AI-Enhanced Detection for Aortic Stenosis

Enrolling by invitation at 3 trial locations

Overseen By Rohan Khera, MD, MS

Age: 65+

Sex: Any

Trial Phase: Academic

Sponsor: Yale University

Disqualifiers: Echocardiogram, Aortic valve replacement, Dementia, others

No Placebo Group All trial participants will receive the active study treatment (no placebo)

Compensation: Varies

Number of Visits: 1

What You Need to Know Before You Apply

What is the purpose of this trial?

This trial aims to evaluate how effectively artificial intelligence (AI) can detect aortic stenosis, a condition where the heart's aortic valve narrows. Using portable devices like a small ECG (a test that checks heart rhythm) and ultrasound, the study assesses whether AI can better predict who might have this heart issue. Participants will join either an experimental group, using AI-guided devices such as the AI-ECG risk algorithm and AI-POCUS (AI-enhanced point-of-care ultrasound), or a control group, using standard methods. This trial suits individuals aged 70 or older attending routine check-ups at specific clinics, provided they haven't undergone certain heart procedures or recent heart screenings.

As an unphased trial, this study offers participants the opportunity to contribute to innovative research that could enhance heart health detection methods.

Do I need to stop my current medications for the trial?

The trial information does not specify whether you need to stop taking your current medications.

What prior data suggests that these AI-based diagnostic methods are safe for detecting aortic stenosis?

Research has shown that both the AI-ECG risk algorithm and AI-POCUS are generally safe and well-tolerated in studies.

The AI-ECG risk algorithm is a software tool that analyzes data from heart activity tests (ECGs) to help detect aortic stenosis, a narrowing of a heart valve. Studies have found no negative effects, as it only uses existing data and doesn't directly impact health.

AI-POCUS assists doctors in identifying severe aortic stenosis by enhancing ultrasound imaging. Research indicates it is effective and safe, as it improves image interpretation without altering the ultrasound process itself.

In summary, both technologies in this trial are software tools that help detect heart issues without adding any risk to patients. ^{1 2 3 4 5}

Why are researchers excited about this trial?

Researchers are excited about using AI-enhanced techniques for detecting aortic stenosis because they could significantly improve early diagnosis. Unlike traditional methods that rely heavily on physical exams and echocardiograms, this approach uses an AI-driven risk algorithm to analyze portable 1-lead ECGs and, if needed, point-of-care ultrasounds (POCUS). This combination aims to streamline the screening process, making it faster and more accessible. By leveraging AI, the hope is that these techniques will catch aortic stenosis earlier, potentially improving patient outcomes and reducing the strain on healthcare resources.

What evidence suggests that this trial's AI-based diagnostic methods are effective for detecting aortic stenosis?

This trial will compare the effectiveness of AI-ECG and AI-POCUS in detecting aortic stenosis. Research has shown that the AI-ECG risk algorithm effectively detects aortic stenosis (AS), a heart condition. One study found this AI model to be 85% accurate in identifying moderate to severe AS. Another study reported an even higher accuracy of 98.6%, demonstrating excellent detection capabilities. For the AI-POCUS method, research indicates it can accurately identify AS using portable ultrasound devices. One report highlighted an impressive accuracy rate of over 99% for AI-guided detection of heart conditions with ultrasound. Overall, combining AI-ECG and AI-POCUS provides a powerful tool for effectively identifying aortic stenosis. ^{1 3 4 6 7}

Who Is on the Research Team?

Rohan Khera, MD, MS

Principal Investigator

Yale University

Are You a Good Fit for This Trial?

This trial is for individuals aged 70 or older attending routine outpatient primary care clinics at specific sites. It's not for those with prior aortic valve replacement/repair, implantable cardiac devices, moderate/severe AS history, recent echocardiograms, heart transplants, non-English speakers, dementia or less than a year to live.

Inclusion Criteria

Attending a routine outpatient primary care clinic at one of the three enrollment sites

I am 70 years old or older.

