From June 2008 to November 2008, 729 consecutive symptomatic patients suspected of or with known coronary artery disease (M:F = 366:363; mean age, 60.5 years ± 11.0; age range, 23-89 years) underwent 64-slice DSCT coronary angiography in our hospital: those patients with Kawasaki disease, previous coronary artery bypass surgery or percutaneous coronary intervention, and those patients without any cardiac symptoms were excluded. Among the 729 patients, 131 patients (M:F = 85:46; mean age, 64.5 years ± 8.9; age range, 38-83 years) had undergone invasive coronary artery angiography with a mean time interval between DSCT and invasive coronary angiography of 26.4 ± 23.7 days (range: 0-122 days).

All the CT examinations were performed using a dual-source scanner (Somatom Definition; Siemens Medical Solutions, Forchheim, Germany). Among the 729 enrolled patients, 588 patients with a prescanning heart rate of 65 beats per minute (bpm) or higher were given 50-100 mg of oral metoprolol (Betaloc; AstraZeneca, Sweden) 45-60 min prior to the CT examination unless the subject had a contraindication to beta-blockers: 50 mg of beta-blocker in 184 patients with a prescanning heart rate of 65 to 69 bpm, and 100 mg in 404 patients with a prescanning heart rate of 70 bpm or higher. However, additional beta-blockers were not given to any patients who underwent CT scanning 45-60 min after administration of the first dose of betablocker, and even if the heart rate did not decrease below 65 bpm. Among the 131 patients who underwent invasive coronary angiography, 102 patients took beta-blockers.

Precontrast scanning for calcium scoring was performed using a prospective ECG-gating protocol with 80% of the R-R interval (3-mm section thickness and collimation, 120 kVp, tube current-time: 100 mAs, collimation: 32 × 0.6 mm, section acquisition: 64 × 0.6 mm with the z-flying focal spot technique, gantry rotation time: 330 msec, pitch: 0.2-0.5, depending on the heart rate).

Sublingual nitroglycerin (0.6 mg) (Nitroquick; Ethex, St. Louis, MO) was administered for coronary vasodilation after completion of the calcium scoring scanning in all patients, except nineteen who had a contraindication to nitroglycerin. Sixty mL of a nonionic contrast medium (Ultravist 370; Bayer Schering Pharma, Berlin, Germany) was injected into an antecubital vein at 5 mL/sec, followed by additional 20 mL of a nonionic contrast medium and 50 mL of an 8:2 mixture of normal saline and contrast medium and both at a flow rate of 4 mL/sec with the use of a dual power injector (Stellant; Medrad, Indianola, PA). The bolus triggering method was used to determine the beginning of CT acquisition by monitoring the signal density of the contrast medium in the mid ascending aorta. CT scans were started 8 seconds after a threshold trigger of 150 HU above the baseline was reached. In general, cardiac CT scans were performed using the retrospective ECG-gated mode with ECG pulsing. The ECG-pulsing window for radiation dose reduction was applied: 65-75% of the R-R interval in patients with ≤ 65 bpm and 25-80% for the patients with > 65 bpm. Outside the ECG-pulsing window, the tube current was reduced to 4% of the full current (Mindose; Siemens Healthcare). The pitch was automatically adapted to the heart rate: 0.20 (40 bpm), 0.26 (60 bpm), 0.32 (70 bpm), 0.37 (80 bpm), 0.43 (90 bpm) and 0.50 (100 bpm). The scanning parameters used in our study were as follows: a detector collimation of 32 × 0.6 mm, a section collimation of 64 × 0.6 mm with a z-flying focal spot, a gantry rotation time of 330 msec, a tube voltage of 100 or 120 kVp, and a maximum tube current of 320 mAs per rotation. Scans were usually performed from the diaphragm to the level of the tracheal bifurcation in the caudocranial direction. The mean heart rate and range of the heart rate during CT scanning were recorded for all the patients.

Standard reconstruction algorithms were applied using the absolute reverse or percentage technique to obtain data sets during end systole and/or mid- to end diastole according to the heart rate. In the cases for which the standard reconstruction algorithm provided data sets with insufficient image quality of one or more coronary segments, additional data sets were manually reconstructed. If necessary, multiple data sets of a single patient were used separately to obtain optimal image quality for all the coronary segments. Image reconstruction windows were manually repositioned to achieve high image quality for the patients with arrhythmia, as was previously described (8).

