Step-and-shoot prospectively ECG-gated versus retrospectively ECG-gated with tube current modulation coronary CT angiography using the 128-slice MDCT: comparison of image quality and radiation dose

Many reports have been published on the high diagnostic accuracy of 64-slice multidetector computed tomography (MDCT) for detection of significant coronary stenosis in both per-patient and per-segment analyses (1–3). Additionally, many other reports indicate that the negative predictive value for the classification of segments and patients with or without coronary artery disease (CAD) is high (97–100%) (3–8). These results indicate that coronary artery disease can be excluded through a less invasive method than traditional coronary angiography. Recently, retrospective ECG-gated coronary CT angiography (CCTA) has been accepted as a reliable non-invasive method for detecting coronary artery stenosis under appropriate heart rate (HR) control, with optimal contrast enhancement and in the absence of extensive calcification. The major obstacle for the common use of retrospective CCTA, however, is a high radiation dose that is associated with a risk of cancer (9). To minimize the exposure to radiation during CCTA examination using MDCT, several strategies have been developed, including retrospective electrocardiography (ECG)-dependent tube current modulation (TCM), tube voltage reduction, attenuation-based TCM, and prospective ECG triggering (step-and-shoot) techniques (10–12).

Prospective ECG triggering techniques are categorized under low-dose CCTAs, but these techniques do not provide information regarding left ventricular (LV) function and volume because they do not cover 100% of the cardiac phase, although the radiation dose is reduced. Recently developed CT scanners, such as dual-source MDCT and 128-slice MDCT, have enabled retrospective CCTA with low levels of patient radiation exposure by decreasing the tube current to 4% during the non-reconstructed phase in all scanned phases (13). However, relatively few studies have explored the relationship between image quality and radiation dose in step-and-shoot and retrospective CCTA with TCM in 128-slice MDCT coronary angiography. Therefore, the purpose of our study was to compare the image quality and patient radiation dose between the step-and-shoot and retrospective CCTA with TCM techniques for 128-slice MDCT coronary angiography.

Material and Methods

Patients

The Institutional Review Board of our hospital approved this protocol. We included 160 consecutive patients who underwent 128-slice MDCT for evaluation of stable angina with suspected CAD or atypical chest pain. The exclusion criteria for MDCT were as follows: unstable angina, irregular heartbeat, hypersensitivity to contrast media, contraindication for β-blockers, previous history of myocardial infarction, elevated serum creatinine levels (>1.5 mg/dl), and inability to hold breath for 15 seconds.

We used the step-and-shoot protocol for all patients examined between January and February in 2009 and retrospective CCTA with TCM for all patients examined between March and April in 2009. Before the examination, the HR of each patient was measured. Patients with a pre-measured HR of >65 beats per minute (bpm) were administered a β-blocker (10–20 mg propranolol) orally 1 h before the scan. No additional intravenous β-blocker was administered at the time of examination. Consequently, 15 patients were excluded. Among the remaining patients, 68 patients underwent step-and-shoot (41 men, 27 women; mean age 54.6 ± 14.1 years), and 77 patients underwent retrospective CCTA with TCM (40 men, 37 women; mean age 56.9 ± 10.6 years) (Table 1).

OPEN IN VIEWER

Table 1.

Characteristics of patients examined in two 128-slice MDCT coronary angiography protocols

Characteristic	Prospective CT angiography with step-and-shoot mode (n = 68)	Retrospective CT angiography with tube current modulation mode (n = 77)	P value
HR (bpm)	59.3 ± 6.8	59.1 ± 9.8	0.89
Actual z-axis length (cm)	13.7 ± 0.1	14.9 ± 0.7	< 0.01
Actual z-axis rad dose (mSv)	1.74 ± 0.70	4.51 ± 1.18	< 0.0001
Normalized 12 cm z-axis rad dose (mSv)	1.52 ± 0.11	3.60 ± 0.70	< 0.0001
Assessable coronary segment	927/947 (97.15)	1051/1071 (98.13)	0.144
Image quality score	1.67 ± 0.74	1.44 ± 0.71	> 0.05

