Pediatric Cardiac CT

Abstract

Technological developments have significantly advanced the role of CT for noninvasive imaging of the cardiovascular system in children. This chapter provides an overview of the technical considerations that are essential for performing high quality pediatric cardiac CT using the lowest possible radiation dose. The clinical utility and application of CT compared with MRI is discussed for a wide range of congenital and acquired pediatric cardiovascular diseases involving the systemic and pulmonary vasculature, coronary arteries, heart chamber morphology and function, and thoracic airways. Postoperative considerations following repair of congenital heart disease are also addressed.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

Recent technical developments in CT imaging have significantly advanced the role of multidetector CT (MDCT) in noninvasive imaging of the cardiovascular system. The number of detector rows has increased from 2 to 320 and the gantry rotation time has decreased down to 270 ms. These two technical improvements result in shorter scan time, longer z-axis coverage, and less motion artifacts on pediatric cardiac CT, and enhance the its diagnostic value in evaluating extracardiac vessels, lungs, and airways in children with congenital heart disease (CHD) (Westra et al. 1999; Kim et al. 2002; Goo et al. 2003, 2005a). Electrocardiography (ECG)-synchronized scan acquisitions are almost always required for evaluating intracardiac structures, ventricular function, and coronary arteries (Goo et al. 2005b, 2010b; Tsai et al. 2007; Goo 2010a, 2011a). New CT applications for children with CHD include dual-energy lung perfusion (Goo 2010b) and four-dimensional (4D) airway (Greenberg 2012) assessments. In this chapter, we will focus on the technical considerations essential for pediatric cardiac CT, discuss the relative merits and demerits of cardiac CT and cardiac MRI, and describe clinical applications of pediatric cardiac CT in children with CHD.

2 Technical Considerations

2.1 Spatial and Temporal Resolution

Thanks to the improved longitudinal (z-axis) spatial resolution (0.5–0.625 mm) of modern MDCT systems, isotropic spatial resolution vastly improving the quality of multiplanar reformatted and three-dimensional (3D) CT images can be achieved (Mahesh 2002). Pediatric cardiac CT particularly demands high temporal resolution because children may not cooperate with a breathing instruction and a child’s heart rate is usually high. In this respect, a dual-source scanner is quite helpful to decrease motion artifacts on pediatric cardiac CT by increasing temporal resolution. Only a quarter of gantry rotation (70–83 msec) is necessary for cardiac CT using dual-source scanning and half-scan reconstruction techniques (Petersilka et al. 2008). Recently introduced high-pitch (up to 3.0–3.4) dual-source helical scanning with or without ECG triggering allows substantial reduction of motion artifacts on pediatric cardiac CT (Han et al. 2011; Nie et al. 2012). The benefit of higher temporal resolution of a dual-source CT system has also been proven by improved coronary artery visibility on prospectively ECG-triggered sequential scan (Goo 2010b) and retrospectively ECG-gated helical scan (Ben Saad et al. 2009).

An alternative way to increase temporal resolution of ECG-synchronized scanning includes multi-segment reconstruction. However, multi-segment reconstruction, in which multiple segments with higher temporal resolution acquired over multiple heart beats are used to generate a single image, is not commonly used in clinical practice because it requires much lower pitch lengthening scan time and resulting in degraded image quality caused by increased heart rate variability. On the other hand, beta-blockers may be used to lower heart rates and improve image quality of ECG-synchronized CT by increasing the mid-diastolic cardiac rest period, the so-called “diastasis.” However, satisfactory reduction of heart rates is not always achieved and longer preparation time considerably delays patient throughput. The most effective method to obtain the best image quality of ECG-synchronized CT at higher heart rates (e.g., >75 bpm) is to acquire CT data at the end-systolic phase. Compared with mid-diastolic data acquisition, another benefit of end-systolic data acquisition is that the image quality is less affected by arrhythmia. Scan time and radiation dose of retrospectively ECG-gated CT scanning can be reduced by using heart rate-adapted pitch (e.g., 0.17 for <40 bpm to 0.38 for >100 bpm) (McCollough et al. 2007). However, it should be kept in mind that gaps in the image data set may occur and result in image degradation if the pitch is too high for a given heart rate.

2.2 Retrospective ECG Gating and Prospective ECG Triggering

Retrospective ECG gating and prospective ECG triggering are two methods for ECG-synchronized CT scanning. Retrospective ECG gating is used with helical CT scanning. In this scan mode, helical CT data are continuously acquired with low pitch and images are then retrospectively reconstructed at a desired cardiac phase. The low pitch required for this scan mode, to avoid gaps in the image data, results in high radiation dose and long scan time. Multi-phase image reconstruction (e.g., 10 phases by 10 % of the RR interval) though the entire cardiac cycle allows the assessment of ventricular function and motion of valves. Although adaptive algorithms are incorporated to ameliorate image degradation of retrospectively ECG-gated helical CT images caused by arrhythmia, severe and persistent arrhythmia, such as atrial fibrillation, may result in substantial deterioration of image quality.

In contrast, prospective ECG triggering is used with sequential or step-and-shoot CT scanning. In this scan mode, axial CT data covering beam collimation are acquired at a predefined cardiac phase without table movement. The time for table feeding is mandatory for next axial CT data acquisition. This scan mode delivers a very low radiation dose because there are no data overlapping for image reconstruction. However, total scan time is considerably prolonged and multi-phase image reconstruction usually cannot be obtained. Recently, prospective ECG triggering may be used with high-pitch dual-source helical CT scanning (Han et al. 2011; Nie et al. 2012).

Initially recognized differences between retrospective ECG gating and prospective ECG triggering have become somewhat vague due to further technical developments. As a result, radiation dose of retrospectively ECG-gated helical scanning can be dramatically reduced by means of aggressive ECG-controlled tube current modulation, and multi-phase image reconstruction is also possible for prospectively ECG-triggered sequential scanning by using extended gantry rotation.

