Dania A. Bani Hani, Jafar A. Alshraideh, Hamza Alduraidi, Akram Saleh, Nathan T. Riek, Karam Daoud, Rui Qi Ji, Samir Saba, Clifton W. Callaway, Salah Al-Zaiti
Abstract
The 12-lead ECG is central to the evaluation of patients with suspected coronary artery disease (CAD) but has limited diagnostic accuracy for identifying significant obstructive disease prior to elective coronary angiography. We previously developed an artificial intelligence–enhanced ECG (AI-ECG) model for detecting acute coronary occlusion. Whether this model generalizes to identifying angiographically significant CAD in symptomatic patients undergoing elective evaluation remains unknown
Introduction
Coronary artery disease (CAD) remains the leading cause of mortality worldwide, accounting for nearly 9 million deaths annually [1]. The disease is characterized by a prolonged subclinical phase, often progressing silently until patients present with symptomatic or catastrophic events such as myocardial infarction [2]. Accordingly, early detection and accurate risk stratification of clinically significant CAD are essential to prevent irreversible myocardial injury and guide timely intervention.
Methods
Study design and population
This prospective cohort study enrolled consecutive adult patients (≥18 years) undergoing elective coronary angiography at Jordan University Hospital between February and July 2024. Eligible patients were referred for angiography based on clinical suspicion of obstructive CAD. Consecutive sampling was employed to minimize selection bias.
Results
Sample characteristics
A total of 363 patients undergoing elective coronary angiography were included. The mean age was 58.9 ± 11 years, and 200 (55.1%) were male. Cardiovascular risk factors were prevalent, including hypertension (62.5%), diabetes mellitus (44.4%), active smoking (35.0%), prior CAD (33.9%), and prior revascularization (43.8%); 65.3% reported a family history of CAD. Chest pressure was the most common presenting symptom (76.9%). Overall, 131 patients (36.1%) had significant coronary artery narrowing (≥70% stenosis or ≥50% left main).
Discussion
In this prospective cohort of symptomatic outpatients undergoing elective coronary angiography, we evaluated whether an AI-enhanced ECG model originally developed for acute coronary occlusion could improve estimation of pretest probability for angiographically significant coronary artery disease in a stable clinical setting.
Conclusion
In this prospective cohort of symptomatic outpatients, the ECG-SMART AI-ECG score demonstrated good discrimination for detecting significant coronary artery disease and remained independently associated with angiographic findings after adjustment for clinical risk factors. These results support its role as a noninvasive electrical biomarker that generalizes beyond acute care settings to patients undergoing elective evaluation.
Citation: Bani Hani DA, Alshraideh JA, Alduraidi H, Saleh A, Riek NT, Daoud K, et al. (2026) Diagnostic performance of the ECG-SMART AI score for detecting angiographically significant coronary artery disease in patients undergoing elective coronary angiography. PLOS Digit Health 5(10): e0001701. https://doi.org/10.1371/journal.pdig.0001701
Editor: Kaisaierjiang Kadier, Xinjiang Medical University Affiliated First Hospital, CHINA
Received: May 7, 2026; Accepted: August 16, 2026; Published: October 8, 2026
Copyright: © 2026 Bani Hani et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The minimal anonymized dataset necessary to replicate the findings reported in this study has been provided as Supporting information accompanying this manuscript. Additional study materials are available from the corresponding author upon reasonable request. For a persistent, non-author institutional point of contact regarding data access and governance, requests may also be directed to the University of Rochester Data Governance and Support Office (datagovernanceoffice@rochester.edu). Study data are maintained within University-managed systems in accordance with institutional requirements for research data retention, security, and governance (https://www.rochester.edu/provost/university-data/data-governance-overview/).
Funding: This work was supported by the National Institutes of Health (NIH) (grant number R01HL137761) awarded to [SAZ]; the Fulbright Program (U.S. Department of State) awarded to [SAZ]; and the University of Jordan awarded to [JAA and HA]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have read the journal’s policy and have the following competing interests to declare: Drs. Salah Al-Zaiti and Clifton Callaway are co-founders and equity shareholders of Intellicardio Inc. (Pittsburgh, PA, USA). The remaining authors declare that they have no competing interests. This competing interest does not alter the authors’ adherence to PLOS Digital Health policies on sharing data and materials.
https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0001701
