| Interethnic validation of electrocardiogram image analysis software for detecting left ventricular dysfunction in an emergency department population |
Haemin Lee1,2
, Woon Yong Kwon3
, Kyoung Jun Song4
, You Hwan Jo1
, Joonghee Kim1,2
, Youngjin Cho2,5
, Ji Eun Hwang1
, Yeongho Choi1
|
1Department of Emergency Medicine, Seoul National University Bundang Hospital, Seongnam, Korea
2ARPI Inc, Seongnam, Korea
3Department of Emergency Medicine, Seoul National University Hospital, Seoul, Korea
4Department of Emergency Medicine, Seoul Metropolitan Government Seoul National University Boramae Medical Center, Seoul, Korea
5Cardiovascular Center, Department of Internal Medicine, Seoul National University Bundang Hospital, Seongnam, Korea |
| Correspondence
Ji Eun Hwang Email: 65903@snubh.org, Yeongho Choi Email: d2uk@snubh.org |
|
Received: October 22, 2024. Revised: January 6, 2025. Accepted: January 8, 2025. Published online: April 30, 2025. |
|
|
| ABSTRACT |
Objective
We previously developed and validated an artificial intelligence-based electrocardiogram (ECG) analysis tool (ECG Buddy) in a Korean population. This study investigated the performance of this tool in a US population, specifically assessing the left ventricular (LV) dysfunction score and LV ejection fraction (LVEF)-ECG feature for predicting LVEF <40%. The study used N-terminal pro-B-type natriuretic peptide (NT-ProBNP) as a comparator.
Methods
We identified emergency department (ED) visits from the MIMIC-IV dataset with information on LVEF <40% or ≥40% and matched 12-lead ECG data recorded within 48 hours of the ED visit. The performance of ECG Buddy’s LV dysfunction score and the LVEF-ECG feature was compared with those of NT-ProBNP using area under the receiver operating characteristic curve (AUC) analysis.
Results
A total of 22,599 ED visits was analyzed. The LV dysfunction score had an AUC of 0.905 (95% confidence interval [CI], 0.899–0.910), with a sensitivity of 85.4% and specificity of 80.8%. The LVEF-ECG feature had an AUC of 0.908 (95% CI, 0.902–0.913), sensitivity of 83.5%, and specificity of 83.0%. NT-ProBNP had an AUC of 0.740 (95% CI, 0.727–0.752), with a sensitivity of 74.8% and specificity of 62.0%. The ECG-based predictors demonstrated superior diagnostic performance compared to NT-ProBNP (all P<0.001). In the sinus rhythm subgroup, the LV dysfunction score achieved an AUC of 0.913 and LVEF-ECG had an AUC of 0.917, both outperforming NT-ProBNP (AUC, 0.748; 95% CI, 0.732–0.763; all P<0.001).
Conclusion
ECG Buddy demonstrated superior accuracy compared with NT-ProBNP in predicting LV systolic dysfunction, validating its utility in a US ED population. |
| Keywords:
Heart failure; Electrocardiography; Artificial intelligence; Ejection fraction; Emergency departments |
|
|
|
PDF Links
Full text via DOI
Download Citation 
Print 
