Chapter 10: Cardiovascular Positron Emission Tomography

Cardiac positron emission tomography (PET) myocardial perfusion imaging is an outstanding tool for diagnosis and risk stratification of patients with known or suspected coronary artery disease with many important differences compared to SPECT imaging. Its clinical use in assessing patients for CAD as well as other nonperfusion indications is expanding rapidly. Cardiac PET offers excellent diagnostic accuracy and image quality, low radiation exposure as well as a noninvasive means of measuring myocardial blood flow reserve. Cardiovascular PET is becoming an important tool for nuclear cardiologists to consider not only for perfusion imaging, but also for a growing number of other cardiovascular conditions such as myocardial viability, infection, and inflammation. Its role in nuclear cardiology has expanded with the increased and continuous availability of PET radiopharmaceuticals and PET camera systems (see Chapters 3 and 4 on radiopharmaceuticals and instrumentation). This chapter will review the principles of cardiovascular PET, data on diagnostic accuracy, image quality, radiation exposure, indications, and addition of myocardial blood flow (MBF) for myocardial perfusion imaging (MPI), as well as exciting new nonperfusion applications.

PET imaging provides superior temporal and spatial resolution of radioactive atoms as they decay. A PET radioactive tracer, which has been engineered to be taken up in to the organ of interest, is injected into the patient. After it reaches the target organ, the radioactive agent begins to decay and emits a positron (Fig. 10-1). This positron then collides with a nearby electron. The resulting collision causes annihilation of both an electron and a positron, creating a high-energy discharge of 1.02 MeV. This energy is split into two gamma rays of 511-keV energy, which are emitted 180 degrees from each other. For image collection, multiple detectors encircle the patient; absorption from both emissions simultaneously occurs, and events that are not 180% are eliminated. The distance to the annihilation event impacts ultimate image quality. The processing of these simultaneous events forms the basis of PET imaging.¹

Cardiac PET instrumentation is discussed in detail in Chapter 4. A PET camera appearance is similar to a computed axial tomography (CT) system. However, in the gantry of the PET camera are multiple detectors placed circularly around the patient (Fig. 10-2). These detectors are grouped into cassettes filled with a specified number of crystals (4,000–24,000). Data from the crystals are then transmitted to photomultiplier tubes similar to Single Photon Emission Computed Tomography (SPECT) imaging. In contrast to SPECT acquisition which requires multiple acquisitions from the same crystals rotating spatially around the patient, PET data acquisition occurs simultaneously surrounding the patient. Most scanners are full-ring design but some have partial-ring detectors with a rapidly rotating gantry.² An attenuation correction scan with either PET radionuclide line source or computed tomography (CT) is an integral part of the acquisition and is applied to each study.

While there are several myocardial tracers available for PET imaging, only three are presently used for clinical imaging: rubidium-82 (Rb-82), N-13 ammonia (N-13), and fluorine-18 (¹⁸F) fluorodeoxyglucose (FDG). For complete review of PET isotopes, see Chapter 3.

Rubidium-82

Rubidium (Rb)-82 is a potassium analog radiopharmaceutical that is produced with commercially available generators which can be stored on site. It is produced by the decay of strontium-82 (⁸²Sr) which has a half-life of 25.5 days. Rb-82 has a half-life of 75 seconds. The extraction fraction of Rb-82 is estimated to be 65% to 75%, substantially higher than SPECT technetium agents.

The use of Rb-82 is provided by an on-site generator and delivery system, which can be stored conveniently in the camera room. The generator is replaced every 4 to 6 weeks and supplied at a fixed cost, independent of the number of studies performed. Because of the cost involved with the generator, consideration for PET use in a facility should include adequate patient volumes. When in use, the generator is replenished quickly, within 10 minutes via ⁸²Sr to Rb-82 decay. This permits minimal downtime between stress and rest imaging, allowing efficient utilization of the generator. The protocol for stress imaging requires the use of a pharmacologic stress agent because of the short half-life of Rb-82.

