Nuclear Medicine Imaging of Myocardial Perfusion

Published on 26/02/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 2751 times

CHAPTER 54 Nuclear Medicine Imaging of Myocardial Perfusion

Olga James, Kenneth J. Nichols, Bennett Chin

Radionuclide myocardial perfusion imaging (MPI) is a well-established, highly accurate, and reproducible noninvasive method to diagnose and assess functionally significant coronary artery disease (CAD). Extensive development and validation have brought this modality to the forefront of quantitative, noninvasive assessment of CAD. High accuracy and reproducibility have also contributed to its widespread adoption as a gold standard to quantify myocardial ischemia. MPI also provides important information regarding myocardial viability and prognosis.

The basic principles of MPI have been validated in animal models and clinical trials over the past several decades.¹^–⁴ Accepted clinical guidelines for use, technical aspects of quality control, and protocols for MPI stress testing and imaging for positron emission tomography (PET) and single photon emission computed tomography (SPECT) have been summarized in joint statements from the American Heart Association, American Society of Nuclear Cardiology, and American College of Cardiology,⁵ and the European Association of Nuclear Medicine and the European Society of Cardiology.⁶ This chapter highlights and summarizes the current clinical aspects of MPI. Emphasis is placed on describing the rationale and principles of MPI. The authors acknowledge variations in clinical practice based on local expertise and preference; however, space limitations do not permit a detailed discussion of all issues.

The first section reviews current clinical MPI using SPECT. Topics include radiotracers, instrumentation, procedures, and data analysis, including quantification and interpretation. Test performance and prognosis are reviewed under the general topic of interpretation. Assessment of myocardial viability with MPI also is briefly summarized (this is discussed more extensively in Chapter 55). Gated MPI to assess ventricular function is discussed in Chapter 56. Planar imaging produces lower image contrast and is reserved for special circumstances such as claustrophobic patients or patients unable to undergo SPECT imaging. An extensive review has been published.⁷ The second section reviews MPI with PET. The same general outline is followed in an abbreviated format with salient differences discussed between PET and SPECT.

MYOCARDIAL PERFUSION IMAGING WITH SPECT

Technical Aspects

Radiotracers

SPECT radionuclides in widespread clinical use include thallium 201 and Tc 99m radiotracers. Thallium 201 is a cyclotron-generated, monovalent cation with biologic properties analogous to those of potassium (K⁺). After intravenous administration, thallium 201 rapidly diffuses from the blood pool into the extravascular space where it is highly extracted by myocytes via the Na⁺,K⁺-ATPase pump. The initial myocardial uptake of thallium 201 is proportional to myocardial blood flow (MBF) and extraction fraction. Thallium 201 has the advantage of a high extraction fraction (approximately 85% to 88%), which is maintained up to MBF of approximately 2.5 mL/min/g. Myocardial territories distal to physiologically significant stenoses (i.e., ischemic myocardial regions) are unable to compensate sufficiently at stress with increased blood flow, and a perfusion defect can be detected as a relative reduction in radiotracer activity in this territory.

After administration, thallium 201 begins to redistribute significantly as it equilibrates with the extracellular concentration. In ischemic territories of lower MBF at rest, lower washout of radiotracer also contributes to equilibration. This more uniform myocardial radiotracer distribution at delayed (typically 4 hours) time points (“redistribution”) is deemed “reversible” compared with stress perfusion, and a functionally significant coronary stenosis can be noninvasively diagnosed. Previous studies have verified that administration of a small reinjection of thallium 201 of 37 MBq (1 mCi) at rest improves the detection of viable myocardium.⁸ This is a widely accepted clinical procedure to enhance the detection of viable myocardium.

The major disadvantages of thallium 201 are related to its physical properties compared with the properties of Tc 99m radiotracers. The long physical radiotracer half-life (73 hours) limits the administered activity to 148 to 167 MBq (4 to 4.5 mCi) per day because of radiation dose considerations (148 MBq [4 mCi] of activity delivers 24 mSv [2.4 rem] of whole body radiation dose equivalent). Because of radiation dose limitations, low photon flux contributes to lower true count images compared with Tc 99m radiotracer studies. The photopeak of thallium 201 (69 to 83 keV; 95% abundance) is also relatively low; this increases the amount of radiation absorbed and the scatter fraction. Also, spatial resolution and energy resolution of gamma camera imaging is better for the higher 140 keV gamma ray energy of Tc 99m than for the lower energies of thallium 201. The combination of these effects, especially in large patients, adversely affects thallium 201 image quality and can make interpretation difficult.

Tc 99m-sestamibi and Tc 99m-tetrofosmin are the most commonly used clinical MPI radiotracers. Common properties include high lipophilicity, relatively high first-pass extraction, and insignificant radiotracer redistribution. Conceptually, they can be viewed as radiotracer “microspheres” that define MBF at the time of administration with no significant washout or redistribution at delayed imaging. Their first-pass extraction fractions are lower than that of thallium 201. Their linear relationships with blood flow are maintained up to MBF rates of approximately 2 mL/min/g. Theoretically, this can result in reduced sensitivity for detection of CAD; however, large clinical studies have shown similar overall accuracies compared with examinations performed with thallium 201. These radiotracers enter myocytes by passive diffusion, the diffusion rate of which is proportional to the regional myocardial flow. They are then sequestered in the mitochondria because of the electrochemical gradient, which is maintained in viable myocytes.

