CHAPTER 53 Applications of Cardiovascular Magnetic Resonance and Computed Tomography in Cardiovascular Diagnosis
Cardiovascular magnetic resonance (CMR) and cardiovascular computed tomography (CCT) are increasingly used in the diagnosis and management of cardiovascular disease.1–3 Both of these imaging modalities have overcome similar challenges posed by cardiac and respiratory motion and the demand for high temporal and spatial resolution to enable noninvasive imaging that aids in the diagnosis and management of a variety of cardiovascular disorders. In addition, CMR and CCT have unique capabilities that permit great flexibility, precision, and reproducibility in the acquisition and display of anatomic and functional data that are useful for surgical diagnosis, planning, and follow-up.
CMR imaging is performed with the use of static and dynamic magnetic fields and does not employ any ionizing radiation. Images are generated from induced radiofrequency signals arising from water and fat protons in the body. Differences in proton density, magnetic relaxation times (T1, longitudinal relaxation time; T2, transverse relaxation time), blood flow, and other parameters produce intrinsic signal contrast among tissues. CMR approaches can be broadly classified into spin echo (black blood) and gradient echo (bright blood) sequences, often modified with prepulses. Spin echo imaging is particularly useful for defining anatomic structure and tissue characterization (e.g., fat replacement or iron deposition). Gradient echo techniques can produce single-shot (displaying a single phase during the cardiac cycle) or cine (displaying multiple phases at one level during the cardiac cycle) images. Cine images demonstrate motion of structures (such as cardiac chambers and valves) during the cardiac cycle, permitting qualitative and quantitative assessment of motion. Both spin echo and gradient echo CMR techniques are flow sensitive. Because of the inherent contrast between the blood pool and surrounding tissue, administration of an exogenous CMR contrast agent is generally not required for general evaluation of cardiac anatomy. However, administration of an extracellular magnetic resonance–specific intravenous contrast agent, such as gadolinium–diethylenetriaminepentaacetic acid (Gd-DTPA), enables certain applications such as contrast-enhanced magnetic resonance angiography (CE-MRA) and assessment of myocardial perfusion and viability. Gadolinium induces T1 shortening, which is detected as increased signal in contrast-enhanced T1-weighted images (the signal enhancement is not linearly related to contrast agent concentration). Flow velocity encoding (also known as phase contrast) is an additional CMR modality that enables quantitation of blood flow. This method enables determination of regurgitant fraction and shunt flows.
CCT uses ionizing radiation to generate images based on the attenuation of the tissues of the body. Today, CCT is typically performed with third-generation multislice scanners. These CCT units are arranged with a radiation tube opposed to a series of detectors attached to a gantry that rapidly rotate as the patient is advanced through the scanner. The detector arrays typically obtain 4, 8, 16, 32, 40, 64, 256, or 320 axial slices during a single rotation of the gantry. A higher number of slices permits greater coverage with each rotation of the gantry, allowing shorter imaging times, less contrast material, and potentially less radiation exposure. Cardiac imaging is usually performed with 64-, 256-, or 320-slice machines. Images can be acquired in one of two modes: helical (or spiral) or step-and-shoot. In the helical mode, data are acquired in a helical path as the patient is advanced through the scanner. The speed at which the patient is advanced through the scanner is called the pitch. A high pitch (faster speed) is associated with a lower radiation exposure; low pitch (slower speed) is associated with a higher radiation exposure. Cardiac imaging is generally performed with a relatively low pitch. The radiation exposure can be reduced somewhat by electrocardiogram (ECG) dose modulation, varying the intensity of radiation exposure during the cardiac cycle. With the step-and-shoot mode, imaging is performed without advancing the patient through the gantry. After each acquisition, the patient is then advanced a fixed distance, and images are obtained contiguous to the prior image data set. This is repeated until the entire area to be imaged is covered. Imaging can be completed in one or two steps with 256- and 320-slice machines.
The acquisition of most cardiac images requires gating or triggering during a specific portion of the cardiac cycle. Thus, imaging is usually performed during regular sinus rhythm. The duration of image acquisition with CCT is generally less than 10 to 15 seconds, allowing breath-holding to be used for suppression of respiratory motion. Although breath-holding can also be used for most CMR acquisitions, free-breathing navigator gating is sometimes used for longer image acquisitions. Navigator gating is a CMR technique that identifies the interface between the lung and diaphragm, most commonly at the dome of the right hemidiaphragm. Data about the position of the lung-diaphragm interface from the navigator can then be used for respiratory gating.
