Author + information
- Received July 23, 2008
- Revision received December 11, 2008
- Accepted December 24, 2008
- Published online April 1, 2009.
- Hyuk-Jae Chang, MD, PhD,
- Richard T. George, MD,
- Karl H. Schuleri, MD,
- Kristine Evers,
- Kakuya Kitagawa, MD, PhD,
- João A.C. Lima, MD and
- Albert C. Lardo, PhD⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Albert C. Lardo, Johns Hopkins University School of Medicine/Division of Cardiology, 1042 Ross Building, 720 Rutland Avenue, Baltimore, Maryland 21205
Objectives This study sought to determine whether low-dose, prospective electrocardiogram (ECG)-gated delayed contrast-enhanced multidetector computed tomography (DCE-MDCT) can accurately delineate the extent of myocardial infarction (MI) compared with retrospective ECG-gated DCE-MDCT.
Background For defining the location and extent of MI, DCE-MDCT compares well with delayed enhanced cardiac magnetic resonance. However, the addition of a delayed scan requires additional radiation exposure to patients. MDCT protocols using prospective ECG gating can substantially reduce effective radiation dose exposure, but these protocols have not yet been applied to infarct imaging.
Methods Ten porcine models of acute MI were imaged 10 days after MI using prospective and retrospective ECG-gated DCE-MDCT (64-slice) 10 min after a 90-ml contrast bolus. The MDCT images were analyzed using a semiautomated computed tomography density (CTD) threshold technique. Infarct size, signal-to-noise (SNR) ratios, contrast-to-noise (CNR) ratios, and image quality metrics were compared between the 2 ECG-gating techniques.
Results Infarct volume measurements obtained by both methods were strongly correlated (R = 0.93, p < 0.001) and in good agreement (mean difference: −0.46 ml ± 4.00%). Compared with retrospective ECG gating, estimated radiation dosages were markedly reduced with prospective ECG gating (930.1 ± 62.2 mGy×cm vs. 42.4 ± 2.3 mGy×cm, p < 0.001). The SNR and CNR of infarcted myocardium were somewhat lower for prospective gated images (22.0 ± 11.0 vs. 16.3 ± 7.8 and 8.8 ± 5.3 vs. 7.0 ± 3.9, respectively; p < 0.001). However, all examinations using prospective gating protocol achieved sufficient diagnostic image quality for the assessment of MI.
Conclusions Prospective ECG-gated DCE-MDCT accurately assesses infarct size compared with retrospective ECG-gated DCE-MDCT imaging. Although infarct SNR and CNR were significantly higher for the retrospective gated protocol, prospective ECG-gated DCE-MDCT provides high-resolution imaging of MI, while substantially lowering the radiation dose.
The ability to distinguish dysfunctional but viable myocardium from nonviable tissue after myocardial ischemia has important implications for the therapeutic management of patients with coronary artery disease (CAD) (1,2). The assessment of both myocardial viability and infarct morphology using delayed contrast-enhancement (DCE)-cardiac magnetic resonance (CMR) has been well validated over the past several years (3,4); however, the limited spatial resolution of CMR has the potential to overestimate infarct size due to partial volume effects (5), and long exam times exclude its application in the setting of acute chest pain. Furthermore, with the proliferation of intracardiac devices (i.e., implantable cardioverter-defibrillators and biventricular pacers), alternatives to DCE-CMR are required in this population of patients, who are potential candidates for viability imaging.
The potential to visualize nonviable myocardium using computed tomography (CT) was first reported in late 1970s (6,7). Because iodinated contrast agents for CT show similar delayed contrast kinetics as gadolinium-based CMR contrast agents (8) and new-generation detectors are providing images with unsurpassed spatial resolution, the direct assessment of myocardial viability by DCE multidetector computed tomography (MDCT) has evoked increasing interest and initial results are promising. Recently, we showed that the spatial extent of acute and healed myocardial infarction (MI) can be determined and quantified accurately with DCE-MDCT (9). However, the need to increase X-ray exposure with additional late scans is a major obstacle to the widespread use of MDCT for this application.
