IMPACT-AI: IMaging to Personalise coronary Artery disease management using Computed Tomography and Artificial Intelligence in Adults referred for CT coronary angiography

The intervention is a personalised model of care guided by an artificial intelligence (AI) enabled, computed tomography coronary angiography (CTCA)-based clinical decision support tool (AI-CTCA), compared with usual care based on standard visual CTCA interpretation alone. In the intervention arm, all participants undergo clinically indicated CTCA on a 320-detector scanner using a standardised acquisition protocol; images are then analysed with FDA-cleared deep learning software (Autoplaque v3.0) that automatically segments the coronary arteries and quantifies total, calcified, and non-calcified plaque volumes and stenosis severity for all coronary segments greater or equal to 1.5 mm. The software also generates an Ischaemia Risk Score (0–100) per vessel, integrates plaque volume into age- and sex-specific percentile categories, and produces a one-page AI-CTCA report with recommendations for downstream functional testing or invasive coronary angiography and for medical therapy based on a predefined AI-CTCA Treatment Algorithm. Reporting clinicians in the intervention arm first perform standard visual CTCA interpretation (Society of Cardiovascular Computed Tomography 18-segment model), then review the AI-CTCA report and repeat their assessment, documenting plaque burden, and confidence scores in electronic questionnaires. Reporting clinicians will undergo an individualised training session for 1 hour with the study coordinators and provided a Clinician Information Sheet prior to commencing analysis for the trial. Subsequent care for these patients is transferred, with the referrer’s permission, to a cardiologist-led team trained in AI-CTCA interpretation, who use the AI-CTCA recommendations and Ischaemia Risk Score to guide referral for functional non-invasive imaging or invasive angiography and to intensify risk-factor management and lipid-lowering therapy according to the AI-CTCA Treatment Algorithm, while retaining clinician discretion. Participants in the intervention arm attend a structured follow-up pathway, including a baseline clinic visit, telehealth review at 6 months, an in-person clinic visit at 12 months, and standardised remote follow-up at 6, 12, 18, and 24 months (to document any interventions, medication changes, and adherence), with repeat CTCA (with individualised AI-CTCA report) and blood sampling at 24 months to quantify change in coronary plaque volume. The control arm (comparator) receives usual care directed by the referring clinician based on standard visual CTCA reporting without access to the AI-CTCA output, but with the same schedule of research assessments, blood sampling, and serial CTCA to permit blinded quantification of plaque progression or regression over 24 months.