Advantages of Rb-82 include the ability to have rapid acquisition times for rest and stress imaging and continuous availability. Another advantage was studied by Parkash et al.³ They used Rb-82 with PET/CT imaging and were able to detect smaller perfusion abnormalities and more accurate detecting multivessel disease. This suggests that PET may be able to detect multivessel disease more accurately than with SPECT-imaging techniques.⁴

There are also disadvantages to Rb-82 imaging. Due to the half-life of ⁸²Sr, the same generator is used for 4 to 6 weeks, but with reduced activity during the last weeks, which can impact image quality, with cameras using 2D acquisition. This limitation, however, can be overcome by 3D imaging.⁵ The short half-life of rubidium limits its clinical use to pharmacologic stress and not exercise.

N-13 Ammonia

N-13 has been used as a cardiac tracer for over 20 years. Implementation requires the presence of an on-site cyclotron as it has a half-life of 9.96 minutes. N-13 has an extraction fraction of up to 80%. The overall trapping of N-13 is dependent on a properly functioning metabolism and trapping may be decreased in myocytes that are ischemic.^6,7 Advantages of N-13 ammonia include the high tomographic counts and a clear blood-to-myocardial delineation. Because of the longer half-life, exercise is possible, in contrast to rubidium. Limitations include the finding that even in normal subjects, the retention of N-13 in the lateral wall is 10% less than other segments creating false-positive results.⁸ Prominent liver activity in some studies can complicate interpretation of the inferior wall.⁹ Finally, the necessity for on-site or close proximity cyclotron availability due to short half-life limits its practical use. This has severely limited its use, with perhaps 15 to 20 laboratories using this agent according to recent data. This may change with the availability of a small cyclotron designed for ammonia production only that is much more manageable and can be placed onsite.

Fluorine-18 Fluorodeoxyglucose

FDG is fluorine-labeled with 2-deoxyglucose. It is cyclotron produced and decays into O-18 oxygen. When it decays, its positron range is very short (0.5 mm). This fact, combined with its half-life of 109.8 minutes, allows it to provide images with the highest spatial resolution of any of the commonly used radionuclides.

FDG is a glucose analog and is brought into the myocardium via diffusion across glucose-specific transporters that are located on the myocytes. Once in the cell, it is transformed into FDG-6-P. There are low levels of glucose-6-phosphatase to reverse this step; there the FDG-6-P is essentially trapped in the myocardial cell. The use of F-18-FDG is further elaborated on the viability discussion later in the chapter as well as Chapter 21. Active research is underway to validate F-18 Flurpiridaz, a blood flow agent for the detection of CAD.

The increasing clinical use of cardiac PET MPI is due to several features that clinicians find important for their laboratories as well as patient care. This section will outline demonstrated advantages including imaging quality, diagnostic accuracy, radiation exposure, assessment of MBF, and conclude with a section on which patients are either recommended or preferred for PET MPI.¹⁰

Excellent Image Quality

Myocardial PET perfusion imaging utilizes high-energy level radiopharmaceutical tracers, with short half-lives. This translates into relatively higher doses being administered (but with lower radiation exposure) and better acquisition characteristics. The result is high myocardial counts, high spatial and contrast resolution, and high signal-to-noise ratio. Improved PET image quality was demonstrated by Yoshinaga et al., in which patients with equivocal SPECT studies were referred for cardiac PET imaging.¹¹ In a very high percentage of patients, the PET study resulted in good to excellent image quality. Using a comparison of similar but matched patients undergoing SPECT or PET, Bateman et al. also reported a significant improvement in image quality with PET over SPECT, with fewer equivocal reports.¹²

The routine use of attenuation correction plus higher-energy levels of PET tracers substantially reduces attenuation artifacts in images, reducing the potential for false-positive studies for interpretation. This may be particularly useful in obese patients however, in a study by Bateman et al. the specificity of PET was significantly higher in comparison to SPECT, regardless of body habitus.¹² Similarly, patients who had equivocal SPECT studies were found to have diagnostically useful PET studies in 98% of the referred patients. Thus, while obese patients or patients with prior equivocal SPECT studies are excellent candidates for PET imaging, the high accuracy among all patients suggests even broader application.¹⁰ In addition, the incidence of gut activity affecting interpretation was significantly and substantially reduced with PET in relation to pharmacologic SPECT.¹²