High temporal resolution is needed to take advantage of these gating techniques. CMR can typically be performed with very high temporal resolution, with CMR data acquired in less than 40 msec. The temporal resolution of CCT is limited by gantry rotation speed. Typical gantry rotation speeds are 330 to 400 msec per rotation. The use of half-cycle reconstruction allows an image to be acquired in one half-rotation, with an effective temporal resolution of approximately 165 to 200 msec. Intravenous or oral beta-blockers are typically administered for coronary artery CCT to prolong the period of diastasis during which the coronary artery can best be imaged. The temporal resolution of CCT can be improved to approximately 83 msec with dual source technology, in which two sets of radiation tubes and detectors are mounted on the gantry. Each set acquires data over one quarter-rotation, and the data are combined to form a single image. Multicycle reconstruction can also be used to improve the temporal resolution of CCT to approximately 40 msec by acquiring data for a single image over multiple cardiac cycles, but this is infrequently used because of the requirement for a very low pitch and a resultant very high radiation exposure.
The development of high spatial resolution has also been important for CMR and CCT, particularly for coronary artery imaging. Spatial resolution of CMR has been improved with specific imaging sequences that increase acquisition time and with imaging at higher field strengths (e.g., 3 T) and is typically 1 to 2 mm in-plane spatial resolution, whereas the spatial resolution of CCT has improved with the development of smaller detectors. Routine CCT generally has a higher spatial resolution compared with CMR, with isotropic spatial resolution of 0.5 mm. Although this spatial resolution is technically achievable by CMR, scan length would be further prolonged and image quality would suffer from reduced signal-to-noise ratio.
Image contrast for CMR is generally created with specific imaging sequences and prepulses. The use of a contrast agent is generally not required except for specific applications as noted before. CCT does not have the inherent contrast resolution of CMR; thus, iodinated contrast material is typically required for imaging. The timing of administration of the contrast agent relative to image acquisition is critical in producing high-quality images. Imaging is typically performed during passage of contrast material in the ascending aorta and coronary arteries. The correct timing is determined by a small timing bolus given just before imaging or by automated detection of contrast material in the ascending or descending aorta. A similar process is used for CE-MRA.
The advantages and limitations of CMR and CCT complement those of other imaging techniques, such as echocardiography, x-ray angiography, and radionuclide imaging. Compared with echocardiography and radionuclide imaging, CMR and CCT offer superior anatomic scope and spatial resolution. CMR is the reference standard for evaluation of left ventricular cavity size, systolic function, and mass, providing highly reproducible measures and enhancing noninvasive follow-up of disease processes.4 CCT measures of left ventricular cavity size and systolic function compare favorably with CMR.5 In contrast to echocardiography and nuclear imaging, CMR permits unrestricted image acquisition orientation, which can be readily adjusted to particular patient and study requirements. CCT acquisitions are always in the axial plane. Isotropic resolution facilitates post-processing reconstruction in any desired orientation. There is also much advanced post-processing software for CCT. In contrast, echocardiography offers the advantages of portability, lower cost, lack of ionizing radiation, widespread availability, greater ease of patient monitoring, and greater sensitivity for structures with chaotic motion, such as vegetations. Whereas CMR and CCT myocardial perfusion techniques have been shown to provide useful qualitative and quantitative data, they have not yet been clinically validated to provide the diagnostic and prognostic information proven for radionuclide techniques. Further comparisons between CMR, CCT, and other techniques will be made in the following sections dealing with specific types of examinations.
Precautions generally applicable to magnetic resonance imaging are applicable to CMR. Before imaging, all patients must undergo detailed screening for any potential contraindications to magnetic resonance scanning. In addition to general concerns of metallic implants and severe claustrophobia, patients should be screened for the presence of any incompatible material. Excluded devices include some that are relatively common among those with cardiovascular disease, such as pacemakers, retained permanent pacemaker leads, and implantable cardioverters-defibrillators. Bioprosthetic and mechanical heart valves, sternotomy wires, thoracic vascular clips, and intracoronary stents are generally considered CMR safe at field strengths up to 3 T (see www.mrisafety.com), although they may produce local artifacts that reduce image quality.6
Because of bulk cardiac motion during systole and diastole, most CMR protocols require ECG triggering with images composed from data collected during multiple successive cardiac cycles. Despite this, good functional image quality can frequently be obtained among patients with atrial fibrillation,7 although image quality may be impaired among subjects with frequent and irregular premature beats. Among patients with irregular rhythms, non–ECG-gated real-time imaging CMR (which permits real-time image acquisition analogous to two-dimensional echocardiography but at lower spatial and temporal resolutions than those attained with a gated CMR technique) can provide useful information.8 CCT data are generally of poor quality in the setting of atrial fibrillation and frequent ectopy,9,10 although acquisitions with 256- or 320-slice machines over a single heart beat may provide better image quality.