2.3 Radiation Dose Optimization

It is well known that children are more sensitive to the carcinogenic hazards of ionizing radiation than adults, and have a longer expected lifespan leading to a greater lifetime attributable risk of radiation-induced malignancies. Therefore, the radiation dose of pediatric cardiac CT should be minimized while maintaining diagnostic quality. As a first step of the optimization process, we should understand not only important scanning parameters affecting image quality and radiation dose, but also available dose-reducing strategies (Donnelly and Frush 2003; Kalra et al. 2004; Goo 2012). The establishment of body size-adapted CT protocols is particularly important in children. As body size parameters, body weight is most commonly used in our current practice (Yang and Goo 2008) but cross-sectional dimensions, such as diameter, circumference, or area, are better for radiation dose adaptation to body habitus (Menke 2005; Goo 2011b). In addition, all recently manufactured MDCT scanners are now equipped with automatic exposure control using tube current modulation. This dose reduction technique adjusts the tube current along the x-y plane (angular modulation) or along the z-axis (longitudinal modulation) or both (combined modulation) depending on the size, shape, and density of the scanning region (Kalra et al. 2004) (Fig. 1). Algorithms used for tube current modulation are vendor-specific and vary in the image quality reference parameter and in the preferred order of acquisition of the localizer radiographs. For optimal dose reduction, the patient should be positioned in the isocenter of the CT gantry (Lee et al. 2008). When scanning the pediatric thorax for cardiac applications, excluding the arms from the scan range and using optimal tube voltage will allow an even greater tube current reduction by combined tube current modulation (Goo and Suh 2006a, b; Greess et al. 2002).

Fig. 1

Automatic exposure control with MDCT. Lateral topogram (scout view) for thoracoabdominal CT in a 6-year-old child. The curve represents the automatically adapted milliampere-seconds value as a function of z-axis position during scanning with spiral CT. Although the standard adult protocol was used, the average milliampere-seconds value throughout the scan was adjusted to 38 mAs with automatic exposure control (Flohr et al. 2005, used with permission)

The benefit of the use of lower tube voltage for pediatric cardiac CT is a higher iodine contrast-to-noise ratio at a given radiation dose (Yu et al. 2011). When we use low tube voltage with tube current modulation, we should recognize that the tube current may be saturated to its maximum level resulting in excessive image noise and adversely affecting the image quality (Goo and Suh 2006b; Israel et al. 2008). The most practical solution to this problem is to use a lower pitch value. Other important measures to decrease radiation dose of pediatric cardiac CT include confining the study to the anatomical area of interest, avoiding multiphase examinations, and using iterative or other noise-reducing reconstruction algorithms (Goo 2012). In the contemporary MDCT era, a greater contribution of unnecessary overranging effect should be seriously considered particularly in CT examinations with a shorter scan range, such as pediatric cardiac CT, (Tzedakis et al. 2007) and volumetric axial scan modes with wide array detector configurations (Kroft et al. 2010) or adaptive section collimation should be used to reduce the overranging if applicable (Deak et al. 2009).

ECG-controlled tube current modulation is a technique available to reduce the radiation dose associated with retrospective gating (Jakobs et al. 2002). In those cardiac phases that are not needed for image reconstruction with high quality, the tube current is reduced to 4–20 % of the initial setting. Cardiac phases demanding the full tube current for high image quality should be appropriately selected according to a patient’s heart rate (Weustink et al. 2008).

By using available dose-reduction strategies simultaneously, the radiation dose of pediatric cardiac CT now can often be reduced to less than 1.0 mSv (Goo 2010a, 2011a).

2.4 Planning Scan Technique and Intravenous Contrast Injection

It is vital for the radiologist performing and interpreting pediatric cardiac CT to review any information pertaining to the child’s form of CHD and surgical repair or palliation prior to scanning. The knowledge helps the radiologist determine scan and intravenous (IV) injection protocols to best visualize anatomic substrates of the child. In addition, the ability of a child to follow breathing instructions is important to determine scan technique. The ultimate goal of IV contrast injection in pediatric cardiac CT is to obtain homogeneous enhancement of cardiovascular structures included in the scan range as much as possible because single-phase scanning should demonstrate all of them clearly to minimize radiation dose. However, homogeneous enhancement on pediatric cardiac CT is not always attainable, particularly in cases with complicated anatomy such as a Fontan pathway (Fig. 2). For optimal enhancement of a Fontan pathway, simultaneous IV injection of 50 % diluted or undiluted contrast agent via upper and lower extremities has shown good results (Greenberg and Bhutta 2008; Prabhu et al. 2009; Goo 2011b). Inhomogeneous cardiovascular enhancement may be contradictorily helpful to demonstrate hemodynamic findings, such as collateral vessels and contrast jet through a defect (Goo 2011c).

Fig. 2

Lack of opacification of the Fontan pathway during arterial phase of scanning. Coronal and axial oblique MIP images demonstrate dense contrast in the LPA via a left sided SVC after a left arm injection (a). The Fontan pathway from the IVC (b, d) and the distal RPA (c) are incompletely opacified. Coronal oblique images reconstructed from the venous phase demonstrate homogeneous opacification of the LPA (e) Fontan pathway (f), and RPA (g)

IV injection rates in children should necessarily be adjusted depending on the size of the IV catheter able to be placed and the amount of contrast to be injected (usually in the range of 1–2 ml/kg). Table 1 provides a guideline for contrast injection rates by size of the IV catheter. The optimal scan delay from the start of IV injection is usually determined by using bolus tracking. As compared with a biphasic IV injection protocol, a triphasic protocol, in which undiluted contrast agent is followed by 50–60 % diluted contrast agent and then by a saline chaser, can not only provide improved enhancement of the right heart, but also reduce perivenous contrast artifacts (Litmanovich et al. 2008).