An advantage of Tc 99m radiotracers is the short half-life (6.02 hours). This short half-life permits a higher administered activity (approximately 8 to 10 times higher; 14 mSv [1.4 rem] for 1.11 GBq [30 mCi] activity), higher photon flux, and, consequently, higher count images compared with thallium 201. In addition, the higher photopeak of 140 keV results in less soft tissue attenuation, less scatter, and improved spatial resolution. These factors contribute to improved image quality compared with thallium 201, and reduce the time required for image acquisition. This is particularly important because ECG gating with Tc 99m radiotracers is commonly performed. Reduced imaging time also reduces patient discomfort and the likelihood of motion artifacts, which may compromise image quality. Tc 99m is also commercially available from generators, which are widely available at a low cost.

Tc 99m-sestamibi is a lipophilic cationic compound. Tc 99m-tetrofosmin is a diphosphine lipophilic cationic complex of Tc 99m. Tetrofosmin is reported to have faster clearance from the liver and lungs; however, the clinical significance with respect to improved diagnostic accuracy has yet to be established. A large retrospective study has shown that Tc 99m-tetrofosmin scans are essentially equivalent to Tc 99m-sestamibi in determining prognosis in high-risk patients.⁹ Examples of normal Tc 99m-sestamibi studies are shown in Figures 54-1 and 54-2. A gated study showing the utility of aiding in identifying a breast attenuation artifact is shown in Figure 54-3.

FIGURE 54-1 Normal Tc 99m-sestamibi exercise MPI shows inferior diaphragm attenuation in a male patient. Decreased activity inferiorly is typically mild, gradually increasing in severity from apex to base (arrows). BP, blood pressure; HLA, horizontal long axis; MPHR, maximum predicted heart rate; SA, short axis; VLA, vertical long axis.

FIGURE 54-2 Normal Tc 99m-sestamibi exercise MPI shows anterior breast attenuation in a female patient. Decreased activity anteriorly is typically mild, may gradually increase in severity from apex to base (arrows), and may not follow the typical coronary distribution of a perfusion defect in the left anterior descending artery territory. BP, blood pressure; HLA, horizontal long axis; MPHR, maximum predicted heart rate; SA, short axis; VLA, vertical long axis.

FIGURE 54-3 Normal Tc 99m-sestamibi exercise MPI shows anterior breast attenuation in a female patient. The rest gated study shows normal wall motion and normal wall thickening anteriorly. The rest perfusion study (ungated, second panel, left column) shows mild decreased activity anteriorly that does not involve the apex (noncoronary distribution for a left anterior descending artery stenosis). The stress perfusion study (ungated, second panel, right column) shows no perfusion defect on stress imaging. ED, end-diastole; ES, end-systole; HLA, horizontal long axis; SA, short axis; VLA, vertical long axis.

Instrumentation

Clinical gamma cameras used in SPECT have until more recently been similar in design and performance. The overall system resolution is limited by the lowest resolution component of the entire system. The limiting factor in clinical SPECT spatial resolution is collimator design. Typical Tc 99m full-width half-maximum SPECT spatial resolution is approximately 8 to 9 mm. This resolution can be improved by using “ultra-high” resolution collimation, but at the expense of reduced photon collection efficiency (i.e., lower camera sensitivity). For the same imaging time, higher resolution parallel hole collimation results in lower acquired counts. This can be compensated for by increasing imaging time.

A modest improvement in image resolution may not be deemed necessary if the diagnostic accuracy is not significantly improved, particularly if increasing imaging time results in a higher frequency of patient motion artifacts. Advanced instrumentation and software are currently being tested, with the goal of reducing imaging time while preserving image quality. The specific details of typical clinical instrumentation, including gamma cameras, collimators, and quality assurance (QA), have been well described in a detailed review article.¹⁰

Techniques in SPECT

Indications

The American College of Cardiology Foundation and the American Society of Nuclear Cardiology jointly published guidelines for appropriateness criteria.¹¹ The indications deemed most appropriate were those in which patients presented with intermediate or high pretest probabilities in the categories described subsequently. MPI as a screening tool in very low pretest probability patients is generally considered inappropriate.

Appropriate patient populations include the following categories: (1) detection of CAD in symptomatic or asymptomatic patients (chest pain, newly diagnosed heart failure or diastolic dysfunction, newly diagnosed arrhythmias including atrial fibrillation and ventricular fibrillation); (2) risk assessment in patients with intermediate or high pretest probability of CAD; (3) risk assessment in patients with known coronary disease; and (4) evaluation of myocardial viability. Although other indications may be warranted, the above-listed indications were judged to be most appropriate in most referred cases. Typical MPI patterns are shown in Figures 54-4 through 54-6.

FIGURE 54-4 Myocardial ischemia. A, Tc 99m-sestamibi exercise-induced myocardial ischemia in the left anterior descending coronary artery (LAD) distribution confirmed at cardiac catheterization. Moderate and severe distal anterior (large yellow arrow), anteroseptal (white arrows), apical (red arrows), inferoseptal, and inferior (small yellow arrows) defects are not present on resting delayed images (mild inferior diaphragm attenuation also present). Bull’s-eye confirms reversibility of defects. LAD and right coronary artery (RCA) stenoses were confirmed at cardiac catheterization. BP, blood pressure; HLA, horizontal long axis; SA, short axis; VLA, vertical long axis. B, Gated Tc 99m-sestamibi MPI shows normal wall motion despite multivessel CAD. End-diastolic (left) and end-systolic (middle) volumes show concentric left ventricular contraction in all regions. Volume curves (right) show normal left ventricular end-diastolic and end-systolic volumes, and normal left ventricular ejection fraction of 58%.