So far, infarct detection by MDCT has most often been based on the description of perfusion deficits during the first pass of contrast. MDCT first-pass imaging correlates well with infarct size in nonreperfused MI (10,11); however, in the setting of reperfusion, it shows insufficient accuracy and significant underestimation of infarct size (12,13).
Several studies reported various low-dose DCE-MDCT protocols based on modulation of tube current and/or reductions in tube voltage (14–16). However, these methods come at the cost of greater image noise, and dose reductions were not enough to apply DCE-MDCT into routine clinical practice, when combined with coronary computed tomography angiography (CCTA).
More recently, new CCTA acquisition protocols have been proposed with prospective electrocardiogram (ECG) gating (17). With this technique, radiation is only administered at predefined time points of the cardiac cycle, rather than the entire cardiac cycle as in the case of retrospective ECG-gated helical MDCT acquisition. The former is associated with a substantial reduction of radiation dose without compromise of diagnostic image quality.
Accordingly, the purpose of the present study was to explore the hypothesis that low-radiation DCE-MDCT viability imaging with prospective ECG gating is accurate for the measurement of infarct size and microvascular obstruction (MVO) compared with high-dose retrospective ECG-gated DCE-MDCT imaging.
All animal studies were approved by the Johns Hopkins University Institutional Animal Care and Use Committee and comply with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health publication no. 80-23, revised 1985).
Myocardial infarction was created by temporary balloon occlusion of the left anterior descending (LAD) coronary artery for 120 min, followed by reperfusion. Ten Göttingen minipigs (Marshall BioResources, North Rose, New York) (mean weight 28 ± 3 kg) were subjected to induction of anesthesia with ketamine (35 mg/kg intramuscularly), and initially anesthesia was achieved with pentobarbital (20 to 60 mg/kg intravenously). Minipigs were intubated and mechanically ventilated with inhaled isoflurane during catheterization and MDCT scanning. A midline neck incision was made, and the right carotid artery was cannulated with an 8-F sheath. Heparin 5,000 U was administered intravenously. Left ventriculogram was performed with either a 7-F pigtail catheter with side holes or a Judkins right 4 guide catheter in both the left anterior oblique views. The left coronary artery was cannulated with a Judkins right 4 or hockey stick guiding catheter, and a baseline coronary angiogram was obtained to show the patency of artery. Then, a 0.014-inch angioplasty guidewire was inserted into the LAD under fluoroscopic guidance and 2.5 × 12-mm Maverick balloon (Boston Scientific, Natick, Massachusetts) was inflated to 4 atm just distal to the second diagonal branch of the LAD. Complete cessation of flow in the LAD was confirmed by angiography. The balloon was left inflated for 120 min with an intravenous lidocaine infusion throughout this time. After balloon deflation, restoration of flow in the LAD was confirmed by angiography and any spasm of the coronary artery was treated with intracoronary nitroglycerin 100 μg. Repeat left ventriculogram confirmed the presence of an anterior wall MI. The catheters and sheath were removed, and the carotid artery ligated.
MDCT imaging and reconstruction
MDCT imaging was performed with a 0.5 mm × 64-detector scanner (Aquilion 64, Toshiba Medical Systems Corporation, Otawara, Japan). Animals received intravenous metoprolol (2 to 5 mg) and amiodarone (50 to 150 mg) to achieve a heart rate <100 beats/min for delayed-enhancement studies. Mean heart rate during the MDCT examination was 96 ± 8 beats/min. After scout acquisition and slice prescription, a 90-ml bolus of iodixanol (Visipaque 320, Amersham Health, Amersham, United Kingdom) was injected intravenously at a rate of 5 ml/s, followed by a 30-ml saline chaser. During image acquisition, mechanical ventilation was suspended and delayed enhanced images were acquired 10 min after contrast delivery. Immediately after scanning with a prospective ECG-gated MDCT protocol, a retrospectively gated scan was performed with preset parameters as in Table 1. The mean time interval between both scans was 45 ± 15 s (Fig. 1).