Table 1 Maximal intravenous contrast injection rates

3 Complementary Role of MDCT and MRI for Cardiac Imaging

The complementary role of MDCT and MRI for noninvasive cardiac imaging in children with CHD should be clearly recognized. Relative merits and demerits of the two imaging modalities are summarized in Table 2. Recent technical developments in pediatric cardiac CT have changed the complementary role between MDCT and MRI to some extent. Although cardiac MRI techniques have also advanced, major limitations of cardiac MRI particularly in young children are still remained to be overcome: limited accessibility to MRI scanners for pediatric cardiovascular examinations in many hospitals and institutions, longer examination time, more requirement of sedation or anesthesia, and technical expertise to perform routine quality examinations. For the initial evaluation of CHD in infants and young children, echocardiography often provides a complete assessment of intracardiac morphology, flow, and ventricular function. However, echocardiography is limited in evaluating extracardiac structures and may not adequately show intracardiac and coronary artery anatomy. Cardiac CT may be helpful to compensate for these blind spots of echocardiography. Consequently, pediatric cardiac CT is increasingly used as a complementary imaging modality in many institutions before and after surgical correction in young children with CHD. On the other hand, cardiac MRI or invasive cardiac catheter angiography is seldom mandatory for surgical planning in these patients.

Table 2 Complementary role of MDCT and MRI for cardiac imaging

As evaluation with echocardiography becomes increasingly more difficult in older and larger patients who have had multiple cardiothoracic surgeries, cardiac MRI is regarded as the noninvasive imaging method of choice for serial follow-up examinations in patients with repaired CHD. Examples include patients with repaired tetralogy of Fallot (TOF), patients with a systemic right ventricle following the Senning or Mustard operation, and functional single ventricle patients following the Fontan operation. Unfortunately, the image quality of cardiac MRI may be considerably compromised by susceptibility artifact from the previously placed embolization coils, stents, and occlusion devices. In addition, indwelling pacemakers and AICD devices remain contraindications for MRI. On these occasions, cardiac CT may be considered as an alternative imaging method. In fact, cardiac CT is the diagnostic imaging method of choice in assessing vascular stent patency (Eichhorn et al. 2006) (Fig. 3).

Fig. 3

TOF and coarctation status post repair. Sagittal oblique MIP image (a) shows recurrent stenosis at the aortic isthmus (large arrow) and proximal left subclavian artery (small arrow) following an end-to-end anastomosis for coarctation repair. Oblique axial MIP images show patent bilateral proximal branch PA stents with narrowing of the RPA (arrow in b) and LPA (arrow in c) distal to the stents

4 Clinical Applications

4.1 Pulmonary Vasculature

4.1.1 Pulmonary Arteries

In patients with TOF with pulmonary atresia, precise preoperative delineation of the presence, size and confluency of the pulmonary arteries, and major aortopulmonary collateral arteries (MAPCAs) is necessary for surgical planning. This diagnostic task can be readily accomplished with cardiac MDCT (Goo et al. 2005a; Greil et al. 2006) (Fig. 4). As a result, the procedure time of catheter angiography and possibility of overlooked MAPCAs can be substantially reduced. Nonetheless, conventional catheter angiography is necessary for identifying communications between pulmonary arterial feeders. Abnormalities of the branch pulmonary arteries that are well depicted on MDCT include an abnormal origin or course such as in truncus arteriosus or pulmonary artery sling (Fig. 5). The branch pulmonary arteries can be atretic, stenotic, or hypoplastic related to decreased blood flow during growth (Fig. 6), extrinsic compression, or as a result of a surgically altered course or anastomosis such as a palliative shunt between the systemic and pulmonary artery circulation (Fig. 7).

Fig. 4

Tetralogy with pulmonary atresia and MAPCAs. Segmented 3D surface model from MDCT angiography demonstrates a collateral (APC) from the descending aorta (Dao) to the left and right lung connecting to the hypoplastic left (LPA) and right (RPA) pulmonary arteries (a). Corresponding X-ray angiogram (b) (Greil et al. 2006, used with permission)

Fig. 5

Pulmonary artery sling. Axial MDCT images demonstrate the LPA (a) arising from the RPA (b) and passing posterior to the carina. A 3D volume rendered posterior image (c)

Fig. 6

Branch pulmonary artery stenosis. Axial oblique MIP image (a) and volume rendered image (b) demonstrate a juxta-ductal stenosis (arrow) and kinking of the LPA

Fig. 7

Functional single ventricle status post right modified BT shunt. Coronal oblique MIP image (a) and volume-rendered image (b) show a stenosis of the RPA (arrow) proximal to the shunt

4.1.2 Pulmonary Embolism

Pulmonary embolism (PE) is an uncommonly diagnosed condition in children. The clinical presentation is often subtle because symptoms are nonspecific and can be masked by the underlying clinical condition. Delays in diagnosis are frequent because definite signs of associated pulmonary or cardiac dysfunction appear to be less common in children than in adults. Specific risk factors for PE in the pediatric patient include associated deep venous thrombosis, indwelling central venous catheters, cardiac surgery, thrombotic disorders, vascular malformations, and malignancy and multiple factors are often present in the same patient (Babyn et al. 2005). PE may complicate CHD particularly after surgical treatments, such as cavopulmonary connections (Fig. 8). Diagnostic strategies for detection and treatment of PE in children are mostly extrapolated from evidence that has been compiled in the adult literature.

