REVIEW ARTICLE Proceedings of the cardiac PET summit meeting 12 may 2014: Cardiac PET and SPECT instrumentation Ernest V. Garcia, PhDa a
Department of Radiology and Imaging Sciences, Emory University, Atlanta, GA
Received Mar 2, 2015; accepted Mar 2, 2015 doi:10.1007/s12350-015-0114-7
Advances in PET and SPECT and imaging hardware and software are vastly improving the noninvasive evaluation of myocardial perfusion and function. PET perfusion imaging has benefitted from the introduction of novel detectors that now allow true 3D imaging, and precise attenuation correction (AC). These developments have also resulted in perfusion images with higher spatial and contrast resolution that may be acquired in shorter protocols and/or with less patient radiation exposure than traditional PET or SPECT studies. Hybrid PET/CT cameras utilize transmission computed tomographic (CT) scans for AC, and offer the additional clinical advantages of evaluating coronary calcium and myocardial anatomy but at a higher cost than PET scanners that use 68Ge radioactive line sources. As cardiac PET systems continue to improve, dedicated cardiac SPECT systems are also undergoing a profound change in their design. The scintillation camera general purpose design is being replaced with systems with multiple detectors focused on the heart yielding 5 to 10 times the sensitivity of conventional SPECT. As a result, shorter acquisition times and/or lower tracer doses produce higher quality SPECT images than were possible before. This article reviews these concepts and compares the attributes of PET and SPECT instrumentation. Key Words: Digital PET detectors Æ time of flight Æ 3D PET Æ hybrid PET/CT Æ CZT detectors Æ cost of PET instrumentation INTRODUCTION
PRINCIPLE OF 511 KEV PHOTON EMISSION
The accuracy, safety, and convenience of PET invariably arises from its inherent methods of detecting the intensity, location, and timing of counts emanating from injected radiopharmaceuticals and is related to the physics and technology of this imaging modality.
A positron is an antimatter particle and a positive electron. When emitted, it travels through matter and loses its kinetic energy. The positron (antimatter) and its corresponding particle matter with opposite charge, an electron, are then attracted and drift toward each other with zero momentum. When the matter and antimatter particles touch each other through the pair annihilation process, the mass of these two electrons is converted to energy via the equation E = mc2. The resting mass of the positron and its opposing electron are each 511 keV; because there is zero momentum when they touch, two photons are discharged along the same ray in directly opposite directions (Figure 1).1-3 The gamma rays resulting from the annihilation can thus be localized without lead collimation with the principles of coincidence detection. For this reason, PET cameras do not require mechanical collimators, and since there is no lead to stop photons, PET records or counts thousands
The current state of nuclear cardiology instrumentation for both PET and SPECT was recently discussed at the ASNC Cardiac PET Summit meeting held in Baltimore, MD on 12 May 2014. This article represents the information presented at that meeting. Reprint requests: Ernest V. Garcia, PhD, Department of Radiology and Imaging Sciences, Emory University, 101 Woodruff Circle, Room 1203, Atlanta, GA, 30322; [email protected] J Nucl Cardiol 2015;22:563–70. 1071-3581/$34.00 Copyright Ó 2015 American Society of Nuclear Cardiology.
563
564
Garcia Cardiac PET and SPECT instrumentation
F18
Journal of Nuclear CardiologyÒ May/June 2015
511 keV photon
ß+
e511 keV photon E=mc 2
Figure 1. Principle of positron annihilation. PET cameras detect paired gamma rays (511 keV of energy each), produced by the positron annihilations. The paired gamma rays travel in opposite directions discharged at a 180-degree angle to each other. Thus, positron decay can be localized without collimation with the use of the principles of coincidence detection. Since PET cameras do not require collimators, these systems have a much higher count rate than SPECT systems. Inset shows how the two 511 keV photons are used to localize the radioactive event using electronic collimation.
more photons per mCi than SPECT systems. Thus, one of the PET’s major attributes is that it has a superior count sensitivity than SPECT. LIMITATIONS ON PET RESOLUTION The higher the kinetic energy of an emitted positron, the longer it travels in the patient before it interacts with an electron in an annihilation event converting its combined masses into energy. When the positron and electron pair touches, they still produce a total of 2 9 511 keV, but because of residual kinetic energy both the 511 keV photons may not necessarily be discharged in exactly opposite directions. An energetic positron will transfer more than 511 keV, and the resulting photons will not be discharged along the same ray. Depending on the residual momentum, the photons may be emitted as much as a half-degree
away from 180° of each other (Figure 2). Resolution of PET images is ultimately determined by the distance the positron travels from the emitted nucleus to the point of annihilation (the range),4,5 the residual kinetic energy at the time of pair annihilation, random counts, and the finite thickness of the detectors. Their physical properties predict that fluorine (F18) would provide higher resolution than rubidium because fluorine’s positron range (0.6 mm) is considerably lower than that of rubidium (4 mm). Theoretically, imaging in a strong magnetic field can reduce the positron range and kinetic energy of the emitted photon thus increasing resolution,5 a potential benefit of hybrid PET/MR systems. Figure 3 demonstrates how the path of an emitted positron in a strong magnetic field spirals inward significantly reducing its range and thus potentially increasing the spatial resolution of the system.
Journal of Nuclear CardiologyÒ Volume 22, Number 3;563–70
Garcia Cardiac PET and SPECT instrumentation
565
1-4 mm + randoms
180 ± 0.5º
0-5 mm Ave. Range in Water F-18 = .64 mm Rb-82 = 4.29 mm Figure 2. Limitations in PET resolution. Resolution losses are due to the travel of the positron from the emitted nucleus, the residual kinetic energy at the time of pair annihilation, the finite thickness of the detectors, and random radiation. Inset shows how the reconstructed ray (dashed line) does not cross the location of the emitted positron.
