Dr. Shi is a fourth-year Resident at Thomas Jefferson University Hospital, Philadelphia, PA. She received her medical degree from Beijing Medical University in 1991, and PhD degree from Brown University in 1998. She will begin a Body MR Imaging fellowship at Beth Israel Deaconess Medical Center in Boston, MA in 2003.

Dr. Mitchell is a Professor of Radiology and the Director of the Magnetic Resonance Imaging Division, Department of Radiology, Thomas Jefferson University Hospital, Philadelphia, PA.

Magnetic resonance (MR) imaging of coronary artery disease has emerged during the past decade because of high-field strength magnets, improved high-gradient hardware, and fast imaging techniques. It provides crucial information on cardiac anatomy and function including wall motion, response to stress, myocardial perfusion, viability, and coronary morphology. This report reviews the current MR applications in ischemic heart disease.

Cardiovascular disease is the leading cause of death in the United States, ¹ accounting for 22% of all deaths. In patients who have myocardial infarction, approximately one-third die from the event. Owing to associated high morbidity and mortality, early detection and treatment is imperative. The mainstream treatments for coronary artery disease are coronary angioplasty, stent, or coronary bypass graft, along with medical management. Clinical decision of coronary angioplasty or bypass surgery relies on information of regional tissue perfusion, viability, and coronary arterial morphology. Traditionally, evaluation of ischemic heart disease requires multiple imaging modalities, including radionuclide stress test, stress echocardiogram, positron-emission tomography (PET), and coronary artery angiography. The stress studies provide functional information such as wall motion abnormality, perfusion defect, and viability of the myocardium, whereas invasive coronary angiography offers information about the coronary vascular flow compromise. In spite of all these tests, distinguishing between viable and infarcted myocardium within the region of abnormal wall motion has remained elusive.

During the past decade, the rapid advances in magnetic resonance (MR) imaging hardware and techniques made MR evaluation of ischemic heart disease feasible today. MR imaging is potentially superior to the traditional modalities, such as echocardiography and single-photon emission computed tomography (SPECT), because of its high spatial and temporal resolution. It allows the most comprehensive anatomical and functional evaluation of coronary artery disease, including ventricular morphology, wall motion, response to stress, perfusion, myocardial viability, and coronary artery stenosis. The purpose of this report is to provide an overview of the current MR imaging techniques and their clinical applications.

Techniques

The greatest challenge of cardiac MR imaging is motion artifact from the heart, adjacent vascular structures, and respiration. The obstacles of cardiac MR imaging have been overcome by cardiac gating, suspension of breathing, or adaptation to diaphragmatic motion, along with the availability of high field strength magnets, advanced gradient hardware, and fast imaging pulse sequences.

Most cardiac MR imaging is currently performed on 1.5 T scanners. Although different coils or coil combinations can be used, a torso or cardiac phase-array coil is generally applied to the chest. Electro-cardiography triggering is essential to ensure data acquisition at the same appropriate point of each cardiac cycle in order to offset cardiac motion and match each image to the desired cardiac phase. Breath-holding is frequently required, although non­breath-holding techniques have been developed that are based on precise monitoring of diaphragmatic motion. ^2,3 For improved breath-holding, nasal cannular oxygen should be supplied to allow longer and steadier breath-hold.

Prescribing imaging planes

Cardiac imaging planes are prescribed using the axes of the heart as references. The main imaging planes are the short-axis, 4-chamber, and vertical long-axis views. After a coronal sub-second T2-weighted single-shot fast-spin-echo localizer, standard axial images through the chest are obtained. Then, an oblique plane is positioned parallel to the ventricular septum to obtain a 2-chamber scout, which depicts the left ventricle and the left atrium. From the 2-chamber scout axial views, slices perpendicular to the long axis of the heart in both planes are used to prescribe the short-axis view. From the short-axis view, the long-axis 4-chamber view can be prescribed by indicating a slice that bisects the left ventricle. Lastly, the vertical long-axis 2-chamber view is prescribed by choosing a vertical plane that bisects the left ventricle.

