Diagnostic performance of coronary CT angiography by using different generations of multisection scanners:...

To retrospectively compare sensitivity and specificity of four generations of multidetector computed tomographic (CT) scanners for diagnosing significant (>or=50%) coronary artery stenosis, with quantitative conventional coronary angiography as reference standard.

The institutional review board approved this study. All patients consented to undergo CT studies prior to conventional coronary angiography, after they were informed of the additional radiation dose, and to the use of their data for future retrospective research. Two hundred four patients (157 men, 47 women; mean age, 58 years +/- 11 [standard deviation]), classified in four groups of 51 patients each, underwent coronary CT angiography with four-section, first- and second-generation 16-section, and 64-section CT scanners. Patients in sinus rhythm scheduled for conventional coronary angiography (stable angina, atypical chest pain) were included. Patients with bypass grafts and stents were excluded. Two readers unaware of results of conventional coronary angiography evaluated CT scans. Coronary artery segments of 2 mm or larger in diameter were included for comparative evaluation with quantitative coronary angiography. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for detection of significant stenoses (>or=50% luminal diameter reduction) were calculated.

Image quality was rated poor for the following percentages of coronary artery segments: 33.1% at four-section CT, 14.4% at first-generation 16-section CT, 6.3% at second-generation 16-section CT, and 2.6% at 64-section CT. Sensitivity, specificity, PPV, and NPV, respectively, were as follows: 57%, 91%, 60%, and 90% at four-section CT; 90%, 93%, 65%, and 99% at first-generation 16-section CT; 97%, 98%, 87%, and 100% at second-generation 16-section CT; and 99%, 96%, 80%, and 100% at 64-section CT. Diagnostic performance of four-section CT was significantly poorer than that of second-generation 16-section CT (odds ratio = 4.57) and 64-section CT (odds ratio = 2.89).

Diagnostic performance of coronary CT angiography varies among scanners of different generations. Earlier-generation scanners (four sections) had significantly poorer performance; performance of 16- compared with 64-section CT scanners showed progressive, although not significant, improvement.