Guidelines recommend the use of non-invasive tests for the diagnosis of coronary artery disease (CAD) in stable angina patients with intermediate pre-test probability. Knuuti at al. assessed with a meta-analysis the performance of non-invasive diagnostic modalities in identifying anatomically (defined by invasive coronary angiography, ICA) or functionally significant (defined by FFR) lesions.
A total of 133 and 23 studies using ICA and FFR as a reference standard, respectively, were included. Concerning anatomical significance, stress ECG showed the poorest performance in ruling-out (LR- = 0.68) and ruling-in (LR+ = 1.53) CAD, while coronary CT angiography (CCTA) (LR- = 0.04) and PET (LR+ = 5.87) were the best tests for ruling-out and ruling-in the disease, respectively. Regarding functional significance, CCTA, PET and CMR performed similarly and best for ruling-out (LR- = 0.13), while functional imaging (PET, SPECT, CMR) was the best approach for ruling-in significant lesions (LR+ range: 3.87-7.1). ICA had the poorest performance for ruling-out and, together with CCTA for ruling-in CAD.
Sechtem, in his editorial, supports the selection of a diagnostic test based on its range of optimal performance, but also warrants design of new studies, since some of the older studies included in the meta-analysis may not be representative of the current clinical scenario.
LR-: Negative likelihood ratio; LR+: Positive likelihood ratio
Author: Lorena Casadonte