Author + information
- Received December 3, 2007
- Revision received February 19, 2008
- Accepted February 29, 2008
- Published online May 1, 2008.
- Phalla Ou, MD, PhD⁎,†,⁎ (, )
- David S. Celermajer, MBBS, DSc, FRACP‡,
- Davide Marini, MD†,
- Gabriella Agnoletti, MD, PhD†,
- Pascal Vouhé, MD, PhD†,
- Francis Brunelle, MD⁎,
- Kim-Hanh Le Quan Sang, MD§,
- Jean Christophe Thalabard, MD, PhD¶,
- Daniel Sidi, MD, PhD† and
- Damien Bonnet, MD, PhD†
Reprint requests and correspondence:
Dr. Phalla Ou, Hôpital Necker-Enfants Malades, Department of Pediatric Radiology, 149 rue de Sèvres, 75743 Paris Cedex 15, France.
Objectives We investigated the accuracy of 64-slice computed tomography (CT) angiography, as compared to invasive angiography, to evaluate reimplanted coronary arteries in children after arterial switch operation (ASO) for transposition of the great arteries (TGA).
Background Assessment of the integrity of reimplanted coronary arteries is crucial for long-term outcome after ASO for TGA. Noninvasive tests have limited accuracy for detecting significant coronary lesions, and invasive coronary angiography is usually required in this setting.
Methods One hundred thirty consecutive children, after ASO for TGA (age 5.6 ± 1.1 years), underwent conventional invasive coronary angiography and coronary CT angiography using a 64-slice scanner. The ability of CT to detect significant coronary stenoses (>30% diameter reduction) of the coronary ostia and proximal segments, and other abnormalities of the coronary arteries was analyzed by blinded comparison to the invasive coronary angiogram.
Results The CT was fully evaluable in 126 of 130 patients (97%), allowing assessment of ostia and proximal segments of all coronary arteries. The CT correctly detected all 12 patients (9.2%) in whom invasive coronary angiography had identified significant coronary lesions, with a sensitivity, specificity, and negative predictive value of 100%. In addition, CT showed nonsignificant coronary lesions (<30% luminal narrowing) in 6 patients and allowed determination of the underlying reasons for coronary luminal narrowing, such as stretching or compression of the re-implanted coronary arteries caused by their anatomic relationship to the adjacent great vessels.
Conclusions 64-slice CT coronary angiography performs as well as invasive angiography for detecting significant coronary lesions in the majority of children who have undergone the arterial switch procedure for TGA. CT also provides information on the underlying mechanism of coronary luminal narrowing.
During the arterial switch operation (ASO) for transposition of the great arteries (TGA) (1), reimplantation of the coronary arteries presents the principal difficulty in the surgical procedure, particularly in cases of abnormalities of coronary origin or course (2,3). Mortality and long-term outcome after ASO mainly depend on the perfusion provided by the transferred coronary circulation (4–6). Both immediate post-operative survival (1,2,7) and sudden death several years after surgery (1,6,8) are mainly related to coronary artery pathology, which leads to myocardial ischemia and/or infarction.
Assessment of the integrity of the coronary circulation is therefore of importance during the follow-up after ASO. Recently, Legendre et al. (9) reported a prevalence of 7.2% for coronary events during a mean follow-up period of 59 months after ASO in a large series of 1,198 survivors. Importantly, the investigators showed that classic noninvasive tests—electrocardiogram, echocardiography, treadmill test, myocardial scintigraphy—are not sensitive enough to detect significant coronary stenosis. Therefore, they concluded that children surviving the ASO require coronary artery angiography, consistent with the recommendations of previous investigators who had shown that significant coronary lesions may be subclinical and cause sudden death (4,5). Coronary lesions after ASO almost always involve the ostial and proximal segments and usually result from compression, kinking, or stretching of the coronary arteries in relation to the aorta and/or pulmonary artery during somatic growth (4,5).
