Medical Physics and Informatics Original Research

Transcription

1 Medical Physics and Informatics Original Research Jadhav et al. CT Angiography of Neonates and Infants Medical Physics and Informatics Original Research Siddharth P. Jadhav 1 Farahnaz Golriz 1 Lamya A. Atweh 1 Wei Zhang 1 Rajesh Krishnamurthy 1,2 Jadhav SP, Golriz F, Atweh LA, Zhang W, Krishnamurthy R Keywords: congenital heart disease, CT angiography, neonates and infants, prospective ECG gating, radiation dose, target mode, volumetric scanning DOI: /AJR Received March 14, 2014; accepted after revision June 17, EB Singleton Department of Pediatric Radiology, Texas Children s Hospital, 6621 Fannin St, Houston, TX Address correspondence to R. Krishnamurthy (rxkrishn@texaschildrens.org). 2 Departments of Radiology and Pediatrics, Baylor College of Medicine, Houston, TX. WEB This is a web exclusive article. AJR 2015; 204:W184 W X/15/2042 W184 American Roentgen Ray Society CT Angiography of Neonates and Infants: Comparison of Radiation Dose and Image Quality of Target Mode Prospectively ECG-Gated 320-MDCT and Ungated Helical 64-MDCT OBJECTIVE. The purpose of this study was to evaluate the radiation dose and image quality of target mode prospectively ECG-gated volumetric CT angiography (CTA) performed with a 320-MDCT scanner compared with the radiation dose and image quality of ungated helical CTA performed with a 64-MDCT scanner. MATERIALS AND METHODS. An experience with CTA for cardiovascular indications in neonates and infants 0 6 months old was retrospectively assessed. Radiation doses and quantitative and qualitative image quality scores of 28 CTA examinations performed with a 320-MDCT scanner and volumetric target mode prospective ECG gating plus iterative reconstruction (target mode) were compared with the doses and scores of 28 CTA examinations performed with a 64-MDCT scanner and ungated helical scanning plus filtered back projection reconstruction (ungated mode). All target mode studies were performed during free breathing. Seven ungated CTA examinations (25%) were performed with general endotracheal anesthesia. The findings of 17 preoperative CTA examinations performed in target mode were also compared with surgical reports for evaluation of diagnostic accuracy. RESULTS. All studies performed with target mode technique were diagnostic for the main clinical indication. Effective doses were significantly lower in the target mode group (0.51 ± 0.19 msv) compared with the ungated mode group (4.8 ± 1.4 msv) (p < ). Quantitative analysis revealed no statistically significant difference between the two groups with respect to signal-to-noise ratio (of pulmonary artery and aorta) and contrast-to-noise ratio. Subjective image quality was significantly better with target mode than with ungated mode (p < ). CONCLUSION. Target mode prospectively ECG-gated volumetric scanning with iterative reconstruction performed with a 320-MDCT scanner has several benefits in cardiovascular imaging of neonates and infants, including low radiation dose, improved image quality, high diagnostic accuracy, and ability to perform free-breathing studies. T echnical advances have led to increased use of CT for the evaluation of children with congenital heart disease (CHD). Because of its noninvasive nature, CT angiography (CTA) is considered an attractive alternative to cardiac catheterization for the evaluation of CHD [1]. However, despite the undisputed benefits of CT, it has come under increasing scrutiny because of high radiation exposure. Children with CHD often undergo multiple CT scans, which results in a high cumulative radiation dose. In addition, pediatric patients are particularly sensitive to radiation, and they have a longer life span to manifest radiation-induced stochastic effects. Unfortunately, most of the approaches to lowering CT radiation dose lead to the undesirable consequence of increased image noise. Therefore, it is impor- tant to explore methods that achieve a balance between radiation dose and image quality. Previous studies [1 4] have supported using low voltage settings for pediatric contrast-enhanced CT. Newer image reconstruction techniques based on iterative reconstruction have also shown promise for reducing radiation dose in adult and pediatric CT [5 8]. Scanning duration and temporal resolution are critical parameters to radiologists imaging neonates and infants, who have high heart rates and cannot cooperate with a breath-hold or lie still without sedation and general endotracheal anesthesia. Fast acquisition is crucial to obtaining motion-free images without sedation. With the introduction of 320-MDCT scanners, volumetric scanning of a 16-cm ( mm) z-axis length during a single gantry rotation ( ms) is possible. With W184 AJR:204, February 2015