HARDWARE APPROACHES TO IMPROVE PET Improvements in gantry design have been focused on hybrid imaging systems and in detector design. Improvements in detector design have consisted of using new scintillation crystals4 and solid-state detectors6 and in moving from two-dimensional (2D) to three-dimensional (3D) imaging.7 Compared to conventional bismuth germanate (BGO) scintillation crystals, newer crystals like lutetium oxyorthosilicate (LSO), lutetiumyttrium oxyorthosilicate (LYSO), and cerium-doped Gd2SiO5 (GSO) have a higher light output and significantly shorter decay constant. The higher light output results in improved energy and spatial resolution, while the faster decay reduces random coincidences which affect the counting rate capabilities of the system. DIGITAL PET DETECTORS Similar to the use of solid-state detectors in SPECT, PET cameras have started replacing conventional photomultiplier tubes with solid-state silicone photomultipliers
coupled to LYSO crystals.6 When these detection modules are stacked up and placed angularly around the patient they allow not only the recognition of which angular detector detected the photon (one of the 511 keV pairs) but also the depth of the interaction (DOI). This additional information allows for a more accurate reprojection of the projected counts and is predicted to result in a factor of two improvement in volumetric resolution and thus quantitative accuracy (Figure 4).
2D VS 3D PET SYSTEMS 2D PET systems (Figure 5A), equipped with lead septa, only accept coincidences from crystals in the same ring of detectors. 3D PET systems7 (Figure 5B), by removing the septa, accept coincidences in any ring and greatly increase count rate and sensitivity, usually by a factor of 5 in the central ring. This increase in count sensitivity can be traded for a lower tracer dose and thus less radiation to the patient. Moreover, this attribute is also used to improve the quality of studies at the end of
566
Garcia Cardiac PET and SPECT instrumentation
Journal of Nuclear CardiologyÒ May/June 2015
detector ring
+
ß+ Path annihilation photons
xxxxx xxBxx xxxxx Figure 3. Path of positron in strong magnetic field improves spatial resolution in PET systems. Spiral path of a positron (b?) under the influence of a strong magnetic field (B) perpendicular to the page (x) as recorded in a bubble chamber. Note that the range of the positron is significantly reduced by the force exerted perpendicular to its path from by the magnetic field causing it to spiral and annihilate much closer to the nucleus (depicted by the green ball) than if there was no magnetic field where the positron path would look more like shown in Figure 1.
+
Figure 4. Improvement in spatial resolution using depth of interaction (DOI) digital PET detectors. As the positron annihilation takes place further from the center of the field of view the reconstructed line of response (red) is increasingly further away from the true line of response (green) in systems that locate the annihilation photons in the middle of the detector rather than at the true depth of the interaction as with DOI detectors.
TOF systems compared to conventional non-TOF PET systems (Figure 6). PET ATTENUATION CORRECTION
82
the usable cycle for Rb generators, particularly in obese patients. However, the difficulties associated with removing the septa are as follows: (1) it greatly increases scatter; (2) it greatly increases random events; and (3) it greatly increases count rate, and so greatly increases dead time.8 These problems are now effectively compensated for the use of 3D PET in cardiac imaging. TIME OF FLIGHT (TOF) The coincidence electronics in advanced PET scanners with TOF electronics are capable of measuring the time interval between one photon hitting one detector and the second photon from the same annihilation event hitting an opposing detector.4,9 That difference in time multiplied by the speed of light estimates the location of the annihilation event along the coincidence ray between the two detectors. This allows TOF scanners to localize an annihilation event to a much smaller directional ray than conventional PET scanners, and results in increased spatial resolution. To date, this improvement has not been applied to cardiac imaging because of cardiac motion during contraction and chest motion during breathing. Nevertheless, the advent of respiratory gating10 and freeze-motion11 algorithms should eventually improve the image quality of cardiac studies acquired on
The energy levels of PET radiotracers (511 keV) are considerably higher than those of the 99mTc-labeled radiopharmaceuticals (140 keV) and thallium-201 (7090 keV) used in SPECT imaging. The attenuation of imaging single photons in SPECT is compounded by having to image two photons in PET in order to locate an event.2 When they pass through soft tissue, the two 511 keV photons must be detected in coincidence before the event can be recorded along with the direction of the radiation necessary to create images. Thus, even though each 511 keV photon attenuates less than a single 140 keV Tc-99m photon, the fact that two photons are needed to be recorded in PET translates to more attenuation in PET than SPECT. Figure 7 shows how because the two annihilation photons project along the same ray, they must travel through the same total amount of tissue regardless of where along the ray they were emitted and therefore the measured attenuation in PET is exactly the same along this line, whereas in SPECT the attenuation would vary exponentially. These phenomena translate to a very simple algorithm for correcting attenuation from PET images and a very complicated one for SPECT. Unlike SPECT, PET data can be accurately corrected for attenuation by simply multiplying each projection line with the appropriate attenuation correction (AC) factor determined from a
Journal of Nuclear CardiologyÒ Volume 22, Number 3;563–70
Garcia Cardiac PET and SPECT instrumentation
567
A 2D PET Coincidence Circuit
Pb Septa
B 3D PET Coincidence Circuit
Pb Side shields
Figure 5. 2D vs 3D PET Acquisition. 2D PET systems, equipped with lead septa, only accept coincidences from crystals in the same ring of detectors. 3D PET systems, by removing the septa, accept coincidences in any ring and greatly increase count rate and sensitivity. However, the difficulties associated with removing the septa are as follows: (1) it greatly increases scatter; (2) it greatly increases random events; and (3) it greatly increases count rate, and so greatly increases dead time. These problems must be effectively compensated for the use of 3D PET in cardiac imaging.
