Cardiac Magnetic Resonance

Magnetic resonance imaging is without doubt one of the most sophisticated and versatile noninvasive diagnostic tools. With the use of electrocardiogram (ECG)–gated acquisition coupled to dedicated surface coils, high-resolution static and moving images of the heart and cardiovascular system may be obtained. A powerful helium-cooled superconductor generates a high-strength magnetic field in which nuclei with unpaired numbers of protons and electrons are forced to behave like small magnets, aligning their poles in the direction of the magnetic field (T1 or longitudinal relaxation) and precess at a frequency proportional to the magnetic field strength (T2 or transverse relaxation). The most common magnetic field strength used for cardiac applications is 1.5 Tesla (1500 times the earth’s magnetic field strength).

To generate images, surface coils emit sequences of radio signal pulses that excite the nuclei to briefly flip out of alignment, while receiving coils measure the magnetic signals produced by the nuclei as they return to their baseline state of relaxation. The magnitude of the signals received depends on the magnetic field strength, the concentration of odd nuclei (primarily the hydrogen ion [H⁺]), and the T1 and the T2 relaxation times, which are tissue-specific. Paramagnetic contrast agents, such as gadolinium compounds, specifically alter the T1 and T2 times and are used to highlight the blood pool and areas of fibrosis or edema, where they gradually accumulate. Field magnetic gradients permit identification of the spatial location of each signal being received, allowing acquisition of images in any location and plane within the center of the magnet. Different pulse sequences are designed to acquire images that are preferentially based on the distribution of tissue. Cardiac images may be acquired from a single heartbeat and displayed in near-real time or derived from a series of consecutive heartbeats acquired during a 20- to 40-second breath-hold or respiratory gating. The spatial and temporal resolution of CMR images depends on the magnitude of the magnetic field and the acquisition time. Therefore, real-time imaging is done at low resolution and is limited to specific applications or to patients who have severe heart rate variability and/or are unable to comply with breath-holding.

A particularly useful CMR application called phase-contrast velocity–encoded mapping may be applied to interrogate flow across the cardiac valves. After excitation, moving targets produce a signal that is proportional to the velocity at which they travel through an imaging plane. Flow may be calculated as the product of area times velocity within a conduit, and pressure gradients across valves may be estimated, as in echocardiography, with use of the Bernoulli equation.

A typical CMR study performed for the evaluation of valvular disease may include 30 to 60 acquisitions obtained from different imaging planes and using varying pulse sequences. Thus, a complete study may take 30 to 90 minutes, making the examination of clinically unstable patients difficult.

Cardiac Computed Tomography

Electron-beam computed tomography (EBCT) and multidetector computed tomography (MDCT) allow ECG-gated images to be obtained with sufficient temporal and spatial resolution to visualize the beating heart. The most common applications of cardiac CT are evaluation of coronary calcification and detection of coronary stenosis. Both EBCT and MDCT systems basically contain an X-ray source and a detector array of crystals that receive the X-rays after traveling through the body and convert photons into electrical impulses. Unlike plain fluoroscopy, in CT imaging the x-ray data need to be collected for a minimum of 180 degrees. From the precise spatial and temporal location of each x-ray photon detected, a three-dimensional (3D) data set is reconstructed using filtered back-projection. In addition, CT imaging requires ECG-gated temporal registration to match data collected at the same time during the cardiac cycle, given the continuous motion of the heart.

With EBCT, a rapidly rotating electron beam generates x-rays, which are electronically steered to sweep over tungsten targets arranged in a stationary semicircular array under the patient table. Even though this system has the advantage of high temporal resolution (33 to 50 ms) for each slice acquisition because of the lack of rotating parts, it has limited speed to cover the z-axis (moving from one slice to the next) and has limited x-ray power, producing poor-quality images in large patients. MDCT imaging involves a rotating x-ray source that emits a fan-shaped beam of x-rays that passes through the body and is detected by a moving detector array panel set at the opposite end of a doughnut-shaped gantry. Coverage in the z-axis is obtained by moving the table through the gantry during image acquisition in a “spiral” or “helical” movement. The superior spatial resolution and x-ray power have established MDCT as the preferred CT technology for cardiac imaging. Nevertheless, temporal resolution is limited because of the weight of the moving components (100 to 250 ms for single-source and 50 to 100 ms for the newer dual-source scanners). Moreover, radiation exposure may be considerable (8 to 25 mSv), and the use of intravenous iodine contrast material is required, limiting the utility of this method for routine cardiac valve evaluation. In selected patients, however, CT may provide very useful information, such as the extent of valvular calcification and dimensions of stenotic orifices.

A typical CT study has one to three data sets (noncontrast, contrast-enhanced, and postcontrast) acquired over a 5- to 15-second period each. Each set contains data acquired over multiple sequential cardiac cycles. Each cardiac cycle corresponds to a different series of craniocaudal axial images. Thus, significant changes in heart rate may produce misregistration artifacts that may be relevant to accurate visualization of the cardiac valve anatomy.

