Automated AI reading of echocardiograms to detect coronary artery disease
Automated Echocardiographic Detection of Coronary Artery Disease Using Artificial Intelligence Methods
Beijing Hospital · NCT06314295
This project will test whether artificial intelligence can read heart ultrasound videos to detect coronary artery disease in people who are planning to have coronary angiography.
Quick facts
| Study type | Observational |
|---|---|
| Enrollment | 1500 (estimated) |
| Ages | 18 Years to 90 Years |
| Sex | All |
| Sponsor | Beijing Hospital (other gov) |
| Locations | 1 site (Beijing) |
| Trial ID | NCT06314295 on ClinicalTrials.gov |
What this trial studies
The project uses new ultrasound measures such as non-invasive myocardial work combined with automated image analysis to improve non-invasive detection of coronary artery disease. Researchers will train and apply deep-learning models to classify cardiac ultrasound views, perform automatic segmentation of the left ventricular cavity, myocardium, and left atrial wall using a CLAS model, and track motion to generate velocity vector maps of intracardiac flow. The automated diagnostic model will be validated and optimized by comparing AI outputs with coronary angiography results in patients with suspected coronary artery disease. The work is observational and focuses on cases with adequate image quality while excluding patients with major valvular disease, hypertrophic cardiomyopathy, or severe arrhythmias.
Who should consider this trial
Good fit: Ideal candidates are adults with suspected coronary artery disease who are planning to undergo coronary angiography and can obtain good-quality transthoracic echocardiogram images.
Not a fit: Patients with poor-quality ultrasound images or excluded conditions such as aortic stenosis, prior aortic valve replacement, hypertrophic cardiomyopathy, severe valve disease, severe arrhythmia, severe cardiomyopathy, or severe congenital heart disease are unlikely to benefit from the AI model.
Why it matters
Potential benefit: If successful, this could provide a faster, non-invasive ultrasound-based way to screen and triage people for coronary artery disease, reducing dependence on invasive testing for some patients.
How similar studies have performed: Prior work applying AI to echocardiographic segmentation and automated measurements has shown promising results, but fully automated ultrasound-video diagnostic models for coronary artery disease remain early and not widely validated.
Eligibility criteria
Show full inclusion / exclusion criteria
Inclusion Criteria: * Patients with suspected coronary artery disease * Patients plan to undergo coronary angiography Exclusion Criteria: * Patients with aortic valve stenosis * Patients with aortic valve replacement surgery * Patients with hypertrophic cardiomyopathy * Patients with severe heart valve disease * Patients with severe arrhythmia * Patients with severe cardiomyopathy * Patients with severe congenital heart disease * The quality of ultrasound images is poor
Where this trial is running
Beijing
- Beijing Hospital — Beijing, China (RECRUITING)
How to participate
- Review the eligibility criteria above with your treating physician.
- Visit the official trial page on ClinicalTrials.gov for the most current contact information and recruitment status.
- Contact the listed study coordinator or principal investigator to request pre-screening. Pre-screening is free and never obligates you to enroll.
Conditions: Coronary Artery Disease