Applied Radiology

Dr. Chan is an Assistant Professor, and Dr. Williamson is a Cardiovascular Imaging Fellow, Department of Radiology, Stanford University Medical Center, Stanford, CA.

This article is based on the authors' July 2003 presentation at the Second International Radiology Symposium on State-of-the-Art Imaging, organized by the Stanford Radiology Postgraduate Program at Stanford University.

Coronary artery disease (CAD) is the leading cause of death in developed countries. In the United States, CAD accounts for more than 30% of deaths. The economic burden of CAD is estimated to be more than $60 billion per year. ¹ In the past 30 years, great strides have been made in understanding the molecular and genetic underpinnings of this disease. Technologies in diagnostic imaging and therapeutic intervention have also progressed tremendously. Together, these advances help clinicians optimize care of their patients.

Various imaging techniques have been developed to evaluate various aspects of CAD. In particular, the assessment of myocardial viability and cardiac function is important for the following purposes: 1) to determine the optimal interventional and surgical treatment strategies for CAD, including revascularization, scar reduction surgery, cardiac resynchronization with biventricular pacemaker, and cardiac transplant; 2) to optimize medical therapy by monitoring treatment response; 3) to predict long-term outcome; and 4) to assess entrance criteria for clinical trials.

Definitions

The ability of the myocardium to perform a certain level of mechanical work is dictated by the balance between the supply and demand of oxygen and other metabolites. The term ischemia generally refers to the condition in which the oxygen demand of the myocardium exceeds oxygen delivery, defined as the product of perfusion and oxygen content. When oxygen delivery is adequate for the demand at rest but not during stress, the affected myocardium is called myocardium-at-risk . If ischemia worsens to the point that oxygen delivery is inadequate at rest, the myocardium may undergo infarction or hibernation. Myocardial infarction , or cell death, is a series of irreversible steps that begins with apoptosis. The threshold of infarction as a function of ischemia varies depending on the duration of ischemia and previous exposures to ischemic insults, known as preconditioning . Myo-cardium that suffers from a gradual onset of ischemia or episodic but brief instances of ischemia may avoid infarction and instead enter an altered cellular state in which the contractile function is depressed. This is likely a preservation mechanism in an attempt to lower the metabolic demands. The depressed function may recover after revascularization, and the corresponding myocardium is called hibernating myocardium . Hibernating myocardium often coexists with infarcted tissue. The goal of myocardial viability imaging is to identify and to differentiate the two.

Cardiac function is a nonspecific term that generally refers to the pumping ability of the heart to move blood forward in the vascular system. Overall, cardiac function incorporates a number of aspects, including: diastolic function, which describes the ability to fill a ventricle; systolic function, which describes the contractility of the ventricle; cardiac output, which describes the actual amount of blood delivered by the heart; and valvular function, which describes the efficiencies of the valves as flow
regulators. Each of these functions is quantitatively defined by physical parameters. For instance, diastolic function is often estimated by measuring the rate of pressure drop in the ventricle during diastole. Systolic function is commonly evaluated by measuring the ejection fraction. Cardiac output is measured by calculating the average rate of blood flow through the aortic valve. Valvular function is assessed by measuring valvular pressure gradients and calculating regurgitant fractions. Traditionally, cardiac function has been narrowly identified with systolic function and is sometimes used synonymously with ejection fraction. Clearly, the concept of cardiac function is multifactorial. An effective imaging technique should investigate as many aspects of cardiac function as possible.

Imaging methods

Several existing imaging techniques are used to evaluate myocardial viability. Thalium rest-redistribution perfusion scintigraphy is perhaps the best-known and the best-studied method. ² A myocardial region that does not demonstrate thallium uptake on redistribution images is considered nonviable scar. Positron emission tomography (PET) imaging of fluorine-18 fluorodeoxyglucose (FDG) determines metabolically active myo-cardium. ³ A reduction of FDG uptake indicates nonviable tissue. Although, traditionally, these nuclear medicine methods do not resolve cardiac motion, new generations of single-photon emission computed tomography (SPECT) and PET scanners incorporate electrocardiography (ECG)-gating to produce cardiac cine images for volumetric analysis. ⁴

Echocardiography relies on the visual impression of myocardial wall thinning, dyskinetic wall motion, and a lack of increase in contractility under stress as indications of infarcted tissue. ⁵ Echocardiography has excellent temporal resolution, and in patients whose physiques are favorable to ultrasound imaging, subtle changes in wall motion can be detected by experienced operators. Although qualitative and semiquantitative assessments of ejection fraction are routinely performed, echocardiography is intrinsically a two-dimensional (2D) technique in which accurate volume quantification is difficult.