FIGURE 54-5 MI. A, Tc 99m-sestamibi adenosine pharmacologic stress study shows MI without ischemia. The left ventricular cavity is markedly dilated. Severe distal anterior (small white arrows), anteroseptal (small yellow arrow), apical (red arrows), septal (large yellow arrow), and inferoseptal and inferior (large white arrow) defects are present on stress and rest images (i.e., fixed defects). Bull’s-eye confirms irreversibility of defects. HLA, horizontal long axis; ICD, implantable cardioverter-defibrillator; LV, left ventricle; SA, short axis; VLA, vertical long axis. B, Gated Tc 99m-sestamibi MPI shows markedly abnormal wall motion with severe global hypokinesis. End-diastolic (left) and end-systolic (middle) volumes show severe global hypokinesis in all regions. Volume curves (right) show abnormal left ventricular end-diastolic and end-systolic volumes, and abnormal left ventricular ejection fraction of 16%.

FIGURE 54-6 MI and myocardial ischemia. Tc 99m-sestamibi stress study shows MI and myocardial ischemia. Severe distal inferolateral (IL) and inferior defect is fixed (white arrows). Anterior (large yellow arrows) and anterolateral (AL) (red arrows) mild defects are reversible. Inferoseptal region (small yellow arrows), at the periphery of the infarct, shows partial improvement (compare stress with rest images). Bull’s-eye confirms reversibility and irreversibility of defects. HLA, horizontal long axis; LAD, left anterior descending coronary artery; LADD, left anterior descending diagonal coronary artery; LCx, left circumflex coronary artery; RCA, right coronary artery; SA, short axis; VLA, vertical long axis.

A unique clinical scenario for MPI is in the evaluation of acute coronary syndromes in patients presenting to the emergency department.¹² In patients without a history of prior myocardial infarction (MI) and an intermediate probability of CAD, the sensitivity for detection is very high. In this population actively having chest pain at the time of radiotracer administration, the negative predictive value of normal MPI is 99% to 100%. If the chest pain has resolved at the time of radiotracer administration, test sensitivity is modestly reduced, and current guidelines recommend repeating radiotracer administration within 2 hours of symptom abatement.¹² Because Tc 99m perfusion agents do not have significant redistribution for 6 hours, imaging can be performed after resolution of chest pain and still reflects the myocardial perfusion at the time administration. In patients presenting to the emergency department with chest pain that has resolved, and in whom recent myocardial injury has been excluded by a chest pain protocol including ECG and cardiac enzymes, a subsequent stress MPI study can be safely performed to exclude functional CAD.

Another unique use of MPI is for risk assessment. A large body of knowledge has shown that the defect size and severity, and the amount of reversibility by MPI are highly predictive of subsequent cardiac events. The negative predictive value of a normal or mildly abnormal study is also very high; these patients have a cardiac mortality of less than 1% per year, which is approximately the same as that of the general population. A highly abnormal study with severe and extensive perfusion abnormalities can have a cardiac event rate of 6% to 8% per year. This significantly higher incidence of adverse cardiac events, including subsequent MI, can have an impact on subsequent clinical management decisions.

As previously mentioned, an important specific indication is for the evaluation of myocardial viability. In most cases, SPECT MPI can accurately assess myocardial viability with a resting thallium 201 study followed by redistribution imaging, or with a stress-redistribution-reinjection thallium 201 protocol.

Contraindications

The contraindications for stress MPI are the same as those for exercise and pharmacologic stress testing without the radionuclide component with the additional concerns related to radiation exposure and radionuclide administration (e.g. pregnant or nursing woman). The radionuclides are administered in tracer quantities, and thus, they have no known pharmacologic side effects. Contraindications for exercise stress testing are well established and include ongoing MI or myocardial ischemia, unstable angina, hypotension, and unstable cardiovascular conditions which could result in exercise induced cardiac fatality. Additional relative contraindications may include those which limit the patient’s ability to exercise such as stroke, severe debilitation, severe peripheral vascular disease or underlying respiratory disease limiting exercise capacity, and altered mental status. Contraindications for pharmacologic stress testing are related to the specific pharmacologic stress agents administered, and are described in the next section under pharmacologic stress testing.

Pitfalls and Solutions

Quality Assurance

A QA program for gamma camera SPECT operation is essential to avoid artifacts, which could result in misinterpretation of test results. Proof of the establishment and diligent adherence to an equipment QA program is an integral part of laboratory accreditation by the American College of Radiology and the Intersocietal Commission for the Accreditation of Nuclear Medicine Laboratories. Conventional planar and specific SPECT and PET imaging QA procedures should be performed and recorded periodically. Daily QA includes ensuring correct isotope energy peak, and “daily floods” to assess gamma camera imaging field uniformity. Weekly QA includes resolution and linearity checks with “bar phantoms.” Many manufacturers include software that can automatically compute planar measurements including differential and integral flood field uniformity and intrinsic linear resolution from bar phantoms.¹⁰

SPECT requires additional calibrations and gamma camera measurement to avoid artifacts specifically related to tomographic imaging. Calibrations to correct for the center of rotation and high count extrinsic (with a collimator) flood field uniformity are typically acquired on a weekly or quarterly basis. SPECT imaging of multipurpose thermoplastic (Plexiglas) phantoms is also highly recommended on a quarterly basis, and is mandated by laboratory accrediting agencies. These phantoms have solid “cold” spheres of differing sizes, and are filled with background radioactivity. The phantom assesses three-dimensional spatial resolution, uniformity, and tomographic image contrast. Full QA testing is also typically performed with initial equipment acceptance testing, and after each major servicing or software upgrade to ensure proper functioning. Errors in center of rotation or nonuniformities in the flood field may result in loss of spatial resolution (blurring), or SPECT artifacts (“ring”), or both, which could affect clinical interpretation of images and can be particularly troublesome in MPI.

