In this enduring activity, Deepthy Varghese, MSN, ACNP, FNP, of Northside Hospital is joined by Dr. Matthew Kalscheur, MD, of the University of Wisconsin and Dr. Hawkins Gay, MD, MPH, of Northwestern University Feinberg School of Medicine discuss the study aimed to validate the use of artificial intelligence (AI) in identifying paroxysmal atrial fibrillation (AF) from 12-lead electrocardiograms (ECGs) in sinus rhythm (SR). The AI algorithm was trained on ECG data from a single center, including patients with and without AF. The dataset comprised 494,042 ECGs from 142,310 patients. Testing the model on the first ECG of each patient showed an accuracy of 78.1%, an area under the receiver-operating characteristic curve of 0.87, and an area under the precision-recall curve of 0.48. Performance varied with AF prevalence, demonstrating higher precision in high-risk groups (30% AF prevalence) and lower precision in low-risk groups (3% AF prevalence). The approach was robust when externally validated in another hospital. In conclusion, the AI-enabled ECG algorithm effectively identified paroxysmal AF in patients in sinus rhythm, showcasing potential clinical utility, particularly in high-risk populations.
- Provider:Heart Rhythm Society
- Activity Link: https://www.heartrhythm365.org/URL/TheLeadEpisode46
- Start Date: 2024-01-19 06:00:00
- End Date: 2024-01-19 06:00:00
- Credit Details: AMA PRA Category 1 Credit™️: 0.25 hours
- MOC Credit Details: ABIM - 0.25 Point; Credit Type(s): Medical Knowledge (ABIM)
ABP - 0.25 Point; Credit Type(s): Lifelong Learning and Self-Assessment (ABP) - Commercial Support: No
- Activity Type: Enduring Material
- CME Finder Type: Online Learning
- Fee to Participate: No, it's free
- Measured Outcome: Learner Competence
- Provider Ship: Directly Provided
- Registration: Open to all
- Specialty: Clinical Cardiac Electrophysiology, Pediatric Cardiology