2 CT Angiography of Neonates and Infants A use of half-scan reconstruction, the temporal resolution of a 320-MDCT scanner can be improved to ms [9]. The long z-axis coverage and high temporal resolution allow scanning of the entire chest in children within one cardiac cycle. The short scan duration freezes respiratory motion and makes CTA feasible during free breathing. Moreover, volume scanning of the entire heart in a single cardiac cycle produces temporally uniform images with homogeneous contrast enhancement. Volume scanning with a widedetector scanner also has the advantage of reducing radiation dose compared with helical scanning owing to lack of low-pitch overscanning and minimal overbeaming [10]. ECG-synchronized CTA is more accurate than ungated techniques for the assessment of coronary arteries and intracardiac structures [11]. The major drawback of retrospectively ECG-gated CTA compared with ungated CTA is increased radiation dose. Prospective ECG gating offers an advantage over retrospective ECG gating by reducing the radiation dose. In prospective ECG gating, data acquisition occurs during a lowmotion cardiac phase, which corresponds to mid- to end-diastole in patients with low heart rates and to end-systole in patients with high heart rates [12, 13]. One limitation of prospective data acquisition with a 64-MDCT scanner is image artifacts due to heart rate variability and ectopic beats [14]. With substantial variability in heart rates, accurate prospective centering of the image window within the low-motion phase is not possible. In addition, heart rate variability forces data acquisition during different cardiac phases from one cardiac cycle to another, leading to stair-step artifact. Temporal padding overcomes this limitation by extending the acquisition before and after the determined acquisition window. This technique allows retrospective modification of the reconstruction window to ensure image reconstruction within the low-motion period and in identical cardiac phases from one cardiac cycle to another [15]. However, temporal padding increases the radiation dose to the patient. Acquisition of a temporally uniform volumetric dataset with a 320-MDCT scanner allows the use of techniques such as PhaseXact and ImageXact (Toshiba Medical Systems). These postacquisition data-processing methods perform half-scan reconstruction from the best motion-free cardiac phase within the prospectively triggered raw data that spans the rotation time of milliseconds (Fig. 1). This is referred to as target mode prospectively ECG-gated acquisition because the optimal image is acquired without padding. This technique maximizes image quality by allowing pulsation-related motion control without increasing the radiation burden to the patient. The purpose of this study was to assess our experience with the evaluation of CHD in neonates and infants by use of free-breathing target mode prospectively ECG-gated volumetric CTA performed with a 320- MDCT scanner and iterative reconstruction. We also retrospectively compared radiation dose and image quality in this group with those in patients who underwent ungated helical CTA with a 64-MDCT scanner and filtered back projection (FBP) reconstruction. Materials and Methods This study was approved by the institutional review board and was HIPAA compliant. The requirement for written informed consent was waived owing to the retrospective nature of the study. Patients and CT Techniques The study population consisted of neonates and infants 0 6 months old who underwent CTA for B known or suspected congenital cardiovascular disease or postoperative follow-up. Target mode group Between September 2011 and October 2013, 29 volumetric CTA examinations with target mode prospective ECG gating were performed on 27 patients under the supervision of one attending cardiac radiologist at our hospital. All examinations were performed with a 320-MDCT scanner (Aquilion ONE, Toshiba Medical Systems) with a gantry rotation time of 350 milliseconds. All CTA images were obtained without a breath-hold and without sedation for the study. One neonate who was already sedated in the cardiovascular ICU was included in the study. Tube voltage was set at 80 kv and tube current at ma according to patient age and clinical indication. A first-pass contrast enhancement technique was used to maximize contrast in the vasculature of interest. The nonionic contrast agent ioversol (320 mg I/mL, Optiray, Mallinckrodt Imaging) was injected through a 22-gauge or smaller catheter with a dual-syringe injector at a dose of ml/kg. The injection rate was ml/s depending on catheter size, patient size, and the structure of interest. Contrast medium was injected via a lower-extremity peripheral catheter in 27 examinations and via an umbilical catheter in one patient. One patient with contrast injection via a peripherally inserted central catheter was excluded from the study owing to slow speed of injection, nonoptimal contrast enhancement, and subsequent poor image quality. A total of 28 examinations of 26 patients were included in the study. All images were obtained during one cardiac cycle and during the first pass of contrast medium through the anatomic structures of interest as determined with a bolus-tracking technique. The targets for the center of the acquisition window were set at 40% of the R-R interval in patients with heart rates of 90 beats/min and higher and 75% of the R-R interval for those with heart rates lower than 90 beats/min. Data were reconstructed Fig. 1 2-day-old girl with tetralogy of Fallot. Example of use of ImageXact reconstruction technique (Toshiba Medical Systems). A, CT image obtained at 40% of cardiac cycle shows right coronary artery (RCA) apparently arising from main pulmonary artery (MPA) and crossing ventrally to right ventricular outflow tract. With ImageXact technique, several phases of cardiac cycle from 18% to 58% were reconstructed from this slice. B, CT image reconstructed at 28% of cardiac cycle shows RCA arising from left anterior descending coronary artery (LAD) and crossing right ventricular outflow tract ventrally. AJR:204, February 2015 W185

