AI-Assisted Interpretation of Cardiac CT in the Emergency Department

" This prospective, pragmatic, randomized controlled trial is designed to evaluate the impact of an artificial intelligence (AI)-based coronary computed tomography angiography (CCTA) interpretation tool (Angiomics) on emergency physicians' diagnostic performance and clinical decision-making in patients presenting with acute chest pain.

CCTA is a critical diagnostic modality for suspected acute coronary syndrome (ACS) in the emergency department (ED). Accurate interpretation often requires experienced radiologists, who may not always be available, particularly during off-hours. The introduction of AI-based interpretation tools into clinical workflow has the potential to enhance diagnostic accuracy, increase physician confidence, reduce delays in decision-making, and improve efficiency of resource utilization. However, evidence regarding the real-world effectiveness of such AI tools in the ED setting remains limited.

Eligible participants will include adults aged 18 years or older presenting to the ED with chest pain and classified as intermediate risk (HEART score 4-6). Participants will be randomized into two groups: (1) AI-assisted CCTA interpretation, in which emergency physicians interpret scans with access to AI results; and (2) standard interpretation, in which emergency physicians interpret CCTA without AI support. In both groups, physicians will document the presence of stenosis in the four major coronary arteries (LM, LAD, LCX, RCA) and report diagnostic confidence on a 5-point Likert scale.

The primary outcome is the negative predictive value (NPV) of CCTA interpretation at the patient level, comparing AI-assisted versus standard interpretations against the reference standard of blinded consensus readings by board-certified radiologists. Secondary outcomes include sensitivity, specificity, positive predictive value (PPV), accuracy, diagnostic confidence, vessel-level diagnostic performance, and agreement with radiologist consensus using Cohen's Kappa.

The study aims to enroll approximately 530 participants (276 in the control arm and 254 in the intervention arm, accounting for an expected 10% dropout). Enrollment and follow-up will be conducted at Severance Hospital and Gangnam Severance Hospital over a 24-month period following IRB approval. The results are expected to provide evidence for the clinical utility and effectiveness of AI-based CCTA interpretation in the ED and to guide integration of AI into emergency care in order to optimize patient outcomes and healthcare efficiency.