Assessment of anatomy is done with one or more of the following techniques: ECG-gated multislice spin-echo (black blood images); breath-hold single-slice or multislice turbo spin-echo (black blood); balanced SSFP (white blood); and noncontrast- and contrast-enhanced three-dimensional (3D) MRA. The latter technique is usually applied for the evaluation of anomalies of the great arteries, pulmonary veins, and surgical shunts.

The techniques used in the evaluation of right and left ventricular function are generically called cine MRI. There are a multitude of cine MRI sequences; the most frequently applied one now is a balanced SSFP sequence and realtime cine MRI. Because of optimized homogeneous contrast between cavitary blood and myocardium, balanced SSFP sequences (balanced fast field-echo [FFE]; true fast imaging with steady precession [FISP]; and fast imaging employing steady-state acquisition [FIESTA]) are now preferred. The so-called real-time sequence may be advantageous in infants and young children who cannot hold their breath.

Cine MR images are used to quantify ventricular volumes and function. Right and left ventricular volumes are indexed to body surface areas (EDV/m² and ESV/m²). A distinct advantage of MRI compared to echocardiography is its precision and reproducibility for quantifying right ventricular (RV) volume and function. For ventricular volumetrics, images are acquired in the cardiac short-axis plane.

Cine MR images are also used to evaluate valvular function. Planes approximately parallel to the valve leaflet or cusps can demonstrate valve motion, such as the horizontal long-axis plane for assessing mitral valve motion and RV outlet plane for visualizing infundibular and pulmonary valvular stenosis. The high-velocity jet caused by valvular stenosis and regurgitation may be identified on cine MR images as a flow void (Fig. 29.1). However, flow voids may not be discernible if the echo time (TE) is less than 6 milliseconds on GRE cine and on balanced SSFP cine. On the latter cine, high flow velocity may cause increased signal.

Quantification of blood flow is done using velocity-encoded (VEC; phase-contrast) MRI. In phase-contrast imaging, the signal intensity signifies the velocity of blood at each pixel. ECG-gated phase-contrast images can be produced in which each image shows the velocity at a different time in the cardiac cycle. Since these are cine images, they can be referred to as VEC cine MRI. A region of interest drawn around a blood vessel will give the mean velocity within the vessel at that point in the cardiac cycle. The cross-sectional area of the vessel can be multiplied by the spatial mean velocity to obtain the flow in the blood vessel.

VEC cine MRI is used to measure valvular regurgitant volume (Fig. 29.2); differential flow in central pulmonary arteries; systemic to pulmonary shunt flow; flow through conduits (Rastelli and Fontan conduits); and collateral flow in coarctation. By determining the voxel with the peak velocity just beyond a stenosis, this technique can estimate the peak gradient across a valvular or conduit stenosis.

Echocardiography provides essential diagnostic information in most left-to-right shunts. MRI is used for a few specific purposes. The major indications are for the following suspected lesions: Sinus venosus atrial septal defect (ASD), partial anomalous pulmonary venous connection (PAPVC), and supracristal ventricular septal defect (VSD). VEC cine MRI may also be employed to measure the systemic-to-pulmonary flow ratio (1,2,3 and 4). A new indication in patients with ASD is to provide precise measurement of the defect and its location for planning of transcatheter ASD closure device placement (5,6).

These terms denote a spectrum of abnormalities that have in common an abnormal septation between the atria and ventricles (13). They are also called endocardial cushion defects because the defects are considered to be abnormalities of the embryologic endocardial cushions, which grow together in the center of the heart and divide the atria from the ventricles.

In the normal heart, the atrioventricular septum separates the right atrium from the left ventricle. The atrioventricular septum lies between the more apically located normal tricuspid valve and mitral valve (Fig. 29.15). In all cases of atrioventricular septal defect, the tricuspid and mitral valves lie at the same level, and the atrioventricular septum is defective. This abnormal relationship can be shown on MRI in the transaxial or horizontal long-axis plane (14). In the mildest form of atrioventricular septal defect, the shunt is from the left ventricle to the right atrium, which can be evident on cine MRI. In other cases, the atrial septum adjacent to the atrioventricular valve orifice may also be absent, resulting in an ostium primum ASD (Fig. 29.16). Some patients have an inlet VSD, usually located in the same axial image as the atrioventricular valve or valves. In the most severe cases, both the atrial and ventricular portions of the septum around the valve origins are absent, a condition referred to as complete atrioventricular canal. This creates a common

atrioventricular valve orifice with continuous, common atrioventricular valve leaflets (Fig. 29.17).

MRI is the procedure of choice for definitive diagnosis and assessment of the severity of coarctation. MRI has been shown to be effective for preoperative assessment of coarctation and for postoperative evaluation of recurrent or persistent hypertension (15,16,17,18 and 19). To evaluate coarctation, spin-echo or cine MR sequences are obtained in the axial and oblique sagittal planes. The diameter of the narrowing can be precisely measured with MRI, especially with the use of thin (3-mm) sections through the center of the stenosis. Thin oblique sagittal images through the long-axis plane of the aorta show the diameter of the stenosis and provide an accurate measurement of its length (Fig. 29.18).