3 Jadhav et al. with an iterative reconstruction algorithm (AIDR, Toshiba Medical Systems) in 0.5-mm-thick slices. The best motion-free cardiac phase adjacent to the predefined target phase was individually selected from the raw data for each patient by the attending radiologist using ImageXact software, which allows half-scan reconstruction of data for a selected slice across the entire spectrum of cardiac phases spanned by the rotation time. Ungated mode group A comparison study group of 28 patients who underwent ungated helical CTA was assembled through review of consecutive cardiac CTA examinations of neonates and infants. Studies were selected in reverse chronologic order, beginning with scans obtained in February 2010 and concluding with scans obtained in February All ungated CT examinations were performed with a 64-MDCT scanner (Light- Speed VCT XT, GE Healthcare). Seven studies were performed with general endotracheal anesthesia, and 21 studies without general anesthesia (with or without cardiovascular sedation). Ioversol 320 mg I/mL (Optiray) was injected at a dose of ml/kg through a peripheral vein catheter (upper or lower extremity) in 27 patients and an umbilical catheter in one patient. Images were obtained in the early equilibrium phase. Scanning was performed in a craniocaudal direction. The following acquisition parameters were used: collimation, mm; pitch, 0.98; slice thickness, mm; tube voltage, kv; tube current, ma according to patient age and weight. All images were reconstructed with FBP technique. In both groups, the scan range depended on the area of interest and patient size (Table 1). TABLE 1: Comparison of Patient Characteristics, Contrast Volume, Scanning Parameters, and Radiation Dose Between Target Mode and Ungated Mode Acquisitions Characteristic Target Mode (n = 28) Ungated Mode (n = 28) p Sex Girls 12 (43) 12 (43) Boys 16 (57) 16 (57) Age (d) 40.8 ± 46.7 (1 184) 14 ± 33.2 (1 182) 0.01 Weight (g) 3600 ± 1492 ( ) 2550 ± 923 ( ) Contrast volume (ml/kg) 2.26 ± 0.21 ( ) 2.21 ± 0.20 ( ) 0.5 Tube voltage (kv) ± 11.3 (80 120) < Tube current (ma) 66.4 ± 18.3 (40 100) 214 ± 87.1 (50 350) < Tube current time setting (mas) 26.6 ± 7.3 (16 40) ± 43 (25 202) < Scan length (cm) 8.7 ± 1.7 (6 12) 9.1 ± 1.8 (6 14) 0.37 Volume CT dose index (mgy) 1.5 ± 0.38 ( ) 13.9 ± 5.1 (4.6 25) < Dose-length product (mgy cm) 13.6 ± 5 (6 21.8) ± 37.3 (46 180) < Effective dose (msv) 0.51 ± 0.19 ( ) 4.8 ± 1.4 (1.7 7) < Note Except for sex (no. of patients with percentage in parentheses), values are mean ± SD with range in parentheses. Image Processing Axial images for all cardiac CT examinations were transferred to a postprocessing workstation (Vitrea, Toshiba America). Various image reformatting techniques, including multiplanar reformation, curved planar reformation, thin-slab maximum intensity projection, and volume rendering, were used depending on the structure of interest (Fig. 2). Radiation Dose Calculation Target mode group The volume CT dose index (CTDI vol ) and dose-length product (DLP) displayed on the CT console were recorded for each CTA examination. The CTDI vol and DLP were obtained for a 32-cm phantom. The CTDI vol and DLP based on the 32-cm phantom were multiplied by a factor of 2.3 to obtain the radiation doses based on a 16-cm phantom. The ratios reported in the literature for converting CTDI vol and DLP based on 32-cm phantom size to 16-cm phantom size range TABLE 2: Objective Image Quality Assessment Between Target Mode and Ungated Mode Acquisitions Parameter Target Mode (n = 28) Ungated Mode (n = 28) p Attenuation in aorta (HU) 720 ± ( ) 429 ± ( ) < Noise in aorta (HU) 24 ± 5.6 (13 36) 11.6 ± 3.7 (6 22) < Signal-to-noise ratio in aorta 31.4 ± 10.4 (13 58) 37.1 ± 15.6 (10 69) 0.07 Attenuation in pulmonary artery (HU) 695 ± ( ) 426 ± ( ) < Noise in pulmonary artery (HU) 22.1 ± 4.8 (16 41) 12 ± 3.6 (7 22) < Signal-to-noise ratio in pulmonary artery 31.6 ± 10.8 (13 52) 36 ± 15.4 (15 67) 0.12 Contrast (HU) 680 ± ( ) 372 ± ( ) < Mean noise a (HU) 22.6 ± 4.1 (16 33) 11.5 ± 2.7 (6 16) < Contrast-to-noise ratio 30 ± 11.5 (10 57) 34.2 ± 13.2 (12 61) 0.26 Note Values are mean ± SD with ranges in parentheses. a Mean of noise in aorta, pulmonary artery, and pectoral muscle. from 2 to 2.5 for different scanners [16 18]. Using scanner manufacturer recommendations, we plugged in the same scanning parameters (tube voltage, tube current, collimation) on the scanner console initially for a body scan (32-cm phantom) and then for a head scan (16-cm phantom), which yielded a conversion factor of 2.3 for our scanner. Ungated mode group The CTDI vol and DLP were calculated with CT-Expo software (version 2.1, Stamm and Nagel) on the basis of scanning parameters for a 16-cm phantom. For both groups, the