Magnetic resonance imaging (MRI) is a relative newcomer to myocardial viability imaging. ⁶ Interpretation of viability is based on results from two different techniques: wall-motion cine and delayed-enhancement imaging. Wall- motion cine identifies myocardial wall thinning and abnormal wall contraction. Delayed-enhancement imaging identifies infarcted myocardium. Quantification of ventricular volumes and ejection fractions is also possible using wall motion cine images.

MR evaluation of wall motion
and cardiac function

Although cardiac cine was one of the first techniques implemented for MRI, it has progressed tremendously in terms of scan speed, volume coverage, and image quality over the last 10 years. Currently, the method of choice for visualizing myocardial wall motion is the balanced steady-state free-precession (SSFP) technique. ⁷ Different manufacturers implement this technique under different names: FIESTA for GE (GE Medical Systems, Milwaukee, WI), true-FISP for Siemens, and balanced-FFT for Philips (Philips Medical Systems, Best, The Netherlands). For cardiac cine, balanced-SSFP is usually designed to be a breath-held, 2D-multiplanar, cardiac-gated, segmented k-space sequence. It offers T2-weighted contrast, which yields exquisite native contrast between the myocardium and the ventricular blood pool without exogenous contrast material. It is also relatively immune to in-flow effects and spin-saturation, allowing good image quality in any cardiac plane. The excellent delineation of the blood-myocardial boundary permits detailed observation of wall contractility and accurate boundary segmentation for volume analysis. Additionally, ventricular mass can be calculated from the measured myocardial volume multiplied by myocardial density, which is approximately the density of water. A typical protocol includes contiguous short-axis slices covering the ventricles plus the principle long-axis slices. The former slices are needed for volume quantifications, and the latter slices are needed to evaluate the ventricular apices.

Although regional wall contractility can be quantified, it is more commonly assessed qualitatively. A normal contraction of the left ventricle should show uniform thickening and inward motion of the myocardium. A hypokinetic region demonstrates decreased thickening or inward motion compared with other regions. An akinetic region shows an absence of thickening and motion. A dyskinetic region shows an absence of thickening and paradoxical motion during systole and diastole (Figure 1).

MRI can quantify cardiac output using the phase-contrast technique. A number of implementations are available. A non­breath-held, cardiac-gated, respiratory-compensated, short echo time (TE) and short repetition time (TR) 2D-cine approach is preferred, since the act of breath-holding may alter cardiac output. ⁸ Typically, quantification of the left ventricular cardiac output is performed by selecting a plane above the aortic valve perpendicular to the aorta (Figure 2). Velocity is encoded in the through-plane direction. Flow is calculated as the sum of velocities across the vessel lumen. Reversal of flow in the cardiac output time-curve reflects valvular regurgitation. ⁹ The pressure gradient across the aortic valve can be estimated by applying the Bernoulli equation to the peak velocity similar to echocardiography. By combining flow and volume measurements, mitral regurgitation can be quantified by subtracting forward flow volume per cardiac cycle from the ventricular stroke volume. Similar techniques can be employed to determine cardiac output of the right-side of the heart at the pulmonary trunk (Figure 3) and the valvular functions at the pulmonary valve (Figure 4) and tricuspid valve. An average shunt ratio can be calculated from the ratio between the pulmonary output and the aortic output. Thus, MRI alone can quantify many aspects of the cardiac function.

