Over the past few decades, advances in pediatric cardiology and cardiac surgery have revolutionized the prospect for patients with adult congenital heart disease (ACHD). Although cardiovascular magnetic resonance imaging (CMR) and transthoracic echocardiography (TTE) remain the techniques of choice for their routine assessment and follow-up, advances in cardiac computed tomography (CCT) have led to its emergence as both a complementary technique and an alternative to CMR and TTE when these are unavailable or contraindicated. Current CT scanner technology allows cardiac assessment without blurring from cardiac motion and offers superior spatial resolution compared with that of both CMR and TTE. CCT images comprise near-isotropic voxels that look identical irrespective of the plane in which they are viewed, allowing rotation of the three-dimensional (3D) dataset in any desired plane even after the completion of acquisition, thus rendering prespecification of imaging planes unnecessary. Using high pitch-scan modes, anatomic coverage of the thorax in less than a second is feasible; this reduces the need for sedation and anesthesia. Although the temporal resolution of CCT remains inferior to both CMR and TTE at present, wider detector arrays and dual-source radiographic technology offer resolutions as low as 66 ms on newer generations of scanners. Exposure to iodine contrast (contraindicated in patients with severe renal impairment or contrast allergy) and ionizing radiation remain limiting factors in the widespread application of CCT. Most modern CT scanners incorporate dose-reduction algorithms into their cardiac packages (eg, iterative reconstruction, dose modulation) and wider detector arrays (256- and 320-detector scanners) as well as improved detector sensitivity; these have led to further reductions in radiation dose. CCT has the advantage of being able to assess the coronary arteries and extracardiac anatomy (eg, lung parenchyma, airways, skeletal abnormalities) in addition to CHD, but it cannot assess valvular and shunt flow (because it is a first-pass technique), parameters readily measured by both CMR and TTE. However, CMR and TTE also have important limitations. Acquisitions may be time-consuming, especially in those with ACHD, and may require anesthesia; therefore these studies may not be tolerated by critically ill patients or those with high risk for adverse events with anesthesia. Furthermore, CMR is often limited by its availability, and claustrophobia may prevent successful acquisition in as many as 1 in 20 patients. Importantly, the ever-increasing use of pacemakers or implantable cardiac defibrillators (ICDs) usually precludes assessment by CMR; CCT is an appropriate alternative in these cases. Furthermore, CCT is preferable for the assessment of stents and occlusion devices because images do not suffer from the signal void that these devices create in imaging with CMR. Regardless of the technique selected, all of these methodologies require substantial training and expertise and should be used only by operators with the appropriate experience. The reporting of CCT images should follow the standardized segmental approach described elsewhere in this book.
Technical Considerations
Contrast Protocols
Standard retrospectively gated or 30% to 80% R-R prospectively gated CT coronary angiography (CTA) images usually gives clear information about both left ventricular function and coronary lumenography, if both are required. However, few methodologic studies look at CTA in ACHD. Although most contrast protocols are suitable for all patients, certain considerations should be taken into account in timing the administration of a contrast agent for those with ACHD. A manual test bolus tracked to determine the time to peak concentration at the aortic root is recommended owing to the variable transit time and venous hemodynamics of ACHD patients; this also allows early identification of other late-filling structures. Particular care should be taken in those with presumed or likely pulmonary arterial hypertension in whom transit times may be especially challenging to calculate despite the use of bolus tracking. In patients who have undergone Fontan repair, imaging may be especially difficult because the contrast bolus may pool and become diluted in the passive right-sided circulation. Additionally, consideration should be given to the limb through which the contrast agent is injected, because delivery from either the superior or inferior vena cava may lead to preferential perfusion of one lung. In ACHD, right ventricular function is often of interest; although reduced pulmonary transit time is likely to be of benefit in right ventricular analysis, it may be detrimental to analysis of the left ventricle. Although it is possible to change the scan timing or CT protocol to optimize right ventricular opacification, this, in turn, limits left ventricular opacification and coronary artery assessment, thus preventing complete cardiac assessment within a single breath-hold. However, using specific intravenous contrast protocols, it is possible to combine CTA with CCT within a single scan protocol to allow comprehensive assessment of the pulmonary and coronary arteries, biventricular function, and valvular anatomy without fundamentally altering the region of interest or the basic scan protocol. Finally, because CCT involves intravenous iodinated contrast, often in excess of 70 mL (eg, dual- or triple-phase CTA/CCT protocols), the technique is best avoided in those patients with renal dysfunction when alternative techniques are available.
