Adding risk factors and CAC score to clinical likelihood improved AUC from 0.69 to 0.85 for predicting CAD, accurately identifying 30% with <5% risk.
Does the ESC 2024 risk factor weighted clinical likelihood model, with or without CAC score, improve the prediction of significant CAD compared to traditional symptom-based scores in patients referred for CCTA?
The ESC 2024 risk factor-enhanced clinical likelihood model, particularly when combined with CAC scoring, significantly improves the prediction of significant CAD compared to traditional symptom-based models, allowing safe deferral of CCTA in low-risk patients.
Absolute Event Rate: 0% vs 0%
Abstract Introduction Current ESC-guidelines highlight the importance of using a clinical symptom-based score where pre-test probability (PTP) for coronary artery disease (CAD) is enhanced by incorporating risk factors and coronary artery calcium (CAC) score. Aim To assess the performance of the new risk factor weighted clinical likelihood model (RF-CL) proposed by the ESC 2024 Guidelines in predicting CAD. Method ROCAMBOLE was a prospective observational study of consecutive patients referred for cardiac coronary tomography (CCTA) from September 2019 to December 2023. Online questionnaires related to health and self-reported symptoms were filled out prior to the CCTA. PTP was estimated based on risk factors, chest pain characteristics (typical, atypical, non-specific symptoms or dyspnea at exertion), and CAC score. We calculated the PTP and used table analysis and logistic regression to compare traditional clinical score, adding risk factors and Agatston CAC score. We divided patients into low (5%), intermediate (5% - 15%), and high (15%) PTP for significant CAD (defined as stenosis grade ≥50% in any epicardial vessel) on CCTA. Results Of the 466 patients responding to the questionnaire (91.9% of 507 patients in total), 235 (50.4%) were men and the mean age was 59 ± 11 years. Low, intermediate and high PTP was present in 147 (31.6%), 213 (45.7%) and 106 (22.8%) patients. In total, 69 (14.8%) patients had significant CAD in any of the epicardial vessels, including 6 (4.1%) patients with low PTP, 36 (16.9%) with intermediate PTP and 27 (25.5%) with high PTP. A stenosis grade ≥70% in any vessel was present in 28. Of the 6 patients with significant stenosis in the low PTP group, only 3 (2.0%) had clinically significant stenosis ≥70% in the proximal left anterior descending artery (LAD). The performance in detecting significant CAD by the former Diamond Forrester PTP based solely on chest pain symptoms was AUC 0.69 (95%CI 0.63-0.75). The risk estimation according to the new guidelines with adding risk factors demonstrated significant improvement o (AUC 0.75, 95%CI 0.69-0.80), with additional significant improvement by adding CAC-scoring (AUC 0.85 95%CI 0.82-0.89). The enhanced risk score did not identify 21 patients (14%) with non-stenotic coronary artery disease in the low-risk group, who therefore could have been missed for the purpose of intensive LDL-cholesterol lowering. Conclusion Consistent with the 2024 ESC guidelines, the CL-RF model demonstrates superior predictive performance for CAD in low to intermediate risk groups. Notably, the addition of CAC-scoring significantly enhances these predictions, surpassing both models. Utilizing the risk factor-enhanced score alone, approximately 30% of the referred patients could be accurately identified as having less than 5% risk of CAD and thereby could safely be deferred from CCTA.
Moen et al. (Sat,) reported a other. Adding risk factors and CAC score to clinical likelihood improved AUC from 0.69 to 0.85 for predicting CAD, accurately identifying 30% with <5% risk.