transmission germanium-68 line source or a CT scan. Advanced software effectively manages these challenges. For both PET and SPECT, measurement of patient-specific attenuation maps is required for accurate AC. PET imaging must always include attenuation correction. Attenuation correction during PET myocardial perfusion imaging has traditionally been performed with a germanium-68 radioactive line source but more recently is being performed with CT using hybrid PET/CT systems.12 Transmission scans with 68Ge are acquired measuring the attenuation of 511 keV photons, but the 68 Ge lines decay and are expensive to replace. The availability of hybrid PET/CT systems offers the advantage that CT produces the same quality transmission images over time with a fixed initial cost. However, CT and PET data are not acquired simultaneously, and miss-registration of the emission vs perfusion scans can result in significant errors in attenuation correction.
Moreover, since the transmission scan is measured with the continuous energy spectrum of x-rays, there is a small but measureable error in converting the x-ray attenuation to the mono-energetic 511 keV photons that are attenuated in PET. Experimental hybrid PET/MR systems under development offer the promise of improved registration if acquired simultaneously. Unfortunately the MR scan does not directly provide a map of attenuation coefficients in the thorax and thus must be assigned via segmentation algorithms.13 High-Quality PET Systems Most new high-quality PET systems today come from GE,14 Siemens,15 and Philips.16 Almost all new systems come as PET/CT hybrid systems and use 3D geometry, although Positron Corporation17 still offers PET scanners as an option for a 2D geometry, non-CT, Ge-68 sources transmission system which cost less than
568
Garcia Cardiac PET and SPECT instrumentation
Non-TOF
Journal of Nuclear CardiologyÒ May/June 2015
TOF
Figure 6. Non-time-of-flight (TOF) vs time-of-flight imaging. In TOF, the difference in the time when the two detectors recorded the event is multiplied by the speed of light to position the event along the back projected line. In TOF, instead of backprojecting an entire line only the line segment corresponding to the time window of the event is projected. This results in increased lesion contrast in TOF vs non-TOF imaging.
new PET/CT systems. Each manufacturer has key differentiating features trading-off price with accurate high-resolution images and scan efficiency in terms of dose and time. Philips offers models with TOF and digital detectors described here. Each manufacturer offers several models with price depending on many variables including the number and type of PET detectors determining the axial FOV, the type of specialized electronics and the number of slices in the CT scanner. Although the dollar cost of a new PET/CT system can go into the millions the cost of excellent refurbished PET systems capable of cardiac imaging can be found between $250,000 (Ge-68 transmission lines) and $750,000 (PET/CT) depending on the number of slices offered by the CT system.18 Cardiac imaging centers must be aware that many of these PET/CT systems come without the needed cardiac ECG gating and/or cardiac software unless itemized in the purchase. This omission can create delays when cardiac imagers want to use scanning time on a PET/CT scanner used by radiology for oncologic imaging. Comparison of PET and SPECT As cardiac PET systems continue to improve, dedicated cardiac SPECT systems are undergoing a profound change in their design for the first time in fifty years.19 The scintillation camera general purpose design
is being replaced with systems with multiple detectors focused on the heart yielding 5 to 10 times the sensitivity of conventional SPECT.20 Some of the designs also replace the NaI(Tl) crystal with solid-state CZT electronic detectors with superior energy resolution and thus reduced scatter. There are also significant innovations in reconstruction software incorporated into these newly designed systems that take into account the true physics of the SPECT reconstruction geometry to gain at least a factor of 2 in count sensitivity. These innovations are resulting in shorter study time and/or reduced radiation dose to the patient. Some of these new systems are also ideally suited for dynamic applications allowing the possibility of flow measurements with SPECT. The fast acquisition also makes the hybrid SPECT/CT systems more practical since it allows the CT scanner to be used for a longer part of the day. Several technical issues are highlighted when comparing cardiac SPECT to cardiac PET3. Cardiac PET was developed as a quantitative cardiac measurement tool since its inception. This commitment included performing (a) attenuation correction in all studies, (b) calibrating the scanners with known activity concentrations for measuring absolute counts, (c) acquiring the first pass of the perfusion tracer through the heart to capture an input function to measure absolute flow, and (d) to use perfusion tracers with high enough cellular extraction fractions to permit the measurement of flow at
Journal of Nuclear CardiologyÒ Volume 22, Number 3;563–70
Garcia Cardiac PET and SPECT instrumentation
569
Annihilation event
γ1 γ1 γ2
γ2
AC Corrected Un-Corrected Figure 7. Need and simplicity of attenuation correction with PET. Demonstration of how the two annihilation photons (c1 and c2) project along the same ray, they must travel through the same total amount of tissue regardless of where along the ray the annihilation took place. Therefore, the measured attenuation in PET is exactly the same along this line for the annihilation that took place in the left panel or in a different location but along the same ray in the right panel. The dramatic differences in quality between non-corrected 18F-fluorodeoxyglucose (FDG) transverse axial slice and after applying AC correction factors of 9.6 to the septal wall and 6.3 to the lateral wall are readily apparent in the image inset.