Fig. 8

Coronal reformatted (a) and axial (b) images of a Fontan patient demonstrate extensive thrombus (arrows) extending from the Fontan pathway into the main and segmental branches of the LPA

CT has become the first choice of imaging modalities for detection of PE in symptomatic patients. MDCT has led to improved visualization of peripheral pulmonary arteries for detection of small emboli (Lee et al. 2011), and conventional pulmonary angiography is now rarely performed. Once regarded as the best first noninvasive study for the diagnostic work-up of PE, nuclear medicine perfusion scintigraphy is also now infrequently requested because as many as 73 % of studies are interpreted as indeterminate (PIOPED 1990) and have poor interobserver correlation (Blachere et al. 2000), and there is limited ability to make alternative diagnoses. Despite excellent diagnostic accuracy of MDCT in detecting PE, thromboembolic risk factors should be used as a first-line triage tool to guide more appropriate use of CT pulmonary angiography in children, with associated reductions in radiation exposure and costs (Lee et al. 2012).

MDCT angiography findings of acute pulmonary embolism include intraluminal filling defects in the main and branch pulmonary arteries that can partially or completely fill the lumen (Fig. 9). When the embolus completely fills the lumen of a branch pulmonary artery, the artery can enlarge relative to similar-sized arteries in the hilum. Lung parenchymal findings with PE include peripheral wedge-shaped opacities, hyperlucency, and mosaic perfusion. Findings of acute right ventricular failure, such as right ventricular enlargement and septal flattening, may be present in severe cases.

Fig. 9

Acute pulmonary embolism. Axial images from an MDCT angiogram in a pediatric patient with respiratory distress demonstrate a well-defined filling defect in the distal RPA (arrow in a) extending into the upper lobe pulmonary artery (arrow in b)

A classic finding of chronic PE is an intraluminal filling defect that makes an obtuse angle with the vessel wall and creates an appearance of asymmetric wall thickening (Fig. 10). Contrast-enhanced peripheral arteries can have irregular wall thickening related to recanalization and the residual thrombus may be calcified. Enlarged bronchiolar and systemic collateral arteries can also be seen in association with chronic PE.

Fig. 10

Repaired tetralogy of Fallot status post pulmonary valve replacement with recurrent bacterial endocarditis. Oblique coronal CT images (a, b) demonstrate hypodense clot (arrow) making obtuse angle with the vessel wall in the descending branch of the RPA consistent with chronic pulmonary embolism

Recently, dual-energy CT scanning enables us to evaluate PE and lung perfusion defects at the same time (Fig. 11). In addition to the accurate diagnosis of PE, pulmonary blood volume assessment using dual-energy CT can predict right heart strain and clinical outcome (Bauer et al. 2011). This emerging CT imaging technique for evaluating PE may also be useful in pediatric patients (Goo 2010b; Zhang et al. 2012).

Fig. 11

Dual-energy lung perfusion CT in a child with leukemia and pulmonary embolism. Coronal weighted-average CT image (a) shows an acute embolus (arrow in a) in the RPA. Coronal iodine map (b) reveals a well-circumscribed perfusion defect (arrows in b) in the right lower lung

4.1.3 Pulmonary Veins

MDCT angiography is quite useful and accurate in evaluating anomalous pulmonary venous connections when echocardiography is limited in fully identifying types and obstructions of these total or partial anomalous pulmonary venous connections (Kim et al. 2000). When associated with CHD, pulmonary vein stenosis (PVS) is most often extrinsic due to compression by other vascular structures, or associated with the site of a prior surgical anastomosis (Fig. 12). PVS can rarely be intrinsic and rapidly progressive and refractory to all forms of treatments (Latson and Prieto 2007; Devaney et al. 2006). Progressive PVS can occur in children with or without CHD and MDCT can assess PVS noninvasively and accurately (Ou et al. 2009).

Fig. 12

Total anomalous pulmonary venous return status post repair with pulmonary venous stenosis. Coronal oblique MIP images demonstrate focal stenosis (arrows) of the right (a) and left (b) pulmonary veins. (Courtesy of Sjirk Westra)

Congenital pulmonary venolobar syndrome or scimitar syndrome is a heterogeneous group of congenital anomalies of the thorax that may occur singly or in combination. The main components of the congenital pulmonary venolobar syndrome are hypogenetic lung (lobar agenesis, aplasia, or hypoplasia), partial anomalous pulmonary venous return, absence of a pulmonary artery, pulmonary sequestration, systemic arterialization of the lung, absence of the inferior vena cava (IVC), and accessory diaphragm (Konen et al. 2003). Horseshoe lung may be rarely associated with this syndrome (Goo et al. 2002). MDCT provides a complete evaluation of all pulmonary and systemic vascular, tracheobronchial and pulmonary parenchymal anomalies, necessary in patients under consideration for surgical repair (Fig. 13).

Fig. 13

Scimitar syndrome. Volume-rendered CT image (a) shows partial anomalous pulmonary venous return of the right upper pulmonary vein to the right atrium (long arrow in a). Systemic collateral arteries supplying the right lower lung (short arrows in a) is also depicted. Shaded surface display CT image (b) reveals right lung hypoplasia, horseshoe lung (long arrow in b), and right diaphragmatic hernia (short arrows in b)

4.2 Aorta

4.2.1 Coarctation of the Aorta

Coarctation of the aorta is a congenital obstructive aortic arch anomaly presenting with arch hypoplasia and focal narrowing of the aortic isthmus at the junction of the ductus arteriosus and the aorta. MDCT can demonstrate anatomic features of the anomaly and collateral arteries, if present (Fig. 14), that are helpful for optimal surgical planning. It should be noted that the anomaly is a dynamic process showing progressive obstruction in young infants when the patent ductus arteriosus (PDA) is present. Following surgical repair of coarctation, MDCT can be used to detect residual stenosis, recoarctation, or aneurysm formation at the repair site. In order to avoid radiation exposure in children, MRI may be favored for long-term follow-up after surgical repair. MDCT is the imaging method of choice for evaluating associated airway abnormalities and in-stent stenosis after stent placement. In addition, MDCT is also useful for assessing early post-procedural complications, such as pseudoaneurysm formation and dissection (Fig. 15).