After acquisition of the dataset, prospectively and retrospectively gated images were reconstructed at a 3-mm slice thickness using identical convolution kernel (FC 43) by half-segment and multisegment reconstruction algorithms, respectively. The source images on the same phase of the cardiac cycle with prospectively gated scan were selected for reconstruction from retrospectively gated scan (Table 1). For exposure dose calculation, an acrylic phantom with a diameter of 32 cm was scanned with a 32-cm field of view to measure weighted computed tomography dose index (CTDIw). A calibrated ion chamber dosimeter was used for measurement. The result showed that the normalized CTDIw (nCTDIw) per 100 mAs at 120 kV was 9.4 mGy/100 mAs. Based on this result, the exposure dose was calculated based on the scan conditions shown in Table 1 (18).
The reconstructed data set was then transferred to a custom cardiac software package (Cine Display Application, GE Medical Systems, Milwaukee, Wisconsin) for image analysis.
Using hand planimetry, after definition of endocardial and epicardial contours of left ventricle (LV), normal myocardial computed tomography density (CTD) was determined by drawing a region of interest (ROI) in the myocardium remote from the infarct in each slice. These results were used to define a hyperenhancement CTD threshold as myocardium having an CTD 1 SD above the mean CTD of the remote region. Infarct area was then calculated using a step-based algorithm to: 1) detect voxels within the defined CTD range; 2) check for voxel continuity in the X, Y, and Z directions; 3) generate clusters of voxels that meet the hyperenhancement definition; and 4) define the hyperenhancement as the largest cluster detected. Using this method, the software determined the mean CTD of hyperenhancement in Hounsfield units and calculated total infarct volume and total percent volume. MVO was defined as the hypoenhanced regions within the hyperenhanced area (Fig. 2). The MDCT images were also analyzed by 2 independent experienced operators to calculate interobserver and intraobserver agreements in a random subset of 5 animals.
In each animal, visually matching corresponding subsets of 5 image slices were chosen for comparison of quantitative and qualitative assessments of image quality according to scan protocols. The CTDs were measured on the raw-data images by drawing 10-mm2 ROI in LV blood pool (CTDblood), infarct (CTDinf), and normal myocardium (CTDmyo). The signal-to-noise ratio (SNR) was defined as the ratio of mean CTD in the ROIs to the SD of the CTD in the air (i.e., noise) (SDnoise) lateral to the thorax. The contrast-to-noise ratio (CNR) was calculated according to the following formula (18): CNR = (CTDinf – CTDmyo) / SDnoise. Image quality was assessed on a 4-point scale: 1) not useful for diagnostic purposes; 2) compromised diagnostic image quality; 3) good diagnostic image quality; and 4) excellent diagnostic image quality (18).
Agreement of the results was determined by the Pearson correlation coefficient (19) and Bland-Altman method calculated as mean ± SD difference between the 2 methods at 95% confidence interval (CI) (20). According to the normality of distribution, paired t test (21) or Wilcoxon signed rank test (22) was performed to determine whether there were significant differences in parameters between scan protocols. Intraobserver and interobserver variance was calculated as mean difference ± SD. Statistical significance was defined as a 2-sided probability value <0.05, and all analyses were performed using MedCalc for Windows, version 188.8.131.52 (MedCalc Software, Mariakerke, Belgium).
MDCT images were successfully obtained using both protocols and DCE in the infarct territory was confirmed in all animals (Fig. 3). Total LV myocardial volumes did not differ between prospective and retrospective protocols (mean absolute difference: 0.54 ± 5.81%) and linearly correlated with ones of retrospective scan (R = 0.69, 95% CI: −0.10 to 0.92, p = 0.03).
Comparison of infarct and MVO absolute and percent volume between protocols showed good correlations and only small mean differences (−0.46 ± 4.00% and −0.99 ± 1.45%, respectively) (Table 2 and Fig. 4).
For retrospectively gated data sets, interobserver and intraobserver measurements of infarct percent volumes were closely correlated (R = 0.97, 95% CI: 0.59 to 1.00, p = 0.01, and R = 0.98, 95% CI: 0.67 to 1.00, p = 0.01, respectively) and mean differences in infarct volumes were small (1.98 ± 3.10% and 0.45 ± 2.44%, respectively). Similar results were obtained for prospectively gated acquisitions (R = 0.96, 95% CI: 0.55 to 1.00, p = 0.01, and R = 0.98, 95% CI: 0.62 to 1.00, p = 0.005; mean differences, 2.40 ± 3.30% and 0.31 ± 3.58%, respectively).