Many SPECT systems include attenuation correction components, using radioactive scanning line sources, low-end CT devices, or diagnostic-quality CT. These attenuation-correcting devices require their own set of daily, weekly, and annual QA procedures. Likewise, PET and PET/CT scanners have specific QA requirements, as specified by the equipment manufacturers and as mandated by laboratory accreditation agencies. Typical PET and PET/CT QA procedures have been reviewed in a more recent publication.¹⁰

Description of Techniques and Protocols

Radionuclide Imaging Protocols

Stress Protocols—Pharmacologic versus Exercise Protocols

The most common types of stress protocols in the United States are exercise treadmill and pharmacologic stress test. Exercise stress testing is preferred over pharmacologic testing because it allows a more physiologic assessment of the patient’s functional capacity and symptoms and the hemodynamic response to stress. Additionally, exercise generally results in lower radiotracer uptake in the gastrointestinal tract, improving image quality. The pharmacologic agents routinely used in clinical practice are adenosine, dipyridamole, and dobutamine. Pharmacologic stress tests can be performed on patients who cannot tolerate physical exercise. Patients with left bundle branch block, a known cause of false-positive septal perfusion defects with exercise, may undergo stress testing performed with dipyridamole to decrease the potential for this artifact. Other limiting factors in exercise stress testing are physical deconditioning, peripheral vascular disease, history of stroke, lower extremity amputation, and severe chronic obstructive pulmonary disease.

Exercise Protocols

In normal individuals, peak exercise increases heart rate and myocardial oxygen demand. At peak exercise, MBF typically increases approximately three to four times over that of rest. In patients with significant CAD, this mechanism fails to increase blood flow adequately distal to an obstructive lesion, resulting in an imbalance between oxygen supply and demand.

Before exercise, patients should have nothing per mouth after midnight, and be instructed to wear loose, comfortable clothing. We recommend abstinence from caffeinated substances because of the possibility that a pharmacologic stress test may be needed if an inadequate maximal heart rate is reached during maximal exercise stress testing. Intravenous access is obtained typically in the antecubital vein, ECG leads are securely placed, and the patient’s hemodynamic status is closely monitored and recorded throughout the procedure. The radiotracer is administered at peak or target heart rate while exercising. Hemodynamic and ECG monitoring are continued for at least 5 minutes into the recovery period, or until symptoms or ECG changes resolve. Patients with good functional capacity can typically exercise on the Bruce protocol, which rapidly increases in speed and incline. Patients who are older or debilitated may use a modified Bruce exercise protocol, which more gradually increases in speed and incline to achieve target heart rate. In individuals with significant limitations of physical tolerance or in patients 3 to 5 days post-MI, a Naughton protocol or other low-level exercise protocol may be used at the physician’s discretion.

Pharmacologic Protocols—Adenosine

Adenosine is a coronary vasodilator commonly used in combination with MPI.¹³ This is a purine base, endogenously produced by myocardial smooth muscle and vascular endothelium. It is derived through extracellular dephosphorylation of adenosine triphosphate (ATP) and adenosine diphosphate (ADP). There are four known receptor subtypes specific for adenosine. A2A is considered a cardiac specific receptor, through which coronary vasodilation is initiated after intravenous adenosine administration.

Adenosine causes a vasodilation without direct chronotropic or inotropic responses in myocardium. Secondary hemodynamic changes in response to vasodilation include a modest decrease in systolic and diastolic blood pressure, and a compensatory increase in heart rate with modest increase in cardiac output. The coronary vasodilation increases MBF three to five times above that of baseline resting MBF in normal coronary vessels; this is comparable to the increase in MBF with maximal exercise. Coronary arterial segments distal to a significant coronary artery lesion (i.e., stenoses ≥70%) typically are vasodilated as a basal condition to maintain normal blood flow at rest. Adenosine causes vasodilation and an increase in blood flow to territories with normal coronary arteries, but cannot comparably increase flow to regions distal to stenoses that are already dilated in a resting condition. A relative perfusion abnormality can be seen in regions distal to a stenosis.

Adenosine has a very short half life of 10 to 15 seconds and is administered at the rate of 140 µg/kg/min over 4 minutes. Protocols using either 6 minutes or 3 to 4 minutes are equally effective. The stress radiotracer is administered after 4 minutes of infusion to define coronary blood flow at maximal vasodilation. Side effects are similar to dipyridamole, but also include atrioventricular block. The most common side effects are intermittent atrioventricular block, flushing, chest pain, and dyspnea. Because of the very short physiologic half-life, these side effects are very short-lived and resolve within 1 to 2 minutes of discontinuation.

Pharmacologic Protocols—Dipyridamole

Dipyridamole is also a commonly used vasodilator for MPI stress testing.¹⁴ It is an indirect vasodilator that acts by blocking the cellular metabolism of adenosine. This results in a high local interstitial concentration, which subsequently results in coronary vasodilation. Dipyridamole is infused at a rate of 142 µg/kg/min over 4 minutes. The most common side effects include headache, flushing, hypotension, nausea, and chest discomfort. Because of a relatively longer biologic half-life of dipyridamole, the radiotracer is administered at 6 minutes after the start of the dipyridamole infusion. In addition, side effects may last for several minutes, but can be rapidly reversed by intravenous infusion of aminophylline, which acts by competitively inhibiting adenosine binding to its receptors.

Compared with adenosine, dipyridamole is equally effective in producing coronary vasodilation, and has a lower overall incidence of side effects, including induction of atrioventricular block. Dipyridamole can be used in patients with prolonged P–R interval and in elderly patients with poor tolerance to side effects.

