AI-Assisted Heart Monitoring for Heart Disease

(ECG-AID Trial)

Recruiting at 2 trial locations

Overseen By ECG-AID Study

Age: 18+

Sex: Any

Trial Phase: Academic

Sponsor: Tempus AI

Disqualifiers: Pacemaker, Atrial fibrillation, Cardiac surgery, others

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

Compensation: Varies

Number of Visits: 1

1000 Participants Needed

Trial Summary

What is the purpose of this trial?

Atrial fibrillation is an abnormal beating of the heart that can lead to stroke or heart failure. Structural heart diseases are conditions that affect the heart valves or heart muscle and can cause permanent heart damage if left untreated. Sometimes people have atrial fibrillation or structural heart disease and do not know it. The purpose of this study is to evaluate two devices that can predict who has or may develop atrial fibrillation or structural heart disease based on the results of an electrocardiogram.

Will I have to stop taking my current medications?

The trial information does not specify whether you need to stop taking your current medications. Please consult with the study team or your healthcare provider for guidance.

What data supports the effectiveness of the treatment Electrocardiogram-based Artificial Intelligence for heart disease?

Research shows that using artificial intelligence with electrocardiograms (ECGs) can effectively identify heart problems like left ventricular dysfunction, which can lead to heart failure. In a study with over 52,000 patients, the AI model accurately detected heart issues with high sensitivity and specificity, making it a powerful tool for early screening. ^{1 2 3 4 5}

Is AI-assisted heart monitoring safe for humans?

The research on AI-assisted heart monitoring, particularly using electrocardiograms (ECGs), suggests it is a safe, non-invasive tool for detecting heart issues like left ventricular dysfunction. It has been tested on large groups of patients and adapted for use with wearable devices, showing promise as a reliable screening method without reported safety concerns. ^{4 5 6 7 8}

How does the AI-Assisted Heart Monitoring treatment differ from other heart disease treatments?

This treatment is unique because it uses artificial intelligence to analyze electrocardiograms (ECGs) for real-time heart monitoring, which can detect heart issues like arrhythmias and ventricular dysfunction more efficiently and accurately than traditional methods. It allows for continuous monitoring in a home environment, providing immediate and actionable insights to both patients and clinicians. ^{4 5 9 10 11}

Research Team

John Pfeifer, MD

Principal Investigator

Tempus AI, Inc.

Eligibility Criteria

This trial is for adults aged 40 or older who have had an ECG during routine care. For the atrial fibrillation group, participants must be 65 or older and able to identify a healthcare provider to receive patch monitor results. For structural heart disease, they need to be at least 40 and can't have severe valve issues or poor heart pump function.

Inclusion Criteria

I am over 40 and had an ECG for care between the study dates. I can name a healthcare provider for my echocardiogram results.

I am 65 or older and can name a healthcare provider to get my heart monitor results.

I am 40 or older and have had an ECG as part of my regular health care.

Exclusion Criteria

You cannot finish the follow-up studies on time, are currently in the hospital, have a permanent pacemaker or implanted cardiac defibrillator, have a history of certain heart conditions, had recent or planned heart surgery, or are allergic to adhesive.

I can follow the study schedule and am not currently hospitalized. I don't have severe heart valve issues, weak heart muscle, thick heart walls, or an allergy to ultrasound gel.

Retrospective Phase: Patients who have previously requested that their data not be involved in any secondary use application such as a research study

Timeline

Screening

Participants are screened for eligibility to participate in the trial

2-4 weeks

Initial Assessment

Participants undergo initial assessment including echocardiogram for SHD cohort and ECG patch monitor for AF cohort

2 weeks

1 visit (in-person)

Monitoring

Participants in the AF cohort wear an ECG patch monitor for up to 3 times over 12 months

12 months

Follow-up

Participants are monitored for safety and effectiveness after initial assessment

6 months

Treatment Details

Interventions

• Electrocardiogram-based Artificial Intelligence

Trial Overview The study tests two devices: an echocardiogram and Zio Patch Monitor, which may help predict atrial fibrillation or structural heart disease using electrocardiogram results. It includes looking back at past patient data as well as monitoring new patients.

Participant Groups

Group I: SHD Cohort Experimental Treatment 1 Intervention

Will be comprised 500 participants at increased risk for Structural Heart Disease (SHD) will be referred for a single echocardiogram.

Group II: AF Cohort Experimental Treatment 1 Intervention

Will be comprised of 500 participants predicted to be increased risk for Atrial Fibrillation (AF) will receive a 2-week ECG patch monitor to wear (up to 3 times over 12 months),

Who Is Running the Clinical Trial?

Tempus AI

Lead Sponsor

Trials

18

Recruited

20,700+

Tempus Labs

Lead Sponsor

Trials

17

Recruited

20,200+

Findings from Research

A study involving 221,846 ECGs from four institutions aimed to develop AI models for detecting left ventricular systolic dysfunction (LVSD) with an ejection fraction (EF) <40%, showing promising internal accuracy but variable external validation results.