In addition to its invasive nature and attendant risk, conventional selective coronary angiography has two important limitations in assessing children after ASO. First, the selectively engaged coronary catheter may straighten or pass through an ostial kink or stenosis, which might then be more difficult to visualize. In addition, the images of the coronary vessels are not seen in the topographic context of the adjacent great arteries, which is often the underlying reason for coronary lesions (because of stretching, compression, and/or kinking) (10). In contrast, multislice computed tomography (CT) technology allows direct noninvasive visualization of the coronary artery lumen, and the course of the vessel in relation to the nearby major cardiovascular structures (11,12).
The aim of this study was therefore to assess the efficacy of 64-slice CT for the detection of coronary complications in children having undergone ASO for TGA, in comparison with conventional invasive coronary angiography.
At our institution, children surviving the ASO for TGA undergo routine conventional coronary angiography for clinical purposes, at approximately age 5 to 6 years. Between May 2005 and January 2007, 130 consecutive such children (mean age 5.6 ± 1.1 years, weight 19.7 ± 9 kg, follow-up 5.3 ± 1 years) underwent 64-slice CT angiography in the 24 h before or after their conventional coronary study. Exclusion criteria were known hypersensitivity to iodine, creatinine clearance <80 ml/min, or a history of asthma. All parents or guardians gave written, informed consent, and the hospital ethics committee approved this study.
All examinations were performed with a 64-slice CT system (LightSpeed VCT, GE Medical Systems, Milwaukee, Wisconsin). All patients were in sinus rhythm, and if not contraindicated, received oral beta-blocker medication (propranolol 1 to 2 mg/kg) 1 h before CT examination with the aim of lowering resting heart rate to <80 beats/min. No sedation was given. Thirty minutes before the CT scan, all children were trained by a technician to hold their breath for 5 to 7 s.
The CT scanning was performed with intravenous contrast enhancement; 1.5 ml/kg iodine contrast medium (iohexol 300 mgI/ml) was injected at a rate of 2.5 to 3 ml/s with a power injector (Medrad En Vision CT, Pittsburgh, Pennsylvania), followed by a chaser bolus of 10 ml saline into an antecubital vein. The scan delay was determined using a bolus tracking technique: continuous low-dose fluoroscopy (70 kV, 20 mA) at the level of the ascending aorta was initiated 8 s after the start of contrast injection. The CT attenuation was then measured every 2 s within a circular region of interest in the ascending aorta. When the attenuation value reached a pre-set value of 300 HU during 2 consecutive sampling intervals, acquisition of the CT angiography data set was initiated.
The standard CT angiography acquisition protocol included a 350-ms rotation time and collimation of 64 × 0.625 mm. The scan volume extended from the pulmonary trunk to just below the base of the heart. All scans were performed in a craniocaudal direction. The CT parameters were adapted to the patient's weight. Images were acquired using 80 kV, and a fully automated real-time, anatomy-based dose regulation algorithm modulated the effective tube current (range from 150 to 350 mA). The effective dose of CT coronary angiography was calculated by the method proposed by the European Working Group for Guidelines on Quality Criteria in CT (13). The effective dose was derived from the product of the dose-length product and conversion coefficients for the chest taking account patient age, as proposed by Shrimpton and Wall (14).
A first set of images was reconstructed systematically at 75% of the R-R interval, with a smooth kernel and mediastinal windows. In the case of insufficient image quality, additional sets of images were reconstructed during various time instants during the cardiac cycle, and the data set with the least motion artifact was selected for further analysis.
Conventional coronary angiography
Conventional coronary angiography was performed with oral sedation (Hydroxyzine, 1 mg/kg 1 h before the examination, maximum 100 mg) and local anesthesia using 4-F catheters, as previously described (4). Biplane angiograms were acquired in lateral, left, and right anterior oblique views (20°) after selective injection into the coronary arteries.
The effective dose from conventional angiography was calculated with a PC-based X-ray Monte Carlo program (Radiation and Nuclear Safety Agency, Helsinki, Finland) (15). The software is flexible and can be adjusted to patient morphology.
Analysis of data
All conventional angiograms were analyzed by an experienced interventional pediatric cardiologist. The CT images were assessed by 2 other independent observers with experience in congenital heart disease, blinded to the conventional coronary angiography findings. Coronary arteries were analyzed using axial slices and, if necessary, with the aid of post-processing tools such as multiplanar reconstruction, maximum-intensity thin-slab projection and 3-dimensional reconstruction.
