Artificial Intelligence Software for Heart Disease

(DAISEA-ECG Trial)

Overseen By Marie-Gabrielle Lessard, MSc

Age: 18+

Sex: Any

Trial Phase: Academic

Sponsor: Montreal Heart Institute

Disqualifiers: Pediatrics, Unable to follow guidelines

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

Compensation: Varies

Number of Visits: Unknown

2000 Participants Needed

Trial Summary

Will I have to stop taking my current medications?

The trial information does not specify whether you need to stop taking your current medications. It is best to consult with the trial coordinators for specific guidance.

What data supports the effectiveness of the DeepECG platform treatment for heart disease?

Research shows that the DeepECG platform, which uses artificial intelligence to analyze ECGs (electrocardiograms), can effectively identify heart problems like left ventricular dysfunction. This AI tool has been shown to accurately screen for heart issues, potentially helping doctors catch problems early and improve patient outcomes. ^{1 2 3 4 5}

Is the Artificial Intelligence Software for Heart Disease safe for humans?

The research articles do not provide specific safety data for the AI software, but they discuss its use in predicting cardiovascular risks and improving patient outcomes, which suggests it is being integrated into medical decision-making processes. ^{6 7 8 9 10}

How is the DeepECG platform treatment for heart disease different from other treatments?

The DeepECG platform is unique because it uses artificial intelligence to analyze electrocardiogram (ECG) data for diagnosing heart conditions, offering a fully automated and accurate interpretation that mimics human-like analysis. This approach can potentially improve diagnostic accuracy and efficiency compared to traditional methods, which rely on manual interpretation by healthcare professionals. ^{1 11 12 13 14}

What is the purpose of this trial?

The DAISEA-ECG project aims to improve the diagnosis of heart diseases in primary care through the DeepECG platform, which combines ECG-AI and ECHONeXT algorithms. This study uses a stepped wedge design, where each Family Medicine Group acts as its own control. The FMGs will gradually transition from the control period (without AI recommendations) to the intervention period (with AI recommendations activated) in a randomized sequence. The primary objective is to compare the sensitivity of family physicians in detecting cardiac pathologies, with and without the assistance of the DeepECG platform. Sensitivity is defined as the proportion of patients correctly referred to cardiology or for transthoracic echocardiography (TTE) among those who indeed required cardiovascular evaluation, as confirmed by an independent adjudication committee.

Eligibility Criteria

This trial is for individuals with heart disease being seen by primary care providers. Participants must be part of a Family Medicine Group (FMG) that's included in the study. Specific eligibility criteria are not provided, but typically participants would need to meet certain health conditions.

Inclusion Criteria

Family physicians who have given their free and informed consent

I am 18 or older and do not have ongoing heart issues requiring specialist care.

Family Physicians or Nurse Practitioners practicing in one of the participating FMGs

Exclusion Criteria

I am a family physician who only treats children.

Family physicians unable to follow the project guidelines

Timeline

Screening

Participants are screened for eligibility to participate in the trial

2-4 weeks

Control Period

Family Medicine Groups operate without AI recommendations to establish baseline data

Varies by group

Intervention Period

Family Medicine Groups transition to using the DeepECG platform with AI recommendations

18 months

Follow-up

Participants are monitored for the effectiveness of AI recommendations in improving diagnostic sensitivity

4 weeks

Treatment Details

Interventions

• DeepECG platform

Trial Overview The DAISEA-ECG project tests if the DeepECG platform can help family doctors better detect heart diseases using AI algorithms. The study compares doctor's diagnostic sensitivity with and without AI assistance over different periods in a randomized sequence.

Participant Groups

Group I: DeepECG plateform diagnosis & recommendations Experimental Treatment 1 Intervention

Group II: No DeepECG plateform diagnosis & recommendations Active Control 1 Intervention

Who Is Running the Clinical Trial?

Montreal Heart Institute

Lead Sponsor

Trials

125

Recruited

85,400+

Findings from Research

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]

Using electronic health records from 21,460 patients, a new personalized algorithm called ML4CAD was developed, which significantly improves health outcomes for managing Coronary Artery Disease (CAD) by predicting adverse events with an 81.5% accuracy.

The ML4CAD algorithm increases the expected time from diagnosis to adverse events by 24.11%, particularly benefiting male and Hispanic patients, and provides physicians with an interactive tool for personalized treatment recommendations.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Personalized treatment for coronary artery disease patients: a machine learning approach. Bertsimas, D., Orfanoudaki, A., Weiner, RB.[2021]

An AI model was developed to automatically quantify coronary artery calcium (CAC) using deep learning on ECG-gated cardiac CT images, trained on 560 images and validated on 409 additional images, achieving excellent accuracy with a Cohen's kappa of 0.95 in the test cohort.

The AI model demonstrated minimal differences in CAC scores compared to manual evaluations, indicating its potential for reliable and efficient assessment of coronary artery health, which could enhance clinical decision-making.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Fully automated coronary artery calcium quantification on electrocardiogram-gated non-contrast cardiac computed tomography using deep-learning with novel Heart-labelling method. Takahashi, D., Fujimoto, S., Nozaki, YO., et al.[2023]

References

1. England pubmed.ncbi.nlm.nih.gov

Deep learning of ECG waveforms for diagnosis of heart failure with a reduced left ventricular ejection fraction. [2022]

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

Deep-learning-assisted analysis of echocardiographic videos improves predictions of all-cause mortality. [2023]

3. Korea (South) pubmed.ncbi.nlm.nih.gov

Development and External Validation of a Deep Learning Algorithm for Prognostication of Cardiovascular Outcomes. [2020]

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

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

5. Germany pubmed.ncbi.nlm.nih.gov

Deep learning-based prognostic model using non-enhanced cardiac cine MRI for outcome prediction in patients with heart failure. [2023]

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

Handling missing values in machine learning to predict patient-specific risk of adverse cardiac events: Insights from REFINE SPECT registry. [2023]

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

A deep learning-based electrocardiogram risk score for long term cardiovascular death and disease. [2023]

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

Interpatient Similarities in Cardiac Function: A Platform for Personalized Cardiovascular Medicine. [2021]

9. England pubmed.ncbi.nlm.nih.gov

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

10. Netherlands pubmed.ncbi.nlm.nih.gov

Personalized treatment for coronary artery disease patients: a machine learning approach. [2021]

11. England pubmed.ncbi.nlm.nih.gov

Fully automated coronary artery calcium quantification on electrocardiogram-gated non-contrast cardiac computed tomography using deep-learning with novel Heart-labelling method. [2023]

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

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

13. Switzerland pubmed.ncbi.nlm.nih.gov

Artificial Intelligence-Enhanced Smartwatch ECG for Heart Failure-Reduced Ejection Fraction Detection by Generating 12-Lead ECG. [2022]

14. China pubmed.ncbi.nlm.nih.gov

Automated detection of cardiovascular disease by electrocardiogram signal analysis: a deep learning system. [2020]