Pulse sequences

Today many different cardiac sequences are used to evaluate ischemic heart disease. The two main categories are black-blood and white-blood sequences. The black-blood technique is typically an electrocardiogram (ECG)-gated fast-spin-echo technique with double or triple inversion recovery (IR). The optimal soft-tissue contrast with double IR fast-spin-echo sequence is best for evaluation of anatomy. One particularly interesting application is to detect vessel wall atherosclerosis using high-resolution technique before narrowing can be depicted by X-ray angiography. ⁴

The white-blood technique refers to a variety of gradient-echo cine images obtained throughout the cardiac cycle. These sequences are useful for evaluating parameters such as ventricular mass, chamber volumes, and ejection fraction, etc. Recently, balanced steady-state free precession technique (SSFP) was applied to cardiac imaging because of its high temporal and spatial resolution as well as its motion-independent depiction of blood as high signal intensity, allowing a consistently high blood myocardial contrast-to-noise ratio (CNR). ^5,6

The SSFP pulse sequence is designed to maintain the transverse magnetization in a steady state from one repetition time (TR) to the next ⁶ (Figure 1). It uses ultra-fast TR (<3 ms) and echo time (TE) (1.5 to 1.6 ms), with TE exactly half that of TR, which allows fast imaging and real-time evaluation of the beating heart. In comparison with other fast gradient-echo sequences, balanced SSFP imaging has a shorter acquisition time, which renders breath-holding more feasible, which is of particular interest for patients whose respiratory function is impaired.

Myocardial function can be assessed by careful viewing of cardiac motion on cine-loop display. More precise evaluation of transmural motion is facilitated by the use of MR tagging. MR tagging uses a grid of magnetic saturation produced by applying a sequence of radiofrequency (RF) pulses, with magnetic field gradients switched on while they are applied. ⁷ The saturation RF pulses are followed by a bright blood cine-imaging sequence. This creates a low-signal grid of intersecting low-signal intensity lines superimposed on the bright blood images. This grid is deformed as the myocardium moves; therefore, this technique can be used to track myocardial motion. With a high-density grid, regional myocardial contractility can be determined. Postprocessing software may be used to obtain highly accurate estimates of tag displacement--as little as 0.1 mm. ⁸ Tagging with rapid breath-hold imaging allows 3D reconstruction with improved accuracy if orthogonal planes are acquired from a sufficient number of imaging planes. ⁹

New technical advances

Several other important technical advances have been made to accelerate the data acquisition, maintain high image resolution, and eliminate the motion artifact. Spiral imaging samples k-space along trajectories that start at the center of k-space, where the power spectrum is highest, and spiral outward to create high sample density in the center. Spiral imaging is robust for cardiac imaging, particularly coronary angiography, which is attributed to its insensitivity to flow and motion artifact and high data acquisition efficiency. Parallel imaging, ie, sensitivity encoding and simultaneous acquisition of spatial harmonics, uses arrays of simultaneously operated receiver coils to reduce scan time by a factor of 2 or higher without compromising spatial resolution, although at a cost of lower signal-to-noise ratio (SNR). Parallel imaging and SSFP are a good combination because the latter has high SNR and CNR. Using parallel imaging real-time acquisition, it is possible to measure wall motion at rest and stress without ECG or respiratory trigger. ^10,11 For longer acquisitions, navigator technique overcomes limitations of breath-holding. It uses a navigator MR device to track the diaphragmatic position, which is shown to correlate closely with the position of the heart. This can be used to gate coronary MR angiography acquisition without breath-holding.

Magnetic resonance coronary angiography (MRCA)

The early MRCA studies used 2D spoiled gradient-recalled echo with segmented acquisition and breath-holding. ¹² Although a non­contrast-enhanced 2D time-of-flight technique has been used, intravenous injection of gadolinium provides better vessel contrast. Spiral imaging, balanced SSFP, ^13,14 and/or parallel imaging can improve acquisition quality and efficiency. Imaging acquisition of MRCA is performed during diastole when cardiac motion is minimal and coronary arterial flow is maximal.

Evaluation of ischemic heart disease

Myocardial morphology and wall motion abnormality

Myocardial injury due to coronary artery disease can cause ventricular wall thinning, chamber dilatation, and abnormal wall motion. MR imaging offers superior anatomical detail and soft-tissue contrast compared with echocardiography, especially when cine techniques, real-time imaging, and tagging are used.

Dobutamine cardiac MR stress test

Dobutamine or adenosine stress echocardiography and stress nuclear cardiography have been used routinely for evaluating ischemic coronary artery disease by detecting wall motion abnormality during stress.

Dobutamine stresses the myocardium by increasing both heart rate and contractility, thereby inducing ischemia in regional myocardium supplied by stenotic coronary artery. Dobutamine stress echocardiography has been shown to have sensitivity of 72% to 89% and specificity of 83% to 85%. ^15-17 One limitation of dobutamine stress echocardiography is that 10% to 15% of patients have suboptimal or nondiagnostic images, primarily of the basal lateral and inferior segments. The quality of imaging is especially poor in patients who have emphysema or are obese. Additionally, the study interpretation is operator dependent.