Gating
The improved temporal resolution of current CT scanners (66 to 165 ms) coupled with simultaneous electrocardiographic recording allows image acquisition during multiple phases of the cardiac cycle. This allows selection of the interval of minimum cardiac motion (usually end-diastole) and enables the resolution of structures as small as 0.5 mm. Patients requiring evaluation of structures prone to cardiac motion artifact—such as intracardiac anatomy, coronary arteries, and the aortic root—and those requiring functional assessment should be scanned using ECG gating. Newer ultra-high-pitch acquisitions, which allow for a more limited assessment of cardiac and aortic structures without ECG-gating, can be used for all patients who do not require evaluation of the coronary anatomy. Gating is unnecessary when the predominant clinical question centers on assessment of major extracardiac vascular structures because cardiac motion is less important. Ungated acquisitions are usually used in imaging infants because the scans are quicker to perform, easier to process, and involve lower exposure to ionizing radiation. However, rapid cardiac motion prevents adequate assessment of smaller structures, such as the coronary vessels, in ungated studies. If coronary angiography is required, acquisitions should use either prospective (end-diastolic or end-systolic) or retrospective electrocardiographic triggering. Prospective acquisitions involve the emission of radiation only during a predefined phase of the cardiac cycle, thus reducing radiation dose. End-diastolic acquisitions are suitable for patients with stable heart rhythms in whom the interval of minimum cardiac motion can be predicted reliably. However, at faster and less predictable heart rates, systolic imaging should be used if technically feasible. It should be noted, however, that because prospective gating provides information on only one phase of the cardiac cycle, functional information cannot be obtained and interpretation of the resultant images is thus limited to anatomy. In retrospective acquisitions, radiation is emitted throughout the cardiac cycle. Retrospective gating is useful in patients who do not have a stable heart rate and thus have an unpredictable interval of minimum cardiac motion as well as patients who need functional assessment of the ventricles and heart valves. In patients with ACHD, systolic acquisition or, if not possible, retrospective gating is used most often because the incidence of arrhythmia is higher, and the functional information obtained is helpful.
Cardiac Computed Tomography in Clinical Practice
Recent CCT research and practice have extended beyond noninvasive coronary lumenography to structural and preprocedural assessment. With ACHD patients now surviving longer, they are at equal or increased risk of common cardiac conditions that present in adulthood, such as coronary artery disease. Coronary CTA thus retains the same indications as in patients without ACHD. However, the CTA dataset contains substantially more information than that of the coronary arteries alone and a far broader assessment of cardiac anatomy and function is possible from a single acquisition. In essence, any patient who is unable or unwilling to undergo CMR can be assessed by CCT; and although flow data cannot be obtained, most other aspects of a CMR study are available from within the CCT dataset.
Coronary Artery Assessment
Because of the high incidence of abnormal resting electrocardiograms, stress electrocardiography is often unhelpful for the diagnosis of coronary artery disease in those with ACHD. Abnormal ventricular anatomy also leads to difficulties in the interpretation of myocardial perfusion scans. Many are therefore investigated by invasive coronary angiography, although this may in itself be complicated by the presence of aortic root dilation, variation in the site of the coronary ostia, and unusual coronary anatomy. Furthermore, once these technical issues have been overcome, there is often no evidence of obstructive coronary artery disease. CTA offers excellent negative predictive value for the exclusion of coronary artery disease and is a powerful alternative to invasive coronary angiography in this setting. The use of CTA is especially relevant outside of CHD centers, where operators experienced in invasive coronary angiography for patients with ACHD may be not be readily available. Beyond coronary artery disease, CTA is especially helpful in assessing the origin and course of anomalous coronary arteries, which are seen frequently in those with abnormal cardiovascular anatomy ( Fig. 9.1A ). Aside from common anomalies, such as left coronary artery from right coronary sinus, CTA may provide the first diagnosis in patients with anomalous left coronary artery from pulmonary artery (ALCAPA) on the rare occasions that this presents in adulthood. Patients who have undergone surgical coronary artery manipulation such as reimplantation usually require postsurgical CT angiographic assessment. CTA is also useful in patients with Kawasaki disease, where the site, size, and number of coronary artery aneurysms can be measured, as can the extent of calcification, thrombus, and contrast enhancement within aneurysms. These features are seen with comparable accuracy to conventional coronary angiography (see Fig. 9.1B ). CCT is also a well-established technique for identifying and fully delineating coronary fistulas (see Fig. 9.1C ) and cardiac venous anatomy (see Fig. 9.1D ). The latter may be of particular importance in planning cardiac resynchronization therapy, a technique that is finding greater use in patients with CHD. Delineation of the course and relationship of the coronary arteries to the right ventricular outflow tract (RVOT) and sternum is another indication for CT prior to RVOT intervention—that is, prior to percutaneous pulmonary valve implantation in patients with tetralogy of Fallot (TOF) (see Fig. 9.1E ).
