An augmented reality-based method to assess precordial electrocardiogram leads: a feasibility trial

Area of Science:

• Cardiology
• Biomedical Engineering
• Medical Technology

Background:

• Smartwatch electrocardiograms (ECGs) show promise for detecting cardiac conditions like myocardial ischemia.
• Accurate placement of bipolar chest leads is a significant hurdle for smartwatch ECGs in outpatient settings.

Purpose of the Study:

• To develop and assess an augmented reality (AR)-based smartphone application for guiding correct smartwatch ECG lead placement.
• To evaluate the feasibility and accuracy of AR-guided, multichannel smartwatch ECG recordings in an outpatient environment.

Main Methods:

• An AR smartphone app was developed, utilizing the front camera to guide users in placing a smartwatch on predefined chest positions (V1-V6).
• A machine learning model (MobileNet_v2) was trained to identify lead positions and project them onto the user's chest.
• Patients recorded 9-lead smartwatch ECGs with app assistance; a conventional 12-lead ECG served as a comparison.
• Blinded cardiologists compared smartwatch ECGs to standard unipolar Wilson leads based on visual characteristics.

Main Results:

• All 50 patients successfully recorded a 9-lead smartwatch ECG with app assistance.
• Twelve patients independently recorded all limb and chest leads using the app.
• At least 86% of smartwatch ECG leads were correctly assigned to standard unipolar Wilson leads by blinded cardiologists, indicating high similarity to conventional ECGs.

Conclusions:

• An AR-based method enables independent, multichannel smartwatch ECG recording in outpatient settings.
• The developed AR application significantly improves the accuracy and ease of smartwatch ECG lead placement.
• This technology holds potential for enhanced remote cardiac monitoring and early detection of cardiac pathologies.