Values are numbered with percentages or mean ± SD

HR = heart rate, bpm = beats per minute, rad = radiation

MDCT scanning protocol and reconstruction

MDCT coronary angiography was performed using a 128-slice MDCT scanner (Definition AS + ; Siemens Medical Solutions, Forchheim, Germany) with a rotation time of 300 m/sec and slice collimation of 64 × 0.600 mm with z-flying focal spot technology; the tube voltage was set at 100 kV for patients with a body mass index (BMI) of <25 kg/m² and 120 kV for patients with a BMI of >25 kg/m². The scanning direction was craniocaudal and extended from the level of the carina to the diaphragm. A bolus-tracking method was used to synchronize the arrival of contrast at the level of the coronary arteries with the start of acquisition. For all CT examinations, a dual-head power injector (Stellant D; Medrad, Indianola, PA, USA) was used to administer a 2-phase bolus at a rate of 5 mL per second. First, 70 mL of iopromide (Ultravist 370; Bayer Schering Pharma, Berlin Germany) was administered. Thereafter, 30 mL of a 20–80% blend of Ultravist 370 and saline was administered. Retrospective CCTA with TCM was performed with the following parameters: helical scanning direction, 0.16–0.22 pitch, and use of dose modulation (tube current of 160 mAs/rot during 50–80% of R-R interval and minimal tube current of 6.4 mAs/rot during the remainder of the R-R interval). Images were first reconstructed at 65% of the R-R interval; images were then reconstructed at 50–80% of the R-R interval in 5% increments. Step-and-shoot was performed with the following parameters: 200 msec X-ray exposure time (two-thirds of the gantry rotation time), 130 mAs/rot, and the center of the imaging window was set at 65% of the R-R interval. An additional 100-msec temporal padding was not used for all patients to reduce radiation dose. Images were reconstructed at 65% of the R-R interval. The scan was prescribed by using 3 or 5 of the 40-mm slabs (z-axis length < 12 cm, 3; 12 ≤ z-axis length < 16 cm, 4; and 16 cm ≤ z-axis length, 5) requiring 2–4 incremental table movements of 35 mm with a 5-mm overlap. For reconstructing axial images obtained using both CT techniques, we used a slice thickness of 0.75 mm with 0.4-mm increment and the medium/hard algorithm for reconstruction.

Image post-processing and evaluation

All CT data-sets were transferred to a dedicated workstation (Leonardo; Siemens Medical Solutions, Forchheim, Germany). Image sets available on the workstation included axial source images, curved maximum intensity projection (MIP) images rotated through 360° along the course of each coronary artery and major side branches, thin-slab MIP images, and volume-rendered images. Two experienced radiologists (who were blinded to the objective image quality) independently reviewed all CT images of both groups. For any discrepancy between the two readers in data analysis, a consensus reading was performed. Coronary artery segments of the three main coronary arteries and their major side branches with a luminal diameter of 1.5 mm or larger were classified according to the 15-segment American Heart Association model (14). We excluded segments that were anatomical variants or non-evaluable due to small vessel size (vessel origin diameter, <1.5 mm).

CT image quality analysis

CT images obtained with both step-and-shoot and retrospective CCTA with TCM were randomly presented during each reading session. For step-and-shoot images, the reviewers examined images reconstructed at 65% of the R-R interval and scored each coronary artery segment for image quality. For retrospective CCTA with TCM, the reviewers examined images reconstructed at 50–80% of the R-R interval in 5% increments and scored each coronary artery segment for image quality by using the reconstructed percentage judged best for that segment. Coronary artery segment image quality of cardiac CT was scored at the optimal phase in each technique by two independent radiologists using the 4-point Likert scale (1, excellent image quality; 2, good image quality; 3, fair image quality; 4, poor image quality) (15–17). Excellent image quality (Fig. 1a) was attributed to vessels showing a continuous course, without stair-step artifacts on MPR or MIP images, and appearing as bright circular or oval areas on transverse CT scans, without motion artifacts and surrounded by low-attenuation fat tissue. Good image quality (Fig. 1b) was classified as the presence of discrete blurring of the vessel margin in any planar orientation, with minor motion artifacts seen as a discrete tail or streak emitting shadows on transverse images and no stair-step artifact on MPR images or MIP images. Moderate image quality (Fig. 1c) was classified as noticeably blurred vessel or plaque margins, distinctly broader motion artifacts extending less than 5 mm from the vessel center, and stair-step artifact of less than 25% of the vascular diameter. Poor image quality (Fig. 1d) was defined as an inadequate delineation between the vessel and surrounding tissue, the presence of streak artifacts extending at least 5 mm from the center of the vessel, and stair-step artifacts of more than 25% of the vessel diameter. Images with a score of 4 were considered to be of non-diagnostic image quality.