The performance of AI models varied significantly between institutions, emphasizing the need for external validation and careful consideration of patient characteristics and ECG abnormalities when using AI for LVSD detection.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Importance of external validation and subgroup analysis of artificial intelligence in the detection of low ejection fraction from electrocardiograms. Yagi, R., Goto, S., Katsumata, Y., et al.[2023]

The AI-ECG algorithm demonstrated strong performance in detecting left ventricular systolic dysfunction (LVSD) in an external population of 4277 adults, achieving an area under the receiver operating curve of 0.82, indicating good accuracy.

While the AI-ECG showed high specificity (97.4%) and accuracy (97.0%), its sensitivity was lower at 26.9%, suggesting that population-specific cut-offs may be needed for optimal clinical use, especially given the differences in patient characteristics compared to the original study.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

External validation of a deep learning electrocardiogram algorithm to detect ventricular dysfunction. Attia, IZ., Tseng, AS., Benavente, ED., et al.[2021]

A study involving 44,959 patients demonstrated that an AI model applied to ECG data can effectively identify asymptomatic left ventricular dysfunction (ALVD), achieving high accuracy (85.7%) and sensitivity (86.3%).

Patients who screened positive for ALVD using the AI model were found to be four times more likely to develop future ventricular dysfunction, highlighting the potential of AI-enhanced ECG as a proactive screening tool for early intervention.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Screening for cardiac contractile dysfunction using an artificial intelligence-enabled electrocardiogram. Attia, ZI., Kapa, S., Lopez-Jimenez, F., et al.[2022]

The study emphasizes the need to expand cardiovascular safety assessments beyond just electrocardiogram events to include a wider range of cardiovascular parameters, using real-life case studies to demonstrate progress.

To enhance the prediction of cardiovascular events in patients, the development of more relevant humanized models is essential, allowing for better translation of findings to novel therapeutic approaches.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Current and future approaches to nonclinical cardiovascular safety assessment. Collins, TA., Rolf, MG., Pointon, A.[2021]

A comprehensive database of 396,985 drug-related cardiac adverse effects (AEs) linked to 1,632 approved drugs was created, revealing significant clusters of AEs affecting heart function, such as arrhythmias and heart failure.

The study identified that cardiac AEs are highly correlated with drugs targeting cardiovascular functions and specific mechanisms of action, including interactions with various receptors like alpha- and beta-adrenergic receptors, which may help predict and understand these adverse effects.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Prediction of drug-related cardiac adverse effects in humans--A: creation of a database of effects and identification of factors affecting their occurrence. Matthews, EJ., Frid, AA.[2013]

AI-enhanced electrocardiograms (AI-ECG) can effectively track changes in cardiac structure and function in patients with obstructive hypertrophic cardiomyopathy (HCM) undergoing treatment with mavacamten, as shown in a study involving 13 patients and 216 ECGs.

Both AI-ECG algorithms demonstrated significant reductions in HCM scores during treatment, correlating well with echocardiographic measures and laboratory markers, suggesting that AI-ECG could be a valuable tool for monitoring therapeutic responses in HCM.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Patient-Level Artificial Intelligence-Enhanced Electrocardiography in Hypertrophic Cardiomyopathy: Longitudinal Treatment and Clinical Biomarker Correlations. Siontis, KC., Abreau, S., Attia, ZI., et al.[2023]

References

1. United States pubmed.ncbi.nlm.nih.gov

Why recording of an electrocardiogram should be required in every inpatient and outpatient encounter of patients with heart failure. [2011]

2. England pubmed.ncbi.nlm.nih.gov

Artificial Intelligence-Augmented Electrocardiogram Detection of Left Ventricular Systolic Dysfunction in the General Population. [2023]

3. England pubmed.ncbi.nlm.nih.gov

Importance of external validation and subgroup analysis of artificial intelligence in the detection of low ejection fraction from electrocardiograms. [2023]

4. Netherlands pubmed.ncbi.nlm.nih.gov

External validation of a deep learning electrocardiogram algorithm to detect ventricular dysfunction. [2021]

5. United States pubmed.ncbi.nlm.nih.gov

Screening for cardiac contractile dysfunction using an artificial intelligence-enabled electrocardiogram. [2022]

6. England pubmed.ncbi.nlm.nih.gov

Current and future approaches to nonclinical cardiovascular safety assessment. [2021]

7. England pubmed.ncbi.nlm.nih.gov

Detection of left ventricular systolic dysfunction from single-lead electrocardiography adapted for portable and wearable devices. [2023]

8. Netherlands pubmed.ncbi.nlm.nih.gov

Prediction of drug-related cardiac adverse effects in humans--A: creation of a database of effects and identification of factors affecting their occurrence. [2013]

9. United States pubmed.ncbi.nlm.nih.gov

Patient-Level Artificial Intelligence-Enhanced Electrocardiography in Hypertrophic Cardiomyopathy: Longitudinal Treatment and Clinical Biomarker Correlations. [2023]

10. United States pubmed.ncbi.nlm.nih.gov

An artificial intelligence-enabled ECG algorithm for comprehensive ECG interpretation: Can it pass the 'Turing test'? [2022]

11. United States pubmed.ncbi.nlm.nih.gov

The Role of Artificial Intelligence in Arrhythmia Monitoring. [2021]