The dobutamine cardiac MR stress test applies the same pharmacologic and pathophysiologic principles, but provides improved spatial resolution and tissue contrast. Unlike echocardiography, in which the image quality is limited by the acoustic viewing window, MR images can be obtained with good reproducibility, independent of the patient's condition and the experience of the examiner. Direct comparison of stress echocardiography and stress MR cardiography by Nagel et al ¹⁸ showed that dobutamine MR stress cardiography had improved sensitivity (83%) and specificity (86%) for detecting ischemic heart disease compared with those of stress echocardiography, which were 74% and 70%, respectively.

The protocol for dobutamine cardiac MR stress tests has been described. ¹⁸ To perform the test, dobutamine is infused intravenously from 0 mg/kg per minute to 5, 10, 20, and 40 mg/kg per minute, with 3 minute intervals at each level. At each stress level, multiple MR images can be acquired. The location and orientation of repetitive images are highly reproducible because identical spatial coordinates, rather than visual assessment, are used for repetitive imaging. The patient is monitored with ECG, and blood pressure is checked frequently during the study. The stress test is terminated if the patient has new onset of symptoms including chest pain or dyspnea, decrease in systolic pressure >40 mm Hg, arterial hypertension (blood pressure >240/120 mm Hg), severe arrhythmia, or if the target heart rate of 85% of the maximum predicted heart rate has been reached. In addition, cardiac wall motion is monitored by the physician throughout the examination, and the test is considered positive if wall motion abnormality occurs. Monitoring of ventricular systolic thickening is particularly important because the ST segment is altered by the magnetic field, and monitoring ECG for ischemia is therefore not possible. Atropine may be administered in 0.3 mg increments every 30 seconds up to a total dose of 1.5 mg to augment heart rate. Regional wall motion is assessed in a 16- or 17-segment model as performed by echocardiography. Figure 2 shows hypokinesia at the septal, apical-lateral, and mid-lateral wall induced by dobutamine.

Contrast-enhanced MR cardiography

The most frequently used contrast agents for evaluating enhancement of ischemic or infarcted myocardium are gadolinium chelate, such as gad-opentetate dimeglumine. Gadolinium enhances the proton relaxation, shortening T1. After intravenous administration of gadolinium, normal myocardium will show increased signal intensity on the first pass, and then decreasing signal intensity on delayed images after washout of the contrast. The dose of the gadolinium is usually 0.2 mmol per kilogram body weight.

MR detection of perfusion defect

Clinical decisions regarding revascularization rely on the evaluation of coronary anatomy, regional myocardial tissue perfusion, and assessment of the amount of the ischemic and infarcted myocardium. Single-photon emission computed tomography has played a key role in assessment of perfusion, although it is limited by attenuation artifact and radiation to the patient. Attenuation artifact can be corrected on improved PET imaging, which also allows for the quantification of perfusion; however, this modality is not widely available. MR imaging has been shown to detect the region of perfusion defect during the first pass of contrast, correlating with ischemia or infarction, where there is a slow upslope, decreased peak intensity, or lack of uptake of the gadolinium ^19,20 (Figure 3).

The sensitivity and specificity of first-pass bolus MR imaging for depiction of perfusion defects are 74% to 92% and 87% to 96%, respectively, using results of conventional coronary angiography as a reference. This is superior to the sensitivity of 65% to 82% and the specificity of 75% to 81% found with the scintigraphic studies. ^21,22 The MR imaging perfusion measurements can be performed during rest as well as during maximal hyperemia induced with either adenosine or dipyridamole. MR first-pass measurements can also discern collateral flow in myocardium and are able to identify small changes in myocardial blood flow and myocardial perfusion reserve (the ratio of stress blood flow over resting).

Accurate characterization of regional myocardial perfusion requires repeated imaging of the entire heart during the first pass of gadolinium. Current technology permits 2 to 6 image acquisitions in 1 cardiac cycle, sufficient to track the contrast bolus through the myocardium. Moreover, the spatial resolution (in-plane spatial resolution <3 mm) is sufficient to differentiate between subendocardial perfusion and subepicardial perfusion. In SPECT, however, the wall perfusion defect is an 'all or none' phenomenon, and subendocardial or subepicardial defects cannot be distinguished.

Examination of myocardial viability

Determining the viability of ischemic myocardium is important in patients with left ventricular dysfunction due to coronary artery disease. Ischemic but viable myocardium will likely benefit from coronary revascularization; however, revascularization of nonviable tissue will fail to improve the ventricular function, and surgery bears significant morbidity and mortality.

Single-photon emission CT, stress echocardiography, and PET scan are currently used noninvasively to evaluate myocardial viability. One common limitation of these studies is the inability to detect the transmural extent of ventricular wall viability.