All subjects require appropriate monitoring during the imaging study. Basic monitoring modalities include ECG monitoring for rate and rhythm (CMR magnetic fields distort ST segment appearance, rendering it uninterpretable), intercom voice contact, and visualization (by direct view or camera). For patients requiring greater intensity of monitoring, automated cuff blood pressure monitoring and pulse oximetry can be added.
As noted earlier, CMR generally does not require the use of an exogenous contrast agent. When needed, however, gadolinium-containing CMR contrast agents have a much more favorable safety profile in regard to both nephrotoxicity and anaphylaxis compared with iodinated agents used in x-ray angiography and computed tomography.11 Administration of gadolinium contrast to patients with severe renal dysfunction may result in nephrogenic systemic fibrosis, a very rare but severe disorder that may result in death.12,13 Patients are typically screened with a questionnaire to identify those who may potentially have renal disease and who require a determination of the estimated glomerular filtration rate. Patients with mild renal impairment (estimated glomerular filtration rate of 30 to 60 mL/kg/1.73 m2) can be imaged with a reduced dose of contrast agent. Alternative imaging modalities are usually used in those patients with moderate to severe renal dysfunction (estimated glomerular filtration rate of less than 30 mL/kg/1.73 m2), especially in the setting of dialysis therapy.
Because of the general requirement for iodinated contrast agents, CCT must be performed with caution in patients with diabetes or mild renal insufficiency. Alternative noninvasive imaging methods are preferable in the setting of moderate to severe renal insufficiency unless dialysis has already been instituted. The use of saline, bicarbonate solution, and N-acetylcysteine may reduce the incidence of acute renal failure due to iodinated contrast material14–16 but may not be effective in patients with severe renal dysfunction.17,18
Radiation exposure is a significant consideration in the use of CCT. Typical helical acquisitions with a 64-slice machine are associated with an effective dose of 15 to 21 mSv, which is associated with a nontrivial risk of cancer that is higher in women and in younger patients.19 Dose modulation (reducing the radiation dose during ventricular systole) reduces the effective dose to 5 to 10 mSv.19 Imaging with lower energy radiation reduces the dose an additional 15% to 20%.20 Prospective gating with use of a step-and-shoot mode of imaging is associated with a relatively low radiation exposure of 4 to 6 mSv, but it does not perform imaging over the entire cardiac cycle.21 Increasing awareness of the radiation exposure associated with CCT has led to the adoption of one or more of these techniques at most centers. For middle-aged and younger patients with chronic disorders, potential radiation exposure for the expected multiple tests over their lifetimes should be considered.
CMR and CCT are widely used clinically for the assessment of thoracic aortic aneurysm and dissection. With CMR, the structure of the aorta is delineated by a combination of the following protocol components in the transverse, coronal, sagittal, and oblique planes: (1) ECG-gated spin echo imaging, which reveals the aortic wall with rapidly flowing blood appearing black and thrombus and slowly moving blood appearing gray; (2) ECG-gated steady-state free precession (SSFP) imaging, which produces bright blood images in single-shot as well as in cine acquisitions; and (3) three-dimensional (3D) CE-MRA with a gradient echo acquisition. Temporally resolved CE-MRA is particularly useful to minimize motion artifacts that would otherwise cause nondiagnostic or false-positive results. With CCT, imaging of the aorta involves a larger image acquisition volume to include the entire thoracic aorta and can be performed with or without ECG gating. Gating is preferred for the evaluation of dissection to avoid motion artifacts that can mimic a dissection flap.21
CMR and CCT are also superior methods for identification of true and false thoracic aortic aneurysms. In true aneurysms, the aneurysmal aortic wall is composed of intima, media, and adventitia. False aneurysms represent a contained rupture of the intima and media, with only the adventitia and periadventitial connective tissue limiting the hemorrhage (Fig. 53-1). False aneurysms generally have a narrow “neck” or communication with the main aortic lumen. True aneurysms are more commonly fusiform (bulge aligned along the long axis of the aorta; Fig. 53-2) than saccular (sacklike bulge extending from a side of the aortic wall). CE-MRA reveals the presence and extent of these lesions as well as any associated thrombus. CCT is recommended as the imaging modality of choice for most patients in the assessment of acute disease; 3D CE-MRA is recommended for most patients with chronic disease.22 As with aortic dissection, advantages of CMR and CCT assessment compared with x-ray angiography include the capability to evaluate for associated complications, such as hemopericardium and left ventricular dysfunction. CMR can also assess for the presence of associated aortic regurgitation. After composite graft replacement of the ascending aorta, CMR is useful for detection of postoperative complications, such as leakage or hematoma formation.23,24
Figure 53–1 Ruptured aorta. Non–ECG-gated contrast-enhanced CCT axial image of the aorta at the level of the aortic arch. There is a contained rupture of the ascending aorta with blood and thrombus filling the mediastinum (white arrow) and a hemorrhagic left pleural effusion (black arrowheads).