Diagnostic Accuracy with Cardiac PET Perfusion

The literature suggests that PET has high sensitivity and specificity. A meta-analysis of Nandalur et al. demonstrated high diagnostic accuracy from the available PET literature in over 1400 patients (Fig. 10-3).¹³ Two subsequent and larger studies have confirmed these findings, including a systematic review by Mc Ardle et al. which demonstrated that the sensitivity and specificity utilizing Rb-82 was 90% and 88%, respectively.¹⁴ A separate meta-analysis of 11,862 patients by Parker et al., PET MPI, also demonstrated a higher diagnostic accuracy for coronary artery disease than SPECT MPI (Figs. 10-4 and 10-5).¹⁵ Both of these studies also compared PET to SPECT literature and found significantly higher diagnostic accuracy for PET. Many of these reported studies in these two analyses used dedicated PET cameras with line source attenuation correction. A study by Sampson et al. also confirmed high diagnostic accuracy with the PET/CT instrumentation.¹⁶

A direct clinical comparison of SPECT to PET in similar patients was performed by Bateman et al.¹² Over 120 age-, body habitus-, and risk factor-matched patients underwent SPECT or PET imaging and cardiac catheterization. Significantly improved diagnostic accuracy was identified in patients undergoing PET perfusion imaging. Importantly, the ability to identify multivessel ischemia with PET was also significantly higher (74% vs. 41%) than with SPECT perfusion imaging, confirming the findings of Parkash.³ In sub-study analysis, diagnostic accuracy held regardless of body mass index or gender (Fig. 10-6).

Minimizing radiation exposure is an important consideration when ordering a diagnostic test. Efforts have been made on a national level to encourage the lowest radiation exposure possible in patients undergoing diagnostic testing with either CT or nuclear imaging, without sacrificing on image quality and diagnostic accuracy. With these concerns, the American Society of Nuclear Cardiology (ASNC) published an information statement in 2010 recommending radiation exposure for a patient study to be less than 9 mSv.¹⁷ However, a report by Jerome et al. subsequent to this recommendation, found that <1% of SPECT studies met these criteria and the mean radiation exposure was 14.6 mSv, far exceeding doses recommended.¹⁸ In contrast, radiation exposure for cardiac PET studies is substantially less than ASNC recommendations. For example, in a study by Senthamizhchelvan et al., radiation exposure from an Rb-82 PET study was calculated to be 9 mSv/20 mCi, translating to 3 to 6 mSv per patient depending on instrumentation.¹⁹ These findings were confirmed by Hunter et al. in a single-center study in which the PET radiation exposure was calculated to be as low as 1 to 3 mSv.²⁰ On a national basis, a recent study by Lundbye et al.²¹ confirmed that radiation exposure to patients using either Rb-82 or NH₃ ammonia was an average of 3.59 mSv in 122 laboratories undergoing accreditation evaluation by the Intersocietal Accreditation Commission (Fig. 10-7). These data support the ASNC information statement recommendation of using cardiac PET for radiation reduction in patients undergoing nuclear cardiology studies when available.^17,22 In summary, cardiac PET perfusion imaging radiation exposure to patients is low and generally consistent with annual background levels (3–4 mSv), substantially lower than current SPECT practices and is lower than ASNC recommendations.

Risk Stratification with Cardiac PET Perfusion Imaging

Risk stratification with SPECT imaging has become a very important aspect of the success of this modality. There now is growing evidence supporting the substantial prognostic value of cardiac PET. Several small and limited studies demonstrated the concept that the size and severity of the PET perfusion study predicts subsequent cardiac events. The largest study to date was that of Dorbala et al. published in 2013 (Fig. 10-8).²³ This multicenter study of over 7000 patients, evaluated outcomes in patients undergoing pharmacologic stress Rb-82 PET imaging. In a 2- to 3-year follow-up evaluation, the ability of the rubidium PET study to characterize patients into low, intermediate, and high risk for either cardiac death or all-cause mortality was demonstrated (Panels A, B). Two important aspects of that study emerged. First, the patients in the moderate- to high-risk category had a much higher event rate than similar SPECT studies, most likely due to the ability of the imaging procedure to more accurately identify multivessel CAD. Second, those in the lower-risk group had annualized event rates of less than 1%. This finding is in contrast to a pharmacologic SPECT study in which the risk of events is higher, 1% to 2%.²⁴ This is a reassuring finding for patients undergoing pharmacologic stress imaging with PET.