an acceptable error level. In cardiac SPECT, on the other hand, fewer than 15 % of laboratories perform attenuation correction and the acquisition of a first pass has all but disappeared.21,22 Moreover, perhaps the most limiting factors are the low extraction fraction of the Tc99m tracers currently used commercially and the loss in count sensitivity due to the use of SPECT lead collimators (compared to electronic collimation in PET). Both of these limiting factors do not preclude the ability of SPECT to measure relative flow, absolute flow, and flow reserve but they do increase the errors of their measurements when trying to separate normal patients from patients with clinically treatable flow impairments. Whether these issues affect the diagnostic performance of SPECT as compared to PET is the topic of another section in this series of articles. PET Instrumentation Obstacles to Overcome Perhaps the main obstacle to cardiac PET instrumentation is the limited number of PET scanners
available for cardiac imaging. As described above, PET scanners are already quite sophisticated tools to image and quantify cardiac functional parameters. In fact some would say that they are too sophisticated and sophistication brings up scanner cost with an appeal to industry to offer non-hybrid dedicated cardiac PET systems without CT to reduce the cost of the scanner and thus promote more widespread purchases and availability. Such instrumentation is available generally in the form of refurbished cameras on the secondary market. While these cameras are several years old, the crystal technology provides more that adequate PET images. The cost is similar to new instrumentation for SPECT imaging. Another approach to increase availability of cardiac PET is to increase the utilization of oncologic PET scanners by partnering with radiologists and/or nuclear medicine physicians to use their empty slots. In summary, both cardiac PET and SPECT instrumentation continue to make important advances in all technical aspects for improving image quality and quantitative accuracy. The present state of the art
570
Garcia Cardiac PET and SPECT instrumentation
instruments offers a definite advantage to PET over SPECT in these attributes. This is due to both the superior inherent principles of PET imaging and the effort and investments made in PET by the imaging industry. These advantages of cardiac PET imaging warrant that our field finds a solution to the issue of limited availability. Disclosure The authors have indicated that they have no financial conflict of interest.
References 1. Cherry SR, Sorenson JA, Phelps ME. Physics in nuclear medicine. 3rd ed. Philadelphia: Saunders; 2003. p. 511. 2. Garcia EV, Galt R, Faber TL, Chen J. Atlas of nuclear cardiology (3rd edn), In: Vasken D, Jagat N, editors. Current medicine LLC, Chapter 1. 2009. pp 1-34. 3. Garcia EV, Faber TL. Advances in nuclear cardiology instrumentation. Clinical potential of SPECT and PET. Curr Cardiovasc Imaging Rep 2009;2:230-7. 4. Lecomte R. Novel detector technology for clinical PET. Eur J Nucl Med Mol Imaging 2009;36:S69-85. 5. Hammer BE, Christensen NL, Heil BG. Use of magnetic field to increase spatial resolution of positron emission tomography. Med Phys 1994;21:1917-20. 6. Schaart DR, van Dam HT, Seifert S, Vinke R, Dendoover P, Lohner H, Beekman FJ. SiPM-array based PET detectors with depth-of-interaction correction. IEEE Nucl Sci Symposium Conference Record 2008; 3581-3585. 7. deKemp RA, Yoshinaga K, Beanlands RSB. Will 3-dimensional PET-CT enable the routine quantification of myocardial blood flow? J Nucl Cardiol 2007;14:380-97.
Journal of Nuclear CardiologyÒ May/June 2015
8. Votaw JR, White M. Comparison of 2-dimensional and 3-dimensional cardiac Rb-82 PET studies. J Nucl Med 2001;42:701-6. 9. Karp JS, Surti S, Daube-Witherspoon ME, Muehllehner G. Benefit of Time-of-Flight in PET: Experimental and clinical results. J Nucl Med 2008;49:462-70. 10. Livieratos L, Rajappan K, Stegger L, Schafers K, Bailey DL, Camici PG. Respiratory gating of cardiac PET data in list-mode acquisition. Eur J Nucl Med Mol Imaging 2006;33:584-8. 11. Meunier LL, Slomka PJ, Dey D, Ramesh A, Thomson LEJ, Hayes SW, et al. Motion frozen F-18-FDG cardiac PET. J Nucl Cardiol 2011;18:259-66. 12. Koepfli P, Hany TF, Wyss CA, Namdar M, Burger C, Konstantinidis AV, et al. CT attenuation correction for myocardial perfusion quantification using a PET/CT hybrid scanner. J Nucl Med 2004;45:537-42. 13. Nensa F, Poeppel TD, Beiderwellen K, Schelhorn J, Mahabadi AA, Erbel R, et al. Hybrid PET/MR imaging of the heart: Feasibility and initial results. Radiology. 2013;268:366-73. 14. http://www3.gehealthcare.com/en/products/categories/pet-ct/petct_scanners/. Accessed Aug 11, 2014. 15. http://usa.healthcare.siemens.com/molecular-imaging/pet-ct. Accessed Aug 11, 2014. 16. http://www.healthcare.philips.com/main/products/nuclearmedicine/ products/pet/. Accessed Aug 11, 2014. 17. http://www.positron.com/attrius. Accessed Aug 11, 2014. 18. How much does a PET/CT scanner cost? 2012/2013 update. http://info.blockimaging.com/bid/90851/How-Much-Does-a-PETCT-Scanner-Cost-2012-2013-Update. Accessed Aug 11, 2014. 19. Anger HO. Scintillation camera. Rev Sci Instrum 1958;29:27-33. 20. Garcia EV, Faber TL. New trends in camera and software technology in nuclear cardiology. Cardiol Clin 2009;27:227-36. 21. Garcia EV. Are absolute myocardial blood flow PET measurements ready for clinical use? J Nucl Cardiol 2014;21:857-8. 22. Garcia EV. Are SPECT measurements of myocardial blood flow and flow reserve ready for clinical use? Eur J Nucl Med Mol Imaging 2014;41:2291-3.
