Study Protocol for the Artificial Intelligence-Driven Evaluation of Structural Heart Diseases Using Wearable...

Area of Science:

• Cardiology
• Medical Devices
• Artificial Intelligence

Background:

• Portable electrocardiogram (ECG) devices offer potential for early structural heart disease (SHD) detection via AI-ECG algorithms.
• Real-world performance of AI-ECG for SHD screening is currently unknown.
• This study evaluates a wearable-adapted AI algorithm for SHD detection using single-lead portable ECGs.

Purpose of the Study:

• To assess the validity of a wearable-adapted AI-ECG algorithm for identifying SHD.
• To evaluate AI-ECG performance in a real-world clinical screening setting.
• To compare AI-ECG results with transthoracic echocardiogram (TTE) findings.

Main Methods:

• Cross-sectional study protocol at Yale New Haven Hospital (YNHH).
• Enrollment of 585 patients undergoing routine transthoracic echocardiograms (TTE).
• 1-lead ECG acquisition using Apple Watch and another portable device, linked to EHR data.
• AI-ECG algorithm performance assessed against TTE as the gold standard for SHD, including LVSD, valvular disease, and LVH.

Main Results:

• The study protocol is designed to assess AI-ECG algorithm performance in identifying SHD.
• Comparison of AI-ECG results with TTE data will determine algorithm validity.
• Data linkage with EHR and secure data maintenance are established.

Conclusions:

• The study aims to validate a wearable AI-ECG algorithm for real-world SHD screening.
• Findings will inform the utility of portable ECG devices and AI in cardiovascular disease management.
• This research addresses a critical gap in understanding AI-ECG performance in clinical practice.