In a subset of patients separated by gender, similar risk stratification was demonstrated for both male and female patients.²³ These data are particularly important in an era in which women undergo evaluation at an earlier stage than previously and are expected to have a longer longevity than men, placing radiation exposure reduction of considerable importance.

Rb-82 left ventricular ejection fraction reserve (peak stress minus rest ejection fraction) provides incremental prognostic value to clinical data, rest left ventricular ejection fraction, and myocardial perfusion data including the identification of multivessel disease.^25–27 In a study by Lertsburapa et al. in 1600 patients, the relationship between all-cause mortality, perfusion-summed stress score, and ejection fraction was evaluated. The results indicate that the highest-risk patients were those with a high-summed stress score and a low ejection fraction (Fig. 10-9),²⁸ as also noted with SPECT.

An important and unique capability of cardiac PET perfusion imaging is the ability to measur myocardial blood flow (MBF). Typically, both PET and SPECT evaluate differences in regional uptake between rest and stress to identify patterns consistent with coronary stenosis. While this approach has been very useful, and this chapter has emphasized the important advantages of PET MPI, there are still limitations. The presence of normal perfusion does not exclude "balanced" ischemia ever with PET the abnormality present may underestimate the extent of CAD, and with pharmacologic stress one does not know whether the stress agent is successful at increasing blood flow. Currently, PET is unique as a noninvasive tool in its documented ability to measure regional and global MBF. MBF is commonly provided as a relationship between rest and stress conditions as MBF reserve, although proper use of blood flow should include examination of both the rest and stress data. There are several commercially available software tools for measuring MBF with both Rb-82 and N-13 ammonia which have been extensively validated. Data collection generally occurs within the first 2 minutes after injection of tracer, thus not extending the acquisition time.

The evaluation of MBF has now moved into the clinical arena adding both diagnostic and prognostic value to the perfusion imaging data. An important notion is that abnormal MBF can occur in the presence of normal or abnormal perfusion, and if abnormal in the presence of a single localized stenosis, multiple stenosis, diffuse epicardial disease, endothelial inflammation, isolated microvascular dysfunction, or combinations of all. Irrespective of perfusion abnormality, MBF offers incrementally significant improvements in diagnosis of hemodynamically significant CAD.

The clinical significance of the addition of MBF to perfusion data with cardiac PET is well established. The information of normal MBF in addition to normal PET perfusion substantially reduces the likelihood of obstructive CAD (several studies indicate 99% negative predictive value) and future cardiac events (see Fig. 10-10). Abnormal MBF can reflect other cardiac conditions such as endothelial dysfunction, diffuse microvascular disease and assess the efficacy of the stress agent.²⁹ Prognostic data from several sources suggest worse cardiac outcomes occur in patients with perfusion abnormalities and a severe reduction of MBF and myocardial flow reserve. Thus, the addition of MBF to PET perfusion imaging has added another dimension to the study that contributes considerably to the clinical information and subsequent decisions following a study. Any laboratory now performing or entering into the PET perfusion arena should strongly consider the value of MBF.

Multiple publications have demonstrated the prognostic value of MBF. Normal MBF values are associated with low coronary risk, compared to abnormal values which predict higher cardiac risk. These measurements additionally predict the presence and extent of perfusion defects and outcomes. Studies have demonstrated the prognostic value of flow measurements beyond information from the relative perfusion images in consecutive patient populations, in diabetics, in patients with renal failure, and in patients following heart transplants.^30–33 Thus, a patient with a single-vessel PET perfusion pattern but with markedly low myocardial blood flow reserve (difference between rest and stress) indicates more severe disease, and also high-grade stenosis that may require intervention. Another important scenario is that of lack of augmentation of blood flow in the presence of normal perfusion. This indicates that the pharmacologic stress agent has failed, perhaps due to caffeine ingestion or some other factor.