Although the SNR and CNR of infarct were somewhat higher for the retrospective technique, prospective ECG-gated DCE-MDCT provided high-resolution imaging of MI and lowered the radiation dose (930.1 ± 62.2 mGy×cm vs. 42.4 ± 2.3 mGy×cm, p < 0.001). On a slice by-slice analysis, the CTDs of LV blood pool, normal myocardium, and infarcted myocardium were not different between methods (Table 3).
The infarct SNR and CNR were lower for prospectively gated images, however all examinations achieved either good or excellent diagnostic image quality(62%: good, 38%: excellent on a per-slice analysis, p < 0.001; 70%: good, 30%: excellent on a per-subject analysis, p = 0.04).For retrospectively gated images, 52% were graded as good while 48% were rated as excellent on a per-slice analysis and 50%: good, 50%: excellent on a per-subject analysis.
The major finding of this study is that prospective ECG-gated DCE-MDCT myocardial imaging can accurately quantify nonviable and viable myocardium after MI while lowering the radiation dose by an order of magnitude compared with retrospective ECG-gated DCE-MDCT imaging.
Almost 30 years ago, the potential of computed tomography for the detection of MI was reported (6–8). Since this time, improvements in MDCT technology related to spatial and temporal resolution have allowed detailed imaging of myocardial viability. Due to concerns about increasing X-ray exposure with additional late scans; however, infarct detection by MDCT has been based on the appearance of perfusion deficits during the first pass of contrast. Hoffmann et al. (10) performed 4-slice MDCT in a pig model of nonreperfused MI and showed a good correlation between histomorphometric infarct and MDCT regions of hypoenhancement. In the setting of reperfusion (12,13), however, MDCT first-pass imaging shows insufficient accuracy and significant underestimation of infarct size, which have become much more common after the introduction of reperfusion therapy.
The DCE-MDCT can determine and accurately quantify MI by well-delineated hyperenhanced regions within certain time intervals after contrast injection, whereas regions of MVO are characterized by hypoenhancement on early imaging in a time-dependent manner (9,23). Furthermore, iodine concentration on the infarct region reaches a maximum very quickly (5 to 10 min) after contrast injection; therefore, the single contrast bolus can be used to acquire both CCTA and DCE-MDCT to enhance nonviable myocardium. The assessment of both myocardial viability and infarct morphology, in conjunction with current high-resolution coronary artery imaging, represents an undisputed advantage of MDCT over other competing imaging technologies and has important implications for comprehensive assessment of the patients with CAD and MI. Hence, establishment of MDCT protocols with a lower radiation dose while maintaining suitable image quality is an essential prerequisite to assess myocardial viability as a part of a comprehensive study.
Initial studies for the practicability of MDCT viability imaging have been performed with standard 120 kV and 400 to 800 mAs full-dose protocols (15), but there are also promising results with lower-dose late-phase CT scans. Brodoefel et al. (14) reported that myocardial viability can accurately be assessed by MDCT at 80 kV. Closer to the K-edge of iodine, a tube voltage of 80 kV achieves a more optimal absorption of radiation and hence has the potential to allow better contrast enhancement. However, this comes at the cost of greater image noise. Accordingly, studies have used high tube current (800 mAs), which lowers the radiation dose by only 65% when compared with standard acquisition parameters. In practice, the use of a voltage of 80 kV leads to an exponential increase in image noise in thoracic imaging of patients weighing more than 75 kg. Therefore, high patient weight or body mass index remains a considerable limitation for the use of low kilovoltage protocols in routine clinical practice (24).
