AUTHOR=Accorsi Tarso Augusto Duenhas , Pitta Fabio Grunspun , Rompkoski Juliane , Moreira Flavio Tocci , Morbeck Renata Albaladejo , Köhler Karen Francine , Lima Karine De Amicis , Pedrotti Carlos Henrique Sartorato 

TITLE=Evaluation of AI-enhanced tele-ECG response time and diagnosis in acute chest pain patients

JOURNAL=Frontiers in Cardiovascular Medicine

VOLUME=Volume 12 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/cardiovascular-medicine/articles/10.3389/fcvm.2025.1532770

DOI=10.3389/fcvm.2025.1532770

ISSN=2297-055X

ABSTRACT=BackgroundThe impact of artificial intelligence in improving Tele-ECG response times and diagnostic accuracy among emergency patients experiencing acute chest pain remains uncertain. This study assesses the performance of AI-assisted cardiologists’ ECG report generation and characterizes diagnoses derived from examinations conducted at emergency facility without on-site cardiology services.MethodsA retrospective cross-sectional observational study at a Telemedicine Center in São Paulo, Brazil, examined ECG data from patients aged 18 and older with suspected ischemic syndromes at peripheral emergency departments in Goiânia, Brazil. Seventeen cardiologists carefully evaluated ECGs, focusing on identifying critical diagnostic red flags. Advanced AI algorithms enabled the accurate measurement of electrocardiographic segments and intervals, improving the detection of abnormalities and deviations from standard parameters.ResultsOut of 25,346 ECG tracings submitted, 22,159 (87.42%) were analyzed. Unanalyzed tracings included 953 (3.75%) with artifacts, 633 (2.49%) with atrial fibrillation, 506 (1.99%) with inverted leads, and 628 (2.47%) with flat lines. The median age of patients was 49 (30–64) years, with 12,082 (54.52%) females. ST-segment elevation myocardial infarction (STEMI) was diagnosed in 202 (0.9%) cases. Other diagnoses included normal tracings, diffuse ventricular repolarization changes, sinus tachycardia, complete branch block, left ventricular hypertrophy, intraventricular conduction disorders, electrically inactive areas, sinus bradycardia, and atrioventricular conduction disorders. Request times averaged 11:30 AM (±7.07 h). The median response time was 75 (50–125) seconds, with a median of 375 (207–655) seconds for STEMI reports.ConclusionMost ECGs are interpretable, but clearer tracings are needed. Quick response times are likely due to early AI detection of abnormalities. The low occurrence of acute myocardial infarction and other prognostic indicators suggests a low-risk group using the emergency department as their main healthcare access point.