effective dose was calculated by multiplying the DLP by the conversion factor of msv/ mgy cm for patients younger than 4 months and msv/mgy cm for infants 4 6 months old [19] (Table 1). Image Quality Quantitative analysis To obtain objective indexes of image quality, we determined the attenuation, noise, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) for the two scanning techniques (Table 2). Uniform circular ROIs with areas of 5 mm 2 were drawn by a radiologist (who was not an image reviewer) in three anatomic regions: the descending aorta (at the level of the main pulmonary artery), main pulmonary artery (left pulmonary artery in patients with main pulmonary artery atresia), and left pectoral muscle. Attenuation was defined as CT number in HU and noise as the SD of CT number in HU in the ROI. Contrast was defined as the difference between the attenuation of either descending aorta or pulmonary artery (depending on the target structure) and W186 AJR:204, February 2015

4 CT Angiography of Neonates and Infants A D G Fig. 2 3-day-old girl with heterotaxy, asplenia, and total anomalous pulmonary venous return (TAPVR). A H, Axial (A F), coronal maximum-intensity-projection (G), and posterior volume-rendered projection (H) images show TAPVR to left innominate vein (LIV) via meandering vein (purple arrow) within left lung without evidence of obstruction of pulmonary veins along their course. Additional findings include single coronary artery arising from posterior sinus of aorta, left coronary artery (blue arrows) taking posterior and inferior course in relation to proximal main pulmonary artery, and moderate tracheomalacia (A and B). Patient also had double outlet right ventricle with pulmonary stenosis (not shown), unobstructed aorta (Ao), and ductus arteriosus (white arrow) between right innominate artery and right pulmonary artery. Dose-length product for entire study was 21.2 mgy cm, which corresponds to approximately 0.8 msv. Use of target mode prospective gating allows motion-free assessment of wide array of cardiac, coronary, vascular, airway, and pulmonary parenchymal pathologic conditions. SVC = superior vena cava, RV = right ventricle, LV = left ventricle. B E H C F AJR:204, February 2015 W187

5 Jadhav et al. the attenuation of pectoral muscle. In both groups, the pulmonary artery was the most used structure (17 patients in the target mode group, 16 patients in the ungated mode group) for measurement of contrast. SNR and CNR were calculated as follows: SNR A = attenuation A / noise A SNR PA = attenuation PA / noise PA CNR = contrast / mean noise, where mean noise is the mean of noise in the descending aorta (A), pulmonary artery (PA), and pectoral muscle. In the target mode group, all measurements were performed at 0.5-mm slice thickness. In the ungated group, the measurements were performed at or 1.25-mm slice thickness (average, 0.88 mm). Qualitative analysis Each dataset was independently reviewed by two cardiac radiologists. CTA images obtained with the target mode and the ungated mode were randomly presented during each reading session. In each group, the extracardiac vessels (pulmonary vessels), cardiac valves (aortic valves), intracardiac structures, coronary arteries, homogeneity of contrast, and presence of beam hardening were evaluated on a semiquantitative scale (0 1). With the scoring system, 17 parameters were evaluated for each CTA examination (Table 3). Diagnostic Accuracy In the target mode group, the preoperative CTA findings for the main indication of study and the associated findings were compared with the surgical findings (Table 4). Statistical Analysis All statistical calculations were performed with SAS software (version 9.3, SAS Institute). Continuous variables were presented as mean ± SD, and categoric variables were presented as frequencies or percentages. To compare continuous variables between two groups, the Wilcoxon rank sum test was used, and to compare categoric variables between two groups, the Fisher exact test was used. Interobserver agreement for image quality scoring was assessed with Cohen kappa statistics [20] and interpreted according to the Landis and Koch guideline [21]. All analyses were based on two-sided tests, and p < 0.05 was considered statistically significant. Results Patient Population and Scan Parameters Patients in the target mode and ungated mode groups did not differ in terms of sex. The mean age and body weight were higher in the target mode than in the ungated mode group, but the age ranges were similar (Table 1). The distribution of clinical indications TABLE 3: Subjective Image Quality Assessment Between Target Mode and Ungated Mode Acquisitions (Reviewer 1) No. of Patients With Score 1 Target Ungated Structure of Interest Parameter Mode Mode p Left pulmonary artery and vein Edge definition a First to third branches b Third-order branches b Homogeneity of contrast c Arteriovenous separation d Aortic valve Annulus definition a No. of leaflets b 22 9 < Suspension apparatus b Ventricles Endocardial margin definition a Papillary muscles and trabecula b Atria Wall definition a Atrial septum b Coronary arteries Origin of right coronary artery b First order branch of LAD b Second order