Heart Attacks can be potentially detected using a Smartwatch and a little Ai according to a new study

sensors on the back of Apple Watch Series 6

Smartwatches have come a long way in the area of health and wellness monitoring since the introduction of the Apple Watch. FDA Approved ECG (Electrocardiogram) and Atrial Fibrillation detection has become staples on the Apple Watch, Samsung Galaxy Watch, and some Fitbit devices.

Related Reading:

The problem

However, all the current smartwatches offering heart monitoring features have one major drawback. 

They all generate an ECG that is similar to a single-lead ECG.

When you take an ECG in a doctor’s office, a standard 12-lead ECG is usually taken. This 12-lead ECG records electrical signals from different angles in the heart to produce twelve different waveforms.

A single-lead ECG is able to provide information about heart rate and heart rhythm and enables the classification of AFib. 

However, a single-lead ECG cannot be used to identify some other conditions, like heart attacks.

Multiple-lead ECG is necessary for the accurate and robust detection of cardiac disorders, particularly acute myocardial infarction.

The Big Story

According to a new study published by researchers in South Korea, advances in Artificial intelligence based algorithms can help with the detection of myocardial infarction (heart attacks) using a single lead smartwatch.

Apple Watch can be used to generate multi-channel electrocardiograms with some maneuvering according to earlier studies. Last year, researchers found that ECGs recorded on smartwatches might be useful to obtain an earlier diagnosis of acute coronary syndromes. These researchers used an Apple Watch Series 4 for their study. Their findings were published in JAMA Cardiology. 

Apple Watch and other smartwatches with ECG functionality can be used to obtain multi lead asynchronous ECG recordings by placing the watch sensor sequentially on different parts of the body.

Multilead ECG using smartwatch
Example of measuring multi-lead electrocardiogram (ECG) from a smartwatch. Multiple-lead ECG can be obtained from smartwatches by sequentially placing the smartwatch on different parts of the body.

Researchers of this new study noted that asynchronous ECG lead sets can be derived from ECG reports to simulate a situation similar to the sequential recording of ECG leads via smartwatches. For example, a 4-lead subset consisting of leads I, aVR, V1, and V4 from the ECG report is completely asynchronous.

According to the researchers of this study, “Our primary aim was to develop an AI model for detecting acute myocardial infarction from asynchronous ECG signals, which outperforms the automatic ECG interpretation provided by the GE ECG analysis program. 

Our secondary aim was to determine the optimal number of leads required for sufficient diagnostic power.”

For their AI based algorithm, they extracted data for 97,742 patients aged 20 years or older with 183,982 ECGs recorded within 24 hours from each visit to the emergency room.

Major Findings

This study shows the feasibility of multiple lead-based AI-enabled ECG algorithms on smartwatches for the automated diagnosis of cardiac disorders by developing an AI model for detecting acute myocardial infarction with asynchronous ECG signals.

The study results showed that measuring at least 3 leads, and ideally more than 4 leads, is necessary for accurate detection.

For every 30-minute delay in coronary reperfusion, the relative 1-year mortality rate increases by 7.5%. 

According to the researchers, with smartwatches, measuring additional leads would only take less than a minute, and the benefit of doing so would greatly outweigh the risk.

The researchers argue that their study has important medical and economic impacts.

“First, our model can significantly reduce time to diagnosis, and consequently reduce time to reperfusion, which is the elapsed time between the onset of symptoms and reperfusion and is critical to the clinical outcome of the disease”

“With our model implemented on smartwatches, reliable preliminary diagnosis can be made even before contact with emergency services, thereby greatly reducing the time from the onset of symptoms to diagnosis. With the preliminary diagnosis already made, patients can be promptly triaged to the most appropriate form of treatment”.

You can read about the details of this new study along with the researcher’s guidelines on future work, limitations around the current approach, and more.

The ability of Apple’s ECG app to accurately classify an ECG recording into AFib and sinus rhythm was tested in a clinical trial of approximately 600 subjects and demonstrated 99.6% specificity with respect to sinus rhythm classification and 98.3% sensitivity for AFib classification for the classifiable results.

Similar large-scale clinical studies would be required to pursue the efficacy of using multi-lead asynchronous ECG for detecting Heart Attacks with a smartwatch.

With improvements in Edge computing and smartwatch companies adding more processing power as well as battery power, it is not long before we get to see some sophisticated health monitoring technologies on our wrists that could potentially save lives.

Source: Automated Detection of Acute Myocardial Infarction Using Asynchronous Electrocardiogram Signals—Preview of Implementing Artificial Intelligence With Multichannel Electrocardiographs Obtained From Smartwatches: Retrospective Study

I am a technologist with years of experience with Apple and wearOS products and have a Bachelor’s degree in Computer Science. In my day job, I advise fortune 500 companies with their digital transformation strategies and also consult with numerous digital health startups in an advisory capacity. I'm VERY interested in exploring the digital health&fitness-tech evolution and keep a close eye on patents, FDA approvals, strategic partnerships and developments happening in the wearables and digital health sector. When I'm not writing or presenting, I run with my Apple Watch or Fossil Gen 5 LTE and keep a close eye on my HRV and other recovery metrics.

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.