Exclusion Criteria

I have had surgery or a procedure to fix or replace my aortic valve.

Presence of implantable cardiac devices, including permanent cardiac pacer, implantable cardioverter-defibrillator, or left ventricular assist device

Prior history of moderate or severe AS

Timeline for a Trial Participant

Screening

Participants are screened for eligibility to participate in the trial

2-4 weeks

Diagnostic Assessment

Participants undergo sequential screening for aortic stenosis using portable 1-lead ECGs, followed by point-of-care ultrasound (POCUS) if indicated by AI-based risk algorithms

4-6 weeks

1-2 visits (in-person)

Follow-up

Participants are monitored for diagnosis of advanced aortic stenosis via transthoracic echocardiogram (TTE) and electronic health record (EHR) review

12 months

What Are the Treatments Tested in This Trial?

Interventions

• AI-ECG risk algorithm
• AI-POCUS

Trial Overview The DETECT-AS Diagnostic Study is testing AI tools that analyze data from portable ECGs and cardiac ultrasounds to predict the risk of aortic stenosis in patients. The study aims to enhance early detection using these advanced technologies.

How Is the Trial Designed?

Group I: Intervention Experimental Treatment 4 Interventions

The intervention arm will undergo sequential screening for aortic stenosis using portable 1-lead electrocardiograms (ECGs), followed by point-of-care ultrasound (POCUS), if indicated, by artificial intelligence (AI)-based risk algorithms.

Group II: Control Placebo Group 2 Interventions

The control arm will undergo a portable 1-lead electrocardiogram (ECG), with 10% randomly assigned to undergo point-of-care ultrasound (POCUS).

Who Is Running the Clinical Trial?

Yale University

Lead Sponsor

Trials

1,963

Recruited

3,046,000+

The Methodist Hospital Research Institute

Collaborator

Trials

299

Recruited

82,500+

Icahn School of Medicine at Mount Sinai

Collaborator

Trials

933

Recruited

579,000+

Kaiser Permanente School of Medicine

Collaborator

Trials

4

Recruited

1,300+

National Institute on Aging (NIA)

Collaborator

Trials

1,841

Recruited

28,150,000+

Citations

1. academic.oup.com academic.oup.com/ehjdh/article/6/4/656/8180648

External assessment of an artificial intelligence-enabled ... The AI-ECG model achieved an AUC of 0.85 (95% CI: 0.84–0.87) in detecting moderate to severe AS. Sensitivity, specificity, PPV, NPV, and ...

2. pmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov/articles/PMC10373677/

Enhanced detection of severe aortic stenosis via artificial ... The performance of our AI-DSA to detect severe AS is shown in figure 2—the AUROC (95% CI) being close to one overall (0.986 (0.985 to 0.987)) ...

3. sciencedirect.com sciencedirect.com/science/article/pii/S2772963X25005460

Prediction of Aortic Stenosis Progression Using Artificial ... During the follow-up, 1,625 (47%) patients developed severe AS. The model demonstrated strong predictive performance—an area under the curve– ...

4. ahajournals.org ahajournals.org/doi/10.1161/JAHA.119.014717

Deep Learning–Based Algorithm for Detecting Aortic ... The deep learning–based algorithm demonstrated high accuracy for significant AS detection using both 12‐lead and single‐lead ECGs. Clinical ...

5. clinmedres.org clinmedres.org/content/22/3/145

Diagnostic Accuracy of AI Algorithms in Aortic Stenosis ... The present meta-analysis revealed that AI algorithms serve as powerful screening tools for the detection of patients with moderate to severe AS.

6. pmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov/articles/PMC11416525/

AI-Enhanced Prediction of Aortic Stenosis Progression This study demonstrates how an artificial intelligence-guided approach in clinical routine could help enhance risk stratification of AS.

7. nature.com nature.com/articles/s41598-025-19916-5

Incidence and severity of aortic stenosis according to ... Deep learning algorithms can use ECGs to risk stratify for undiagnosed AS27, and chest radiographs to differentiate people with and without AS ...