Use of an artificial intelligence (AI)-based app using electrocardiogram (ECG) signals recorded with an Apple Watch is able to identify left-ventricular dysfunction, research presented at the Heart Rhythm Society’s 2022 annual meeting (HRS 2022, 29 April–1 May, San Francisco, USA) has found.
The tracings, which are recorded in non-clinical environments and transferred to the medical centre for AI analysis, are also presented on an interactive dashboard integrated with the electronic health record to facilitate clinician review. Researchers from Mayo Clinic (Rochester, USA), presented insights into the use of the app during a late-breaking trial session at HRS 2022.
Considered stage B heart failure, left-ventricular dysfunction is often asymptomatic and, if left untreated, can lead to adverse events. Traditional ECGs are unable to identify and diagnose ventricular dysfunction, even with expert human interpretation, according to the researchers.
The introduction of AI to 12 lead ECG analysis in a clinical setting enabled the ECG to identify left ventricular dysfunction, they note. The authors of the study sought to adapt the neural network so that instead of reading a 12 lead ECG, it could read an Apple Watch ECG to assess applying the test remotely and, by using patient-owned equipment, inexpensively.
Mayo Clinic patients with the Mayo Clinic iOS app and an Apple Watch were invited to participate in the study. Of the 3,884 Mayo Clinic patients who reported Apple Watch ownership, 2,454 subjects from 46 states and 11 countries participated and downloaded the study app.
The average age of participants was 53±15 years and 56% were female. The app, developed by Mayo Clinic, sent all previously recorded ECGs for clinician review. ECGs acquired from the wearable devices within one month of a clinically ordered ECG were analysed by AI for the presence of ejection fraction ≤40% using a model adapted for single lead use.
“We have seen how artificial intelligence has revolutionised the already common ECG into a tool that can be used to identify occult cardiovascular diseases. Our team saw vast potential to expand tracking outside of a physician’s office by using popular wearable devices,” said Zachi Itzhak Attia, (Mayo Clinic, Rochester, USA). “We set out to create a platform that could not only provide accurate readings, but also would yield high patient engagement with an easy to navigate, user-friendly process that can be completed from the comfort of a patient’s home.”
Between August, 2021 and August, 2022, patients shared 125,610 ECGs and 92% of patients used the app more than once. Of the participants, 421 had at least one sinus rhythm (NSR) ECG (avg: 17 ECGs, with NSR determined by watch algorithm) within 30 days of an echocardiogram.
Among the participants, 16 of these patients (3.8%) had an EF ≤40%, and 13 out of these 16 patients were identified by the watch AI ECG, which had an Area Under the Curve (AUC) 0.875, sensitivity 81.2% and specificity 81.3%.
“These findings show that the application of AI to a wearable device ECG can effectively monitor left ventricular dysfunction. For patients who might unknowingly have this condition, such as those with hypertension, diabetes, advancing age, and people receiving some forms of chemotherapy, the tool could enable early detection and help physicians optimize treatment options,” said Paul Friedman (Mayo Clinic, Rochester, USA). “This technology has the potential to be scaled and adopted by hospital systems to better serve patients, particularly in remote communities or geographically diverse populations around the world, potentially addressing healthcare disparities, and enabling physicians to offer more coordinated patient care.”
The authors are currently seeking US Food and Drug Administration (FDA) approval for the current algorithm used in this trial, and would like to see further studies test additional AI algorithms developed by their team. They would also like expand this current interface for additional data collection and to screen for other common heart conditions among patients, such as atrial fibrillation.
