Patients were instructed to avoid coffee, tea, and oral beta-blockers in the 24h before the scan. Two venous access sites (18-G for contrast administration, 20-G for adenosine infusion) were placed in the right arm. Heart rate and blood pressure were monitored.

All scans were performed on a 128-DSCT system (Flash Definition®; Siemens, Forcheim, Germany) with the following protocol:

Once scout views from the tracheal bifurcation to the diaphragm in the craniocaudal direction were taken and the scan was planned, adenosine was administered with a perfusion pump (Alaris System, Cardinal Health; Ohio, United States) at a constant rate of 140 μg/kg/min; 3 min afterward, 60mL of iopromide (Ultravist 370, Bayer Schering Pharma; Berlin, Germany) iodine contrast agent was administered, followed by 60mL of saline, at 4mL/s, with an injector (Stellant Dual, Medrad; Pennsylvania, United States). A retrospective electrocardiogram-gated dual-energy scan with tube current modulation was performed with the following technical characteristics: rotation time, 330ms; heart rate-dependent pitch, 0.2-0.43, collimation, 0.6mm; and temporal resolution, 165ms. The tubes operated with 165 mAs/rot at 100 kV and with 140 mAs/rot at 140kV. Tube current modulation was performed with the MinDose technique with full tube current applied between 60% and 75% of the cardiac cycle, which was reduced to 4% outside this time window.

To achieve adequate enhancement, bolus tracking was performed by placing a region of interest in the aortic arch with a trigger threshold of 160 HU. A delay of 10s was added to guarantee adequate myocardial perfusion.

Adenosine infusion was stopped once the acquisition was finished. Blood pressure, electrocardiogram, and clinical symptoms were carefully monitored.

A monosegment reconstruction algorithm that uses data from a complete rotation of both detectors was used for image reconstruction. Data were reconstructed in diastole from 60% to 75% of the R-R interval, with a 3-mm slice thickness, 1.5-mm reconstruction interval, and D30f reconstruction kernel.

After 5min, coronary CT angiography was performed. If there was no contraindication and the heart rate was >65 bpm, beta-blocker (5-15mg intravenous metoprolol) was administered. If the heart rate was stable and <65 bpm, prospective acquisition was performed with the high-pitch technique. If a stable heart rate of <65 bpm could not be achieved, retrospective electrocardiogram-gated spiral acquisition was performed. Sublingual nitroglycerine (0.5-1mg) was administered to all patients.

The technical parameters for the prospective high-pitch acquisition were as follows: collimation, 0.6mm (128×0.6 mm); rotation time, 280 ms; effective tube current, 370 mAs with CARE DOSE modulation; 100 or 120 Kv current, depending on body weight <80 or >80 kg, respectively; pitch, 3.4; and temporal resolution, 75ms. The acquisition was performed in the caudocranial direction beginning at 60% of the R-R interval.

The technical parameters for the retrospective acquisition were as follows: collimation, 0.6mm (128×0.6mm); rotation time, 280ms; 100 or 120 Kv current, depending on body weight <80 or >80kg, respectively; effective tube current, 370 mAs; heart rate-dependent pitch, 0.2-0.43; and MinDose tube current modulation with the maximum current of the tube between 60% and 75% of the cardiac cycle, reduced to 4% outside this time window.

Vascular enhancement was achieved by the administration of 60mL of iopromide (Ultravist 370, Bayer Schering Pharma), followed by 60mL of saline, at a constant rate of 6mL/s. Bolus tracking was performed by placing a region of interest in the aortic arch with a trigger threshold of 100 HU.

In all cases, image reconstruction was performed with a 0.6-mm slice thickness, 0.4mm, reconstruction interval, and B26f reconstruction kernel. When the acquisition was retrospective, the reconstruction was obtained in the phase of the cardiac cycle where the coronary arteries could best be assessed.

Significant stenosis was considered to be present when the stenosis size was >50% of the lumen diameter on quantitative analysis.

The myocardial images of this series were evaluated in resting perfusion.

After 7min, a prospective acquisition was performed with the high-pitch technique, with the same settings as those of coronary CT angiography, except that a lower radiation dose was used with 80kV and 300 mAs/rot. No intravenous contrast agent or any other drug was administered.

Image reconstruction was performed with a 0.6-mm slice thickness, 0.4-mm reconstruction interval, and B26f reconstruction kernel.

Intravenous adenosine was administered via a perfusion pump (Alaris System, Cardinal Health) at a rate of 140 μg/kg/min. Blood pressure, heart rate, and oxygen saturation were monitored at 1-min intervals. After 3min, a bolus of 0.05 mmol/kg gadobutrol (Bayer Schering Pharma) intravenous contrast agent was injected at 4mL/s with 40 mL of saline at 4 mL/s through the antecubital vein in the left arm. The first-pass perfusion was performed with a T1-weighted turbo gradient echo (TGE) sequence (repetition time, 2.2ms; echo time, 1.04ms; angle, 50°; slice thickness, 10mm) in the short-axis, which included the base, the middle third, and apex of the left ventricle. The images were acquired at end-diastole to maximize the intravascular signal.

Approximately 5min after performing the stress perfusion, contrast was administered again and baseline perfusion was assessed with the same technique.

After a 10-min wait, the delayed enhancement images were acquired. Before the viability images were acquired, the optimal inversion time was calculated to best cancel out the myocardial signal with a TGE-echo planar sequence (repetition time, 40ms; echo time, 5ms; angle, 15°; slice thickness, 10mm). For the delayed enhancement, a T1-weighted 3-dimensional-TGE sequence with a tissue preparation pulse was used (repetition time, 3.9-4.3ms; echo time, 1.2-1.3 ms; angle, 15°; slice thickness, 10mm). All images were acquired during breath holding in the short- and long-axes of the left ventricle and with a 4-chamber view.

Following adenosine administration, heart rate increased from 63.7 (9.3) bpm at rest to 75.5 (13.8) bpm. There were no major complications that hampered scan completion, although 9 patients had 2:1 atrioventricular blocks, 3 had chest discomfort, and 4 had headache. All these complications resolved after stopping adenosine infusion. The estimated effective dose was 5.2 (1.2) mSv. Perfusion defects were found in 17 of the 56 patients (30.3%), 29 of the 168 coronary artery territories (17.2%), and 55 of the 952 myocardial segments studied (5.7%).

We performed high-pitch prospective acquisition (heart rate, 57 [6] bpm) in 44 patients and retrospective acquisition (heart rate, 74 [5] bpm) in 12. The estimated effective dose was 2.5 (2.1) mSv. In a per-patient analysis, 27 had normal coronary arteries or <50% stenosis, 15 had significant stenosis (single-vessel disease in 11, 2-vessel in 2, and 3-vessel in 3), and 14 had at least 1 artery that could not be assessed due to an artifact or severe calcification. In a per-artery analysis, the anterior descending, right coronary, and circumflex arteries were normal or without significant stenosis in 37, 34, and 37 patients, with significant stenosis in 5, 8, and 8 patients, and were impossible to assess in 14, 14, and 11 patients, respectively.

Hyperenhancement was seen in 10 of 56 patients (17.8%), 14 of 168 coronary artery territories (8.3%), and 28 of 952 myocardial segments studied (2.9%). The estimated effective dose was 0.48 (0.19) mSv.