branch of LAD b Beam hardening Beam hardening in right atrium e Beam hardening in superior vena cava e Note LAD = left anterior descending coronary artery. Each parameter was scored on the following semiqualitative scale: a 1, good sharpness or minimal blurring due to minor motion artifacts or noise; 0, very poor margin sharpness due to substantial motion artifacts or noise. b 1, clear visualization without any motion artifact or minor motion artifacts or noise; 0, identified but equivocal or no visualization due to substantial motion artifacts or noise. c 1, homogeneous contrast; 0, nonhomogeneous contrast. d 1, arteriovenous separation with high accuracy; 0, arteriovenous separation not possible. e 1, absence of beam hardening; 0, presence of beam hardening. was comparable between the two groups. In the target mode group, it included evaluation of pulmonary artery and aortopulmonary collaterals caused by pulmonary atresia or stenosis (n = 11), truncus arteriosus (n = 1), assessment of pulmonary veins (n = 6), coarctation of the aorta (n = 3), evaluation of the patency of Blalock-Taussig shunt or right ventricle pulmonary artery conduit (n = 5), and exclusion of vascular-mediated airway compromise (n = 2). In the ungated group, the indications for the study were as follows: assessment of the status of pulmonary artery and aortopulmonary collaterals due to pulmonary artery atresia or stenosis (n = 13), assessment of pulmonary veins (n = 6), coarctation of the aorta (n = 4), evaluation of Blalock-Taussig shunt patency (n = 3), and exclusion of vascular ring (n = 2). The mean heart rate in the target mode group was 147 ± 12.7 beats/min (range, beats/min). Because the heart rates of all patients in the target mode group were greater than 110 beats/min, the center of the acquisition window was 40% of the R-R interval for all of the examinations. The average volume of contrast medium injected per patient in the target mode group was 8.8 ± 3.5 ml (range, 4 18 ml), which was significantly greater than the average volume injected in the ungated mode group (6.1 ± 2 ml; range, 2 10 ml) (p < ). However, the doses of contrast medium injected per kilogram of body weight were not significantly different between the two groups (p = 0.5) (Table 1). Radiation Dose The mean z-axis coverage length in the target mode group was not significantly different from the mean z-axis coverage length in the ungated mode group (p = 0.37) (Ta- W188 AJR:204, February 2015

6 CT Angiography of Neonates and Infants TABLE 4: Age, Radiation Dose, Imaging Findings, and Surgical Findings in Target Mode Group Patient No. Age (d) Effective Dose (msv) ble 1). The CTDI vol, DLP, and effective doses were significantly lower in the target mode group than in the ungated mode group (p < ) (Table 1). The mean patient effective dose for the target mode group was 0.51 ± 0.19 msv (range, msv); for the ungated mode group, this dose was 4.8 ± 1.4 msv (range, msv) (p < ). Image Quality Quantitative analysis Both image noise and contrast were significantly higher in the Surgical and Imaging Findings 1 a PA, RVOT narrowing, MAPCAs (n = 3), VSD 2 a PA, ductal dependent flow, VSD, overriding aorta, RAA, mirror-image branching, double SVC Patient 2 after BT shunt, patent BT shunt 4 a PA, ductal dependent flow, DORV, ASD(s), VSD, d-tga, RAA, mirror-image branching, LAD originating from RCA 5 a PA, overriding aorta, MAPCAs (n = 2), VSD, RAA, mirror-image branching 6 a Heterotaxy, PA, MAPCAs (n = 2), ASD(s), VSD, RDCAVC, double SVC 7 a PS, RVOT narrowing, overriding aorta, ASD(s) b, VSD, aberrant RSA, double SVC b 8 a Heterotaxy, PA, ductal dependent flow, ASD(s, p), VSD, RDCAVC, d-tga, TAPVR Patient 8 after BT shunt, patent BT shunt 10 a Heterotaxy, valvular PS, DORV, ASD(s, p), VSD, RDCAVC, d-tga, TAPVR, double SVC, single coronary artery 11 a CoA, ASD(s), VSD 12 a CoA, bilateral stenosis of subclavian arteries 13 a Heterotaxy, PA, ductal dependent flow, ASD(s, p), VSD, RDCAVC, d-tga, RAA, mirror-image branching, double SVC 14 a TA (type 1), ASD(s), VSD, RAA, mirror-image branching, stenosis of RCA origin b 15 a Hypoplasia of aortic arch and CoA, PDA, enlarged MPA, VSD b 16 a PDA, mild stenosis of LPA origin 17 a Heterotaxy, PA, ductal dependent flow, single ventricle, ASD(s, p), RDCAVC, TAPVR, RAA, double SVC 18 a RVOT narrowing, ASD(s), l-tga, TAPVR, RAA PA, ductal dependent flow, overriding aorta, ASD(s), VSD TA after RV-PA conduit, unobstructed RV-PA conduit TAPVR after repair, mild narrowing of common pulmonary anastomosis to atrium HLHS after Norwood procedure, unobstructed RV-PA conduit HLHS after Norwood procedure, patent BT shunt, moderate stenosis of LPA origin DORV after repair, aberrant RSA, no vascular mediated airway compromise PS, RVOT narrowing, DORV, VSD, aftermath of BT shunt, patent BT shunt Pulmonary valve stenosis, no vascular mediated airway compromise Severe dilatation of MPA and PA branches, no pulmonary vein stenosis PS, RVOT narrowing, overriding aorta, VSD, RAA, aberrant LSA, anomalous origin of LCA from right sinus Note PA = pulmonary artery atresia, RVOT = right ventricular outflow tract, MAPCAs = major aortopulmonary collateral arteries, VSD = ventricular septal defect, RAA = right aortic