Figure 53–2 Aortic dissection. Axial images of aortic dissection from the same patient. A, T1-weighted spin echo CMR shows a dissection flap (arrow) in the ascending aorta. Note the increased signal in the false lumen (asterisk) due to slow blood flow. B, ECG-gated contrast-enhanced CCT axial image at the same location as the CMR image. The dissection flap (arrow) is again identified in the ascending aorta.
CMR and CCT, along with transesophageal echocardiography (TEE), are the primary methods used to diagnose and to monitor patients with acute or chronic aortic dissection. Because each of these imaging modalities has high diagnostic accuracy for dissection, the selection among these methods is generally governed by the condition of the patient, institutional access, and local expertise. Both CMR and CCT offer a good combination of sensitivity and specificity (sensitivity above 95% and specificity above 90%) for dissection25 and provide information about involvement of major branch vessels and all segments of the aorta, unlike TEE, which is limited to the thoracic aorta and by the adequacy of acoustic windows (particularly for the segment of ascending aorta anterior to the trachea). All three methods provide useful information about pericardial involvement. CMR can also assess aortic valve integrity, as can TEE. The main disadvantages of CMR in the acute setting are potential obstacles to continuous monitoring and care of an unstable patient during transport and performance of the study and the requirement that the patient remain motionless during the examination. CCT is frequently available in the emergency department and is therefore the most common initial imaging modality chosen to diagnose acute aortic dissection.26 However, CMR is considered the imaging procedure of choice27 for serial follow-up of the medically or surgically treated patient with dissection according to the recommendations of the task force on aortic dissection of the European Society of Cardiology (which have been endorsed by the American College of Cardiology) because of the lack of radiation exposure.28 Follow-up is recommended after hospital discharge at 1 month, 3 months, 6 months, 12 months, and yearly thereafter.28 Accurate interpretation of postoperative images requires knowledge of the surgical procedure performed and the expected range of routine postoperative sequelae, including thickening around the graft and presence of thrombus outside the graft and within the native aortic wrap.29
CCT aortic assessment is typically completed in less than 1 minute and often in less than 30 seconds; CMR aortic assessment can often be completed within 20 minutes. Both can display the location and extent of dissection identified as an intimal flap separating true and false aortic lumens along with sites of intraluminal communication and can readily assess involvement of the aortic root, arch vessels, and renal arteries (see Fig. 53-2). CMR spin echo images may identify relatively bright regions within the true or false lumen attributable to stagnant blood flow or thrombus. Gradient echo images demonstrate flap motion and blood flow in the true and false lumens. 3D CE-MRA is highly sensitive for dissection30 and can be implemented with subsecond temporal resolution to obviate the need for a breath hold (Fig. 53-3).31 Alternatively, SSFP imaging without administration of an exogenous contrast agent can be accomplished within 4 minutes and may suffice (Fig. 53-4).32 Left ventricular function and involvement of the proximal coronary arteries can be assessed by CCT with a gated acquisition or CMR with additional cine and coronary imaging. CMR can also assess the presence of aortic regurgitation (by cine imaging of the left ventricular outflow tract) and the size of the regurgitant fraction (by a phase velocity encoding acquisition at the base of the aortic root).
Figure 53–3 Aortic dissection. The three-dimensional contrast-enhanced CMR oblique image of the thoracic aorta demonstrates a DeBakey classification type 1 dissection (arrows) involving the ascending and descending thoracic aorta.
Figure 53–4 Aortic aneurysm. Non–contrast-enhanced CMR in the transverse (A) and oblique sagittal (B) imaging planes of a patient with a descending thoracic aortic aneurysm (arrow) partially filled with thrombus (arrowhead). Each image was acquired in less than 1 second with the SSFP sequence without the administration of an exogenous contrast agent.
(Courtesy of Tim Leiner, MD, PhD.)