With high spatial resolution, contrast-enhanced MR imaging can detect the necrotic tissue in both acute and chronic myocardial infarction, which correlates with histology in animal study. ²³ On delayed postcontrast MR images, infarcted tissue hyperenhances, reflecting its composition of fibrous tissue and reduced cellularity. In contrast, viable myocardial regions do not hyperenhance, even after severe transient ischemia with concomitant regional dysfunction. ²³ The location and transmural extent of healed Q-wave and non­Q-wave myocardial infarction can also be accurately determined by contrast-enhanced MR. ²⁴ The reported sensitivity for detecting healed infarction by MR was 91% at 3 months and 100% at 14 months, and specificity was 100%. Kim et al ²⁵ showed that regional contractile function recovered in patients who did not have hyperenhancement of the dysfunctional region before revascularization, but did not recover in patients who had extensive hyperenhancement (Figure 4). Furthermore, the transmural extent of hyperenhancement was significantly related to the likelihood of improvement in contractility after revascularization; contractility in myocardium with hyperenhancement of 50% to 75% of regional tissue did not improve after revascularization. Recently Klein et al ²⁶ reported that in severe ischemic heart failure, MR imaging hyperenhancement as a marker of myocardial scar closely agrees with PET data. Although hyperenhancement correlates with areas of decreased flow and metabolism, MR imaging identifies scar tissue more frequently than PET, reflecting the higher spatial resolution.

Although the previously discussed data support the correlation of delayed hyperenhancement and irreversible myocardial injury, other investigators found that hyperenhanced regions had partially reversible dysfunction after acute infarction. ^27,28 Additional research needs to be conducted to elucidate this issue.

MRCA

Invasive X-ray coronary angiography, the current gold standard for assessing coronary patency, is both costly and associated with a low risk of morbidity. Among patients undergoing coronary artery catheterization, only 50% have hemodynamically significant stenosis and are treated with coronary angioplasty, stent, or coronary bypass. Twenty percent of the patients undergoing coronary artery catheterization have normal coronary artery anatomy, and 30% have coronary artery disease but no intervention is performed. ²⁹ A noninvasive assessment of coronary arteries by MRCA would reduce the number of patients undergoing unnecessary coronary catheterization.

The clinical applications of MRCA in ischemic heart disease include coronary lesion detection, assessment of coronary bypass graft patency, and vessel patency after stent placement. The added advantages of MRCA are that an iodinated contrast agent is not required, there is no radiation exposure, and the study can be performed as part of comprehensive MR imaging evaluation of anatomy and ventricular function.

Owing to differences in technique, study design, small sample size from each center, the reported sensitivity and specificity of MRCA are variable, ranging from 60% to 98% for proximal and middle coronary arteries. ³⁰ A large-scale multicenter trial recently showed that MRCA had an overall accuracy of 72% in diagnosing coronary artery disease compared with conventional X-ray coronary angiography (Figure 5). ³¹ Among patients with left main coronary artery or 3-vessel disease, the sensitivity, specificity, and accuracy were 100%, 85%, and 87%, respectively. The reported negative predictive values for any coronary artery disease and for left main artery or 3-vessel disease were 81% and 100%, respectively. This study confirms that noninvasive MRCA provides accurate identification and exclusion of left main coronary artery or 3-vessel disease. However, the accuracy of MRCA for detecting left circumflex artery disease was low, probably related to its small caliber.

Imaging of coronary artery plaque

Acute coronary syndromes often result from rupture of a mildly to moderately stenotic vulnerable plaque. ^32,33 X-ray angiography provides information of luminal diameter, but not the composition of the atherosclerotic plaque. Vulnerable plaques tend to have a thin fibrous cap and a large lipid core, which can be depicted by MR imaging. Using high-resolution black-blood spin-echo technique, coronary vessel wall and plaque morphology can be assessed. ^33,34 This may allow identification of the vulnerable plaques before they rupture.

Limitations

Even though MR imaging can provide more comprehensive examination of coronary artery disease than any other single imaging modality, it has its own limitations. Although the hardware and software are commercially available, all centers do not have the resources to perform cardiac MR imaging. Patients cannot be scanned if they have claustrophobia or metallic implants such as pacemakers, cardioverter-defibrillators, certain types of cardiac valves, ear implants, certain intracranial aneurysmal clips, etc. Optimal imaging quality cannot be achieved in patients who have severe arrhythmia and respiratory compromise.

Conclusion

With fast imaging techniques, noninvasive cardiac MR imaging provides information regarding not only the anatomy of the coronary arteries and cardiac chambers, but also functional information such as wall motion abnormality, myocardial contractility, myocardial perfusion, reaction to stress, and tissue viability. Cardiac MR imaging is expected to become a cost- effective one-stop-shop for the evaluation of coronary artery disease in clinical practice in the near future.