear whether previously used drugs that may have affected the ECG were used for therapeutic purposes. [atrial fibrillation] treatment or treatment of hypertension. "
The data was further based on patients at the American Mayo Clinic, who are very different from the diverse multicultural population of the UK, he noted. "These details really matter in AI-based studies."
* Attia ZI, PA Noseworthy, Lopez, Jimenez F, et al. An ECG algorithm allowed artificial intelligence to identify patients with atrial fibrillation during sinus rhythm: a retrospective badysis of the prediction of the results. The Lancet, August 1, 2019. DOI: 10.1016 / S0140-6736 (19031721-0
**Hendriks JML, Fabritz L. AI can now identify atrial fibrillation by sinus rhythm. The lancet. Posted on August 1, 2019. DOI: 10.1016 / S0140-6736 (19) 31719-2
Source link