OPEN IN VIEWER

Fig. 1

Coronary artery segment image quality on optimal phases in each technique was scored with a four-point Likert scale by two independent radiologists. (a) Excellent image quality (score of 1); (b) Good image quality (score of 2); (c) Fair image quality (score of 3); (d) Non-diagnostic image quality (score of 4)

If the scores of both reviewers differed by only 1 image quality score after their separate reading sessions, the higher numerical score was adopted for statistical analysis. If the scores of both reviewers differed by more than 1 image quality score after reading sessions, the reviewers subsequently interpreted the images together and reached a consensus.

Radiation dose

The dose-length product (DLP, measured in milligray-centimeters [mGy cm]) is defined as the volume CT dose index multiplied by scan length and is an indicator of the integrated radiation dose of the entire CT examination. DLP was displayed on the dose report on the CT scanner and recorded. A reasonable approximation of the effective radiation dose (E) of CCTA was calculated by multiplying the DLP by a conversion coefficient for the chest (κ = 0.017 mSv/mGy cm) (18). Because the patients' heart sizes were variable, the effective dose was also normalized to a 12-cm z-axis length of a typical heart and calculated as the product of 12 divided by the length of the heart examined (measured in centimeters) multiplied by the effective dose of the examination (19).

Statistical analysis

All statistical analyses were performed with statistical software (SPSS, version 12.0 for Windows; SPSS, Chicago, IL, USA and MedCalc for Windows, version 9.6.4.0). Continuous variables were expressed as mean ± standard deviation and categorical variables as frequencies or percentages. The patient data of the two groups were compared using Student's t-test for continuous variables such as average HR and the chi-squared test for categorical variables such as image quality score. Interobserver agreement of image quality assessment was performed by kappa analysis (<0.4, poor agreement; 0.4–0.75, fair to good agreement; >0.75, excellent agreement). The relationship between CT image quality and average HR in each group was analyzed with Pearson's correlation coefficient. The Pearson's correlation coefficient was valued as follows: poor = 0; slight = 0.01–0.20; fair = 0.21–0.40; moderate = 0.41–0.60; good = 0.61–0.80; and excellent = 0.81–1.00. The average HR cut-off to predict non-diagnostic image quality was determined by receiver operator characteristic (ROC) analysis. A P value of less than 0.05 was deemed statistically significant. All statistical tests were two-sided.