arch, SVC = superior vena cava, BT = Blalock-Taussig, DORV = double outlet right ventricle, ASD(s) = atrial septal defect (secundum), d-tga = dextrorotated transposition of great arteries, LAD = left anterior descending coronary artery, RCA = right coronary artery, RDCAVC = right-dominant common atrioventricular canal, PS = pulmonary artery stenosis, RSA = right subclavian artery, ASD(p) = ASD (primum), TAPVR = total anomalous pulmonary venous return, CoA = coarctation of aorta, TA = truncus arteriosus, PDA = patent ductus arteriosus, MPA = main pulmonary artery, LPA = left pulmonary artery, l-tga = levorotated TGA, RV-PA = right ventricle pulmonary artery, HLHS = hypoplastic left heart syndrome, LSA = left subclavian artery, LCA = left coronary artery. a Preoperative CT angiographic examination with findings confirmed at surgery. b Finding missed at CT angiography. target mode group than in the ungated mode group. The SNR (in aorta and pulmonary artery) and CNR tended to be lower in the target mode group, although the difference between the two groups was not statistically significant (Table 2). Qualitative analysis Agreement between the reviewers on image quality scores was excellent (target mode, κ = 0.85; ungated, κ = 0.82). The average score by reviewer 1 for the target mode group was 13.9 ± 2.5 (range, 9 17) and for the ungated group was 9.5 ± 3.6 (range, 4 16) (p < ). The average score by reviewer 2 was 12.7 ± 2.3 (range, 8 17) for the target mode group and 8.5 ± 3.5 (range, 3 15) for the ungated group (p < ). Table 3 summarizes the scores given by reviewer 1 for each of the 17 evaluated parameters in target mode and ungated studies. Compared with ungated CTA, target mode CTA was significantly better for assessment of intracardiac structures and coronary vessels. In addition, homogeneity of contrast and arteriovenous separation were better, AJR:204, February 2015 W189

7 Jadhav et al. and beam hardening was less in the target mode group, although the differences were not statistically significant. Comparison of Target Mode CT Angiographic Findings With Surgical Findings None of the target mode group patients needed diagnostic catheterization. CTA was performed after palliative or corrective cardiac surgery in eight patients, and 20 patients had no previous surgery. Seventeen of these 20 patients underwent surgery, and three died before any intervention was performed. In the 17 patients who underwent surgery, CTA was 100% accurate for the main indication of the study. CTA was also accurate in the diagnosis of all associated extracardiac defects except one left superior vena cava due to nonopacification related to the firstpass technique. Regarding coexistent intracardiac defects, a muscular ventricular septal defect (owing to nasogastric tube artifact) and a secundum atrial septal defect (owing to beam-hardening artifact) were missed at CTA. Both of them were detected at echocardiography and were not hemodynamically significant. Three patients had congenital anomalies of the coronary artery. Two coronary anomalies left anterior descending coronary artery originating from the right coronary artery and single coronary artery were accurately diagnosed with target mode CTA. One patient had stenosis of the origin of the right coronary artery, which was not detected with CTA (Table 4). Discussion A worldwide effort is being made to minimize radiation exposure of children during diagnostic imaging. Several studies have shown higher risk of cancer among children undergoing CT than among adults [22 25]. A retrospective cohort study of 180,000 patients who underwent CT before 22 years of age in the United Kingdom [24] showed excess risk of leukemia and brain tumors in the first decade after CT. According to the results of that study, cumulative radiation doses of 50 mgy can almost triple the risk of leukemia, and radiation doses of 60 mgy can almost triple the risk of brain cancer. Anesthesia during childhood has also gained attention due to its potential effect on neurodevelopmental outcome. There is increasing evidence that multiple exposures to anesthesia are related to neurodevelopmental disability in children [26 28]. Ing et al. [29] reported in 2012 that even a single exposure to anesthesia for surgery or diagnostic testing of a child younger than 3 years is associated with long-term deficits in receptive and expressive language and cognition. In our study, a number of tools were used in the target mode group to decrease radiation dose and improve image quality while allowing unsedated, free-breathing acquisition in children with high heart rates. These included volumetric acquisition, half-scan reconstruction, target mode of prospective ECG gating, low tube current and tube voltage, first-pass contrast enhancement, and iterative reconstruction. Although tube voltage and tube current were significantly lower in the target mode group than in the ungated mode group, no significant difference between the two groups was found with respect to SNR and CNR. This finding suggests that first-pass contrast enhancement and iterative reconstruction in the target mode group played important parts in improving contrast and decreasing noise. The subjective image quality was