Contraindications to Vasodilator Stress Testing

Vasodilators are contraindicated in patients with severe chronic obstructive pulmonary disease and asthma, particularly patients currently receiving treatment with β-agonists. Because high levels of adenosine block the β-agonist activity of bronchodilators, patients with chronic or reactive airways disease may experience bronchoconstriction and respiratory compromise. Unstable angina, persistent hypotension with systolic blood pressure less than 90 mm Hg, high-grade atrioventricular block without a pacemaker, uncontrolled arrhythmias, and severe aortic stenosis are also contraindications. Patients who have consumed caffeine or taken xanthine derivatives (e.g., theophylline, aminophylline) before testing have a potential for a false-negative result from inadequate vasodilation. Because xanthines block adenosine receptors, these should be discontinued at least 24 hours before vasodilator stress testing.

Pharmacologic Protocols—Dobutamine

Dobutamine is a β-adrenergic agent with a biologic half-life of 2 minutes. It acts by increasing myocardial contraction and heart rate, which subsequently increases myocardial oxygen demand and blood flow.¹⁵ It is considered an alternative stress agent when an exercise or pharmacologic vasodilator study cannot be performed. Intravenous administration starts at an initial dose of 5 to 10 µg/kg/min for 3 minutes, with subsequent increases to doses of 20 µg/kg/min, 30 µg/kg/min, and 40 µg/kg/min in 3-minute stages, or until the goal of reaching 85% of maximal heart rate target is achieved. At approximately 1 to 2 minutes after target heart rate is achieved, the radiopharmaceutical is administered intravenously.

Severe hypotension, ventricular tachycardia and hemodynamically unstable sustained supraventricular tachycardia, and atrial fibrillation with rapid ventricular response are indications for termination of dobutamine. Other side effects of dobutamine include chest pain, dyspnea, palpitations, hypertension, anxiety, nausea, and vomiting. Contraindications to dobutamine include ventricular tachycardia, atrial fibrillation with rapid ventricular response, uncontrolled hypertension, hypotension, hypertrophic obstructive cardiomyopathy or aortic stenosis with severe left ventricular outflow obstruction, recent aortic dissection, and coronary artery dissection.

SPECT Imaging Protocols

Several protocols using thallium 201, Tc 99m, or both radiotracers have been well studied and characterized with respect to accuracy. Although modest differences exist, overall accuracy of the most common protocols is comparable. When using same-day protocols with Tc 99m radiotracers, however, the resting study should be performed first to avoid “false-positive” fixed defects.¹⁶ If stress imaging is performed first, the higher MBF at stress combined with a relatively high stress radiotracer injection activity may lead to overestimation of MI because the subsequent resting images also reflect stress tracer distribution. If the resting study is performed first, the subsequent stress activity is much higher (approximately three times higher), and the increased blood flow at stress (approximately 2.5 to 3 times higher) produces a higher “weighting” on the stress images; this effectively reduces the contribution of the previous resting activity on the stress images. The number of potential false-negative results for ischemia is small because the resting activity has a relatively small contribution to the stress imaging.

Image Interpretation

Postprocessing

Tomographic images typically are reconstructed from the raw projection data with filtered back-projection and low-pass filtering. The degree of low-pass filtering depends on the counting statistics and noise. New algorithms have become commercially available that replace filtered back-projection by iterative reconstruction techniques, some versions of which incorporate depth-dependent collimator imaging characteristics and attenuation correcting data, both of which may improve overall image quality, particularly for low count perfusion images.

After tomographic reconstruction, the stress and rest images are normalized to the peak activity within the myocardium. This normalization permits a consistent comparison of regions within the myocardium, and it allows for a relative comparison between the stress and rest studies despite differences in absolute counts between studies.

An important aspect of MPI is quantification of defect size and severity. After normalization, this quantification can be performed automatically by computer analysis. Because the normal radiotracer distribution may vary depending on the patient population (e.g., breast attenuation in women and diaphragm attenuation in men), tomographic images are matched to the appropriate normal database cohort. Myocardial perfusion tomograms are processed automatically to generate a two-dimensional parametric display (bull’s-eye plot) from the three-dimensional count distribution. This display provides a rapid overview of three-dimensional myocardial perfusion for comparison between stress and rest count distributions.

Automated boundary detection programs define endocardial and epicardial borders of the myocardium. Applied to all cycles of gated studies, the left ventricular volumes provide measurements of global and regional left ventricular function and ejection fraction. Several groups and manufacturers have implemented their algorithms into commercially available software. Because of differences in physical modeling assumptions, algorithm implementation, and boundary detection methods, different software packages have different normal values and may produce significantly different results.¹⁷ Direct comparisons between values produced from different software packages should be performed carefully and should take into account algorithm differences, including the use of algorithm-specific databases of normal limits.

Interpretation

Interpretation of projection images (“raw data”) is essential for proper interpretation of the tomographic images. Projection data typically are first viewed in cine format to identify potential causes of decreased activity or artifacts. Identifying patient motion during image acquisition is an important step in identifying artifacts that could potentially be misinterpreted as perfusion defects. The effect of patient motion is often very difficult to predict with respect to location or appearance on reconstructed images. An artifactual perfusion defect may appear with misalignment of the ventricular walls, association with extracardiac activity adjacent to the defect, or a defect in a noncoronary distribution.

If there is a concern that an apparent perfusion defect may represent an artifact owing to patient motion, repeat imaging (Tc 99m radiotracers or resting thallium 201) should be considered to clarify the issue. Although motion correction software may reduce the effect of patient motion, software may be unable to eliminate completely the effect of patient motion on the final reconstructed tomographic images. Any application of motion-correcting software must be reviewed carefully to verify that the algorithm performed the intended correction appropriately, as part of routine patient-specific QA. Generally, vertical motion errors are detected and corrected more reliably than horizontal and bulk body motion errors. An example of a motion artifact is shown in Figure 54-7.