The prognostic value of MBF has been demonstrated by several studies, including a recent publication by Murthy et al. (Fig. 10-11).³⁴ The cumulative incidence of cardiac mortality was calculated by tertiles of MBFR, demonstrating a significant relationship between lowest to highest flow in cardiac mortality.

The incorporation of MBF into clinical reports is now occurring and referring physicians/health care providers are finding benefit. However, unlike perfusion reports in which national standards have been established for several years, there is no national consensus on when and what to report. Recently, guidance was provided by a publication of Bateman et al.²⁹ They recommend three general categories as follows:

1. Clearly normal-flow augmentation. If the study is well above the normal MBF for a particular software program, this finding indicates adequate pharmacologic stress and the risk of a major coronary event is very low.

2. Abnormally low-flow augmentation. This should raise concern for possible significant CAD in an otherwise normal perfusion study, and places the patient in a higher-risk category.

3. No-flow augmentation. This circumstance is either due to a technical error in the study or pharmaceutical failure due to antagonists such as caffeine. This is particularly important in a study in which normal perfusion is identified.

In summary, assessment of MBF during a cardiac PET perfusion study has been found to add considerable diagnostic and prognostic information. It should be strongly encouraged in clinical PET laboratories.

Cardiac PET perfusion imaging can be performed using several available radiotracers; the most common cardiac PET perfusion agent is Rb-82. There are over 200 laboratories using this tracer in the United States. N-13 ammonia is also being used clinically, but is far less prevalent (estimated 20 laboratories in the United States).

Cardiac PET perfusion imaging with rubidium is performed using pharmacologic stress because of the very short half-life of 75 seconds. This can be performed using vasodilator stress (regadenoson and dipyridamole) or inotropic stress (dobutamine). In general, adenosine is not recommended due to excessive patient motion during the acquisition. The protocol for cardiac PET perfusion imaging using the radiotracer Rb-82 is substantially faster than a SPECT protocol. A rest/stress protocol for SPECT imaging, regardless of radiopharmaceutical, requires at least 2.5 to as much as 4 hours for completion. In contrast, cardiac PET perfusion imaging protocols require 25 to 30 minutes for PET/CT camera (Fig. 10-12) and 35 to 40 minutes for a "dedicated" line source PET camera protocol (Fig. 10-13). The acquisition time for Rb-82 at both rest and stress is approximately 5 minutes. The strontium generator can produce a second dose of rubidium within 10 minutes of injection, thus the second dose is available when the stress protocol is completed. Because of the short protocol, it is recommended that the patient stay under the camera during the stress pharmaceutical injection. If MBF is being collected, most recommend the first 2 minutes after injection of both the rest and stress dose, thus not prolonging the procedure.

Cardiac PET perfusion studies offer high diagnostic accuracy and image quality, short protocols, low radiation exposure, and now additional information from MBF data. Thus, this test provides considerable value to the patient, which generally supersedes current SPECT procedures. Recently, a Joint Position Statement¹⁰ was published by the ASNC and the Society of Nuclear Medicine and Molecular Imaging (SNMMI) that addressed the choices between PET and SPECT (Table 10-1). These professional societies stated that PET was the "preferred test" for patients recommended for pharmacologic stress imaging in both patients with suspected or known coronary artery disease that met proper indications. The societies also extended their position on PET as a "recommended test" for several patient conditions beyond pharmacologic stress alone, as noted in Table 10-1. These conditions include challenging body habitus, a previous inconclusive test, "high-risk" patients in whom a diagnostic error test would result in other downstream testing, younger patients in whom radiation exposure is of considerable concern, and those patients in which MBF data would be clinically important. These patient groups are independent of exercise status. This document elucidates the current patients in which cardiac PET would be most useful. If longer-acting agents for PET perfusion such as F-18 based tracers (2-hour half-life) are developed, this may be extended.