Results

Fifteen of the 160 CCTA examinations were excluded from the study group analysis because of a markedly irregular heartbeat (n = 7) or breathing artifacts (n = 8). Thus, the final study groups included 68 and 77 patients in the step-and-shoot and retrospective CCTA with TCM groups, respectively. There was no significant difference between the groups for average HR (P = 0.89). Oral β-blockers were administered to control the HR before performing the CCTA in 40 of 68 patients (59%) who underwent step-and-shoot and in 52 of 77 patients (68%) who underwent retrospective CCTA with TCM. A total of 2018 segments with a diameter of at least 1.5 mm were evaluated in 145 patients (157 segments were missing because of anatomic variations and a diameter of less than 1.5 mm at the origin, as well as the 20 patients who had at least one non-diagnostic coronary artery segment). There was good interobserver agreement (κ = 0.72) for image quality scores between the independent reviewers. Immediate agreement between both readers was achieved for 1432 of the coronary segments (71%), whereas a consensus reading was required for the remaining 586 segments (29%). Of the 947 coronary artery segments imaged with step-and-shoot, 920 (97.2%) were considered diagnostic, whereas 27 (2.8%) were non-diagnostic because of stair-step artifacts (n = 20) or cardiac motion artifacts (n = 7). Twenty (1.9%) of 1071 coronary artery segments imaged with retrospective CCTA with TCM were non-evaluable because of cardiac motion artifacts. Image quality analysis in coronary segments determined that image quality in each group showed no significant difference (Table 2). Image quality scores of coronary artery segments in the step-and-shoot positively related to HR (r = 0.53, P < 0.001) and radiation dose (r = 0.48, P < 0.001). Age was negatively related to image quality scores (r = −0.26, P < 0.05). ROC analysis revealed the average HR that produced a significant effect on the image quality in the coronary segments. The ROC curve that plotted coronary segments (with at least one coronary segment of non-diagnostic image quality) against average HR determined an average HR of 63.5 bpm as the best threshold for the prediction of non-diagnostic image quality in step-and-shoot (Fig. 2a) and retrospective CCTA with TCM (Fig. 2b). Consequently, non-diagnostic coronary segments in step-and-shoot were significantly less frequent (one of 500 coronary segments [0.2%] in one of 36 patients [2.8%]) with an HR of <63.5 bpm than when HR was ≥63.5 bpm (26 of 447 coronary segments [5.8%] in 11 of 32 patients [34%]; P < 0.0001) (Fig. 3). Moreover, non-diagnostic coronary segments in retrospective CCTA with TCM were significantly less frequent (two of 736 coronary segments [0.3%] in one of 53 patients [1.9%]) with an HR of <63.5 bpm than when HR was ≥63.5 bpm (18 of 335 coronary segments [5.4%] in seven of 24 patients [29.2%]; P < 0.0001) (Fig. 4). With the use of step-and-shoot, 48 and 20 patients were scanned with a tube voltage of 100 kV and 120 kV, respectively. In retrospective CCTA with TCM, 52 and 25 patients were scanned with a tube voltage of 100 kV and 120 kV, respectively. The mean z-axis length for step-and-shoot was 13.7 ± 0.1 cm and this length was significantly less than that in retrospective CCTA with TCM (14.9 ± 0.7 cm, P < 0.01). The mean patient radiation dose was determined from the actual examination z-axis length. As a result, the mean patient radiation dose for step-and-shoot was less than that for retrospective CCTA with TCM (1.74 ± 0.70 mSv vs. 4.51 ± 1.18 mSv; P < 0.0001). When normalized to a 12-cm z-axis length of a typical heart, the mean patient radiation dose for the step-and-shoot was also less than that for the retrospective CCTA with TCM (1.52 ± 0.11 mSv vs. 3.60 ± 0.70 mSv) (P < 0.0001) (Table 1). The mean patient radiation dose for both the actual examination z-axis length and for the z-axis length normalized to 12 cm was 61.4% and 57.8% lower, respectively, for step-and-shoot than for retrospective CCTA with TCM. No significant differences were observed in the image quality scores of coronary artery segments in the step-and-shoot and retrospective CCTA with TCM between the obesity (BMI ≥ 25) and non-obesity (BMI < 25) groups (P = 0.056 vs. P = 0.017). However, radiation doses were higher in the obesity group compared to the non-obesity group for both step-and-shoot (1.40 ± 0.40 mSv vs. 2.61 ± 0.51 mSv) and retrospective CCTA with TCM (4.23 ± 0.90 mSv vs. 6.61 ± 0.92 mSv) (P < 0.0001 vs. P < 0.0001).

OPEN IN VIEWER

Fig. 2

Receiver-operator characteristic (ROC) curves for average heart rate (HR) versus non-diagnostic image quality (a) in step-and-shoot mode and (b) in retrospective tube current modulation mode. (a) ROC analysis revealed an average HR threshold of 63.5 beats per minute in the step-and-shoot mode for the prediction of non-diagnostic image quality (circle); (b) ROC analysis revealed an average HR threshold of 63.5 beats per minute in the retrospective CCTA with tube current modulation mode for the prediction of non-diagnostic image quality (circle)

OPEN IN VIEWER

Fig. 3

Curved multiplanar reformation images of coronary artery in the step-and-shoot mode. (a) Image of RCA was of excellent image quality with radiation dose of 1.08 mSv in an 82-year-old woman who had atypical chest pain. Her mean heart rate was 59 bpm; (b) Image of RCA was of poor image quality with radiation dose of 1.30 mSv in an 81-year-old woman who had atypical chest pain. Her mean heart rate was 69 bpm

OPEN IN VIEWER

Fig. 4

Curved multiplanar reformation images of coronary artery in the retrospective with tube current modulation mode. (a) Image of RCA was of excellent or good image quality with radiation dose of 5.1 mSv in a 53-year-old man who had atypical chest pain. His mean heart rate was 45 bpm; (b) Image of RCA was of poor image quality with radiation dose of 6.8 mSv in a 41-year-old man who had atypical chest pain. His mean heart rate was 78 bpm

OPEN IN VIEWER

Table 2.