Fig. 14

Coarctation with transverse arch hypoplasia. Sagittal oblique MIP image (a) demonstrates a severe discrete narrowing (arrow) of the aortic isthmus. Axial reformatted images show the markedly small caliber of the isthmus (arrow in b) and numerous dilated intercostal collaterals joining to the proximal descending aorta (arrows in c) (Courtesy of Shi-Joon Yoo)

Fig. 15

Dissection and pseudoaneurysm of the thoracic aorta following balloon dilatation and stenting of coarctation. Oblique sagittal MIP (a) and volume rendered (b) images demonstrate a stent at the aortic isthmus with a dissection and large pseudoaneurysm (arrow) protruding from the aorta. The patient required an additional covered stent to be placed which closed the entrance of the pseudoaneurysm into the thoracic aorta. The pseudoaneurysm was shown to be thrombosed on subsequent imaging

4.2.2 Interrupted Aortic Arch

Interrupted aortic arch (IAA) is a rare aortic anomaly defined as a complete luminal and anatomic discontinuity of the aortic arch. The anomaly is classified as three types depending on the site of interruption, i.e., distal to the subclavian artery in type A, between the second carotid and ipsilateral subclavian arteries in type B, and between two carotid arteries in type C. Type B is most common in the Western population (Fig. 16), while type A is most common in the Asian population (Lee et al. 2006). Each type may be further divided into three subtypes depending on the origin of the subclavian artery, i.e., normal in subtype 1, aberrant in subtype 2, and isolated in subtype 3. In addition to the anatomic types of IAA, cardiac CT may be used to evaluate the distance between the proximal and distal segments of IAA, the sizes of a PDA, the aorta and the thymus, the presence of subaortic stenosis, and other cardiac defects (Yang et al. 2008). In IAA, a right aortic arch is almost always associated with DiGeorge syndrome and/or chromosome 22q11 deletion (McElhinney et al. 1999b). As in coarctation of the aorta, IAA may be associated with the bicuspid aortic valve and other components of Shone complex including supravalvular mitral membrane, parachute mitral valve, and subaortic stenosis.

Fig. 16

Interrupted aortic arch type B2. Volume-rendered CT image shows interruption of aortic arch between the left common carotid artery (LCCA) and the left subclavian artery (LSCA). Aberrant right subclavian artery (arrow) is noted

4.2.3 Valvular and Supravalvular Aortic Stenosis

Valvular aortic stenosis occurs in approximately 3–6 % of patients with CHD and is often associated with a congenital bicuspid aortic valve. Although the effective valve area can be reduced at birth, the stenosis of a bicuspid valve is progressive and clinical symptoms do not usually develop until young adulthood. MDCT is seldom used for evaluating valvular aortic stenosis because anatomic details of the aortic valve in children are not well-seen on CT and high radiation dose is necessary for the complete assessment of the aortic valve. The surgical approach to aortic valve replacement for severe congenital aortic stenosis in young patients is difficult because placement of a mechanical valve is not a good option because of the risk of long-term anticoagulation. Other options include placement of homograft or xenograft valves. In the Ross procedure, the stenotic aortic valve is replaced with the patient’s pulmonary valve, and a right ventricle to pulmonary artery conduit is placed.

Supravalvular aortic stenosis (SVAS) may be non-syndromic or associated with Williams syndrome. ECG-synchronized cardiac CT can show not only SVAS, but also the bicuspid aortic valve, dilated coronary arteries, coronary ostial stenosis, and left ventricular hypertrophy (Liu et al. 2007). SVAS may be focal (the “hourglass” appearance) (Fig. 17) or diffuse (10–30 % of cases), starting at the sino-tubular junction. Aortoplasty is indicated in symptomatic patients or in those with a transaortic valve gradient greater than or equal to 50 mmHg (Scott et al. 2009).

Fig. 17

Supravalvular aortic stenosis in Williams syndrome. Volume-rendered CT image shows a focal stenosis at the sinotubular junction (arrows). In addition, there is no evidence of coronary ostial stenosis

4.2.4 Connective Tissue Disorders

Marfan syndrome and type IV Ehlers-Danlos syndrome are connective tissue disorders that can have cardiovascular manifestations. Both are associated with cystic medial necrosis of the aortic wall, which adversely affects the ability of the aortic wall to withstand systemic pressures, leading to dilatation. The characteristic findings of the aorta include dilatation of the aortic root and proximal ascending aorta and effacement of the sino-tubular junction. The dilatation of the aortic root results in suboptimal coaptation of the aortic valve cusps, which can lead to aortic regurgitation that can further weaken the aortic wall as more throughput volume is needed to maintain cardiac output. Both echocardiography and MRI can be used for serial follow-up of ascending aorta size and aortic regurgitation in patients with Marfan syndrome. CT may be used in some patients with severe chest wall deformity limiting echocardiographic evaluation or who cannot tolerate lengthy MRI evaluation (Ha et al. 2007). In general, when the maximum diameter of the ascending aorta is 1.5 times that of the descending thoracic aorta at the level of the diaphragm, an aneurysm is considered to be present. In addition to the maximal diameter of the aortic root, its growth rate should be considered in determining the optimal timing for surgical replacement of the aortic root and/or ascending aorta. Serious complications of Marfan’s or type IV Ehlers-Danlos syndrome include dissection and rupture of the ascending aorta (Fig. 18). Although transesophageal echocardiography (TEE) or MRI could be performed urgently if dissection is suspected, MDCT angiography with multiplanar reconstruction is a highly sensitive and specific technique for the detection and characterization of the extent and orientation of the intimal flap, delineation of the true and false lumens, presence of intramural hematoma, and involvement of the major aortic branches and coronary arteries (McMahon and Squirrell 2010).