significantly better in the target mode group owing to decreased cardiac pulsation and respiratory motion artifact, decreased beam-hardening artifact, and uniform contrast enhancement. To our knowledge, there has been no report of the use of target mode prospective ECG gating combined with iterative reconstruction for the evaluation of CHD in children. Al-Mousily et al. [30] conducted a study that included eight children with CHD (mean age, 8.1 months; range, months) undergoing 320-MDCT with prospective ECG gating and FBP that led to a radiation dose of 0.8 ± 0.39 msv (range, msv). A more recent study of 22 children with CHD (mean age, 18 months; range, 14 days 9 years) who underwent prospectively ECG-gated 320-MDCT with FBP technique [31] showed a radiation dose of 0.42 ± 0.08 msv (range, msv). Of note in these studies is the size of the phantom size used to determine radiation dose. In both of these studies, the DLP for body surface area was given on a 32-cm phantom; this DLP is 40 50% of the DLP value based on a 16-cm phantom for the same acquisition. Based on the 32-cm phantom, the average effective dose in our study would be 0.22 ± 0.09 msv. The combination of target mode prospective ECG gating, first-pass technique, and iterative reconstruction allowed the use of lower tube current in this study ( ma), which led to a further dose reduction while maintaining image quality. The target mode CTA technique also has the advantage of avoiding the need for sedation. To our knowledge, there has been only one report of nonsedated, free-breathing CTA for the evaluation of CHD in neonates and infants. In a study by Han et al. [32], 19 neonates underwent nonsedated, free-breathing dual-source CTA with prospective ECG gating in high-pitch helical mode with a 128- MDCT scanner and had a median effective dose of 0.86 msv (range, msv). However, the application of high-pitch helical mode requires a regular heart rate. In patients with marked heart rate variability, which is common in neonates and infants, the starting phase of scanning does not precisely match the preselected starting phase. This reduces image quality and causes phase-related artifacts across the scanned volume, particularly affecting the coronary arteries and intracardiac structures. This is less of a problem in evaluation of extracardiac vessels such as the aorta and pulmonary arteries [18, 33]. This problem could be avoided by use of volumetric scanning of the entire heart in the same phase of the cardiac cycle and postacquisition half-scan reconstruction of the best motion-free cardiac phase. Target mode prospectively ECG-gated CTA with first-pass technique and iterative reconstruction had diagnostic image quality for all target structures and depicted coexistent extracardiac and intracardiac defects. The combination of target mode prospectively ECG-gated CTA and echocardiography was diagnostic for the assessment of intracardiac defects in all patients. This ability to perform sufficiently high-quality diagnostic imaging of CHD at low radiation exposures without the need for sedation may tilt the scales in favor of CT over MRI. In our current practice, this protocol did not ensure diagnostic confidence for the assessment of coronary anomalies in nonsedated freebreathing neonates and infants with high heart rates. Therefore, we resort to retrospective cardiac gating when assessment of coronary arteries is requested. This study had limitations. First, the radiation dose was not obtained directly from the 64-MDCT scanner but was calculated with software on the basis of the scanning parameters. A further limitation was that image noise was measured with different slice thicknesses between the groups, which might have caused image noise overestimation in the target mode group compared with the ungated group [34]. However, the pur- W190 AJR:204, February 2015

8 CT Angiography of Neonates and Infants pose of this study was not to show the superiority of target mode prospective ECG gating over ungated scanning but to assess whether target mode prospectively ECG-gated volumetric scanning with iterative reconstruction provides sufficient image quality at significantly low radiation doses. Conclusion Target mode prospectively ECG-gated volumetric first-pass technique with iterative reconstruction has several advantages in CTA for the assessment of CHD in neonates and infants. It improves image quality compared with that of ungated acquisition, reduces the radiation dose, and obviates breath-holding and sedation. References 1. Lee T, Tsai IC, Fu YC, et al. Using multidetectorrow CT in neonates with complex congenital heart disease to replace diagnostic cardiac catheterization for anatomical investigation: initial experiences in technical and clinical feasibility. Pediatr Radiol 2006; 36: Suess C, Chen X. Dose optimization in pediatric CT: current technology and future innovations. Pediatr Radiol 2002; 32: McCollough CH, Primak AN, Braun N, Kofler J, Yu L, Christner J. Strategies for reducing radiation dose in CT. Radiol Clin North Am 2009; 47: Nievelstein RA, van Dam IM, van der Molen AJ. Multidetector CT in children: current concepts and dose reduction strategies. Pediatr Radiol 2010; 40: Tomizawa N, Nojo T, Akahane