General Principles of PET Imaging

PET imaging is based on the use of radiotracers that decay by positron emission. A positron (a positively charged electron) is emitted from the nuclei of unstable isotopes during radioactive decay.¹⁴ Then, both the positron and an electron undergo a process known as positron annihilation and are converted into two coincident gamma-ray photons of 511 keV that travel in opposite directions. The distance traveled by the positron prior to annihilation constitutes the positron range. Detectors placed on either side of the active volume are connected in a coincidence circuit, so that if both detectors record an event within a very short interval, it is assumed that positron annihilation has occurred.¹² This constitutes the basis of PET.

PET utilizes biological radiotracers labeled with positron-emitting isotopes such as carbon (¹¹C), oxygen (¹⁵O), rubidium (⁸²Rb), and fluorine (¹⁸F) that mimic natural substances. Thus, PET enables the measurement of several biological processes, including cardiac tissue blood flow, metabolism, and neurohormonal and receptor function. Many of the isotopes used for PET imaging have short physical half-lives (T_1/2),¹⁴ making them readily applicable in studies requiring repetitive measurements in the same session.¹⁵

Attenuation effects are significantly higher with PET than with SPECT. Attenuation correction (AC) is relatively straightforward with PET because the length of the path of attenuation for the pair of 511-keV photons is constant and known for PET, whereas it is variable with SPECT.¹³ Traditionally, external transmission sources, such as germanium-68 (⁶⁸Ge) or cesium-137 (¹³⁷Cs), have been used as established means for AC of the PET emission data.¹³ Today this has been replaced exclusively by x-ray transmission scanning in the current generation of PET/CT systems, shortening the total transmission imaging time to under a minute. Since fast helical CT scans acquire images at a single point in the respiratory cycle, whereas PET data are averaged over many respiratory cycles, respiratory motion mismatch artifacts can be produced.¹¹ These artifacts, when present, typically affect the anterior and anterolateral segments of the LV.^16,17

In recent years, PET instrumentation has shown substantial evolution. New crystal materials, such as lutetium oxyorthosilicate (LSO) and gadolinium oxyorthosilicate (GSO) have become available.¹³ These crystals are attractive for PET, owing to faster light decay time and higher light yield than bismuth germanate (BGO) crystals.¹⁸ Furthermore, there is an increasing trend to apply 3D-imaging acquisition mode (septa out) instead of the traditional two-dimensional (2D) mode (septa in), with the potential to improve image quality and reduce injected doses but at the expense of increased background counts and higher reliance on scatter-correction accuracy.^11,19 Notably, several new PET/CT scanners operate only in 3D mode.¹³ The current role of 3D mode versus 2D-mode will be fully described elsewhere in this textbook.

Cardiac PET has several technical advantages over traditional SPECT that should be appreciated:

Myocardial PET Perfusion Tracers

According to their physical properties, myocardial PET blood flow tracers fall into two basic categories: (1) inert, freely diffusible tracers like H₂¹⁵O and (2) physiologically retained tracers like ¹³NH₃ and ⁸²Rb.^4,14 The main aspects of PET flow tracers have been summarized in Tables 19-1, 19-2, and 19-3.

Table 19-1 Myocardial PET Flow Tracers

Table 19-2 Myocardial PET Perfusion Tracers: Principal Features

Table 19-3 PET/CT Studies and Effective Doses

Nitrogen-13 Ammonia

¹³NH₃ (T_1/2 = 9.96 minutes) requires an on-site cyclotron and radiochemistry synthesis capability.^3,22 In the bloodstream, neutral ¹³NH₃ is in equilibrium with the ionic form, ammonium (NH₄⁺). ¹³NH₃ diffuses freely across capillary and cell membranes and is retained in myocardial tissue,²⁴ whereby it can be either incorporated into synthesis of ¹³N-glutamine, or it can diffuse back into the vascular space. The initial extraction is high, even at high flow rates. As such, the uptake rate constant K₁ is a good estimate for quantitative blood flow (Fig. 19-1).