The parameters used for image reconstruction for DSCT coronary angiography included a slice thickness of 0.6 mm, increments of 0.4 mm and a kernel value of B26f. The parameters used for image reconstruction for calcium scoring included a slice width of 3 mm and a kernel of B35f. The DSCT data sets were transferred to an offline workstation (Wizard, Siemens Medical Solutions Software, Germany) with cardiac post-processing three-dimensional software (Rapidia; INFINITT, Seoul, Korea). In addition to the axial images and the oblique multiplanar reconstructions (MPR), the curved MPRs (CMPRs), maximum intensity projections (MIPs) and three-dimensional volume rendering technique (VRT) reconstructions were created for the visualization and analysis of the data.

In the first session, one observer classified a total of 729 patients into four categories through a database of radiologic reports made by one radiologist: (a) patients with neither calcium nor stenosis, (b) patients with insignificant stenosis, (c) patients with indeterminate stenosis and (d) patients with significant stenosis. Lesions with a stenosis of 50% or greater in diameter were considered to be significant. Indeterminate stenosis was defined when a lesion was unevaluable due to artifacts or adjacent calcification. If a patient had a coronary artery segment with significant stenosis and another coronary artery segment was indeterminate at the same time, then he/she was classified as a patient with significant stenosis. On the other hand, if a patient had one coronary artery segment with insignificant stenosis and another coronary artery segment was indeterminate, then he/she was classified as a patient with indeterminate stenosis. As routine practice, the CT data were processed using three-dimensional software (Rapidia; INFINITT, Seoul, Korea) and the CT images were interactively assessed by one experienced radiologist (with 4 years of experience in cardiovascular imaging). The contrast-enhanced dual-source CT was evaluated using 0.6 mm thick-transverse axial images, MIP, CMPR and three-dimensional VRT.

By reviewing the electronic medical records, the results of the treadmill test or myocardial single photon emission CT (SPECT) were also recorded when this was performed within one month of before or after DSCT coronary angiography. Myocardial perfusion was evaluated using dual isotopes of rest thallium SPECT (Tl-SPECT) and stress 99mTc sestamibi SPECT. The observer also checked whether patients had undergone invasive coronary artery angiography and the observer also recorded its results: lesions with a stenosis of 50% or greater in diameter were considered to be significant.

In the second session, one observer selected 131 patients who underwent further invasive coronary angiography within three months of after DSCT coronary angiography, and he recorded the presence of significant stenosis of the coronary arteries based on a modified model of the coronary tree with 15 segments (22) through a database of radiologic reports made by one radiologist: for the right coronary artery, segment 1 was considered proximal, segment 2 was considered middle, segment 3 was considered distal, segment 4 was considered the posterior descending segment and segment 5 was considered the posterolateral branches. Segment 6 was considered the left main stem artery. For the left anterior descending artery, segment 7 was considered proximal, segment 8 was considered middle, segment 9 was considered distal, segment 10 was considered the first diagonal and segment 11 was considered the second diagonal. For the left circumflex artery, 12 was considered proximal, 13 was considered distal, 14 was considered the first obtuse marginal branch and 15 was considered the second obtuse marginal branch. In routine practice, the stenoses were visually evaluated and described as the percentage of lumen diameter reduction. The observer classified each segment as significant (≥ 50% lumen diameter reduction) or nonsignificant (< 50% lumen diameter reduction) based on the radiologic reports. All the coronary artery segments with a diameter of 1 mm or larger were included.

In the third session, two readers working in consensus retrospectively evaluated the overall image quality of the CT images of 131 patients in terms of the degree of coronary or respiratory motion artifact, beam hardening artifacts due to coronary calcification, coronary artery enhancement and quantum noise on a 4-point scale as excellent: 1, good: 2, fair: 3 or poor: 4 (10).

The conventional coronary angiography was performed by three experienced cardiologists according to the standard procedure of using the transfemoral or transradial Judkins technique at our hospital. To visualize the right coronary artery, at least two projections were obtained; for the left coronary artery, at least six projections were obtained. The severity of stenosis was retrospectively evaluated, using quantitative coronary analysis (QCA, version 3.3; Philips, Eindhoven, the Netherlands), by a single observer who was blinded to the CT results. Segmental disease was analyzed in each vessel by using the same 15-segment model that was employed for the dual-source CT analysis. The severity of stenosis was classified on the projection according to the maximal luminal diameter stenosis. Lesions with a stenosis of 50% or greater in diameter were considered to be significant.

All conflicting results between the two modalities were retrospectively analyzed by two observers, and the results were investigated for the reasons for disagreement. The causes of the unevaluable segments were also evaluated. In the case of discordance between DSCT and coronary angiography in defining the coronary artery segment according to the American Heart Association classification, disagreement was resolved via a consensus reading (23).