FIGURE 54-7 Motion artifact causing a defect owing to patient motion during image acquisition. Top row, Initial images acquired with significant patient motion suggest an anteroseptal defect (arrow) with misalignment of the ventricular walls and extracardiac activity outside of the normal myocardial boundaries (HLA view). Bottom row, Repeat imaging without patient motion shows normal perfusion. HLA, horizontal long axis; SA, short axis; VLA, vertical long axis.

Another common artifact is due to soft tissue attenuation. In women, breast attenuation can result in either a uniform or a nonuniform decrease in detected counts, most commonly in the anterior or anterolateral left ventricular wall region. Similarly, inferior decreased activity from diaphragm attenuation may be seen with a relatively high frequency in men. “Diaphragmatic creep” is an artifact caused by the changing position of the heart immediately after strenuous exercise stress and during the recovery phase. Because deep respiratory excursion gradually subsides after exercise, the location and orientation of the heart may change while imaging after exercise stress. The resulting tomographic images, which are a composite of various positions, may show a lower activity in the inferior wall because of this cardiac motion artifact. Assessment of the projection images for changes in heart position through imaging can determine if the potential for this artifact is present in an individual study.

Three physiologic parameters can be assessed from the projection data. The first is left ventricular cavity size. It is possible to appreciate markedly decreased counts within the left ventricular cavity, and a relatively large cavity compared with the myocardial thickness indicates left ventricular dilation. Measurements of ventricular size from the computer-generated boundaries may also confirm left ventricular enlargement. In patients with prior apical infarction and left ventricular dysfunction, ungated images may show a markedly decreased left ventricular cavity activity because of abnormally and severely decreased wall motion adjacent to infarction. In patients with an area of severely decreased activity at the apex because of prior infarction, severely decreased wall motion or dyskinesis may contribute to an almost entirely absent apical defect of attenuating blood pool (“black hole sign”; Fig. 54-8) that is associated with a higher incidence of apical aneurysm.

FIGURE 54-8 Apical black hole sign. Apical infarct and aneurysm in a patient with a history of CAD, ischemic cardiomyopathy, large anterior apical MI, and left ventricular aneurysm resection. Tc 99m-sestamibi stress study shows severe fixed apical defect (arrow; black hole sign). The severe left ventricular apical cavity dilation is compatible with residual left ventricular aneurysm. HLA, horizontal long axis; VLA, vertical long axis.

The second physiologic sign is increased lung uptake. This increased lung uptake has been well described in thallium 201 myocardial perfusion studies as a poor prognostic indicator, which reflects left ventricular dysfunction. The increase in lung uptake, defined as greater than 50% of the peak myocardial activity, is an indicator of abnormally prolonged pulmonary transit time. Patients with either exercise-induced or resting left ventricular dysfunction have increased pulmonary transit time, and prolonged exposure of the radiotracer to the lungs results in increased lung uptake, which can be assessed visually. This may be seen in patients with prior MI and elevated left ventricular end-diastolic pressure, or exercise-induced left ventricular dysfunction when the pulmonary transit time is increased. These findings may also be seen with Tc 99m radiotracers; however, because Tc 99m imaging may be performed at a time considerably later than the stress testing, some degree of lung clearance may occur, and increased lung uptake may not be apparent on subsequent imaging.

The third physiologic sign relates to right ventricular uptake. In normal patients, right ventricular myocardium is thin and may not be seen reliably because of partial volume effects. In patients with elevated right ventricular pressures or unusually thickened myocardium, abnormal right ventricular size or uptake may also be detected, in at least the thickest parts of the right ventricle, if not the entire right ventricle. These may be indirect signs of right ventricular hypertrophy, volume overload, or pressure overload.

When interpreting and reporting MPI, important aspects include perfusion defect size, severity, and reversibility. Although a visual interpretation may be accurate with respect to the presence or absence of coronary disease, quantitative information has prognostic significance. The size and severity of the perfusion defect are predictive of cardiac events, including MI and cardiac death.¹⁸ These are important for risk stratification and patient management. Small, mild defects are associated with low cardiac mortality of less than 1% per year.¹⁹ Severe defects and extensive ischemia are associated with significantly higher cardiac events and worse short-term prognosis. In addition, the short-term outcome for cardiac events in patients with mild ischemia is worse for patients undergoing revascularization.²⁰ In contrast, the short-term cardiac event rate for patients with severe defects is lower for patients undergoing revascularization compared with medical management. Quantitative measures of myocardial perfusion scintigraphy may determine the most appropriate subsequent therapy in patients with known CAD.²⁰

As discussed in Chapter 55, global and regional left ventricular function are very important parameters for clinical management. Left ventricular functional assessment by left ventricular ejection fraction is highly predictive of mortality. Regional functional assessment may also influence clinical management. A segment of dysfunctional myocardium with preserved viability may be important from the perspective of potential revascularization.

Other Artifacts and Normal Variants

Apical Thinning

The thickness of the myocardium in the apex varies from patient to patient. Physiologic thinning of the myocardium at the apex may show an apparent relative decrease in radiotracer uptake, however, because of partial volume effects.

Left Bundle Branch Block

Reversible septal perfusion defects can be seen in patients with left bundle branch block who undergo exercise stress MPI. This phenomenon has been attributed to asynchronous septal relaxation, which is out of phase with diastolic filling of the remainder of the ventricle. Perfusion defects related to left bundle branch block usually do not involve the apex. Septal asynchrony is more profound with higher heart rates, and it may appear as a reversible perfusion defect. Use of vasodilator stress has been reported to reduce the frequency of this artifact. More recently, the evaluation of patients with left bundle branch block by MPI has been enhanced considerably by the additional computations of ventricular asynchrony made possible by three-dimensional phase distributions measured from gated MPI count distributions.

Long Membranous/Short Muscular Septum

The length of the muscular septum varies among individuals. Absence of uptake in the basal membranous portion of the septum results in the appearance of a short basal septum.