Nonperfusion Indications for Cardiovascular PET

Within the last several years, indications for cardiovascular PET imaging beyond myocardial perfusion imaging have increased considerably. Many of these indications are addressed in more detail on subsequent chapters (Chapters 21 and 24). Currently, many of these indications utilize 2-[18F]-fluoro-2-deoxy-d-glucose (FDG). FDG is a glucose analog used extensively for oncology studies in identifying presence, severity, location, and treatment success of a wide range of tumors. This glucose agent is now being used in a variety of cardiovascular conditions described below.

Congestive heart failure is one of the most common cardiac reasons for admission to a hospital in the United States. This disease affects close to 6 million people with an incidence that approaches 10 per 1000 people who are older than age 65 years.²⁶ Heart failure is associated with a very high morbidity and mortality. Any improvement in left ventricular function would be beneficial to such patients. Cardiac revascularization is extremely beneficial in those patients who demonstrate myocardial viability and potentially detrimental to those patients who do not.

The identification of heart failure patients who would benefit from revascularization can be aided by conducting viability studies. It is known that under normal conditions the cardiac myocytes preferentially undergo fatty acid metabolism for energy. In contrast, myocytes that are under stress or that are damaged change to glucose metabolism for their energy source. As FDG is a glucose analog, this shift allows the use of FDG-labeled isotopes to test for myocardial viability. The FDG uptake can be enhanced by glucose loading a patient prior to the administration. This increase in serum glucose causes a reflexive increase in insulin levels that then causes a decrease in fatty acid metabolism. This results in an imaging mismatch between perfusion and metabolic imaging, resulting in an increase in glucose uptake into the myocytes. Myocardial viability assessment using PET procedures is fully discussed in Chapter 21.

Sarcoidosis is a multi-organ disease with no known etiology. It can affect many organ systems including the lymph nodes, skin, eyes, and the nervous, musculoskeletal, renal, and endocrine systems, and importantly the heart.³⁵ Patients inflicted with this systemic disease carry a good prognosis overall. In sharp contrast, however, patients with cardiac sarcoid involvement carry a considerably higher morbidity and mortality.

Cardiac sarcoidosis can affect any part of the heart, including both the left and right ventricles. The clinical severity depends on the location of the sarcoid granuloma, the degree of infiltration, and the amount of inflammation. The most clinically significant manifestations are heart failure and fatal and nonfatal arrhythmias, which account for the high incidence of death in these patients. The presenting signs and symptoms warranting evaluation are varied and may also include heart block, chest pain, pleural effusions, dyspnea, and new onset of heart failure. Arrhythmias range from asymptomatic conduction abnormalities to fatal ventricular tachycardia or fibrillation.

The diagnosis of cardiac sarcoid remains a difficult one to make, short of biopsy. The use of FDG has become a novel way to identify active cardiac sarcoid occurrence. The images obtained provide the clinician with evidence of location and severity of infiltration. A recent review demonstrates the value of this identification both for diagnosis as well as measuring treatment success versus failure.³⁶ For a full discussion of this procedure and its clinical value, see Chapter 24.

Cardiac sarcoid imaging takes advantage of the ability of FDG to identify in-glucose activity in inflammation conditions. More recently, this agent has been extended to identify infectious processes within the cardiovascular system, such as prosthetic valves, implanted devices or pockets, and graft infections.³⁷ There are emerging data on various vasculitis conditions as well, although this should be considered preliminary. Finally, the evaluation of coronary plaque is also becoming possible with the use of F-18 sodium fluoride (NaF). These conditions and applications are also discussed in Chapter 24.

A growing body of literature supports the value of cardiac PET perfusion imaging in clinical practice. Advantages include high diagnostic accuracy, excellent image quality with minimal attenuation artifact, rapid protocols, and low radiation exposure. This chapter presented the supporting clinical data, procedures, and radiopharmaceuticals available for cardiac PET studies; recent recommendations from ASNC/SNMMI support cardiac PET as the "preferred" imaging study in all patients undergoing pharmacologic stress. Optimal patients also include those with equivocal SPECT studies, obese patients, those with a high suspicion for coronary artery disease, and patients with known coronary artery disease. Emerging nonperfusion indications for cardiovascular PET including viability assessment, inflammation, and infection imaging have expanded the role of this modality.