Figure 19-1 A, Tracer uptake rate K₁ versus FLOW from a one-tissue-compartment model of tracer kinetics. K₁ is equal to the product of FLOW times the unidirectional extraction fraction. B, One-tissue-compartment model diagram and equation. The myocardial curve C_M(t) is equal to the arterial input curve C_A(t) convolved with a single decaying exponential function K₁e^−k2t. C, The model parameters K₁, DV (=K₁/k₂), and a partial-volume correction parameter (1-TBV) are estimated by nonlinear least-squares fitting of the model (dark blue line) to the measured PET data (blue dots).

The net tissue retention is approximately 90% in the resting state. For ¹³NH₃, the direct relationship between net tissue retention and blood flow is preserved for values of blood flow up to 2.5 mL/min/g, but at higher flow rates this linear relationship is lost (Fig. 19-2).²⁵ Therefore, it is necessary to correct for flow-dependent changes in net tissue retention.

Figure 19-2 A, Tracer retention rate R versus FLOW from a net retention model of tracer kinetics. R is equal to the product of FLOW times the net retention fraction. B, Net retention model diagram and equation. The myocardial curve C_M(t) is equal to the arterial input curve C_A(t) convolved with a constant retention rate R, assuming there is no washout (dotted line) from the myocardium. C, The model parameter R is computed simply as the ratio of the myocardial uptake activity (dark blue) after time T (e.g., 2 min) divided by the integral area (green) under the arterial input curve, and assuming a constant partial-volume correction factor Rc.

Myocardial retention of ¹³NH₃ may be heterogeneous,²⁶ and the lateral LV wall uptake can be 10% lower than that of other segments. Also, image quality can be hampered by the occasional intense liver activity, which could interfere with the evaluation of the inferior wall. Finally, in patients with severe left ventricular ejection fraction (LVEF) impairment, chronic obstructive pulmonary disease (COPD), or smoking, the sequestration of ¹³NH₃ in the lungs can be abnormally increased. Then it would be necessary to delay the time between injection and image acquisition to enhance image quality.¹³

Rubidium-82

⁸²Rb is produced from a strontium-82 (⁸²Sr)/⁸²Rb generator, which can be eluted every 10 minutes.²⁷ The T_1/2 of ⁸²Sr is 25.5 days, which results in a generator life of 6 to 8 weeks. The short T_1/2 of ⁸²Rb (76 seconds) allows repeated and sequential perfusion studies but requires rapid image acquisition shortly after tracer administration.^{3,12 82}Rb is a monovalent cationic analog of potassium and has similar biological activity to thallium-201 (²⁰¹Tl). Myocardial uptake of ⁸²Rb requires active transport via the sodium/potassium adenosine triphosphate transporter. In animal models, the net retention is approximately 50% at rest and decreases to 30% at peak flow. The retention fraction can be altered by acidosis and acute hypoxia.⁴

A small and mobile generator infusion system is used for eluting ⁸²Rb every 10 to 15 minutes with low radiation exposure.²⁷ Quantitative assessment with ⁸²Rb is quite feasible and clinically practical with this generator as compared to cyclotron-produced compounds (Figs. 19-3, 19-4, and 19-5).^28–31

Figure 19-3 4DM display of ⁸²Rb PET MPI in a patient with normal perfusion.

Figure 19-4 4DM display of ⁸²Rb PET MPI shows a severe perfusion defect in the inferior and inferolateral walls that partially improves at rest. There was a reversible wall-motion abnormality on the gated images. The findings are consistent with mostly moderate ischemia and nontransmural scar in the distribution of the LCX territory.

Figure 19-5 4DM display of ⁸²Rb PET MPI of a high risk scan. Images show a severe perfusion defect in the anterior wall and apex with significant improvement on rest images. This is consistent with severe ischemia in the distribution of the LAD territory. In addition, there is transient ischemic dilation and Gated images showed stress induced wall motion abnormalities.