The oblique sagittal images also display the dimension of aortic isthmus (region between the left subclavian artery and the ligamentum arteriosum) and the aortic arch. In some cases, a single 3-mm slice in this plane may not display the coarctation and arch because of tortuosity of the arch and proximal descending aorta, but the assessment can be done from adjacent images. Gadolinium-enhanced 3D MRA can display the entire thoracic aorta on a single image with the use of maximum intensity projection or volume-rendering reconstruction techniques (Fig. 29.19). Gadoliniumenhanced MRA is also effective for demonstrating collateral vessels (Fig. 29.20).

VEC cine MRI can be applied to demonstrate the presence of and estimate the volume of collateral circulation to the descending aorta below the coarctation site (15,16) (Fig. 29.21). This is accomplished by using two imaging planes perpendicular to the aorta; one is about 2 cm beyond the coarctation and the other at the level of the diaphragm. In the normal aorta, volume flow is greater (about 5% to 7% higher) at the proximal site. On the other hand, in a hemodynamically significant coarctation, volume flow is greater at the diaphragm because of retrograde flow through the intercostal and mammary arteries and other collaterals

into the distal aorta. The presence of greater volume flow at the diaphragmatic level is considered a functional indicator of hemodynamic significance of the coarctation. There is a rough linear relationship between the percentage of stenosis and the volume of collateral circulation (15,16). After stenting of the coarctation, VEC cine MRI has demonstrated reversal of the collateral flow since volume flow at the proximal site becomes greater than at the distal one.

Theoretically, VEC cine MRI can also be used to estimate the gradient across the coarctation (17,20). Using planes perpendicular and parallel to the coarctation, peak velocity of flow can be estimated. Applying the modified Bernoulli formula (peak pressure gradient = 4 × [peak velocity]²), the peak gradient is estimated. The voxel(s) with highest velocity in systole can be identified on phase images in both planes. The highest in either plan is considered the peak velocity (Fig. 29.22).

There are numerous types of arch anomalies resulting from abnormal resorption of the anterior or posterior segments of the embryonic double-arch configuration. However, the only ones encountered with any frequency are complete (patent) double arch, double arch with atretic posterior component of the left arch (Fig. 29.23), right arch with aberrant (retroesophageal) left subclavian artery (Fig. 29.24), and left arch with aberrant right subclavian artery. The first three produce complete vascular rings that narrow the trachea and esophagus. Compression of the trachea is by the aortic component situated between the vertebral body and the esophagus. The vascular ring with right arch and aberrant left subclavian artery is completed by the posterior component that is tethered anteriorly by a left-sided ligamentum arteriosum. With mirror-image right aortic arch, the ligamentum courses between the left innominate artery and the proximal left pulmonary artery. On the other hand, with nonmirrorimage right aortic arch, the ligamentum courses between the descending aorta and the proximal left pulmonary artery. At the site of attachment of the ligamentum, there is an enlargement or diverticulum of the descending aorta. This localized dilatation tethered anteriorly by the posterior left-sided ligamentum is the structure that compresses the trachea and esophagus rather than the left subclavian artery itself. These anatomic arrangements can be readily discerned on MR images in the transaxial plane.

For the evaluation of arch anomalies, spin-echo images are acquired in sagittal (5-mm slice thickness) and transaxial (3-mm slice thickness) planes (see Fig. 29.23). These images display the aortic anomaly and verify that the anomalous component produces airway compression (21,22). Cine MR images may be used to display the pulsatile nature of the airway obstruction. In some cases, the site of maximal airway compression is at the carina or even involves a proximal bronchus.

Contrast-enhanced 3D MRA may be used to demonstrate the arch anomaly. The MRA acquisition is done in either the sagittal or coronal planes.

Pulmonary arterial anomalies are evaluated using spin-echo and cine MRI (Fig. 29.25) and contrast-enhanced 3D MRA (Fig. 29.26) for depicting morphology and VEC (phasecontrast) MRI for measuring blood flow. Spin-echo and cine MR images are acquired in the transaxial plane followed by images along the long axis of the right (oblique coronal plane) and left (oblique sagittal plane) pulmonary arteries. Contrastenhanced 3D MRA is usually acquired in the coronal plane; a short acquisition period is optimal to depict the pulmonary arteries before enhancement of the left heart and aorta. VEC cine sequences are obtained perpendicular to the long axis of the right, left, and main pulmonary arterial segments.

Compared to echocardiography, MRI is more useful for examining the right and left pulmonary arteries, as well as the lobar and segmental arteries. MRI is especially useful for demonstrating central and peripheral pulmonary artery stenoses (see Fig. 29.27), absent pulmonary artery (Fig. 29.28), and pulmonary artery sling (Fig. 29.29). MRI and MRA are the preferred techniques for the evaluation of residual pulmonary arterial stenoses after repair of tetralogy of Fallot (TOF) (Fig. 29.30).

The hemodynamic significance of pulmonary arterial stenoses is assessed by VEC cine MRI measurement of blood



flow separately for each pulmonary artery (2,23) (Fig. 29.30). Since flow is also measured for the main pulmonary artery, the values can be expressed as the percentage of total pulmonary blood flow to each lung. The measurement can be done before and after angioplasty to document therapeutic benefit, although such measurements may not be possible in the presence of stainless steel stents.