Analysis was limited to 6 coronary artery segments: ostia of the left and right coronary artery, left main coronary artery, and the proximal segments of the left anterior descending, left circumflex, and right coronary artery. In both modalities, coronary lesions were graded using visual assessment, and classified as either normal or having a significant stenosis (>30% diameter reduction or occlusion). Coronary pattern was described using the classification of Yacoub and Radley-Smith (2).
Data were stored and analyzed using the JMP software Version 5.0.1a (SAS Institute Inc., Cary, North Carolina). Values are expressed as mean ± standard deviation, range, or median. For each of the 6 coronary segments outlined above, the presence of stenotic lesions in CT was compared with conventional angiography. Sensitivity, specificity, and negative predictive value were calculated for detection of coronary stenoses or occlusions. Interobserver variability for detecting ostial and proximal coronary lesions on CT was determined by κ-statistics. Differences between parameters for CT and conventional angiography were tested by unpaired t tests at the 95% confidence level (2-tailed).
Table 1 lists technical characteristics of CT and invasive coronary angiography. The CT examinations were successfully completed without sedation in all patients. There were no complications. The mean heart rate at rest was 92 ± 19 beats/min (range 80 to 115 beats/min). Effective beta-blockade (mean 74 ± 13 beats/min, range 60 to 78 beats/min) and breath holding during the acquisition resulted in interpretable images from 126 patients (97%) (age 5.9 ± 0.7 years). The other 4 subjects were slightly younger (average age 4.9 years), and scans were acquired with a high heart rate (average 91 beats/min) and/or without breath holding. These 4 patients were thus excluded from further analyses. The duration of the CT examination (including patient preparation, CT acquisition, and post-processing analysis) was shorter than the duration of conventional angiography (20 ± 6 min vs. 45 ± 23 min, p < 0.001). The volume of iodinated contrast agent was similar for both techniques (30 ± 7 ml vs. 37 ± 25 ml, p = 0.6).
Invasive angiography was performed with oral sedation (hydroxyzine, 1 mg/kg 1 h before the examination, maximum 100 mg) in all patients, and general anesthesia was required in 2 patients. Complications were more frequent (n = 6, 4.6%), including bleeding (n = 1), femoral hematoma at the arterial access site (n = 3), femoral artery aneurysm (n = 1), and transient ischemic stroke (n = 1). The effective radiation dose of invasive angiography was lower than CT (3.1 ± 1.6 mSv vs. 4.5 ± 0.5 mSv, p < 0.001).
During invasive angiography, selective coronary visualization was achieved in 120 patients (92.3%) and nonselective angiography was performed in the other 10 patients (7.7%). Twelve patients (9.2%) had significant coronary lesions, including 1 ostial occlusion of the right coronary artery, 3 ostial occlusions of the left coronary artery, 2 ostial occlusions of the reimplanted circumflex artery, 1 stenosis of the right coronary artery, 3 stenoses of the left main coronary artery, 1 stenosis of the left anterior descending artery, and 1 stenosis of the circumflex artery (Table 2). The mechanism of the aforementioned coronary artery lesions were not clearly delineated on conventional angiography. There were no coronary lesions noted in the 4 patients who had been excluded from the analysis because of inadequate CT images.
The CT angiography was able to assess all 6 pre-specified coronary segments in all 126 patients in whom acquisitions were obtained in appropriate conditions of beta-blockade and breath hold. The CT correctly detected all 12 patients and the precise coronary lesions that had been identified by conventional angiography. Interestingly, CT was also able to visualize the topographic relationship of the proximal vessels relative to the great arteries. Mechanisms of coronary lesions were not evident in the 6 patients presenting with complete occlusion of coronary ostia. Occlusion involved both patients with a normal type A coronary pattern and those presenting with a higher-risk pattern. In the 6 other patients with stenosis, but not occlusion of a coronary artery, CT was able to clearly show the mechanism of the stenosis (Table 2): Stenosis related to a coronary course between the aorta and pulmonary artery was seen in 1 patient presenting with a classic high-risk type C coronary pattern (Patient #7, Fig. 1). Stretching of the proximal segment of the left circumflex artery (type D coronary pattern) was seen in 1 patient in whom the coronary vessel performed a long retroartic loop around a dilated neoaorta (Patient #8, Fig. 2). Compression of the ostium and/or the proximal segment of the coronary artery was seen in 4 patients (Patients #9 to 11 presented with a normal type A coronary artery pattern, Patient #12 presented with a type D coronary artery pattern). In all of the 4 latter patients, the coronary artery (1 right and 3 left coronary arteries) had been reimplanted at an anterior position, so that the ostium arose at around 12 o'clock with the patient in supine position (Fig. 3). Thus, the sensitivity, specificity, and negative predictive value of CT in detecting significant coronary artery lesions were of 100% compared with conventional angiography.