Department of Radiology and Imaging Sciences, Emory University, Atlanta, GA
Received Mar 2, 2015; accepted Mar 2, 2015 doi:10.1007/s12350-015-0114-7
Advances in PET and SPECT and imaging hardware and software are vastly improving the noninvasive evaluation of myocardial perfusion and function. PET perfusion imaging has benefitted from the introduction of novel detectors that now allow true 3D imaging, and precise attenuation correction (AC). These developments have also resulted in perfusion images with higher spatial and contrast resolution that may be acquired in shorter protocols and/or with less patient radiation exposure than traditional PET or SPECT studies. Hybrid PET/CT cameras utilize transmission computed tomographic (CT) scans for AC, and offer the additional clinical advantages of evaluating coronary calcium and myocardial anatomy but at a higher cost than PET scanners that use 68Ge radioactive line sources. As cardiac PET systems continue to improve, dedicated cardiac SPECT systems are also undergoing a profound change in their design. The scintillation camera general purpose design is being replaced with systems with multiple detectors focused on the heart yielding 5 to 10 times the sensitivity of conventional SPECT. As a result, shorter acquisition times and/or lower tracer doses produce higher quality SPECT images than were possible before. This article reviews these concepts and compares the attributes of PET and SPECT instrumentation. Key Words: Digital PET detectors Æ time of flight Æ 3D PET Æ hybrid PET/CT Æ CZT detectors Æ cost of PET instrumentation INTRODUCTION
PRINCIPLE OF 511 KEV PHOTON EMISSION
The accuracy, safety, and convenience of PET invariably arises from its inherent methods of detecting the intensity, location, and timing of counts emanating from injected radiopharmaceuticals and is related to the physics and technology of this imaging modality.
A positron is an antimatter particle and a positive electron. When emitted, it travels through matter and loses its kinetic energy. The positron (antimatter) and its corresponding particle matter with opposite charge, an electron, are then attracted and drift toward each other with zero momentum. When the matter and antimatter particles touch each other through the pair annihilation process, the mass of these two electrons is converted to energy via the equation E = mc2. The resting mass of the positron and its opposing electron are each 511 keV; because there is zero momentum when they touch, two photons are discharged along the same ray in directly opposite directions (Figure 1).1-3 The gamma rays resulting from the annihilation can thus be localized without lead collimation with the principles of coincidence detection. For this reason, PET cameras do not require mechanical collimators, and since there is no lead to stop photons, PET records or counts thousands
The current state of nuclear cardiology instrumentation for both PET and SPECT was recently discussed at the ASNC Cardiac PET Summit meeting held in Baltimore, MD on 12 May 2014. This article represents the information presented at that meeting. Reprint requests: Ernest V. Garcia, PhD, Department of Radiology and Imaging Sciences, Emory University, 101 Woodruff Circle, Room 1203, Atlanta, GA, 30322; [email protected] J Nucl Cardiol 2015;22:563–70. 1071-3581/$34.00 Copyright Ó 2015 American Society of Nuclear Cardiology.
563
564
Garcia Cardiac PET and SPECT instrumentation
F18
Journal of Nuclear CardiologyÒ May/June 2015
511 keV photon
ß+
e511 keV photon E=mc 2
Figure 1. Principle of positron annihilation. PET cameras detect paired gamma rays (511 keV of energy each), produced by the positron annihilations. The paired gamma rays travel in opposite directions discharged at a 180-degree angle to each other. Thus, positron decay can be localized without collimation with the use of the principles of coincidence detection. Since PET cameras do not require collimators, these systems have a much higher count rate than SPECT systems. Inset shows how the two 511 keV photons are used to localize the radioactive event using electronic collimation.
more photons per mCi than SPECT systems. Thus, one of the PET’s major attributes is that it has a superior count sensitivity than SPECT. LIMITATIONS ON PET RESOLUTION The higher the kinetic energy of an emitted positron, the longer it travels in the patient before it interacts with an electron in an annihilation event converting its combined masses into energy. When the positron and electron pair touches, they still produce a total of 2 9 511 keV, but because of residual kinetic energy both the 511 keV photons may not necessarily be discharged in exactly opposite directions. An energetic positron will transfer more than 511 keV, and the resulting photons will not be discharged along the same ray. Depending on the residual momentum, the photons may be emitted as much as a half-degree
away from 180° of each other (Figure 2). Resolution of PET images is ultimately determined by the distance the positron travels from the emitted nucleus to the point of annihilation (the range),4,5 the residual kinetic energy at the time of pair annihilation, random counts, and the finite thickness of the detectors. Their physical properties predict that fluorine (F18) would provide higher resolution than rubidium because fluorine’s positron range (0.6 mm) is considerably lower than that of rubidium (4 mm). Theoretically, imaging in a strong magnetic field can reduce the positron range and kinetic energy of the emitted photon thus increasing resolution,5 a potential benefit of hybrid PET/MR systems. Figure 3 demonstrates how the path of an emitted positron in a strong magnetic field spirals inward significantly reducing its range and thus potentially increasing the spatial resolution of the system.
Journal of Nuclear CardiologyÒ Volume 22, Number 3;563–70
Garcia Cardiac PET and SPECT instrumentation
565
1-4 mm + randoms
180 ± 0.5º
0-5 mm Ave. Range in Water F-18 = .64 mm Rb-82 = 4.29 mm Figure 2. Limitations in PET resolution. Resolution losses are due to the travel of the positron from the emitted nucleus, the residual kinetic energy at the time of pair annihilation, the finite thickness of the detectors, and random radiation. Inset shows how the reconstructed ray (dashed line) does not cross the location of the emitted positron.