The Left Ventricle

In patients with chronic valvular disease, changes in left ventricular (LV) size, mass, and function are critical for establishing the timing of surgical intervention. CMR is a superior method for quantifying LV volumes, ejection fraction, and mass. Cine (bright-blood) ECG-gated two-dimensional images may be obtained in any of the conventional two-chamber, three-chamber, or four-chamber long-axis or cross-sectional short-axis planes. For the purpose of quantification, a stack of 10 to 20 contiguous 5- to 10-mm-thick short-axis images are acquired from the base to the apex of the left ventricle, during breath-holding. ¹ These images may be then traced offline in a dedicated computer workstation. Volumetric quantification is based on the Simpson rule, for which the volume of the left ventricle equals the sum of the 10 to 20 cross-sectional discs. LV mass is calculated by subtracting the endocardially from the epicardially traced volumes and multiplying by a muscle weight–specific constant. CMR yields more accurate values than angiographic or echocardiographic planar imaging methods, especially when the ventricular shape deviates from the normal geometric model, because in the latter methods, fewer images are used and most of the data are interpolated.^2,3 In addition, CMR is superior for visualization of the true LV apex, which is often outside the acoustic echocardiographic windows. Several studies have validated CMR measurements of LV volumes, mass, and function using quantitative reference techniques, such as indicator dilution, radionuclide angiography, and studies of postmortem examination.^4–6

CMR is the best method for longitudinal follow-up of LV mass and volumes to determine the effect of therapeutic interventions owing to its excellent interstudy reproducibility with an intertest variability less than 5%.7–10 CMR has been used to evaluate sequential changes in LV and right ventricular (RV) geometry in response to chronic mitral regurgitation. ¹¹ CMR is been used increasingly in clinical trials, allowing reduction in the sample size required to detect a difference between treatment arms.^12–14

CMR is also useful for calculating LV wall stress using measurements of wall thickness and chamber dimensions in conjunction with cuff blood pressures and a carotid pulse tracing. Increased wall stress in patients with chronic valvular regurgitation has been shown to predict adverse events and deterioration of LV function. ¹⁵

CT also allows accurate quantification of LV volumes, ejection fraction, and mass.16,17 In fact, early studies suggested that EBCT was a reference standard for evaluation of other methods of calculating LV volumes in patients with valvular heart disease. ¹⁸ The main advantage of CT is its isotropic, submillimeter spatial resolution, which allows easy, direct 3D volumetric assessment, in many cases with the use of automated border detection algorithms. However, the limited temporal resolution of CT may result in overestimation of the true end-systolic volume, particularly in patients with elevated heart rate. Moreover, the rapid infusion of contrast material and use of beta-blockers may result in slightly larger LV volumes and reduced ejection fractions.

In many patients with valvular disease, in particular those with AS, diastolic dysfunction is an important contributor to symptoms and prognosis. CMR has been used to evaluate diastolic filling on the basis of regional myocardial diastolic velocities and strain. ¹⁹ In addition, the 3D pattern of LV filling and pulmonary venous flow has been evaluated with phase-contrast velocity–encoded data.^20,21 In patients with LV hypertrophy due to AS, CMR-determined early filling and atrial contraction velocities have been shown to correlate with Doppler echocardiographic measurements and have shown a higher atrial contribution to ventricular filling in the patients than in control subjects. ²²

The Right Ventricle

CMR is the most reliable way to assess regional and global RV function quantitatively.23,24 Furthermore, CMR may give an unrestricted view of the RV outflow tract. The end-diastolic and end-systolic velocities are calculated by the manual drawing of endocardial contours at end-diastole and end-systole, respectively, on cine loops, oriented in the axial plane or along the short axis of the left ventricle. Ejection fraction and RV mass are determined in a fashion similar to that used in the left ventricle. Several studies have validated CMR volume measurements of dimensions and function of the right ventricle using quantitative catheterization, postmortem, and radionuclide techniques.^25–28 Thus, this approach offers the potential for more reliable evaluation of RV size and function in patients with valvular disease than with echocardiography. Assessment of RV volumes and function is important in patients with severe tricuspid or pulmonary regurgitation, including assessment after repair of tetralogy of Fallot. ²⁹ In these patients, RV function is an important determinant of long-term prognosis after surgical correction. ³⁰

CT may also be used to evaluate RV volumes and ejection fraction. In order to visualize the RV endocardium, the contrast injection protocol is modified, with use of a biphasic or triphasic contrast injection, depending on what additional information is sought in the study. CT may be useful to evaluate RV size and function in patients with poor acoustic transthoracic echocardiography (TTE) windows and in those with contraindications to CMR, such as implanted left-ventricular assist devices. ³¹

Aortic Stenosis

Echocardiography remains the most useful modality for the evaluation of AS. However, in many patients, the transvalvular gradient determined by continuous-wave Doppler echocardiography may underestimate stenosis severity because of an inadequate Doppler angle of interrogation or reduced cardiac output. The continuity-equation aortic valve area (AVA) may also be inaccurate owing to inadequate measurements of LV outflow tract dimensions or incorrect sampling of pulsed Doppler velocities. ³⁵ Direct planimetry ( Figure 8-1) of the aortic orifice is another method for determination of stenosis severity, but the accuracy is limited by TTE and even by transesophageal echocardiography (TEE) because of the acoustic shadowing produced by heavy leaflet calcification.