Recent Advances in PET Imaging: Novel Myocardial Blood Flow Tracers

The development of novel tracers could help circumvent some of the current limitations observed with the conventional radioisotopes, such as the “roll-off phenomenon” from incomplete retention and short half-lives that entail on-site production, which ultimately increase costs. The principal characteristics of a new class of radiotracers have been evaluated. Despite promising preliminary results, further research is necessary to elucidate whether these agents will have clinical application in humans.

¹⁸F-labeled compounds have relatively long physical T_1/2 (110 minutes). Among these, ¹⁸F-p-fluorobenzyl triphenyl phosphonium cation (¹⁸F-FBnTP) is a member of a new class of positron-emitting lipophilic cations that may act as myocardial perfusion PET tracers.^{43,44 18}F-BMS-747158-02 constitutes another emerging agent that is an analog of the insecticide pyridaben, an inhibitor of mitochondrial complex I (MC-1).⁴⁵ Mitochondria constitute 20% to 30% of the myocardial intracellular volume. Consequently, molecules that target mitochondrial proteins may be enriched and retained selectively in the myocardium. With the longer ¹⁸F T_1/2, these compounds have the potential for wide distribution. However, they may require reinjection or 2-day stress/rest imaging protocol similar to technetium-99 (^99mTc) SPECT agents. On the other hand, this feature may enable routine exercise stress, which has not been widely developed with PET imaging to date. Recent data suggest kinetics may be suitable for quantification (Fig. 19-6).⁴⁶

Figure 19-6 Cardiac images in rat model of left coronary ligation. A, Cardiac PET short- and long-axis images with BMS-747158-02 at 5 to 15 minutes after injection. The no-flow region is clearly identified. B, Ex vivo histologic images of heart short-axis slices from apex to base stained with blue dye. Blue areas indicate well-perfused zones, and no-blue areas indicate no-flow zones. C, In vivo cardiac PET short-axis images from apex to base. The dark areas in the left ventricular wall of PET images with BMS-747158-02 C, match closely with the no-flow zones identified by blue stain ex vivo B.

(Reprinted with permission from Yu M, Guaraldi MT, Mistry M, Robinson SP, et al: BMS-747158-02: A novel PET myocardial perfusion imaging agent, J Nucl Cardiol 14:789–798, 2007.)

A ⁶⁸Ge/⁶⁸Ga generator can provide a convenient source of PET tracers because of the long physical half-life of ⁶⁸Ge (T_1/2 = 271 days) and a suitable daughter half-life (⁶⁸Ga; T_1/2 = 67.7 minutes). Recently, a ligand has been successfully labeled and tested with ⁶⁷Ga for SPECT.⁴⁷ The biodistribution of this novel complex has been assessed in normal and infarcted rat models, whereby it showed a high heart uptake, albeit low cardiac retention and high liver uptake. Further studies of other gallium complexes may improve the feasibility of measuring regional myocardial perfusion using similar ligands labeled with the PET isotope ⁶⁸Ge.

Patient Preparation

Patients should be instructed to fast for at least 6 hours, to abstain from caffeine-containing products at least 12 hours, and to avoid theophylline-containing medications for 48 hours prior to the test.¹³ Diabetic patients should be guided on how to administer insulin before the study. Patients undergoing dobutamine stress tests should discontinue beta-blockers 48 hours prior to the test (only for diagnostic purposes and if it is clinically safe to do so).