HARDWARE APPROACHES TO IMPROVE PET Improvements in gantry design have been focused on hybrid imaging systems and in detector design. Improvements in detector design have consisted of using new scintillation crystals4 and solid-state detectors6 and in moving from two-dimensional (2D) to three-dimensional (3D) imaging.7 Compared to conventional bismuth germanate (BGO) scintillation crystals, newer crystals like lutetium oxyorthosilicate (LSO), lutetiumyttrium oxyorthosilicate (LYSO), and cerium-doped Gd2SiO5 (GSO) have a higher light output and significantly shorter decay constant. The higher light output results in improved energy and spatial resolution, while the faster decay reduces random coincidences which affect the counting rate capabilities of the system. DIGITAL PET DETECTORS Similar to the use of solid-state detectors in SPECT, PET cameras have started replacing conventional photomultiplier tubes with solid-state silicone photomultipliers
coupled to LYSO crystals.6 When these detection modules are stacked up and placed angularly around the patient they allow not only the recognition of which angular detector detected the photon (one of the 511 keV pairs) but also the depth of the interaction (DOI). This additional information allows for a more accurate reprojection of the projected counts and is predicted to result in a factor of two improvement in volumetric resolution and thus quantitative accuracy (Figure 4).
2D VS 3D PET SYSTEMS 2D PET systems (Figure 5A), equipped with lead septa, only accept coincidences from crystals in the same ring of detectors. 3D PET systems7 (Figure 5B), by removing the septa, accept coincidences in any ring and greatly increase count rate and sensitivity, usually by a factor of 5 in the central ring. This increase in count sensitivity can be traded for a lower tracer dose and thus less radiation to the patient. Moreover, this attribute is also used to improve the quality of studies at the end of
566
Garcia Cardiac PET and SPECT instrumentation
Journal of Nuclear CardiologyÒ May/June 2015
detector ring
+
ß+ Path annihilation photons
xxxxx xxBxx xxxxx Figure 3. Path of positron in strong magnetic field improves spatial resolution in PET systems. Spiral path of a positron (b?) under the influence of a strong magnetic field (B) perpendicular to the page (x) as recorded in a bubble chamber. Note that the range of the positron is significantly reduced by the force exerted perpendicular to its path from by the magnetic field causing it to spiral and annihilate much closer to the nucleus (depicted by the green ball) than if there was no magnetic field where the positron path would look more like shown in Figure 1.
+
Figure 4. Improvement in spatial resolution using depth of interaction (DOI) digital PET detectors. As the positron annihilation takes place further from the center of the field of view the reconstructed line of response (red) is increasingly further away from the true line of response (green) in systems that locate the annihilation photons in the middle of the detector rather than at the true depth of the interaction as with DOI detectors.
TOF systems compared to conventional non-TOF PET systems (Figure 6). PET ATTENUATION CORRECTION
82
the usable cycle for Rb generators, particularly in obese patients. However, the difficulties associated with removing the septa are as follows: (1) it greatly increases scatter; (2) it greatly increases random events; and (3) it greatly increases count rate, and so greatly increases dead time.8 These problems are now effectively compensated for the use of 3D PET in cardiac imaging. TIME OF FLIGHT (TOF) The coincidence electronics in advanced PET scanners with TOF electronics are capable of measuring the time interval between one photon hitting one detector and the second photon from the same annihilation event hitting an opposing detector.4,9 That difference in time multiplied by the speed of light estimates the location of the annihilation event along the coincidence ray between the two detectors. This allows TOF scanners to localize an annihilation event to a much smaller directional ray than conventional PET scanners, and results in increased spatial resolution. To date, this improvement has not been applied to cardiac imaging because of cardiac motion during contraction and chest motion during breathing. Nevertheless, the advent of respiratory gating10 and freeze-motion11 algorithms should eventually improve the image quality of cardiac studies acquired on
The energy levels of PET radiotracers (511 keV) are considerably higher than those of the 99mTc-labeled radiopharmaceuticals (140 keV) and thallium-201 (7090 keV) used in SPECT imaging. The attenuation of imaging single photons in SPECT is compounded by having to image two photons in PET in order to locate an event.2 When they pass through soft tissue, the two 511 keV photons must be detected in coincidence before the event can be recorded along with the direction of the radiation necessary to create images. Thus, even though each 511 keV photon attenuates less than a single 140 keV Tc-99m photon, the fact that two photons are needed to be recorded in PET translates to more attenuation in PET than SPECT. Figure 7 shows how because the two annihilation photons project along the same ray, they must travel through the same total amount of tissue regardless of where along the ray they were emitted and therefore the measured attenuation in PET is exactly the same along this line, whereas in SPECT the attenuation would vary exponentially. These phenomena translate to a very simple algorithm for correcting attenuation from PET images and a very complicated one for SPECT. Unlike SPECT, PET data can be accurately corrected for attenuation by simply multiplying each projection line with the appropriate attenuation correction (AC) factor determined from a
Journal of Nuclear CardiologyÒ Volume 22, Number 3;563–70
Garcia Cardiac PET and SPECT instrumentation
567
A 2D PET Coincidence Circuit
Pb Septa
B 3D PET Coincidence Circuit
Pb Side shields
Figure 5. 2D vs 3D PET Acquisition. 2D PET systems, equipped with lead septa, only accept coincidences from crystals in the same ring of detectors. 3D PET systems, by removing the septa, accept coincidences in any ring and greatly increase count rate and sensitivity. However, the difficulties associated with removing the septa are as follows: (1) it greatly increases scatter; (2) it greatly increases random events; and (3) it greatly increases count rate, and so greatly increases dead time. These problems must be effectively compensated for the use of 3D PET in cardiac imaging.