One observer evaluated whether the optimal ECG pulsing and pitch were used according to the heart rates, whether the scan range covered only the heart or the entire thorax, as well as which tube voltage was used in the 729 CT examinations. Four hundred thirty CT examinations were selected in order to assess the effect of heart rate on the radiation dose in the dedicated routine cardiac CT protocol covering only the heart. The volume CT dose index (CTDI_vol) and the dose length product (DLP), which were provided by the scanner system, were recorded. The effective dose was derived from the DLP and a conversion coefficient for the chest as the investigated anatomic region. This conversion coefficient (k = 0.017 mSv × mGy^-1 × cm^-1) was averaged between the male and female models.

The diagnostic performance of CT coronary angiography for making the diagnosis of significant coronary artery disease as compared with the standard of reference and the quantitative coronary angiography on conventional coronary angiography was determined by measuring the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy. The image quality and radiation dose according to the heart rate frequency and heart rate variability were compared using the Student t-test. We used 65 bpm as the cut-off value for the high heart rate group (9) and the mean heart rate difference of 15 bpm as the cut-off value for the high heart rate variability group (10).

All the statistical analyses were performed with statistical packages (SPSS for Windows, version 12.0, SPSS, Chicago, IL; GraphPad Prism, version 3.02, GraphPad Software, San Diego, CA). Differences were considered significant for p values less than 0.05.

Of the 729 patients, 301 (41%) patients had neither coronary artery calcium nor stenosis seen on CT, and 225 (31%) patients showed insignificant coronary artery stenosis on CT. DSCT coronary angiography showed a low rate of non-diagnostic cases (16 [2%] of 729 patients), and definite significant stenosis of the coronary arteries on CT was reported in 187 (26%) patients. Among the 182 patients with significant stenosis seen on CT, single-vessel disease was reported in 97 (52%) patients, two-vessel disease was reported in 52 (28%) patients and three-vessel disease was reported in 39 (21%) patients.

Of the 16 patients with indeterminate stenosis, dense calcification was the most common cause (n = 10), followed by coronary motion artifacts (n = 4) and respiratory motion artifacts (n = 2). Seven of the 16 patients with indeterminate stenosis on CT underwent further invasive coronary angiography, from which five patients showed significant coronary artery stenosis.

Figure 1 showed the clinical practice in our institution for the 729 symptomatic patients who underwent DSCT coronary angiography. Among 526 patients who showed normal or insignificant coronary stenosis on CT, 15 patients underwent invasive coronary angiography despite of negative results on CT because coronary artery disease was still clinically suspected. All 15 patients showed negative results after invasive coronary angiography. For the treadmill test or SPECT, the false positives and false negatives were substantial. Fourteen (37%) of 38 tests were positive for the patients with normal coronary CT angiography. For the indeterminate cases on CT, all four negative cases on the treadmill test or SPECT turned out to be positive after invasive coronary angiography. For the seven cases with significant stenosis seen on invasive coronary angiography, the treadmill test or SPECT showed negative results.

A total of 63 segments showed discrepancy between DSCT and invasive coronary angiography for the detection of coronary artery stenosis of 50% or greater in diameter. Thirty five false-positive segments and 28 false-negative segments were detected on CT.

The causes of false-positive segments were dense calcification in 16 (46%) segments, coronary motion in 10 (29%), stair-step artifact in two (6%) and disagreement in seven (20%) (Fig. 2).

The causes of the false-negative segments were subclassified into two categories: detection failure in 18 (64%) segments and underestimation of the degree of stenosis in 10 (36%) segments. Detection failure occurred due to a small branch in seven segments (39%), perception error in six (33%), calcification in two (11%), motion artifact in two (11%) and poor enhancement in one (6%) (Fig. 3). The causes of underestimation of stenosis were calcification in six segments (60%), small branch in one (10%), motion artifact in one (10%) and perception error in two (20%). A small branch, calcification and perception error were shown to be the major causes of the overall false-negative segments.

Of the 52 unevaluable segments detected on CT, the main causes of non-diagnostic segments were dense calcification in 37 (71%) segments, wall blurring due to coronary motion in seven (14%), poor enhancement in seven (14%), stair-step artifact in one (2%) and a small branch in one (2%).

Table 2 demonstrated radiation dose exposure according to the mean heart rate and heart rate variability. The mean CTDI_vol and effective dose were 23.3 mGy and 5.8 mSv, respectively. The mean values of the estimated radiation dose during CT scans were significantly greater in the patients with a higher heart rate or higher heart rate variability (p < 0.0001).