Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
| Name | Class |
|---|---|
| Medical AI Co., Ltd | UNKNOWN |
Not provided
Not provided
Not provided
Not provided
This study is a prospective multicenter observational study for external validation and model advancement of a deep learning based 12-lead electrocardiogram analysis algorithm targeting adult patients presenting to the emergency department with chest pain and acute myocardial infarction equivalent symptoms.
About 9,000 adult patients will be enrolled at 20 emergency medical centers in Korea. Artificial intelligence algorithms are manufactured by Medical AI Co., Ltd. It is an advanced version based on the model developed and published in 2020. It had the diagnostic performance of area under the receiver operating curve 0.901 and 0.951 for acute myocardial infarction and ST-segment elevation myocardial infarction, respectively. The primary endpoint is a diagnosis of acute myocardial infarction on the day of the emergency center visit, and the secondary endpoint is a 30-day major adverse cardiac event. From March 2022, patient registration will begin at centers that have been approved by the Institutional Review Board.
This is the first prospective multicenter emergency department validation study for a 12-lead electrocardiogram artificial intelligence algorithm to diagnose acute myocardial infarction. This study will give insight into the direction of future development by verifying whether the deep learning algorithm works well for patients visiting the real-world adult emergency medical center.
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
| Label | Type | Description | Intervention Names |
|---|---|---|---|
| Group 1 | Patients with chest pain or who are clinically suspected as acute myocardial infarction with equivalent symptoms. |
Not provided
| Measure | Description | Time Frame |
|---|---|---|
| Diagnosis of acute myocardial infarction (Type 1, 2) | Accuracy metrics include area under the receiver operating characteristics curve, sensitivity, specificity, positive predictive value, and negative predictive value, along with a 95% confidence interval. | Index admission |
| Measure | Description | Time Frame |
|---|---|---|
| Major adverse cardiovascular event (MACE) | MACE is defined as death, myocardial infarction, stroke, target-vessel revascularization, or stent thrombosis occurring within 30 days of index visit. Accuracy metrics include area under the receiver operating characteristics curve, sensitivity, specificity, positive predictive value, and negative predictive value, along with a 95% confidence interval. | 30-day after index admission |
| Measure | Description | Time Frame |
|---|---|---|
| AI ECG analysis versus clinical risk score (HEART score) | HEART score is an acute coronary syndrome risk calculation tool introduced in recent guidelines, and consists of history, electrocardiogram, age, risk factor, and troponin. On a scale of 0 to 10, the higher the score, the higher the risk. Generally, a score of 7 or higher is classified as high risk. The prediction performance for the primary and secondary endpoints of the HEART score will be analyzed by comparing it with AI-ECG analysis. |
Inclusion Criteria:
Exclusion Criteria:
Not provided
Not provided
Not provided
A total of 20 emergency departments in the Republic of Korea will participate in this study. These centers receive about 800,000 patients annually, and all of these hospitals are capable of their own emergency cardiovascular angiography and percutaneous coronary intervention.
Not provided
| Name | Affiliation | Role |
|---|---|---|
| KS Kim, MD, PhD | CHA University School of Medicine | Principal Investigator |
| Facility | Status | City | State | ZIP | Country | Contacts |
|---|---|---|---|---|---|---|
| CHA Bundang Medical Center | Seongnam | South Korea |
| PubMed Identifier | Type | Citation | Retractions |
|---|---|---|---|
| 42312765 | Derived | Lee HS, Kang S, Kwon JM, Shin TG, Lee Y, Kim DH, Choi SH, Cho H, Lee MJ, Jeong KY, Kim WY, Min YG, Han C, Yoon JC, Jung E, Kim WJ, Ahn C, Seo JY, Lim TH, Kim JS, Son JM, Kim KS, Kim K, Lee MS; ROMIAE Study Group. Sex-Consistent Performance of an AI-Enabled ECG for Acute Myocardial Infarction: The ROMIAE Study. JACC Adv. 2026 Jun;5(6 Pt 1):102813. doi: 10.1016/j.jacadv.2026.102813. | |
| 38012820 |
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
| ID | Term |
|---|---|
| D009203 | Myocardial Infarction |
| D002637 | Chest Pain |
| ID | Term |
|---|---|
| D017202 | Myocardial Ischemia |
| D006331 | Heart Diseases |
| D002318 | Cardiovascular Diseases |
| D014652 | Vascular Diseases |
Not provided
Not provided
Not provided
Not provided
Not provided
| Index admission, 30-day after index admission |
| AI ECG analysis versus clinical risk score (GRACE 2.0 score) | GRACE 2.0 score is a tool for estimating short- and long-term risk in acute coronary syndrome. A low score indicates low risk, and a high score indicates a high risk group. Based on recent guidelines, patients with acute coronary syndrome are categorized as low (≤108 GRACE score), medium (109-140 GRACE score) and high risk (>140 GRACE score). The prediction performance for the primary and secondary endpoints of the GRACE 2.0 score will be analyzed by comparing it with AI-ECG analysis. | Index admission, 30-day after index admission |
| AI ECG analysis versus cardiac biomarker | The performance of cardiac biomarker (hs-troponin I or T) for the diagnosis of acute myocardial infarction during index visit will be compared with AI ECG analysis diagnostic performance. | Index admission |
| AI ECG analysis versus physician's ECG score | The attending physician determines a score between 0 and 10 for the probability of acute myocardial infarction based on the results of the initial patient assessment and the first 12-lead ECG. A score of 0 indicates that the probability of acute myocardial infarction is 0%, and a score of 10 indicates a probability of 100%. The diagnostic performance of this score for acute myocardial infarction will be compared with AI ECG analysis. | Index admission |
| Derived |
| Shin TG, Lee Y, Kim K, Lee MS, Kwon JM; ROMIAE study group. ROMIAE (Rule-Out Acute Myocardial Infarction Using Artificial Intelligence Electrocardiogram Analysis) trial study protocol: a prospective multicenter observational study for validation of a deep learning-based 12-lead electrocardiogram analysis model for detecting acute myocardial infarction in patients visiting the emergency department. Clin Exp Emerg Med. 2023 Dec;10(4):438-445. doi: 10.15441/ceem.22.360. Epub 2023 Nov 28. |
| D007238 |
| Infarction |
| D007511 | Ischemia |
| D010335 | Pathologic Processes |
| D013568 | Pathological Conditions, Signs and Symptoms |
| D009336 | Necrosis |
| D010146 | Pain |
| D009461 | Neurologic Manifestations |
| D012816 | Signs and Symptoms |