Fig. 18

Aortic dissection in Marfan syndrome. Extensive dissection of the thoracic aorta in a teenager who subsequently had the ascending and descending thoracic and abdominal aorta replaced. Oblique MIP images demonstrate the dissection involving the transverse arch (arrow in a), extending into the thoracic and abdominal aorta (b), and also extending into the left subclavian artery (arrows in c, d)

Progressive aneurysmal dilatation of the aorta may also be seen in children with other congenital connective tissue disorders, such as Loeys-Deitz syndrome (Kalra et al. 2011) and arterial tortuosity syndrome (Kalfa et al. 2012).

4.2.5 Takayasu Arteritis

Takayasu arteritis is a progressive large vessel vasculitis that affects the aorta and its major branches as well as the coronary and pulmonary arteries. The accurate diagnosis of the disease depends on imaging studies because the clinical and laboratory presentation at disease onset is often nonspecific. In the early, systemic phase of inflammation, both CT (Zhu et al. 2012) and MRI (Choe et al. 2000) can detect wall thickening and enhancement of the involved vessels and can be used to follow-up response to high-dose steroid therapy. If the disease is not detected and treated early, transmural fibrosis of the vessel wall can lead to the characteristic findings of the late phase, including stenosis, occlusion, mural calcification, intraluminal thrombus, or aneurysmal dilatation of the affected artery (Fig. 19). Both MRA and MDCT angiography can be used for noninvasive detection of the sequelae of chronic disease. Takayasu disease is often recurrent and the timing of progression from early to late phase of disease can be variable, so that early and late findings can be detected concurrently. Children with Takayasu arteritis are more often diagnosed and followed with MRI (Aluquin et al. 2002). CT is more useful for diagnosis of early complications following surgical bypass or transcatheter stenting of vasculitis-induced stenosis, including development of pseudoaneurysms, graft infection, thrombosis, and restenosis.

Fig. 19

Takayasu arteritis and coarctation status post balloon dilatation and stenting. Sagittal oblique MIP image demonstrates a stent at the aortic isthmus and diffuse calcification (large arrows) of the aortic wall consistent with chronic arteritis. The proximal left common carotid and subclavian arteries are moderately stenotic and ectatic (small arrows)

4.3 Coronary Artery Anomalies

Congenital anomalous coronary arteries, although rare, are a well-recognized cause of myocardial ischemia and sudden death in children and young adults, with a much higher prevalence in patients with CHD, especially TOF and transposition of the great arteries (TGA). Hemodynamically significant coronary artery anomalies include anomalous origin from the pulmonary artery, anomalous origin from the opposite sinus of Valsalva with an interarterial course, intramural origin, myocardial bridging, and coronary artery fistula (Goo et al. 2009). Hemodynamically benign anomalies, such as high take-off and multiple ostia, often have clinical significance in patients with CHD (Goo et al. 2009; Tsai et al. 2010). In patients with CHD, certain coronary artery anomalies should be specifically recognized before surgery (Goo et al. 2009).

Echocardiography with color Doppler has replaced cardiac catheterization as the standard method of visualizing the proximal coronary arteries in infants and children (Satomi et al. 1984), but further evaluation with CT or MRI is often necessary to increase diagnostic confidence. Albeit useful in assessing coronary artery anomalies, MRI is still limited in assessing coronary arteries in young infants and distal segments in children mainly due to low spatial resolution (Tangcharoen et al. 2011). In addition, MRI evaluation of coronary arteries is relatively time-consuming. In contrast, MDCT considerably increases coronary artery visibility even without ECG synchronization (Goo et al. 2005b). Nevertheless, ECG-synchronized CT scanning should be used for evaluating coronary artery anomalies (Tsai et al. 2007; Ben Saad et al. 2009; Goo and Yang 2010). MDCT is particularly advantageous in patients presenting with acute symptoms including palpitations, dizziness, atypical or typical exertional chest pain, and dyspnea on exertion, especially in young athletes (Deibler et al. 2004). These coronary artery anomalies can be reliably detected on ECG-synchronized cardiac CT (Figs. 20 and 21).

Fig. 20

Anomalous left coronary artery. Axial (a) and coronal (b) MIP images obtained from a retrospectively gated coronary CTA demonstrate an anomalous left coronary artery (large arrow) arising from the right coronary sinus and passing between the ascending aorta (Ao) and main pulmonary artery (MPA)

Fig. 21

Coronary artery fistula in a 9-year-old girl with Noonan syndrome. Oblique coronal (a) and oblique axial (b, c) images from a retrospectively gated coronary CTA demonstrate diffuse enlargement of the right coronary artery (RCA) with a tortuous distal branch that enters a dilated coronary sinus (CS). Flow through the fistulous connection was coil occluded to prevent coronary steal. (LAD) left anterior descending coronary artery

4.3.1 Kawasaki Disease

Kawasaki disease is a vasculitis of unknown origin that occurs most often in young children. It begins as a pancarditis with vasculitis of small vessels (stage 1), progresses to vasculitis of the epicardial coronary arteries (stage 2), followed by resolution of vascular inflammation with decrease in size of the aneurysms (stage 3), and scarring of the coronary arteries with stenoses (stage 4) (Fujiwara and Hamashima 1978). Coronary artery aneurysms can develop in up to 15–25 % of untreated cases and can be associated with thrombotic events leading to myocardial ischemia and infarction later in their life (Kato et al. 1996). In addition to the standard initial therapy with intravenous gamma-globulin and high dose aspirin, infliximab and steroids are promising new therapies potentially further reducing incidence of coronary artery abnormalities (Dominquez and Anderson 2013). Now we have an increasing chance to see adult patients with cardiovascular sequelae from Kawasaki disease because only recently have these patients reached adulthood (Daniels et al. 2012). Serial follow-up of patients with Kawasaki disease is essential because the size of aneurysms and severity of coronary artery stenosis can change over time. Transthoracic echocardiography is usually used to follow small children for the development of aneurysms, but adequate visualization of the proximal coronaries tends to diminish with increasing age and size of the patients. Coronary artery aneurysms and wall thickenings can be assessed with coronary MR angiography (Greil et al. 2007), but current MRI techniques have limited spatial resolution for reliable detection of obstructive coronary artery lesions, and abnormalities of the distal coronary arteries, compared with coronary CT angiography (Goo et al. 2006).

