Can coronary computed tomography angiography replace invasive angiography? Coronary computed tomography angiography cannot replace invasive angiography.

O ne million patients undergo invasive coronary angiogra-phy (CA) in the United States every year. 1 In view of its invasive nature and related costs, noninvasive imaging of coronary imaging has the theoretical appeal to improve comfort for patients, reduce cost, and refine patient selection to limit invasive procedures to patients requiring treatment for ischemic coronary artery disease (CAD). Coronary computer tomography angiog-raphy (CCTA) has been introduced as a noninvasive method for CAD assessment >15 years ago, 2 but was limited by insufficient temporal and spatial resolution of early multi-detector computed tomography (CT) scanners. The advent of 64-slice and dual-source CT scanners has improved the clinical utility of CCTA by shortening the breath-hold duration and improving resolution. 3 More recently, 320-slice CT scanners were introduced further, improving on the performance of earlier generation scanners. 4 Despite these technological advances, the development of CCTA has reached its limits and remains inferior to coronary angiog-raphy. In this article, we aim to clarify that although CCTA is a valuable noninvasive tool in the diagnostic work-up of patients with low to intermediate likelihood for CAD, it has no role in high-risk patients and should therefore be considered complementary rather than competing with invasive CA. The temporal and spatial resolution of CCTA with contemporary 64-slice CT multidetector scanners remains considerably lower compared with invasive CA (temporal: 80–190 ms versus 10 ms; spatial: 300–400 μm versus 150–200 μm). As a result, invasive CA remains the only method to allow for real-time assessment of coronary anatomy, whereas CCTA, despite short acquisition times, requires labor-and time-intensive image processing before arriving at the diagnosis. Moreover, the accuracy of CCTA to determine stenosis severity is inferior compared with invasive CA as determined by quantitative coronary angiography (Figure 1), which in turn results in a frequent overestimation of stenosis severity by CCTA. Along this line, only ≈50% of significant coronary stenoses (≥70%) identified by CCTA are associated with ischemia after further evaluation. As summarized in Table 1, several multicenter studies applying this technology have shown a sensitivity ranging from 81% to 99%, a specificity ranging from 64% to 93%, a positive predictive value ranging from 64% to 92%, and a negative predictive value ranging from 83% to 99% in a per-patient analysis. 6–9 Of note, a prospective multicenter study conducted in Ontario, Canada, with CCTA images analyzed by local expert observers instead of core-laboratories, showed a high variability in terms of …