AI-algorithm Usage
Conditions
Keywords
Artificial Intelligence, Advanced Usage
Brief summary
The goal of this observational study is to learn if an ai-assistive algorithm would be useful in patients who are under suspicion of coronary disease. The main question it aims to answer: What proportion of patients would clinicians see fit to order an ai-assistive algorithm if available for clinical use? Participants will be asked to use clinical judgement as to whether a patient fits a predetermined criteria for use and select them for ai-assistive analysis.
Detailed description
An anonymous and de-identified database to be created over the next 9 months at the Cardiology Consultants of Philadelphia. Clinicians will be given the opportunity to select patients who would be appropriate for the analysis, then the study database will be able to be collected retrospectively after the fact. The second database will be a survey response database collected anonymously and de-identified of a simple random sample of clinicians (see 6.2) who "ordered" ECGio. The EMR will be scraped by CCP after the fact to identify which patients (whether ECGio was "ordered" or not) would be appropriate for ECGio usage based on the criteria defined in section 3.1. Digital (or PDF), anonymous, and de-identified ECG tracings for the cohort will be collected from the MUSE system. We will be provided an example of the ECG tracing as an XML export (or other format acceptable to the study sponsor), lasting 10 seconds with 500 Hz sampling. Within the digital tracing a total of 5000 data points exist for each of 12 standard leads (aVL, I, -aVR, II, aVF, III, V1, V2, V3, V4, V5, and V6).
Interventions
DEVICEAI-Assistive Algorithm
ECGio is the first coronary stenosis detection software utilizing data from just a 10-second electrocardiogram (ECG). ECGs are inexpensive, non-invasive, commonly administered tests, and measure the electrical activity of the heart in "waves". The ECGio diagnostic algorithm is an ECG analytic tool which provides the clinician with information to detect the presence, severity, and location of clinically significant coronary artery disease. This diagnostic algorithm is currently not FDA approved, but once FDA approval is received, the labeling will address the following indication and use.
Sponsors
Heart Input Output Inc
Study design
Observational model
CASE_ONLY
Time perspective
PROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
18 Years to 89 Years
Healthy volunteers
No
Inclusion criteria
* Age ≥ 18 years. * Patients with medical records stored in a digitized format. * Presenting between February 1 2026 and October 31 2026. * Patient meets one of the following criteria: * Hypertension * Hyperlipidemia * Family History of Disease * Diabetes Mellitus * High BMI (\>30) * Smoker (Former or Current) * Presenting for pre-operative clearance
Exclusion criteria
* Patients with acute coronary syndrome (ACS). * Patient with prior Coronary Artery Bypass Grafting (CABG) * Patients whose ECG tracing has extreme noise or artifact to the extent that it would be recommended to redo the tracing. * Age ≥ 90 years.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Usage Proportion | During or within 1 day, on average, of the patient visit | The proportion of patients in which it was deemed appropriate to use the AI-Assistive Algorithm |
Countries
United States
Contacts
CONTACTMichael Leasure
PRINCIPAL_INVESTIGATORVeronica Covalesky, MD
Cardiology Consultants of Philadelphia
Outcome results
None listed