Diagnostic accuracy of computed tomography coronary angiography for evaluation of coronary artery disease: a comparison between high definition versus standard definition scanner

Abstract

Background: The ability of computed tomography coronary angiography (CTCA) to detect quantitatively the stenosis is lower than invasive coronary angiography (ICA) due to the limited spatial resolution and it could be responsible of a false positive rate of up to 35%. Recently, a high-definition CTCA (HDCTCA) scanner, with improved in-plane spatial resolution of 230 μm has been developed. The aim of this study is to compare the diagnostic accuracy by HDCTCA with standard definition 64-slice scanner (SDCTCA) by using ICA as the reference method. Methods: One-hundred-forty consecutive patients (mean age 65±8 years, male 105) scheduled for ICA were randomized to SDCTCA (n= 70, group 1) or HDCTCA-scan protocol (n= 70, group 2) (Discovery CT 750 HD scanner, GE Healthcare, Milwaukee, WI) before ICA. The scanning parameters were: slice acquisition 64x0.625 mm, gantry rotation time 330 msec and prospective ECG-triggering. We evaluated the Likert image quality (score 1: non-diagnostic to score 4: excellent), overall feasibility (Fe), the sensitivity (Se), specificity (Sp), negative predictive value (NPV), positive predictive value (PPV) and accuracy (Ac) versus ICA in a segment-based model and comparing the diagnostic performance between group 1 and group 2. Results: The 2 groups were homogeneous in terms of baseline characteristics. Group 2 showed a higher mean image quality score (3.8 vs 3.1, p<0.001) and overall evaluability (98% vs 92%, p<0.01) versus group 1, respectively. The Se, Sp, NPV, PPV, and accuracy were 89%, 94%, 96%, 83%, 93% and 93%, 97%, 98%, 93%, 96% in group 1 and group 2, respectively, including all segments censoring not-evaluable segments as positive. Group 2 showed a higher sensitivity and specifiivity and accuracy (p<0.05) but mainly a higher PPV (p<0.01) versus group 1. In a sub-analysis on calcified plaque the overall agreement between CTCA and ICA was lower in segments with calcification versus noncalcified segments in group 1 (p<0.01) while in group 2 there was no statistical difference between the overall agreement versus ICA based on the plaque composition. Conclusions: The present study showed an improved overall feasibility, positive predictive value and accuracy mainly in calcified coronary artery lesions in HDCTCA in comparison with SDCTCA due to the better spatial resolution and the consequent reduced blooming effect. Further studies are needed to compare the performance of HDCTCA versus the new dual-energy CT that offers a possible and alternative solution to the problem of heavily calcified coronary arteries reducing the overestimation of calcium volume.