Objective Based on the development of artificial intelligence (AI), an emerging number of methods have achieved outstanding performances in the diagnosis of acute myocardial infarction (AMI) using an electrocardiogram (ECG). However, AI-ECG analysis using a multicenter prospective design for detecting AMI has yet to be conducted. This prospective multicenter observational study aims to validate an AI-ECG model for detecting AMI in patients visiting the emergency department.
Methods Approximately 9,000 adult patients with chest pain and/or equivalent symptoms of AMI will be enrolled in 18 emergency medical centers in Korea. The AI-ECG analysis algorithm we developed and validated will be used in this study. The primary endpoint is the diagnosis of AMI on the day of visiting the emergency center, and the secondary endpoint is a 30-day major adverse cardiac event. From March 2022, patient registration has begun at centers approved by the institutional review board.
Discussion This is the first prospective study designed to identify the efficacy of an AI-based 12-lead ECG analysis algorithm for diagnosing AMI in emergency departments across multiple centers. This study may provide insights into the utility of deep learning in detecting AMI on electrocardiograms in emergency departments.
Trial registration ClinicalTrials.gov identifier: NCT05435391. Registered on June 28, 2022.
Citations
Citations to this article as recorded by
Accuracy of the large language model ChatGPT in adult emergency department triage: a systematic review and meta-analysis Shuang Gao, Miao Yu, Yuanyuan Zheng, Mengjie Zhang, Zhennan Yang, Jianxia Zhang BMC Emergency Medicine.2026;[Epub] CrossRef
Sex-Consistent Performance of an AI-Enabled ECG for Acute Myocardial Infarction Hak Seung Lee, Sora Kang, Joon-myoung Kwon, Tae Gun Shin, Youngjoo Lee, Dong Hoon Kim, Sung Hyuk Choi, Hanjin Cho, Mi Jin Lee, Ki Young Jeong, Won Young Kim, Young Gi Min, Chul Han, Jae Chol Yoon, Eujene Jung, Woo Jeong Kim, Chiwon Ahn, Jeong Yeol Seo, Ta JACC: Advances.2026; 5(6): 102813. CrossRef
Artificial intelligence applied to electrocardiogram to rule out acute myocardial infarction: the ROMIAE multicentre study Min Sung Lee, Tae Gun Shin, Youngjoo Lee, Dong Hoon Kim, Sung Hyuk Choi, Hanjin Cho, Mi Jin Lee, Ki Young Jeong, Won Young Kim, Young Gi Min, Chul Han, Jae Chol Yoon, Eujene Jung, Woo Jeong Kim, Chiwon Ahn, Jeong Yeol Seo, Tae Ho Lim, Jae Seong Kim, Jeff European Heart Journal.2025; 46(20): 1917. CrossRef
Clinical applications of artificial intelligence and machine learning in neurocardiology: a comprehensive review Jade Basem, Racheed Mani, Scott Sun, Kevin Gilotra, Neda Dianati-Maleki, Reza Dashti Frontiers in Cardiovascular Medicine.2025;[Epub] CrossRef
Artificial Intelligence Algorithms in Cardiovascular Medicine: An Attainable Promise to Improve Patient Outcomes or an Inaccessible Investment? Patrícia Bota, Geerthy Thambiraj, Sandeep C. Bollepalli, Antonis A. Armoundas Current Cardiology Reports.2024; 26(12): 1477. CrossRef
Objective Rapid determination of acute coronary syndrome (ACS) in the emergency department (ED) is very important for patients presenting with ischemic symptoms. The aim of this study was to determine the predictive value of HEART score for ACS and significant coronary artery stenosis (SCS).
Methods We retrospectively analyzed data of patients who visited the ED with chest discomfort and were admitted to the cardiology department. Enrolled patients were classified into ACS and non-ACS groups according to their discharge diagnosis. Patients who underwent imaging were further divided into SCS and non-SCS groups according to study results. We compared age, sex, vital signs, risk factors, electrocardiogram, troponin, and HEART score for each group. For ACS and SCS predictive performance, the test characteristics of HEART score was calculated using sensitivity, specificity, predictive value, likelihood ratio, and receiver operating characteristic (ROC) curve analysis.
Results Of 207 patients, 112 had ACS. Among enrolled patients, 155 underwent imaging workup, of whom 67 had SCS. HEART score ≤3 had 93% sensitivity for ACS and 97% for SCS. HEART score ≥7 had 82% specificity for ACS and 83% for SCS. HEART score area under ROC curve for ACS was 0.706 (95% confidence interval, 0.627–0.776) and 0.737 (95% confidence interval, 0.660–0.804) for SCS.
Conclusion HEART score was a fair predictor of ACS and SCS in ED patients who presented with chest symptoms and were admitted to the cardiology department. The predictive power of HEART score was better for SCS than for ACS.