Stress Testing Protocols

PET perfusion imaging is usually performed with pharmacologic stress. Adenosine, dipyridamole, and adenosine triphosphate block transport of adenosine into the cells and/or increase extracellular levels of adenosine, which causes coronary vasodilatation by interacting with the adenosine A₂ receptors in the cell membrane.¹³ Adenosine and dipyridamole increase MBF without increasing oxygen demand. Side effects are somewhat greater with adenosine than with dipyridamole, but the latter more often require reversal with aminophylline (which is routine in some laboratories). More selective agonism of the adenosine A_2A receptor subtype should theoretically result in a similar degree of coronary vasodilation with fewer and less severe side effects. Binodenoson,⁴⁸ regadenoson,⁴⁹ and apadenoson⁵⁰ are highly selective agonists for the adenosine A_2A receptor and are under investigation for clinical use for vasodilator stress imaging.

Dobutamine stress is a feasible alternative in those situations where adenosine, dipyridamole, or ATP are contraindicated.^13,51 Dobutamine increases MBF to meet increasing myocardial oxygen demand. In segments supplied by diseased vessels, the increase in flow is attenuated during high-dose dobutamine administration (20 to 40 μg/kg/min).

Exercise stress may be a valid alternative and provide clinical information unobtainable with pharmacologic stress that is helpful in decision making, particularly for patients unable to tolerate pharmacologic stress. ¹³NH₃ can be used in conjunction with treadmill exercise (TEX) or upright bicycle test. The tracer is administrated at peak exercise, and exercise should be continued for an additional 30 to 60 seconds after tracer injection. The patient is then repositioned in the PET camera to start the acquisition within 4 to 6 minutes.⁵² Accurate repositioning is of key importance to minimize artifact due to incorrect AC. Exercise stress is also available for the ⁸²Rb imaging^13,53,54 but can be logistically demanding; patients must be moved to and positioned in the scanner within 3 minutes of completing exercise.

Myocardial Perfusion Imaging Protocols

Common protocols used for imaging myocardial perfusion with dedicated PET or PET/CT systems involve the following steps:

Figure 19-7 Protocols for clinical cardiac PET/CT. A, Hybrid list-mode PET + CTA protocol. List mode enables simultaneous gating and dynamic acquisition. B, Gated PET/CT protocol. C, Multiframe (dynamic) ⁸²Rb PET/CT + CAC protocol. D, Multiframe (dynamic) ¹³NH₃ PET/CT, + CAC protocol. CAC, coronary artery calcium scan; CTA, CT angiography; CTAC, CT-based attenuation correction (transmission scan); pharm, pharmacologic.

(Reprinted with permission from Di Carli MF, Dorbala S, Moore S, et al: Clinical myocardial perfusion PET/CT, J Nucl Med 48:783–793, 2007.)

The most widely applied tracers in clinical practice are¹³NH₃ and ⁸²Rb.

Image Evaluation for Technical Sources of Errors: Quality Control

This step is of extreme importance for an accurate interpretation of the images. Patient body movements can lead to blurring of contours. Acquisition of a brief scan or scout image may facilitate accurate patient positioning. Body motion or respiratory movement can lead to transmission-emission misalignment and potential AC-induced artifacts (Fig. 19-8). However, most PET/CT systems should include software tools to correct transmission-emission misalignments. Finally, other potential sources of reconstruction artifacts encompass streak artifacts seen in large patients with arms-down imaging, metal implants, and IV contrast, as well as residual radioactivity in the IV line within the field of view (FOV).^3,13

Figure 19-8 A, PET/CT misregistration: The area of myocardium superimposed on the lung has an attenuation value lower than soft tissue (myocardium), thus the PET perfusion exam is undercorrected. B, Correct PET/CT registration using repeat CT attenuation scan. C, PET/CT misregistration: apparent hypoperfusion of the anterolateral wall. D, Correct PET/CT registration: shows normal perfusion. Recent software developments now enable realignment of the PET/CT image prior to reprocessing without needing to repeat the CT acquisition. This has reduced the frequency of misalignment at the time of interpretation. Careful QC continues to be required.

(Images reproduced with permission and courtesy of Kevin Berger, MD, Michigan State University.)