Misregistration of SPECT and Computed Tomography

With the introduction of fast multislice CT, it is possible that a similar “artifactual” defect could be seen on SPECT if corrected with high temporal resolution CT. Cardiac PET studies have shown clearly potential artifactual “perfusion” defects owing to misalignment of the PET emission and the transformed CT transmission image used for attenuation correction. This is due to differences in the respiratory cycle during CT and the average PET emission diaphragm position. Data for a single high-resolution CT transaxial slice can be acquired in a fraction of a second, whereas PET emission data typically are acquired over several minutes to 15 to 30 minutes. Cardiac PET data may be ECG gated, but data are not currently gated with respiratory cycle information. The frequency of misalignment is high (42% in a retrospective study ²¹); however, impact on the emission corrected images is still under investigation.

Test Accuracy

Detection of Coronary Artery Disease

Stress perfusion scintigraphy with SPECT has the highest accuracy among all imaging modalities in the noninvasive detection of functionally significant CAD. Overall, the diagnostic accuracy is approximately 85% to 90%, with sensitivity and specificity varying depending on the patient population and methodology of the study. In women with functionally significant CAD (>70% stenosis), gated Tc 99m-sestamibi reported diagnostic sensitivity is 80%, and specificity is 92%.²² Other studies have reported higher sensitivities and specificities of 90% and 93%.³^,²³

Myocardial Viability

Using the stress-redistribution-reinjection thallium 201 protocol, areas of reversible ischemia can be identified as viable. A resting-redistribution thallium 201 study can assess regions of low resting blood flow and residual viable myocardium (“hibernating myocardium”). Resting images obtained immediately after injection represent MBF. Any perfusion defects seen on initial imaging caused by low resting blood flow show slower washout and “redistribution” compared with defects with normal resting blood flow. This normalization (“reversibility”) of resting perfusion defects on delayed imaging indicates viable myocardium. If a severe perfusion defect on initial rest imaging does not improve on the redistribution study (“fixed”), this correlates well with infarction.

A more recent study comparing rest-redistribution thallium 201 with myocardial delayed hyperenhancement by MRI confirmed high quantitative correlation of infarct size (r = 0.90).²⁴ Thallium 201 rest-redistribution also correlates well with viability by ¹⁸FDG-PET.⁸ Despite being “fixed” perfusion defects, mild and moderate (<50% decreased counts compared with normal areas) defects are considered viable. Viable regions by thallium 201 showed a 98% concordance with viability by ¹⁸FDG-PET. Only a relatively small proportion (approximately 30%) of severe fixed defects by thallium 201 showed discordant viability by ¹⁸FDG-PET.⁸ This detection of myocardial viability can determine the suitability of a candidate for revascularization.

Rest-redistribution and stress-redistribution-reinjection thallium 201 with SPECT are considered highly accurate in evaluating myocardial viability; however, PET has better spatial resolution and attenuation correction, and is generally considered superior for this indication. MRI has shown comparable results for viability determination, and is able to determine the extent of subendocardial infarction with better spatial resolution.²⁵

MYOCARDIAL PERFUSION IMAGING WITH PET

Clinical Indications

The currently approved indications for PET in nuclear medicine include MPI and myocardial viability assessment. The major advantages of PET over SPECT include the routine use of attenuation correction, higher detection sensitivity, and superior spatial resolution. PET perfusion radiotracers are currently relatively expensive, however, and have shorter physical half-lives compared with SPECT radiotracers. Although absolute quantification of MBF is possible, dynamic imaging and modeling are technically demanding, and have not been considered essential for clinical diagnostic imaging.

Patients with a high body mass index are especially well suited for PET because of (1) routine use of high-quality attenuation correction, and (2) higher photon detection sensitivity, owing to a lower proportion of photon scatter and tissue attenuation compared with SPECT. In addition, PET scans generally contain more than 100 times as many counts as do SPECT scans, particularly for three-dimensional data acquisitions, as opposed to two-dimensional PET data acquisitions. An example of an ammonia N 13 dipyridamole PET study is shown in Figure 54-9.

FIGURE 54-9 Normal ammonia N 13 PET myocardial perfusion pharmacologic stress imaging with dipyridamole. Note excellent spatial resolution and more uniform radiotracer distribution compared with traditional SPECT Tc 99m-sestamibi imaging (see Figs. 54-1 and 54-2), despite the patient’s high body mass index (BMI) of 50. BP, blood pressure; HLA, horizontal long axis; LAD/D1, first diagonal branch of the left anterior descending artery; PTCA, percutaneous transluminal coronary angioplasty; SA, short axis; VLA, vertical long axis.

The major clinical disadvantages of PET currently relate to the available myocardial perfusion radiotracers (expense, availability, cost, and physical properties). These physical factors result in numerous technical differences compared with SPECT, which are discussed briefly here.

Technical Aspects

PET Myocardial Perfusion Imaging Radiotracers

Numerous radiotracers have been validated to evaluate myocardial perfusion. The two most commonly used are ammonia N 13 and rubidium 82. Other validated PET myocardial perfusion radiotracers, such as ¹⁵O water and ¹¹C compounds with high first-pass extraction fraction, are currently investigational and not approved for clinical use. Although data indicate superiority of these as perfusion tracers, they have physical half-lives of only 122 seconds (¹⁵O water) and 20 minutes (¹¹C), requiring an on-site cyclotron for synthesis, and these tracers currently are impractical for widespread clinical use.

PET has been well established as an accurate means for the evaluation of myocardial viability. The presence of preserved myocardial metabolism by ¹⁸FDG is highly accurate in the detection of myocardial viability, and its use is briefly discussed here. Aspects of PET imaging involving highly sophisticated radiotracer compartmental modeling quantification are not discussed. Although well validated in many prior studies, these are not routinely used in clinical practice.