Citations
Citations to this article as recorded by
Diagnostic Accuracy Improvement of an Updated HEART Score to Predict Coronary Artery Disease as Detected by Coronary Computed Tomography Angiography Michele Della Rocca, Stefano Ferdico, Nicola Cosentino, Alice Bonomi, Andrea Baggiano, Manuela Muratori, Saima Mushtaq, Laura Salvini, Matteo Biroli, Edona Leka, Gianluca Pontone, Marco Grazi, Emilio Assanelli Journal of Clinical Medicine.2026; 15(4): 1424. CrossRef
Retrospective analysis of the performance of the PROMISE minimal risk tool for patients presenting with recent onset stable chest pain Conor Tuffs, Natasha Khullar, Terry Levy, Vivek Kodoth, Rosie Swallow, Peter O’Kane, Jehangir Din, Chun Shing Kwok, Nick Curzen, Jonathan Hinton Open Heart.2026; 13(1): e003837. CrossRef
Predictive Performance of the HEART Score for Acute Coronary Syndrome with Significant Coronary Stenosis in Patients with Non–ST-Elevation Acute Chest Pain André Martins, Mónica Amado, Adriana Vazão, Joana Pereira, Luís Santos, Margarida Cabral, Célia Domingues, David Durão, João Morais Journal of Cardiovascular Development and Disease.2026; 13(7): 335. CrossRef
A Cross-Sectional Study on the Relationship between the HEART Score and the Severity of CAD using the SYNTAX Score in Patients with NSTEMI Admitted to the Emergency Department: A Study from Turkey MS Pepele, I Aktas, O Demiroz, S Derya, E Yildirim, MN Bilen, B Demir Nigerian Journal of Clinical Practice.2025; 28(6): 758. CrossRef
Role of HEART score in prediction of coronary artery disease and major adverse cardiac events in patients presenting with chest pain Tanja Stojkovic, Eva Stojkovic, Dejan Sakac, Aleksandar Redzek, Anastazija Stojsic-Milosavljevic, Lazar Velicki, Biljana Parapid Srpski arhiv za celokupno lekarstvo.2022; 150(3-4): 149. CrossRef
Objective The quick sequential organ failure assessment (qSOFA) score, which includes mentation, systolic blood pressure, and respiratory rate, was developed to identify serious sepsis in out-of-hospital or emergency department (ED) settings. We evaluated the ability of the qSOFA score to predict poor outcome in South Korean ED patients with suspected infection.
Methods The qSOFA score was calculated for adult ED patients with suspected infection. Patients who received intravenous or oral antibiotics in the ED were considered to have infection. In-hospital mortality rate, admission rate, intensive care unit (ICU) admission rate, length of hospital stay (LOS), and lactate levels were compared between the qSOFA score groups. Receiver operating characteristic curves and area under the receiver operating characteristic curve values for in-hospital mortality were calculated according to qSOFA cut-off points and lactate levels.
Results Of 2,698 patients, in-hospital mortality occurred in 134 (5.0%). The mortality rate increased with increasing qSOFA score (2.2%, 6.4%, 17.5%, and 42.4% for qSOFA scores 0, 1, 2, and 3, respectively, P<0.001). The admission rate, ICU admission rate, LOS, and lactate level also increased with increasing qSOFA score (all P<0.001). The area under the receiver operating characteristic curve values for predicting in-hospital mortality associated with qSOFA score, lactate ≥2 mmol/L, and lactate ≥4 mmol/L were 0.719 (95% confidence interval [CI], 0.670 to 0.768), 0.657 (95% CI, 0.603 to 0.710), and 0.632 (95% CI, 0.571 to 0.693), respectively.
Conclusion Patients with a higher qSOFA score had higher admission, ICU admission, and in-hospital mortality rates, longer LOS, and higher lactate level. The qSOFA score showed better performance for predicting poor outcome than lactate level.
Citations
Citations to this article as recorded by
Clinical utility of biomarkers for outcomes prediction in adults with suspected sepsis presenting to the emergency department: a synthesis of current evidence Mari Imamura, Sinead N Duggan, Thenmalar Vadiveloo, Jamie G Cooper, Callum T Kaye, Paul Manson, Gianni Virgili, Lorna Aucott, Mike Clarke, Miriam Brazzelli Health Technology Assessment.2026; : 1. CrossRef
Outcome-Specific Prognostic Performance of Pneumonia Severity Assessment Tools for Predicting a Spectrum of Multiple Clinical Events Abuzer Özkan, Ahmet Taha Özatak, Soner Yeşilyurt Bratislava Medical Journal.2026; 127(8): 3712. CrossRef
The Role of qSOFA, Derived Neutrophil-to-Lymphocyte Ratio, MEWS, and PIRO Scores in Predicting the Severity of Odontogenic Infections in Young and Adult Patients Serban Talpos Niculescu, Robert Avramut, Tareq Hajaj, Raluca Maracineanu, Antonis Perdiou, Roxana Talpos Niculescu, Marius Pricop, Horatiu Urechescu, Florin Urtila, Roxana Radu, Nicoleta Nikolajevic Stoican, Malina Popa Biomedicines.2025; 13(3): 532. CrossRef
Sodium level in patients with sepsis: what is the predictive value? Zain Sayed, Ahmed A. Obiedallah, Soheir M. Kasem, Abdelraouf M.S. Abdelraouf, Kareem S.M. Ismail, Alaa O. Ahmed Research and Opinion in Anesthesia & Intensive Care.2025; 12(1): 35. CrossRef
Lactate Enhanced Quick Sequential Organ Failure Assessment Score (Lqsofa) as a Tool to Predict the Prognosis in Patients of Sepsis in Surgical Intensive Care Unit Akash Sharma, Nitin Garg, Rahul Patel, Pooja Sharma International Journal of Recent Surgical and Medical Sciences.2025; 11: e015. CrossRef
Comparative analysis of Glasgow Coma Scale, quick Sepsis-related Organ Failure Assessment, base excess, and lactate for mortality prediction in critically ill emergency department patients Gürbüz Meral, Şenol Ardıç, Serkan Günay, Kadir Güzel, Ahmet Köse, Hülya Gençbay Durmuş, Serhat Uysal, Aydın Coşkun Turkish Journal of Emergency Medicine.2024; 24(4): 231. CrossRef
First
Prev