More recently, a new CCTA acquisition protocol has been introduced that utilizes prospective ECG gating (17). Prospective ECG-gated protocols limit radiation dose exposure to only a small portion of the R-R interval, thus significantly reducing X-ray exposure time to patients. This is in contrast to retrospective ECG-gated protocols during helical MDCT imaging that continuously expose patients to radiation throughout the R-R interval over 5 to 10 heartbeats. The former is likely to be associated with a substantial (90%) reduction of radiation dose. In the present study, we used the prospective ECG-gated protocol that we currently use to perform calcium scoring (120 kV/150 mA) and compared with our commonly used 64-slice CCTA protocol (120 kV/400 mA). Although the signal intensity of infarct and normal myocardium were not different between protocols, there was a decrease in the SNRs of infarct and normal myocardium using the prospective ECG-gated protocol, primarily attributed to its lower tube current and longer data windows used for image reconstruction (half vs. segmental scan reconstruction). However, in the qualitative analysis, all the examinations using the prospective ECG-gated protocol achieved sufficient diagnostic image quality for the assessment of myocardial contrast change. In addition to a decrease in dose caused by the prospective nature of the acquisition, there is a further decrease attributable to the use of a wider collimation. Nieman et al. (25) showed a decrease in radiation exposure with a wider detector collimation (24 × 1.2 mm) and lower tube output (100 kV, 800 mAs). However, the calculated radiation dose was 4.5 ± 2.4 mSv, which was still 62% of their standard acquisition protocol. In our study, we showed that prospective ECG-gated DCE-MDCT can provide a substantial radiation dose reduction and seems to be feasible without sacrificing diagnostic accuracy.
In the present study, we did not provide CMR or histopathological correlation with infarcted tissue as defined by DCE-MDCT. However, our previous experimental work (9) did show that the spatial extent of MI can be quantified accurately with retrospective ECG-gated DCE-MDCT imaging. The contrast injection used in the study for 30-kg pigs in average is approximately equivalent to a double contrast dose for an average-sized patient. Such a dose (90 ml) was selected to ensure that the contrast material was not a limiting factor and tissue uptake was maximized to define the ROI accurately.
Prospective ECG-gated MDCT imaging can accurately assess infarct and MVO size when compared with retrospective ECG-gated MDCT imaging. Although SNR and CNR of infarct were significantly higher in retrospective ECG-gated protocol, prospective ECG-gated DCE-MDCT can provide high-resolution imaging of MI while lowering the radiation dose from 930.1 to 42.4 mGy×cm.
This study was supported, in part, by research grants from Toshiba Medical Systems Co., Ltd., Tokyo, Japan. Dr. George was supported by the Donald W. Reynolds Cardiovascular Research Center at Johns Hopkins University. The terms of this arrangement are managed by Johns Hopkins University in accordance with its conflict of interest policies. Drs. Chang and George contributed equally to this paper.
- Abbreviations and Acronyms
- coronary artery disease
- coronary computed tomographic angiography
- confidence interval
- contrast-to-noise ratio
- cardiac magnetic resonance
- computed tomography
- computed tomography density
- delayed contrast-enhancement
- left anterior descending
- left ventricle/ventricular
- multidetector computed tomography
- myocardial infarction
- microvascular obstruction
- region of interest
- signal-to-noise ratio
- Received July 23, 2008.
- Revision received December 11, 2008.
- Accepted December 24, 2008.
- American College of Cardiology Foundation
- Pagley P.R.,
- Beller G.A.,
- Watson D.D.,
- Gimple L.W.,
- Ragosta M.
- Gerber B.L.,
- Rochitte C.E.,
- Melin J.A.,
- et al.
- Wu K.C.,
- Lima J.A.
- Fieno D.S.,
- Kim R.J.,
- Chen E.L.,
- Lomasney J.W.,
- Klocke F.J.,
- Judd R.M.
- Gray W.R.,
- Buja L.M.,
- Hagler H.K.,
- Parkey R.W.,
- Willerson J.T.
- Higgins C.B.,
- Siemers P.T.,
- Schmidt W.,
- Newell J.D.
- Lardo A.C.,
- Cordeiro M.A.,
- Silva C.,
- et al.
- Mahnken A.H.,
- Koos R.,
- Katoh M.,
- et al.
- Husmann L.,
- Valenta I.,
- Gaemperli O.,
- et al.
- Morin R.L.,
- Gerber T.C.,
- McCollough C.H.
- Pearson K. III.
- Westgard J.O.,
- Hunt M.R.
- Conover W.