Additional findings on CT
In 6 other children, CT showed mild coronary lesions (below the threshold of 30% diameter stenosis). In none of these cases had the lesions been identified in invasive angiography. One patient had a mild compression of the right coronary artery between the great arteries (Fig. 4), 1 patient had mild stretching of the left coronary artery, 2 patients had systolic compression of the proximal portion of the left main coronary artery (Fig. 5), and 2 patients had a discrete stenosis at the ostium of the left coronary main coronary artery that followed a tortuous course between the great arteries. Interobserver agreement for detecting the ostial and proximal coronary lesions was good (κ-value 0.89).
The 12 patients with significant coronary lesion as well as the 6 other with mild coronary lesion underwent functional evaluation with myocardial scintigraphy. The 6 patients with ostial coronary occlusion did not have myocardial ischemia. They had extensive development of collaterals from the contralateral coronary artery. In 5 patients with severe stenosis (1 stenosis of the right coronary artery, 3 stenoses of the left main coronary artery, and 1 stenosis of the left anterior descending artery) myocardial perfusion imaging was positive for ischemia. They underwent successful surgical coronary revascularization (2 patients underwent surgical angioplasty and 3 others had a coronary artery bypass graft). The patient with a significant stenosis of the circumflex artery had a mild and localized ischemia on scintigraphy and was managed medically. All 6 patients with mild coronary lesions did not have ischemia on myocardial scintigraphy.
In this study, we show that noninvasive CT angiography can successfully be performed in the large majority (97%) of 5- to 6-year-old children after ASO for TGA, and is both highly sensitive and specific for the detection of coronary stenoses and/or occlusions after this procedure. Furthermore, CT coronary angiography has the ability to elucidate the underlying mechanism of coronary artery narrowing, for example through stretching, compression, and/or kinking caused by the surrounding great arteries. CT may in fact be able to show mild coronary pathology that is not detectable in invasive angiography, but these observations lack validation and their clinical relevance unclear.
A CT coronary angiography can be performed in a shorter time, has a lower rate of complications, and will most likely be associated with a lower cost than conventional invasive angiography. It is our opinion that CT angiography can be adopted as the follow-up method of choice for examining the coronary circulation in children after ASO for TGA. Although there are insufficient data to make clinical recommendations, coronary CT angiography may be especially useful in children who present with a classic high-risk coronary pattern, such as the types B and C with a course between the aorta and pulmonary artery. However, exposure to ionizing radiation is a trade-off and the long-term impact of radiation exposure in pediatric patients must be considered.
In a previously published preliminary study using 4-slice CT scanning, we had documented promising results for detecting proximal coronary abnormalities in patients after the switch operation (16). Using 64-slice technology, we have found a substantially improved ability to detection the location, nature, and mechanism of coronary stenoses. Furthermore, the introduction of advanced post-processing tools allows improved visualization of anomalous coronary distribution patterns associated with increased risk for stretching or compression, such as that of a coronary segment passing between the main pulmonary artery and ascending aorta (Fig. 5). In fact, these visualization tools may be useful for planning surgical interventions.