In a study of adolescents with Kawasaki disease, ECG-gated coronary CT angiography accurately demonstrated not only aneurysms, but also complete occlusions and stenoses (Sato et al. 2003). Recent technical developments in ECG-synchronized cardiac CT have considerably improved diagnostic performance in assessing coronary artery abnormalities in Kawasaki disease, even in small children with higher heart rates (Duan et al. 2012) (Fig. 22). However, it should be kept in mind that coronary CT angiography is limited in assessing heavily calcified lesions and myocardial perfusion state. On these occasions, MRI can show better luminal assessment of heavily calcified lesions, and a study (Tacke et al. 2011) demonstrated that MRI could provide the comprehensive evaluation of myocardial perfusion and viability in patients with Kawasaki disease.

Fig. 22

Kawasaki disease with mild coronary artery stenosis in a 3-year-old boy. Volume-rendered CT image (a) and reformatted (b) CT image demonstrate a newly developed mild narrowing (arrows in a) and wall thickening (arrows in b) of the proximal right coronary artery

4.4 Airway Compromise in Patients with Congenital Heart Disease

The most common types of vascular anomalies to cause symptomatic tracheal and esophageal compression are the right aortic arch with an aberrant left subclavian artery and the double aortic arch (Fig. 23). In infants and children with these anomalies, symptoms can vary from wheezing to frank respiratory failure, related in part to the direct effect of vascular compression as well as secondary tracheobronchomalacia that can result from prolonged compression. If the vascular ring exhibits less compression, it may be diagnosed in the older child with symptoms primarily of esophageal compression. The double aortic arch is less common than right arch with an aberrant left subclavian artery, but it more often results in a tight ring necessitating earlier surgical intervention for airway obstruction. With the double aortic arch, the right arch is more often dominant and cephalic in location compared with the left arch. However, the left arch is occasionally dominant and one of the arches may be atretic or have an associated coarctation.

Fig. 23

Common vascular rings. Axial (a) and sagittal (b) MIP images demonstrate a right aortic arch (RAA) and aberrant left subclavian artery (LSCA) extending posterior to the trachea and causing narrowing and anterior bowing of the trachea (TR). Posterior volume rendered image (c) shows the dilated proximal left subclavian artery consistent with a diverticulum of Kommerell (arrows) related to ductal flow into the subclavian artery in utero. A 3D bronchographic image (d) demonstrates mild narrowing of the distal trachea (arrow). Posterior and superior volume rendered views (e, f) of a double aortic arch with lungs and trachea included demonstrate tracheal narrowing at the level of the left (LAA) and right (RAA) arches passing to either side of the trachea. The left aortic arch is smaller than the right arch, as is typical (Courtesy of Hyun Woo Goo)

MDCT provides imaging in multiple planes to completely characterize the anomalous vasculature and the extent of airway compression. The current trend of performing minimally invasive surgery for repair of vascular ring using video-assisted thoracoscopic or robotic endoscopic techniques has advantages over lateral thoracotomy including a smaller incision, improved visualization inside the chest cavity, reduced postoperative pain, and risk of chest wall deformity. These less invasive techniques require more precise 3D delineation of cardiovascular structures and airways on preoperative CT (Lambert et al. 2005). In addition to the vascular anomalies described above, there are more rare conditions that can result in symptomatic airway and/or esophageal compression, including pulmonary artery sling (Fig. 5), anomalous innominate artery, circumflex aorta, and cervical aortic arch. Patients with TOF and absent pulmonary valve syndrome can have severe pulmonary regurgitation, which can lead to markedly enlarged pulsatile pulmonary arteries that can cause severe bronchial compression associated with bronchomalacia (Taragin et al. 2006) (Fig. 24).

Fig. 24

Tetralogy of Fallot with absent pulmonary valve syndrome status post repair. Axial (a) and sagittal oblique (b, c) MIP images from an MDCTA demonstrate severe dilatation of the main pulmonary artery (MPA) and proximal branch PAs causing severe compression of the distal trachea and proximal main stem bronchi (arrows)

MDCT not only provides a rapid assessment of the ICU patient for potential causes of failed extubation in the early postoperative period in patients with CHD, but also helps surgical planning to prevent postoperative airway compression (Kim et al. 2002; Goo 2004). The airway compression may be related to the patient’s intrinsic anatomy, such as ascending aorta dilatation (McElhinney et al. 1999a) (Fig. 25), or due to surgically reconstructed vessels, such as with arterial switch operation and aortic arch reconstruction (Jhang et al. 2008). MDCT is also helpful in evaluating small airway abnormalities, such as air trapping (Goo and Kim 2006), mosaic lung attenuation, and plastic bronchitis (Goo et al. 2008), in children with CHD. Furthermore, dynamic airway CT can differentiate tracheobronchomalacia from fixed vascular airway compression (Greenberg 2012).