MRI of myocardial infarction

A number of MRI methods, including MR spectroscopy, have been proposed to image infarcted myocardium. ⁶ Currently, the most effective and straightforward means of MR infarct imaging is based on delayed enhancement of the infarcted myocardium after the administration of a gadolinium contrast agent. This process is nonspecific and, in fact, was first observed in computed tomography after the administration of an iodinated contrast agent. ¹⁰ The delayed enhancement is a result of altered pharmacokinetics of increased distribution volume and slowed wash-out of contrast materials in the infarcted region. The most useful MRI technique employs inversion-recovery nulling of the normal myocardial signal, which accentuates the signal intensity of the retained contrast in the infarcted region. The pulse sequence is typically implemented as a breath-held, cardiac-gated, single cardiac phase, segmented k-space, 2D-multiplanar, gradient-recalled echo sequence. Using inversion-recovery nulling, a five-fold difference in signal intensity has been reported between the enhancing and the nonenhancing regions. ¹¹

The spatial resolution of this technique usually suffices to differentiate transmural from subendocardial infarcts. Published experiments used from 0.1 to 0.2 mmol/kg of gadolinium contrast agents. Imaging is typically performed 15 minutes after contrast administration. Using Tc-sestamibi SPECT as a standard of reference, experiments showed that the size of infarction as determined by delayed enhancement is relatively stable from 10 to 30 minutes. ¹² However, the optimal inversion time (TI), typically 200 to 250 msec, increases with contrast delay time. ¹³ This is because the contrast material in the normal myocardium is slowly being washed out, thus increasing the effective T1 value. If the inversion time is not chosen properly, the normal myocardial signal may not be adequately suppressed, and, in the worst case, the delayed-enhancement signal is suppressed in error (Figure 5).

Interpretation of viability

In chronic infarction, myocardial remodeling and subsequent scarring results in regional thinning of the myocardium. This process may take up to 4 months after the initial myocardial infarction to complete. A thinned myo-cardial wall of <5.5 mm at end-diastole together with an absence of wall thickening during systole correlates with decreased FDG uptake on PET imaging, indicating scar. ¹⁴ In a study of 43 patients with chronic infarcts, a wall thickness <5.5 mm at end-diastole had a negative predictive value of 90% for recovery after revascularization. In contrast, the positive predictive value for recovery in patients with wall thickness >5.5 mm was only 62%. ¹⁵ Therefore, the presence of wall thinning is a good indicator of chronic scar, while an absence of wall thinning alone does not necessarily indicate viable myocardium.

In animal studies, the area of delayed enhancement has been shown to correlate closely with areas of infarcts. ¹⁶ In human studies, delayed enhancement is found to correlate with PET and SPECT findings of infarction. ¹⁷ Kim et al ¹⁸ examined 50 patients with MRI before and after revascularization procedures and stratified their results according to percentage of wall thickness show
ing delayed enhancement. An absence of delayed enhancement in segments that exhibit abnormal contractility has a positive predictive value of 78% for recovery after revascularization. A <25% delayed enhancement has a positive predictive value of 71%. In contrast, the presence of transmural delayed enhancement has a negative predictive value of 98% for recovery. The presence of >50% delayed enhancement has a negative predictive value of 92%. An increasing extent of dysfunctional contraction but viable myocardium before revascularization correlates with greater improvement in ejection fraction after revascularization.

In our experience, chronically infarcted walls, as indicated by end-diastolic wall thinning (<5.5 mm) and akinesia, can show variable delayed enhancement. We consider thin, akinetic, or dyskinetic walls to be nonviable. As discussed above, this interpretation is 90% predictive for nonviable tissue. Walls thicker than 5.5 mm but showing a >50% delayed enhancement are also considered nonviable (Figure 6). This interpretation is 92% predictive for nonviable tissue. This group likely represents more acute infarction when remodeling is not complete. Walls thicker than 5.5 mm and showing a <25% delayed-enhancement are considered potentially viable (Figure 7). This prediction is not perfect, however, since up to 30% of these segments will not improve functionally after revascularization.

Conclusion

MRI is a valuable tool for myocardial viability determination and functional assessment. It has a high confidence (>90%) in the determination of nonviable territories. Its ability to predict recoverable function after revascularization procedure is good (70% to 80%), although not perfect. Performance of delay enhancement imaging requires meticulous attention, especially with regard to the choice of the inversion time. MRI is also capable of quantifying many aspects of cardiac function, including ventricular volumes, ejection fraction, cardiac output, shunt ratio, valvular pressure gradients, and regurgitation fractions.