transmission germanium-68 line source or a CT scan. Advanced software effectively manages these challenges. For both PET and SPECT, measurement of patient-specific attenuation maps is required for accurate AC. PET imaging must always include attenuation correction. Attenuation correction during PET myocardial perfusion imaging has traditionally been performed with a germanium-68 radioactive line source but more recently is being performed with CT using hybrid PET/CT systems.12 Transmission scans with 68Ge are acquired measuring the attenuation of 511 keV photons, but the 68 Ge lines decay and are expensive to replace. The availability of hybrid PET/CT systems offers the advantage that CT produces the same quality transmission images over time with a fixed initial cost. However, CT and PET data are not acquired simultaneously, and miss-registration of the emission vs perfusion scans can result in significant errors in attenuation correction.
Moreover, since the transmission scan is measured with the continuous energy spectrum of x-rays, there is a small but measureable error in converting the x-ray attenuation to the mono-energetic 511 keV photons that are attenuated in PET. Experimental hybrid PET/MR systems under development offer the promise of improved registration if acquired simultaneously. Unfortunately the MR scan does not directly provide a map of attenuation coefficients in the thorax and thus must be assigned via segmentation algorithms.13 High-Quality PET Systems Most new high-quality PET systems today come from GE,14 Siemens,15 and Philips.16 Almost all new systems come as PET/CT hybrid systems and use 3D geometry, although Positron Corporation17 still offers PET scanners as an option for a 2D geometry, non-CT, Ge-68 sources transmission system which cost less than
568
Garcia Cardiac PET and SPECT instrumentation
Non-TOF
Journal of Nuclear CardiologyÒ May/June 2015
TOF
Figure 6. Non-time-of-flight (TOF) vs time-of-flight imaging. In TOF, the difference in the time when the two detectors recorded the event is multiplied by the speed of light to position the event along the back projected line. In TOF, instead of backprojecting an entire line only the line segment corresponding to the time window of the event is projected. This results in increased lesion contrast in TOF vs non-TOF imaging.
new PET/CT systems. Each manufacturer has key differentiating features trading-off price with accurate high-resolution images and scan efficiency in terms of dose and time. Philips offers models with TOF and digital detectors described here. Each manufacturer offers several models with price depending on many variables including the number and type of PET detectors determining the axial FOV, the type of specialized electronics and the number of slices in the CT scanner. Although the dollar cost of a new PET/CT system can go into the millions the cost of excellent refurbished PET systems capable of cardiac imaging can be found between $250,000 (Ge-68 transmission lines) and $750,000 (PET/CT) depending on the number of slices offered by the CT system.18 Cardiac imaging centers must be aware that many of these PET/CT systems come without the needed cardiac ECG gating and/or cardiac software unless itemized in the purchase. This omission can create delays when cardiac imagers want to use scanning time on a PET/CT scanner used by radiology for oncologic imaging. Comparison of PET and SPECT As cardiac PET systems continue to improve, dedicated cardiac SPECT systems are undergoing a profound change in their design for the first time in fifty years.19 The scintillation camera general purpose design
is being replaced with systems with multiple detectors focused on the heart yielding 5 to 10 times the sensitivity of conventional SPECT.20 Some of the designs also replace the NaI(Tl) crystal with solid-state CZT electronic detectors with superior energy resolution and thus reduced scatter. There are also significant innovations in reconstruction software incorporated into these newly designed systems that take into account the true physics of the SPECT reconstruction geometry to gain at least a factor of 2 in count sensitivity. These innovations are resulting in shorter study time and/or reduced radiation dose to the patient. Some of these new systems are also ideally suited for dynamic applications allowing the possibility of flow measurements with SPECT. The fast acquisition also makes the hybrid SPECT/CT systems more practical since it allows the CT scanner to be used for a longer part of the day. Several technical issues are highlighted when comparing cardiac SPECT to cardiac PET3. Cardiac PET was developed as a quantitative cardiac measurement tool since its inception. This commitment included performing (a) attenuation correction in all studies, (b) calibrating the scanners with known activity concentrations for measuring absolute counts, (c) acquiring the first pass of the perfusion tracer through the heart to capture an input function to measure absolute flow, and (d) to use perfusion tracers with high enough cellular extraction fractions to permit the measurement of flow at
Journal of Nuclear CardiologyÒ Volume 22, Number 3;563–70
Garcia Cardiac PET and SPECT instrumentation
569
Annihilation event
γ1 γ1 γ2
γ2
AC Corrected Un-Corrected Figure 7. Need and simplicity of attenuation correction with PET. Demonstration of how the two annihilation photons (c1 and c2) project along the same ray, they must travel through the same total amount of tissue regardless of where along the ray the annihilation took place. Therefore, the measured attenuation in PET is exactly the same along this line for the annihilation that took place in the left panel or in a different location but along the same ray in the right panel. The dramatic differences in quality between non-corrected 18F-fluorodeoxyglucose (FDG) transverse axial slice and after applying AC correction factors of 9.6 to the septal wall and 6.3 to the lateral wall are readily apparent in the image inset.