Fig. 25

Malposition of the aortic arch causing severe airway and branch pulmonary artery narrowing in an infant with severe respiratory compromise following repair of tetralogy of Fallot and pulmonary atresia. Coronal (a), axial (b, d) and sagittal (c) oblique MIP images demonstrate a markedly dilated right aortic arch (RAA) positioned posteriorly within the chest, resulting in severe compression of the distal trachea and carina (small arrows), as well as the right pulmonary artery (arrow in d)

4.5 Cardiac Chamber Morphology and Ventricular Function

The current ECG-synchronized cardiac CT imaging techniques with excellent spatial and temporal resolutions allow accurate assessment of cardiac chamber morphology and ventricular function in children with CHD (Goo 2010a, 2011b). With retrospective ECG gating or prospective ECG triggering, multi-phase cardiac imaging can be obtained for quantitative assessment of ventricular volumes, mass and global function at the expense of higher radiation dose (Fig. 26).

Fig. 26

Short-axis reconstructed images through the heart in diastole (a) and systole (b) for quantification of ventricular function

Recently, cardiac CT was found more accurate than biplane cineventriculography and both 2D and 3D echocardiography in assessing left ventricular function with MRI as the reference standard (Greupner et al. 2012), and cardiac CT was accurate as MRI for left ventricular and right ventricular volume measurement (Maffei et al. 2012). Ventricular volumes measured by cardiac CT may be used for the enlarged right ventricle in repaired TOF to determine the optimal timing of pulmonary valve replacement and for the marginally small ventricle to decide the feasibility of biventricular repair (Kim et al. 2013). In addition, cardiac CT has great potential as a complementary imaging modality in characterizing cardiac chamber morphology in patients with CHD. CT has been shown to depict fatty tissue within the ventricular wall in patients with arrhythmogenic right ventricular dysplasia (ARVD) (Kimura et al 2002; Tandri et al. 2004) (Fig. 27).

Fig. 27

MDCT of ARVD. Axial ECG-gated MDCT image demonstrates fatty tissue in the trabeculae of the RV (large arrow) and along the RV side of the interventricular septum (small arrows). (Used with permission from Kimura et al. 2002)

4.6 Postoperative Congenital Heart Disease

The evaluation of surgical results and possible complications involving palliative shunts, conduits, and intracardiac baffles and the patency of the pulmonary arteries has become a major application of noninvasive imaging of postoperative CHD. CT is most often utilized if a full evaluation of the postoperative vascular morphology is limited on MRI due to artifacts from indwelling ferromagnetic materials such as stents, coils, and occlusion devices, or when a patient has claustrophobia, which tends to be more of an issue with older patients.

Extracardiac conduits are prosthetic or homograft tubes used to create venoarterial, ventriculoarterial, and arterioarterial connections when the structures to be connected are too far away from each other to allow a direct anastomosis. There are three different mechanisms of conduit obstruction: formation of a thick endothelial peal, scarring at sites of anastomosis, and relative narrowing of the conduit associated with growth of structures at either end. Both cardiac CT and MRI allow a more complete visualization of conduits in their entirety than echocardiography or angiography due to a wide field of view and 3D imaging capability.

There are two main types of intracardiac baffles used to redirect venous blood flow through the heart: the atrial inversion procedure (Mustard or Senning operations) for TGA and the Fontan procedure for functionally univentricular hearts. In the Mustard or Senning operation, the native intra-atrial septum is excised and a baffle is inserted to direct superior vena cava (SVC) and IVC blood flow toward the mitral valve. The pulmonary venous blood passes around the baffle and is directed toward the tricuspid valve. Both systemic and pulmonary venous pathways have the potential for obstruction, which can be assessed by CT (Fig. 28). In the Fontan operation, a surgically created pathway reroutes the systemic venous return from the IVC and SVC directly to the pulmonary arteries. Possible complications include pulmonary arteriovenous malformations (Fig. 29), pulmonary venous obstruction due to extrinsic compression by an intracardiac or extracardiac baffle, or an enlarged cardiac structure used for Fontan pathway. Pulmonary artery obstruction at the level of cavopulmonary anastomosis or due to distortion from a prior Blalock-Taussig (BT) shunt or prior surgical pulmonary artery reconstruction can also develop. Imaging evaluation is directed to establish the overall patency of the Fontan pathway. In the early versions of the Fontan operation, the right atrium is incorporated into the systemic venous to pulmonary artery connection, and not infrequently, patients can develop thrombus in the pathway due to relative stasis of slow flowing blood (Fig. 8).

Fig. 28

Complete-TGA status post Senning. Sagittal oblique (a) MIP image demonstrates the characteristic morphology of complete TGA with the anterior aorta arising from the RV and posterior main pulmonary artery arising from the LV. Axial MIP image (b) demonstrates the pulmonary venous (PV) pathway that directs oxygenated blood to the systemic RV positioned anteriorly. Coronal reformatted image (c) demonstrates the typical “pant leg” configuration of the IVC and SVC (with stent) pathway directed to the left ventricle separated by an intra-atrial baffle from the PV pathway

Fig. 29

Pulmonary arteriovenous malformations (AVMs) in a patient with tricuspid atresia status post Glenn shunt to the RPA. Oblique axial image in a four-chamber projection (a) demonstrates a dilated right atrium and hypoplastic RV with fatty tissue completely separating them. Sagittal oblique image (b) demonstrates the patent cavopulmonary anastomosis between the SVC and RPA. Coronal images of the chest (c, d) show marked dilatation of the distal branches of the RPA and pulmonary veins (PV) draining the right lower lobe related to multiple small AVMs, some of which have been occluded by embolization coils. The right lung has become hypoplastic due to steal of blood supply through the AVMs

5 Conclusion

In summary, pediatric cardiac CT provides a number of important advantages for morphologic and functional assessment of the cardiovascular system, airways, and lungs. MDCT is now a major diagnostic tool increasingly used in children with CHD, and when performed for appropriate indications with optimized techniques, the benefits far exceed the very small individual risks.