NCT07023510

The VALVE-AI Trial

VALidation of Screening Valvular Heart Disease Using Electrocardiogram Powered by Artificial Intelligence: A Randomized Controlled Trial

• Status: Recruiting
• Phase: N/A
• Study type: Interventional
• Enrollment: 8,648 (estimated)
• Sponsor: National Defense Medical Center, Taiwan · Academic / Other
• Sex: All
• Age: 60 Years – 85 Years
• Healthy volunteers: Not accepted

Summary

The goal of this clinical trial is to learn if an artificial intelligence-powered electrocardiogram (AI-ECG) can help detect moderate or severe valvular heart diseases (VHD) in adults. The main question it aims to answer is: ．Can AI-ECG screening identify patients with significant heart valve diseases who may benefit from early echocardiography? Researchers will compare the rate of moderate or severe VHD detection between the experimental group and the control group to see if AI-ECG improve the detection rate of significant VHD. Participants will: * Be classified as high- or low-risk for VHD using an AI-ECG system * In the experimental group, high-risk participants will receive echocardiography based on AI-ECG results * In the control group, usual clinical care will be provided without routine echocardiography for AI-ECG high-risk results.

Detailed description

This randomized controlled trial investigates the effectiveness of an artificial intelligence-powered electrocardiogram (AI-ECG) system for early screening of moderate or severe valvular heart disease (VHD) in adults receiving routine ECG examinations. The study population consists of adult outpatients undergoing a standard 12-lead ECG for any clinical indication. Each ECG is analyzed by a validated deep learning algorithm that automatically classifies the patient's risk for significant VHD. Participants identified as high-risk by the AI-ECG system are randomized into either an experimental group or a control group. In the experimental group, high-risk participants undergo transthoracic echocardiography to confirm or exclude moderate or severe VHD. In the control group, high-risk participants continue with usual clinical care without additional echocardiographic screening based solely on the AI-ECG result. Low-risk participants in both groups receive routine care without additional intervention. The primary aim is to determine whether AI-guided ECG screening, coupled with targeted echocardiography in the experimental group, increases the detection rate of clinically significant VHD compared to usual care. Secondary objectives include evaluating the impact on timely diagnosis, downstream clinical management, and the feasibility of integrating AI-ECG screening into routine outpatient workflows. The study will follow participants for up to 90 days post-randomization to assess the detection rate and related outcomes.

Conditions

• Valvular Heart Disease Patients

Interventions

Type	Name	Description
DIAGNOSTIC_TEST	AI-ECG driven echocardiography	The intervention utilizes a previously validated deep learning model based on 12-lead electrocardiogram (ECG) data to screen for moderate-to-severe valvular heart diseases (VHD). The model processes raw ECG signals and integrates age and sex to enhance prediction. (doi: 10.18632/aging.205835.) Participants identified as high-risk for any moderate-to-severe VHD by the algorithm of artificial intelligence-powered electrocardiogram (AI-ECG) in this intervention arm will receive transthoracic echocardiography to confirm diagnosis and guide further management.

Timeline

• Start date: 2025-07-01
• Primary completion: 2026-07-01
• Completion: 2026-07-01
• First posted: 2025-06-17
• Last updated: 2025-06-26

Locations

1 site across 1 country: Taiwan