Noninvasive Tests to Rule-In and Rule-Out CAD Stenosis

Study Questions:

Across what range of pretest probabilities can significant coronary artery disease (CAD) be ruled out or ruled in by stress electrocardiogram (ECG), coronary computed tomography angiography (CCTA), stress echocardiography, stress single-photon emission computed tomography (SPECT), stress positron emission tomography (PET), and stress cardiac magnetic resonance (CMR) imaging?


The authors conducted a meta-analysis of the published literature for published papers evaluating with a reference standard of invasive coronary angiography (ICA) with stenosis >50% or ICA and fractional flow reserve (FFR) ≤0.80 with ≥100 patients in each study. A test was considered to rule-in significant CAD if the post-test probability was >85% and rule-out significant CAD if the post-test probability was <15%.


For each modality, there were between 3 and 30 studies on a pooled per modality total of 418-6,551 subjects. Using ICA anatomy as the gold standard, the sensitivity for CCTA was highest at 97%, while the sensitivity for stress ECG was lowest at 58%. With regard to specificity, stress PET and CMR were highest and second highest at 85% and 80%, respectively, while stress ECG was again lowest at 62%. Using ICA with FFR as the gold standard, CCTA again was most sensitive at 93%, while stress ECG was least sensitive at 68%. Specificity was comparable for stress CMR (87%), stress PET (85%), and stress SPECT (83%), but was poorest for CCTA, with a specificity of only 53%.

Using ICA stenosis as the gold standard, stress ECG could only rule-out CAD in patients with pretest probability ≤19%. By contrast, CCTA was able to rule-out CAD with up to 80% pretest probability. Stress PET and stress CMR could rule-out CAD in patients with up to 58% and 56% pretest probability, respectively.

Conversely, stress ECG could only rule-in CAD in patients already at high risk of CAD with pretest probability of ≥80%, while stress PET was effective at ruling-in CAD down to a pretest probability of 51%, with stress CMR, stress echo, CCTA, and stress SPECT requiring at least 56%, 56%, 58%, and 68% pretest probability to rule-in CAD.

When ICA and FFR were used as the gold standard, the results were largely similar except for CCTA, which then could only rule-out CAD with ≤57% pretest probability and rule-in CAD with ≥75% pretest probability.


The authors concluded that selection of optimal noninvasive tests for CAD depend on pretest probability and whether anatomic or functional gold standards are used.


This is a very interesting study, which adds a novel twist to meta-analyses comparing diagnostic accuracy of various noninvasive methods to diagnose CAD. In this case, the authors computed the range of pretest probabilities over which a test could rule-out CAD (<15% post-test risk) or rule-in CAD (>85% post-test risk). They find that, first, performance of stress ECG alone is poor. Second, if anatomic measures are used, CCTA is particularly effective to rule-out CAD. However, if physiologic gold standards are used, CCTA’s performance to rule-out CAD is only comparable to stress PET and stress CMR, and poorer than these functional modalities at ruling-in CAD.

Importantly, rarely do we “rule-in” CAD with noninvasive testing alone, raising some concerns over the analysis related to the rule-in of 85%. Many doctors would refer to ICA patients with post-test probabilities considerably below that threshold. Consequently, the rule-out analyses in this paper are most interesting. Two important take-home lessons should be noted. First, stress PET and stress CMR are likely more effective at ruling out CAD than stress echo and stress SPECT. Second, CCTA is even more effective then these when a purely anatomic gold standard is used, but is only comparable when a functional gold standard is used.

Clinical Topics: Invasive Cardiovascular Angiography and Intervention, Noninvasive Imaging, Stable Ischemic Heart Disease, Atherosclerotic Disease (CAD/PAD), Interventions and Coronary Artery Disease, Interventions and Imaging, Angiography, Computed Tomography, Echocardiography/Ultrasound, Magnetic Resonance Imaging, Nuclear Imaging, Chronic Angina

Keywords: Angina, Stable, Constriction, Pathologic, Coronary Angiography, Coronary Artery Disease, Coronary Stenosis, Diagnostic Imaging, Echocardiography, Echocardiography, Stress, Electrocardiography, Fractional Flow Reserve, Myocardial, Magnetic Resonance Imaging, Positron-Emission Tomography, Tomography, Emission-Computed, Single-Photon, Tomography, X-Ray Computed

< Back to Listings