an acceptable error level. In cardiac SPECT, on the other hand, fewer than 15 % of laboratories perform attenuation correction and the acquisition of a first pass has all but disappeared.21,22 Moreover, perhaps the most limiting factors are the low extraction fraction of the Tc99m tracers currently used commercially and the loss in count sensitivity due to the use of SPECT lead collimators (compared to electronic collimation in PET). Both of these limiting factors do not preclude the ability of SPECT to measure relative flow, absolute flow, and flow reserve but they do increase the errors of their measurements when trying to separate normal patients from patients with clinically treatable flow impairments. Whether these issues affect the diagnostic performance of SPECT as compared to PET is the topic of another section in this series of articles. PET Instrumentation Obstacles to Overcome Perhaps the main obstacle to cardiac PET instrumentation is the limited number of PET scanners
available for cardiac imaging. As described above, PET scanners are already quite sophisticated tools to image and quantify cardiac functional parameters. In fact some would say that they are too sophisticated and sophistication brings up scanner cost with an appeal to industry to offer non-hybrid dedicated cardiac PET systems without CT to reduce the cost of the scanner and thus promote more widespread purchases and availability. Such instrumentation is available generally in the form of refurbished cameras on the secondary market. While these cameras are several years old, the crystal technology provides more that adequate PET images. The cost is similar to new instrumentation for SPECT imaging. Another approach to increase availability of cardiac PET is to increase the utilization of oncologic PET scanners by partnering with radiologists and/or nuclear medicine physicians to use their empty slots. In summary, both cardiac PET and SPECT instrumentation continue to make important advances in all technical aspects for improving image quality and quantitative accuracy. The present state of the art
570
Garcia Cardiac PET and SPECT instrumentation
instruments offers a definite advantage to PET over SPECT in these attributes. This is due to both the superior inherent principles of PET imaging and the effort and investments made in PET by the imaging industry. These advantages of cardiac PET imaging warrant that our field finds a solution to the issue of limited availability. Disclosure The authors have indicated that they have no financial conflict of interest.
References 1. Cherry SR, Sorenson JA, Phelps ME. Physics in nuclear medicine. 3rd ed. Philadelphia: Saunders; 2003. p. 511. 2. Garcia EV, Galt R, Faber TL, Chen J. Atlas of nuclear cardiology (3rd edn), In: Vasken D, Jagat N, editors. Current medicine LLC, Chapter 1. 2009. pp 1-34. 3. Garcia EV, Faber TL. Advances in nuclear cardiology instrumentation. Clinical potential of SPECT and PET. Curr Cardiovasc Imaging Rep 2009;2:230-7. 4. Lecomte R. Novel detector technology for clinical PET. Eur J Nucl Med Mol Imaging 2009;36:S69-85. 5. Hammer BE, Christensen NL, Heil BG. Use of magnetic field to increase spatial resolution of positron emission tomography. Med Phys 1994;21:1917-20. 6. Schaart DR, van Dam HT, Seifert S, Vinke R, Dendoover P, Lohner H, Beekman FJ. SiPM-array based PET detectors with depth-of-interaction correction. IEEE Nucl Sci Symposium Conference Record 2008; 3581-3585. 7. deKemp RA, Yoshinaga K, Beanlands RSB. Will 3-dimensional PET-CT enable the routine quantification of myocardial blood flow? J Nucl Cardiol 2007;14:380-97.
Journal of Nuclear CardiologyÒ May/June 2015
8. Votaw JR, White M. Comparison of 2-dimensional and 3-dimensional cardiac Rb-82 PET studies. J Nucl Med 2001;42:701-6. 9. Karp JS, Surti S, Daube-Witherspoon ME, Muehllehner G. Benefit of Time-of-Flight in PET: Experimental and clinical results. J Nucl Med 2008;49:462-70. 10. Livieratos L, Rajappan K, Stegger L, Schafers K, Bailey DL, Camici PG. Respiratory gating of cardiac PET data in list-mode acquisition. Eur J Nucl Med Mol Imaging 2006;33:584-8. 11. Meunier LL, Slomka PJ, Dey D, Ramesh A, Thomson LEJ, Hayes SW, et al. Motion frozen F-18-FDG cardiac PET. J Nucl Cardiol 2011;18:259-66. 12. Koepfli P, Hany TF, Wyss CA, Namdar M, Burger C, Konstantinidis AV, et al. CT attenuation correction for myocardial perfusion quantification using a PET/CT hybrid scanner. J Nucl Med 2004;45:537-42. 13. Nensa F, Poeppel TD, Beiderwellen K, Schelhorn J, Mahabadi AA, Erbel R, et al. Hybrid PET/MR imaging of the heart: Feasibility and initial results. Radiology. 2013;268:366-73. 14. http://www3.gehealthcare.com/en/products/categories/pet-ct/petct_scanners/. Accessed Aug 11, 2014. 15. http://usa.healthcare.siemens.com/molecular-imaging/pet-ct. Accessed Aug 11, 2014. 16. http://www.healthcare.philips.com/main/products/nuclearmedicine/ products/pet/. Accessed Aug 11, 2014. 17. http://www.positron.com/attrius. Accessed Aug 11, 2014. 18. How much does a PET/CT scanner cost? 2012/2013 update. http://info.blockimaging.com/bid/90851/How-Much-Does-a-PETCT-Scanner-Cost-2012-2013-Update. Accessed Aug 11, 2014. 19. Anger HO. Scintillation camera. Rev Sci Instrum 1958;29:27-33. 20. Garcia EV, Faber TL. New trends in camera and software technology in nuclear cardiology. Cardiol Clin 2009;27:227-36. 21. Garcia EV. Are absolute myocardial blood flow PET measurements ready for clinical use? J Nucl Cardiol 2014;21:857-8. 22. Garcia EV. Are SPECT measurements of myocardial blood flow and flow reserve ready for clinical use? Eur J Nucl Med Mol Imaging 2014;41:2291-3.