Image Analysis and Interpretation of Perfusion Images

Myocardial perfusion defects are usually identified by visual analysis of the reconstructed slices and compared between stress and rest images. Perfusion defect description, including extent, severity, reversibility, location, and specific coronary territories, should be reported routinely.⁶⁰ Defects are typically defined as fixed, suggesting scar formation; reversible, indicating ischemia; or partly reversible, indicating a mixture of scar and ischemia. ACC/AHA/ASNC guidelines recommend semiquantitative analysis using a 17-segment model.^13,61 Using this approach, summed perfusion defect score can be calculated, and this score is useful for cardiac risk stratification.^61,62

Gated images are also acquired and interpreted. Contour definitions used for EF and volume estimates should be confirmed prior to interpreting gated EF data. The change in EF may have diagnostic value in terms of the extent of disease (discussed later).⁶³ Wall-motion abnormalities at rest represent myocardial injury, either scar or otherwise, as may occur in a dilated cardiomyopathy. Wall-motion abnormalities that appear or are worse on stress imaging indicate ischemia-induced wall-motion abnormalities. This is in contrast to post-stress SPECT imaging acquired at least 30 minutes after the tracer injection, whereby wall-motion abnormalities are not reflecting concurrent ischemia but rather postischemic dysfunction or stunning. PET wall motion is acquired shortly after tracer injection, and thus changes reflect ischemic wall-motion abnormalities.

Advantages of Absolute Quantitative Analysis

Atherosclerotic vascular disease has become a leading cause of death worldwide, and there is a need to identify the presence of CAD before the onset of symptoms. PET technology has the potential to image and measure pathophysiologic and molecular processes in vivo and is now recognized as the best noninvasive means by which to quantify MBF in absolute terms (mL/min/g) and CFR. Table 19-4 describes normal values of MBF and CFR in healthy subjects.

Table 19-4 Normal Values in Baseline Myocardial Blood Flow and Coronary Flow Reserve in Normal Subjects

Parkash et al. demonstrated the clinical importance of CFR over conventional visual analysis in patients with three-vessel CAD using ⁸²Rb PET. In this study, the perfusion defect sizes were larger using quantification in comparison with the conventional relative uptake evaluation.²⁹ Yoshinaga and colleagues compared the clinical value of CFR by PET to the relative assessment of MPI by SPECT in a population of 27 patients with overt CAD. Approximately two-thirds of regions with angiographic lesions more than 50% showed normal SPECT perfusion scans but significant impaired CFR by PET.⁷⁰ Further investigations in larger sample size populations are needed to fully understand the clinical utility and added value of this approach. Only a few studies have assessed the prognostic value of abnormal MBF and CFR.^73,74

Models for Flow Quantification

Absolute measurements of physiologic or biochemical function are obtained using tracer kinetic modeling. The time course of radioactivity of the arterial blood input function Ca (t) can usually be measured from image regions within the LV cavity and combined with the myocardial response function Cm (t) to estimate a quantitative rate of transport.

A mathematical model is constructed with parameters that represent the flux of the radiotracer between the compartments. Various tracer kinetic models have been applied and validated for MBF quantification with H₂¹⁵O, ¹³NH₃, ⁸²Rb, and ¹¹C-acetate. Simple one-tissue-compartment models can be applied for flow quantification with each,^{25,32,34,56,75,76} providing a reasonable approach for estimating absolute tissue perfusion (see Fig. 19-1).¹¹

A further simplified retention model assumes that the tracer is retained completely without subsequent washout. A flow-dependent correction for net tracer retention is required to obtain quantitative perfusion estimates from the retention rates. The net retention model has been shown to produce precise estimates with ¹³NH₃⁷⁷ and ⁸²Rb^62,78 in normal subjects and in patients with CAD, but the results depend in part on the time at which the model is evaluated and on the accuracy of retention measurement.¹¹

Stay updated, free articles. Join our Telegram channel

Join