Functional Assessment of the Left and Right Ventricles
By reconstructing CCT data at multiple phases of the cardiac cycle (usually every 5% or 10%), it is possible to calculate both end-diastolic and end-systolic volumes of the left and right ventricle and thus also stroke volume, cardiac output, and ejection fraction. For estimation of both left and right ventricular function, a biventricular injection protocol should be used so that the endocardial borders of both ventricles are clearly defined.
Ventricular volumes may be calculated either through manual delineation of endocardial and epicardial borders or using a threshold technique that identifies voxels above a certain Hounsfield unit number as contrast rather than tissue ( Fig. 9.2A and B ). The latter is quicker and probably more accurate, although both depend on adequate opacification of the ventricle to make an accurate assessment. CCT agrees well with CMR, TTE, and myocardial perfusion scintigraphic measurements of left ventricular ejection fraction. There is good agreement between CCT and CMR for the calculation of left ventricular volumes, although volumes are significantly greater on CCT than on TTE or perfusion scintigraphy. Right ventricular analysis is more challenging owing to its complex geometry, but calculations of right ventricular function compare well with equilibrium radionuclide ventriculography, and volumes assessed using the threshold technique appear to be accurate as compared with CMR. In addition to volumes, ventricular wall motion, thickening, and thickness can also be derived (see Fig. 9.2C ). Measurements of regional wall motion are reasonable when compared with those of perfusion scintigraphy, although the poorer spatial resolution of the latter may explain why these comparisons are not better. In addition to differences in resolution, the use of β-adrenergic blockade before CCT to control heart rate may lead to discrepancies in functional analysis. Although most studies have evaluated ventricular function in patients without ACHD, available data suggest that CT compares well with CMR for the analysis of global and regional left and right ventricular function in those with complex congenital defects.
Cardiac Morphology and Extracardiac Assessment
Although TTE and CMR are widely accepted as the first-line techniques, CCT is often considered because of the ease and rapidity of acquisition. Although axial images are critical for assessment of major vessels, the use of volume-rendered images and the ability to rotate reformatted structures into any plane allows accurate definition of cardiac and vascular anatomy before any planned intervention. The role of CCT in specific conditions is outlined here; fuller descriptions of each condition may be found elsewhere in this book.
Atrial and Ventricular Septal Defects
CCT is able to characterize the location and size of atrial septal defects (ASDs), especially in areas poorly visualized on TTE ( Fig. 9.3A ). Additionally, biventricular size and function may be assessed, along with any associated anomalies such as anomalous pulmonary venous drainage. CCT could be considered prior to device placement for large ASDs with poorly visualized inferior-posterior rims on echocardiography. CCT may be used as a follow-up investigation after surgical or percutaneous ASD closure, either to evaluate right ventricular function or to assess the state of a septal occlusion device. CCT can also provide detailed anatomic information about size and morphologic features of a patent foramen ovale (PFO). The presence of a short PFO tunnel length and septal aneurysms on CCT correlates well with the presence of a left-to-right shunt on TTE. However, CCT is probably less effective at determining the presence of small defects. Just as for ASDs, TTE remains the technique of choice for the detection of most ventricular and atrioventricular (AV) septal defects (VSDs/AVSDs). The high spatial resolution and 3D capabilities of CCT allow straightforward measurement of VSD size and location when there is diagnostic doubt (see Fig. 9.3B and C ).