M, Torigoe R, Kiryu S, Ohtomo K. Adaptive iterative dose reduction in coronary CT angiography using 320- row CT: assessment of radiation dose reduction and image quality. J Cardiovasc Comput Tomogr 2012; 6: Schuhbaeck A, Achenbach S, Layritz C. Image quality of ultra-low radiation exposure coronary CT angiography with an effective dose <0.1 msv using highpitch spiral acquisition and raw data-based iterative reconstruction. Eur Radiol 2013; 23: Tricarico F, Hlavacek AM, Schoepf UJ. Cardiovascular CT angiography in neonates and children: image quality and potential for radiation dose reduction with iterative image reconstruction techniques. Eur Radiol 2013; 23: Zheng M, Zhao H, Xu J. Image quality of ultralow-dose dual-source CT angiography using highpitch spiral acquisition and iterative reconstruction in young children with congenital heart disease. J Cardiovasc Comput Tomogr 2013; 7: Tomizawa N, Maeda E, Akahane M, Torigoe R, Kiryu S, Ohtomo K. Coronary CT angiography using the second-generation 320-detector row CT: assessment of image quality and radiation dose in various heart rates compared with the first-generation scanner. Int J Cardiovasc Imaging 2013; 29: Sorantin E, Riccabona M, Stücklschweiger G, Guss H, Fotter R. Experience with volumetric (320 rows) pediatric CT. Eur J Radiol 2013; 82: Goo HW. State-of-the-art CT imaging techniques for congenital heart disease. Korean J Radiol 2010; 11: Hoffmann MH, Shi H, Manzke R, et al. Noninvasive coronary angiography with 16-detector row CT: effect of heart rate. Radiology 2005; 234: Herzog C, Arning-Erb M, Zangos S, et al. Multidetector row CT coronary angiography: influence of reconstruction technique and heart rate on image quality. Radiology 2006; 238: Leschka S, Wildermuth S, Boehm T, et al. Noninvasive coronary angiography with 64-section CT: effect of average heart rate and heart rate variability on image quality. Radiology 2006; 241: Earls JP. How to use a prospective gated technique for cardiac CT. J Cardiovasc Comput Tomogr 2009; 3: Paul JF, Rohnean A, Elfassy E, Sigal-Cinqualbre A. Radiation dose for thoracic and coronary stepand-shoot CT using a 128-slice dual source machine in infants and small children with congenital heart disease. Pediatr Radiol 2011; 41: Pache G, Grohmann J, Bulla S, et al. Prospective electrocardiography-triggered CT angiography of the great thoracic vessels in infants and toddlers with congenital heart disease: feasibility and image quality. Eur J Radiol 2011; 80:e440 e Nie P, Wang X, Cheng Z, Ji X, Duan Y, Chen J. Accuracy, image quality and radiation dose comparison of high-pitch spiral and sequential acquisition on 128-slice dual-source CT angiography in children with congenital heart disease. Eur Radiol 2012; 22: Thomas KE, Wang B. Age-specific effective doses for pediatric MSCT examinations at a large children s hospital using DLP conversion coefficients: a simple estimation method. Pediatr Radiol 2008; 38: Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas 1960; 20: Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33: BEIR. Health risks from exposure to low levels of ionizing radiation: BEIR VII. Washington, DC: National Academies Press, Andreassi MG. Radiation risk from pediatric cardiac catheterization: friendly fire on children with congenital heart disease. Circulation 2009; 120: Pearce MS, Salotti JA, Little MP, et al. Radiation exposure from CT scans in childhood and subsequent risk of leukaemia and brain tumours: a retrospective cohort study. Lancet 2012; 380: Mathews JD, Forsythe AV, Brady Z, et al. Cancer risk in 680,000 people exposed to computed tomography scans in childhood or adolescence: data linkage study of 11 million Australians. BMJ 2013; 346:f Flick RP, Katusic SK, Colligan RC, et al. Cognitive and behavioral outcomes after early exposure to anesthesia and surgery. Pediatrics 2011; 128:e1053 e Hays SR, Deshpande JK. Newly postulated neurodevelopmental risks of pediatric anesthesia. Curr Neurol Neurosci Rep 2011; 11: Wilder RT, Flick RP, Sprung J. Early exposure to anesthesia and learning disabilities in a population-based birth cohort. Anesthesiology 2009; 110: Ing C, DiMaggio C, Whitehouse A, et al. Longterm differences in language and cognitive function after childhood exposure to anesthesia. Pediatrics 2012; 130:e476 e Al-Mousily F, Shifrin RY, Fricker FJ, Feranec N, Quinn NS, Chandran A. Use of 320-detector computed tomographic angiography for infants and young children with congenital heart disease. Pediatr Cardiol 2011; 32: Zhang T, Wang W, Luo Z, et al. Initial experience on the application of 320-row CT angiography with low-dose prospective ECG-triggered in children with congenital heart disease. Int J Cardiovasc Imaging 2012; 28: Han BK, Overman DM, Grant K, et al. Non-sedated, free breathing cardiac CT for evaluation of complex congenital heart disease in neonates. J Cardiovasc Comput Tomogr 2013; 7: Hetterich H, Wirth S, Johnson TR, Bamberg F. Highpitch dual spiral cardiovascular computed tomography. Curr Cardiovasc Imaging Rep 2013; 6: O Connor OJ, Vandeleur M, McGarrigle AM. Development of low-dose protocols for thin-section CT assessment of cystic fibrosis in pediatric patients. Radiology 2010; 257: AJR:204, February 2015 W191