BACKGROUND The prevalence of obstructive coronary artery disease (CAD) in symptomatic patients referred for diagnostic testing has declined, warranting optimization of individualized diagnostic... Show moreBACKGROUND The prevalence of obstructive coronary artery disease (CAD) in symptomatic patients referred for diagnostic testing has declined, warranting optimization of individualized diagnostic strategies.OBJECTIVES This study sought to present a simple, clinically applicable tool enabling estimation of the likelihood of obstructive CAD by combining a pre-test probability (PTP) model (Diamond-Forrester approach using sex, age, and symptoms) with clinical risk factors and coronary artery calcium score (CACS).METHODS The new tool was developed in a cohort of symptomatic patients (n = 41,177) referred for diagnostic testing. The risk factor-weighted clinical likelihood (RF-CL) was calculated through PTP and risk factors, while the CACS- weighted clinical likelihood (CACS-CL) added CACS. The 2 calculation models were validated in European and North American cohorts (n = 15,411) and compared with a recently updated PTP table.RESULTS The RF-CL and CACS-CL models predicted the prevalence of obstructive CAD more accurately in the validation cohorts than the PTP model, and markedly increased the area under the receiver-operating characteristic curves of obstructive CAD: for the PTP model, 72 (95% confidence intervals [CI]: 71 to 74); for the RF-CL model, 75 (95% CI: 74 to 76); and for the CACS-CL model, 85 (95% CI: 84 to 86). In total, 38% of the patients in the RF-CL group and 54% in the CACS-CL group were categorized as having a low clinical likelihood of CAD, as compared with 11% with the PTP model.CONCLUSIONS A simple risk factor and CACS-CL tool enables improved prediction and discrimination of patients with suspected obstructive CAD. The tool empowers reclassification of patients to low likelihood of CAD, who need no further testing. (C) 2020 by the American College of Cardiology Foundation. Show less
Aims To provide a pooled estimation of contemporary pre-test probabilities (PTPs) of significant coronary artery disease (CAD) across clinical patient categories, re-evaluate the utility of the... Show moreAims To provide a pooled estimation of contemporary pre-test probabilities (PTPs) of significant coronary artery disease (CAD) across clinical patient categories, re-evaluate the utility of the application of diagnostic techniques according to such estimates, and propose a comprehensive diagnostic technique selection tool for suspected CAD.Methods and results Estimates of significant CAD prevalence across sex, age, and type of chest pain categories from three large-scale studies were pooled (n = 15 815). The updated PTPs and diagnostic performance profiles of exercise electrocardiogram, invasive coronary angiography, coronary computed tomography angiography (CCTA), positron emission tomography (PET), stress cardiac magnetic resonance (CMR), and SPECT were integrated to define the PTP ranges in which ruling-out CAD is possible with a post-test probability of <10% and <5%. These ranges were then integrated in a new colour-coded tabular diagnostic technique selection tool. The Bayesian relationship between PTP and the rate of diagnostic false positives was explored to complement the characterization of their utility. Pooled CAD prevalence was 14.9% (range = 1-52), clearly lower than that used in current clinical guidelines. Ruling-out capabilities of non-invasive imaging were good overall. The greatest ruling-out capacity (i.e. post-test probability <5%) was documented by CCTA, PET, and stress CMR. With decreasing PTP, the fraction of false positive findings rapidly increased, although a lower CAD prevalence partially cancels out such effect.Conclusion The contemporary PTP of significant CAD across symptomatic patient categories is substantially lower than currently assumed. With a low prevalence of the disease, non-invasive testing can rarely rule-in the disease and focus should shift to ruling-out obstructive CAD. The large proportion of false positive findings must be taken into account when patients with low PTP are investigated. Show less