Image quality scores

Image quality score	Step-and-shoot (%)	Retrospective with tube current modulation (%)	P value
1	52.80 (500/947)	70.87 (759/1071)
2	30.31 (287/947)	16.99 (182/1071)
3	14.04 (133/947)	10.27 (110/1071)
4	2.85 (27/947)	1.87 (20/1071)	0.144

Data in parentheses are raw data used to calculate the percentages

Discussion

Many reports have proven that the 64-detector cardiac CT with retrospective gating offers good image quality and few non-evaluable coronary artery segments. However, there are some controversies about the usefulness of cardiac CT because of the relatively high radiation dose to patients. Consequently, many medical professionals and CT system manufacturers have sought ways to reduce radiation exposure from CCTA. The two methods used in our study were representative of low-dose techniques. Furthermore, because our study used 128-slice MDCT, our results differed from those of other studies. The present study showed that step-and-shoot represented 97.1% of the coronary segments with diagnostic image quality and low effective radiation dose exposure (1.52 ± 0.11 mSv) in patients with an HR of ≤65 bpm. As stated above, the high radiation dose associated with the use of retrospective CCTA is mostly due to low pitch, and high dose use has been widely debated (9). Among the strategies developed to reduce patient radiation exposure, a lower tube voltage and ECG-dependent tube current dose modulation are useful and common dose reduction methods for retrospective CCTA (10, 12, 20). In our study, we used the TCM technique, which can decrease the radiation dose by up to 4% in the systolic phase compared with those in the diastolic phase. While this mode has the benefit of reducing the tube current, it may cause image noise and beam-hardening artifacts for calcified plaques with reduced detection of soft plaques, and it potentially decreases overall diagnostic image quality. Another benefit of this mode is that it provides functional cardiac information. In contrast, the step-and-shoot mode cannot display functional cardiac information. Consequently, these two methods differ in their usefulness.

With the use of step-and-shoot (19, 21–23), the radiation dose is markedly reduced, and optimal image quality is maintained or improved because the sequential scan mode is used at a predefined time point in the cardiac cycle (24). A single cardiac phase target (commonly mid-diastole) is a major component of step-and-shoot. Usually, imaging of HR higher than 70 bpm is not recommended on step-and-shoot. A regular and steady HR is needed for diagnostic image quality on step-and-shoot, because the incremental movement of the table should match the exact location of the next R-R interval. Even small HR irregularities may lead to stair-step artifacts at the junction of different image blocks. An additional 100-msec padding increases the temporal window where the prospective scanning is centered, normally providing reconstruction between 65–85% of the R-R interval, but the padding also increases the radiation dose (21). In the present study, we observed a significant impact of the average HR on image quality in both step-and-shoot and retrospective CCTA with TCM. Only 2.9% of coronary segments were non-evaluable with the use of step-and-shoot. This percentage was not significantly different compared with the value of 1.9% for non-evaluable segments with the use of retrospective CCTA with TCM. However, in patients with a higher HR, diagnostic image quality is frequently observed on end-systole and the higher HR can decrease the diagnostic image quality of CCTA when an optimal reconstruction interval is chosen from the maximum tube current window (25, 26). The 150-msec temporal resolution in step-and-shoot might not be adequate to delineate the fast-moving coronary artery, particularly in patients with a faster HR (25–27). With the use of dual-source CT scanning in the step-and-shoot mode, diagnostic image quality was obtained for HRs as high as 70 bpm because of the 83-msec high temporal resolution for dual-source CT (28). Earl et al. (21) found an 83% reduction in radiation dose and significantly better image quality for patients who underwent step-and-shoot compared to patients who underwent retrospective CCTA. In our study, the effective dose for the step-and-shoot group was 63% lower than for the retrospective CCTA with TCM group.

There are several limitations of the present study. First, this was not a prospective study, but rather a cross-sectional study conducted by chart review; also, this study was not matched for each characteristic, which makes it hard to generalize the results of this research. Second, the CT image quality scoring we used may have been influenced by subjectivity bias. However, the high κ value determined for interobserver agreement for the coronary artery segments may argue against such a bias. Third, our study did not analyze the effect of HR variability on image quality of coronary artery segments. Previous studies (10, 14, 23, 24) have assessed HR variability, which was calculated as the standard deviation from the average HR. Fourth, no patients with HR > 65 bpm before CT scan were included in CCTA and retrospective CCTA with TCM. Therefore, we did not completely assess the image quality of either CCTA or retrospective CCTA with TCM in patients with an HR > 65 bpm.

In conclusion, step-and-shoot and retrospective CCTA with TCM were both feasible in patients with HR lower than 63.5 bpm. Furthermore, we have shown that the use of both step-and-shoot and retrospective CCTA with TCM with 150-msec temporal resolution in 128-slice MDCT substantially reduced the radiation dose compared with that of a previous study using 64-slice MDCT but did not significantly decrease coronary artery segment assessability and diagnostic image quality. Additionally, step-and-shoot mode has similar subjective image quality scores but has a 63% lower patient radiation dose compared with retrospective CCTA with TCM.

Conflict of interest: None.