News | Artificial Intelligence | November 22, 2021

New Study Shows Cardiologs' Deep Learning AI Outperforms the Apple Watch ECG Algorithm in Detecting Atrial Arrhythmias

Clinical data presented at the American Heart Association Scientific Sessions 2021


Cardiologs, a leader in artificial intelligence (AI) cardiology diagnostics, announced the results of a clinical study that proved Cardiologs’ deep learning AI outperforms the Apple Watch ECG algorithm in the detection of atrial arrhythmias (AA), which are related to stroke risks. The data was accepted for a moderated digital poster presentation at the American Heart Association (AHA) Scientific Sessions 2021. #AHA ECG Monitoring AI

November 22, 2021 — Cardiologs, a leader in artificial intelligence (AI) cardiology diagnostics, announced the results of a clinical study that proved Cardiologs’ deep learning AI outperforms the Apple Watch ECG algorithm in the detection of atrial arrhythmias (AA), which are related to stroke risks. The data was accepted for a moderated digital poster presentation at the American Heart Association (AHA) Scientific Sessions 2021.

The study, which was conducted by a clinical team from the Institut Cardiovasculaire Paris-Sud (ICPS) in Paris, France, assessed the accuracy of AA detection using the original Apple Watch companion application (ECG 1.0 App) compared to Cardiologs' proprietary deep neural network (DNN). The team wanted to determine if the Apple ECG application was missing important atrial arrhythmias when monitoring patients remotely, and if so, could a more intelligent algorithm used in place of the foundational ECG 1.0 App reduce that risk and improve smartwatch ECG results. Findings demonstrated that the Cardiologs DNN increased sensitivity by 34% while maintaining specificity. In addition, Cardiologs reduced the number of unclassified readings from 25% to 1%, virtually eliminating the common challenge of inconclusive results from wearable smartwatch ECGs.

"We have found that having smartwatches, such as the Apple Watch, to monitor our patients for arrhythmias outside of our center is extremely valuable, and to find an algorithm that significantly improves the overall performance for detection of AA from wearable tech recordings is exciting for me, my colleagues and the patients we care for. Using a more intelligent algorithm will reduce the time spent by physicians reviewing inconclusive results, and also reduce the risk of missed, delayed, and incorrect diagnosis," said Dr. Laurent Fiorina, cardiac electrophysiologist at ICPS. "I look forward to exploring this opportunity to enhance smartwatch readings and supporting future studies of the Cardiologs DNN."

"Cardiac monitoring is almost a standard feature of modern wearables like the Apple Watch, but there are still opportunities to improve ECG analysis to get a more accurate diagnosis. It's exciting to see that our deep learning algorithm can provide a more precise reading, promising many benefits for both patients and healthcare professionals. This advance will no doubt clear the path for the next generation of wearable health technology," said Cardiologs CEO and Co-founder Yann Fleureau.

Cardiologs’ presentation, "An Artificial Intelligence-Based Smartwatch ECG Analysis Improves the Detection of Atrial Arrhythmia," will be available on-demand from Dec. 16–31, 2021.

About Cardiologs

Cardiologs is a medical technology company committed to transforming cardiac diagnostics using medical-grade artificial intelligence and cloud technology. Developed in partnership with leading physicians, the Cardiologs Holter Platform empowers clinicians worldwide to deliver expert cardiac care faster and more efficiently. CE-Marked and FDA cleared for detection of 14 cardiac arrhythmias, the Cardiologs Holter Platform is built on a growing database of more than 20 million ECG recordings and is supported by a number of clinical publications.

For more information: https://cardiologs.com/

Find more AHA 2021 late-breaking studies and video

Related Content

Feature | AHA | By Dave Fornell, DAIC Editor

The following are the late-breaking science presentation sessions at the 2021 American Heart Association (AHA) 2021 ...

Home November 23, 2021
Home
News | AHA

September 20, 2021 — The American Heart Association (AHA) announced Sept. 16 it decided to convert from a planned in ...

Home September 20, 2021
Home
Feature | AHA | Dave Fornell, Editor

Here is a list of the American Heart Association (AHA) late-breaking clinical trial presentations at the 2020 Virtual ...

Home November 20, 2020
Home
News | AHA

July 15, 2020 – The last major cardiovascular conference to go virtual in 2020 due to the COVID-19 (SARS-Cov-2) pandemic ...

Home July 15, 2020
Home
News | AHA

November 22, 2019 — Artificial intelligence can examine electrocardiogram (ECG) test results, a common medical test, to ...

Home November 22, 2019
Home
News | AHA

November 21, 2019 — At the American Heart Association Scientific Sessions 2019, Eko, a digital health company applying ...

Home November 21, 2019
Home
News | AHA

November 21, 2019 — People who experience cardiac arrests over the weekend are less likely to survive long enough to be ...

Home November 21, 2019
Home
News | AHA

November 20, 2019 — Heart attack survivors may have an increased risk of developing cancer compared to people without ...

Home November 20, 2019
Home
News | AHA

November 20, 2019 — Ultra-processed foods, which account for more than half of an average American’s daily calories, are ...

Home November 20, 2019
Home
News | AHA

November 20, 2019 — Frequent cannabis (marijuana) use among young people was linked to an increased risk of stroke, and ...

Home November 20, 2019
Home
Subscribe Now