Ammonia N 13

The very short physical half-life of ammonia N 13 (10 minutes) requires an on-site cyclotron and radiochemical synthesis. The first-pass extraction fraction of ammonia N 13 is very high at approximately 95%. This compound rapidly crosses the intravascular capillary membrane into the interstitial space and into myocytes by passive diffusion. It is rapidly metabolized by ATP-dependent processes and trapped intracellularly. The uptake and retention can be altered by metabolic changes within the myocardium. Although ammonia N 13 is widely considered an excellent indicator of MBF, the dependence on the metabolic state of the myocyte may theoretically limit its ability to indicate absolute blood flow. The radiation dose from ammonia N 13 is 1.7 mSv (0.17 rem) for an injected activity of 740 MBq (20 mCi).

Rubidium 82

Rubidium 82 is a monovalent cationic analogue of potassium. It is commercially produced and available from a strontium (strontium 82) generator. Strontium 82 has a relatively long physical half-life (25.5 days); however, rubidium 82 has a very short physical half life (75 seconds). The strontium 82 generator requires replacement approximately every 4 weeks. Rubidium 82 is extracted from plasma with a very high efficiency through the Na⁺,K⁺-ATPase pump. The myocardial extraction fraction of rubidium 82 is similar to that of thallium 201, but is less than that of ammonia N 13. The extraction can be reduced by severe acidosis, hypoxia, and ischemia. The initial radiotracer uptake is a function of blood flow, metabolism, and myocardial cell integrity. The radiation dose is 2.2 mSv (0.22 rem) for an injection of 1.85 GBq (50 mCi) of activity.

Description of Technique and Protocols

The general procedures for PET MPI are outlined next. Differences in physical properties of the PET radiotracers necessitate differences in protocols to optimize imaging.

Stress Protocols

For ammonia N 13, an exercise protocol can be performed. The very short half-life of ammonia N 13 requires close coordination of isotope production, exercise testing, and imaging, however, if treadmill exercise is performed, the cyclotron, exercise facility, and PET imaging facility must be in close proximity. For rubidium 82, pharmacologic stress is typically performed in place of exercise stress because of the very short physical half-life of the isotope.

Resting perfusion is performed first. Because the physical half-lives of PET radiotracers are short, background resting perfusion does not significantly contribute to the subsequent stress radiotracer distribution. For ammonia N 13, a 740-MBq (20-mCi) activity typically is injected followed by approximately 10 minutes of imaging. For rubidium 82, a typical activity of 1.48 to 2.22 GBq (40 to 60 mCi) is administered followed by approximately 5 minutes of imaging. ECG gating can be performed with both radiotracers to assess left ventricular wall motion and left ventricular function.

PET Image Acquisition

The typical imaging time for acquiring PET myocardial perfusion emission data is approximately 10 minutes for ammonia N 13 and 5 minutes for rubidium 82. The attenuation correction scan, either the PET transmission scan or the CT transmission scan, should be acquired in close temporal proximity to the emission study. A single transmission scan can be performed to correct the rest and stress emission scans if the emission scans are done in close temporal proximity, and there is little chance of patient motion. If there is a question of patient motion between the two scans, separate transmission scans can be acquired to correct each emission scan separately. For attenuation correction, older PET scanners used a transmission scan with a positron-emitting source (e.g., ⁶⁸Ga or ¹³⁷Cs). More recent PET scanners, optimized primarily for whole body ¹⁸FDG-PET oncology applications, have incorporated sophisticated multidetector CT devices, which are used for attenuation correction and colocalization. Cardiac PET emission data from these PET scanners use attenuation correction data acquired using the CT scanners, transformed to attenuation maps for 511-keV gamma ray emission.

Image misregistration between the PET emission and the CT transmission scans is a potential cause of artifacts, the clinical impact of which is currently under investigation. Because of the high temporal resolution of CT and the short imaging time during which the CT scan is acquired, the depth of respiration during CT scanning may significantly differ compared with the average diaphragm position during the PET emission scan. This mismatch could potentially result in inaccurate attenuation correction, and a mismatch could potentially be misinterpreted as a perfusion defect. Several methods to correct for this mismatch are currently under investigation; inspection of the degree of mismatch is recommended as part of QA before interpretation of these studies.

For most clinical studies, relative quantification of perfusion and function with PET is performed in a manner that is very similar to that performed with SPECT. The attenuation-corrected PET scans result in a more uniform normal perfusion pattern with PET. Overall, the diagnostic accuracy of PET is higher than that of SPECT. Because of attenuation correction, higher sensitivity, and better spatial resolution, PET may be particularly useful in large patients, and in patients with severe attenuation artifacts.

Myocardial Viability

The diagnostic accuracy of ¹⁸FDG-PET in the evaluation of myocardial viability is well established.²⁶^,²⁷ Revascularization improved left ventricular function in 85% of segments with preservation of ¹⁸FDG uptake; no functional improvement was seen in 92% of segments with matched absence of blood flow and ¹⁸FDG uptake.²⁶ More recently, ¹⁸FDG-PET myocardial viability used to select revascularization candidates was associated with significantly lower in-hospital death rate, improved 12-month survival, and reduced complicated perioperative recovery.²⁸ Fluorine 18 is a positron imaging isotope with a physical half-life of 110 minutes. ¹⁸FDG is a glucose analogue, which is transported intracellularly via glucose transporters. It is subsequently phosphorylated to FDG-6-phosphate, and trapped intracellularly without further metabolism because of its stereochemistry. The radiation dose of typical injected activity of 370 MBq (10 mCi) of ¹⁸FDG is approximately 11 mSv (1.1 rem).