In 6 patients, CT showed coronary lesions of <30% diameter reduction that were not detectable in invasive angiography, the standard of reference in our study. In fact, CT may be superior to invasive angiography for analysis of the coronary ostia and proximal coronary segments. First of all, CT allows visualization of the coronary arteries in the topographic context of adjacent structures, which may cause compression or kinking. Second, CT allows a multitude of viewing angles and is neither restricted to angiographic projections nor subjected to foreshortening or overlap. Finally, a selectively engaged coronary catheter may pass beyond an ostial or proximal coronary lesion and mask a kink or mild stenosis, which might then be nondetectable in conventional angiography, making the invasive angiogram an imperfect reference standard.
Certain prerequisites must be fulfilled to allow the acquisition of technically satisfactory CT images. Good cooperation of the patient is required because data acquisition must be performed during a breath-hold period of approximately 5 s to avoid respiratory motion artifact. In our experience, careful guidance of the patients, including reassurance, thorough explanation of the procedure, and training of the breath-hold maneuvers, allow the examinations to be performed with high quality, even in patients as young as 5 years. Furthermore, beta-blockade is required to lower the heart rate. We achieved the desired effect by oral medication 1 to 2 h before the examination, which reliably resulted in a heart rate <80 beats/min (16). In our experience, CT coronary angiography may be difficult in children under 5 years of age but can almost always be successfully performed in children ages 5 to 6 years old or more.
Certain limitations of the current approach must be mentioned. First, there was a small number of children (4 of 130) in whom coronary CT angiography was unsatisfactory, and such cases would need to proceed with conventional angiography. Second, CT angiography provides excellent images of the coronary ostia and proximal segments but not necessarily of the distal segments. In clinical practice, however, the abnormalities of importance after ASO occur in the proximal segments, and visualization of distal segments is of lesser importance. Third, radiation exposure is a potential disadvantage in this group of children. Although serial follow-up CT scans will certainly not be reasonable, a single, high-quality CT examination at age 5 to 6 years would be expected to exclude important coronary problems. In case of a negative result, it would provide important reassurance. The disadvantage of a higher radiation dose seems to be more than offset by the improved safety in terms of cardiovascular complications, the need for anesthesia in some children, as well as the longer duration and, most likely, the greater cost of the invasive examination. In addition, several parameters can be optimized to ensure a low radiation dose during pediatric cardiac CT imaging. For example, CT acquisitions could be performed at low tube voltage (17); we ensure that tube current never exceeds 80 kV. In the same way, the tube current is adapted to the patient's weight; electrocardiogram-correlated tube current modulation provides for limiting full tube current to the diastolic phase. Magnetic resonance imaging is another modality that may be useful for follow-up imaging after ASO, particularly because it is not associated with any radiation exposure. However, its relatively low spatial resolution, long scan time, and the potential need for anesthesia are substantial drawbacks as compared with CT (18).
In conclusion, we were able to show the feasibility, safety, accuracy, and mechanistic insights that can be obtained by studying the coronary circulation with 64-slice coronary CT angiography in children aged 5 years to 6 years who require visualization of the coronary arteries after the ASO in early childhood. We consider 64-slice coronary CT angiography as a viable alternative to invasive coronary angiography for the follow-up of patients after the ASO.
- Abbreviations and acronyms
- arterial switch operation
- computed tomography
- transposition of the great arteries
- Received December 3, 2007.
- Revision received February 19, 2008.
- Accepted February 29, 2008.
- American College of Cardiology Foundation
- Sidi D.,
- Planche C.,
- Kachaner J.,
- et al.
- Yacoub M.H.,
- Radley-Smith R.
- Pasquali S.K.,
- Hasselblad V.,
- Li J.S.,
- Kong D.F.,
- Sanders S.P.
- Bonnet D.,
- Bonhoeffer P.,
- Piechaud J.F.,
- et al.
- Bonhoeffer P.,
- Bonnet D.,
- Piechaud J.F.,
- et al.
- Budoff M.J.,
- Achenbach S.,
- Blumenthal R.S.,
- et al.
- Menzel H.G.,
- Schibilla H.,
- Teunen D.
- Shrimpton P.C.,
- Wall B.F.
- ↵Tapiovaara M, Lakkisto M, Servomaa A. A PC-Based Monte Carlo Program for calculating Patient Doses in Medical X-ray Examinations. STUK-A139. Helsinki: Radiation and Nuclear Safety Agency, 1997.