Intramural hematoma can be identified by a localized thickening, frequently crescentic or circular, within the wall of the aorta, interposed between intima and media, with characteristics of an acute or subacute collection of blood.33 Whereas both CMR and CCT can identify intramural hematoma,34 CMR has been found to be superior to CCT in distinguishing acute intramural hematoma from atherosclerotic plaque and chronic intraluminal thrombus.35 Acute hemorrhage appears isointense or more intense compared with the aortic wall on T1-weighted images and displays high signal intensity on T2-weighted images.33,36 In contrast, subacute hemorrhage displays high signal intensity on T1 images and less signal intensity on T2 images. The layer of displaced calcified intima overlying the hematoma generally produces a relatively smooth surface concave to the lumen (crescent shape), which may help distinguish this entity from protuberant, frequently irregularly shaped atherosclerotic plaque.
A sinus of Valsalva aneurysm is often first suspected on transthoracic echocardiography and can be readily visualized by CCT, CMR, and TEE as an enlargement or outpouching (frequently creating a “windsock” appearance) of a sinus of the aortic root. CMR and echocardiography offer the advantage of providing information about blood flow (e.g., rupture of the aneurysmal sinus into an adjacent chamber) in addition to anatomy (Fig. 53-5). Shunt flow through the rupture is generally present both in diastole and in systole. The magnitude of the shunt may be calculated by flow velocity encoding data.
CMR, CCT, and TEE can provide qualitative and quantitative information about the presence, thickness, and distribution of atherosclerotic plaque in the aorta.37–39 Complex plaque is generally defined as protuberant plaque at least 4 mm in thickness or plaque with mobile elements because plaque with these characteristics is associated with embolic risk.40 Plaque thickness can readily be ascertained by CMR or CCT, although overlying mobile elements are best assessed by TEE because of their chaotic motion and relatively small size.41 Ascending aortic plaque is a predictor of adverse cerebral outcomes after coronary artery bypass grafting. Preoperative (CMR or CCT) or intraoperative (TEE) identification of such plaque may prompt alteration in surgical strategy or technique.40,42
Aortic ulceration occurs in regions of atherosclerotic plaque. Penetrating ulcers are described as those breaching the internal elastic lamina with associated hematoma formation in the media.43,44 CMR and CCT images visualize the position and shape of such ulcers and accompanying adjacent intramural hematoma.35,39,41 Penetrating ulcers are also frequently associated with aortic aneurysm formation.44,45 Associated chest or back pain is a significant risk factor for progression to pseudoaneurysm or free rupture.44
Takayasu’s arteritis is a chronic idiopathic vasculitis that primarily affects the aorta and its branches. It is most prevalent among Asian women younger than 40 years.46,47 The aortic arch (or distal aorta) and its branches as well as the pulmonary arteries have characteristically tapered narrowings or occlusions with areas of dilatation. The availability of noninvasive imaging is important among these patients, who require long-term follow-up to guide medical and surgical therapy. These lesions have been traditionally detected by conventional x-ray angiography, but CMR and CCT can accurately display these lesions and also provide information about a vessel wall abnormality.48,49 CMR is generally preferred because of the need for repeated imaging in these younger subjects. Evidence of vessel wall edema is common but does not correlate well with subsequent lesion development.50 Therefore, the current role of imaging in this disease is to identify the characteristic angiographic lesions.49
Both CMR and CCT can readily identify and characterize aortic coarctation, patent ductus arteriosus, and other congenital abnormalities involving the great vessels. The choice is based on local expertise or availability and issues related to the patient’s age and renal function.
Aortic coarctation is characterized by a ridge of medial thickening and intimal hyperplasia along the posterolateral aortic wall. Coarctation most commonly presents just distal to the left subclavian, occurring more rarely just proximal to the left subclavian. Anatomic assessment with CMR or CCT reveals the location of the coarctation and associated collaterals, usually assessed in an oblique sagittal view aligned with the descending and ascending thoracic aorta (Fig. 53-6).51 Typical CMR protocols use spin echo, gradient echo, or CE-MRA techniques.52,53 Additional cardiac lesions that frequently accompany coarctation can also be identified, including bicuspid aortic valve (imaged in cross section) and ventricular septal defect (imaged in a horizontal long-axis or four-chamber view). Imaging is also useful for follow-up after surgical repair or balloon angioplasty,52,54 and routine CMR follow-up has been recommended.55 Potential complications that may be visualized include renarrowing and aneurysm or pseudoaneurysm at the repair site.
Whereas patent ductus arteriosus can usually be identified by transthoracic echocardiography, CMR or CCT may be useful when echocardiographic images are nondiagnostic due to poor acoustic windows.56,57 For both patent ductus arteriosus and atrial or ventricular septal defects, CMR can provide an assessment of the pulmonic-to-systemic flow ratio (Qp/Qs) by applying the flow velocity encoding technique.58
CMR imaging is usually preferred to CCT for evaluation of congenital aortic anomalies because of concerns about radiation exposure in these generally younger patients and the potential need for serial studies in this population.