|
 |
|
ORIGINAL ARTICLE |
|
|
|
| Year : 2022 | Volume
: 47
| Issue : 4 | Page : 374-380 |
| |
Investigation of response of the pixelated CZT SPECT imaging system and comparison with the conventional SPECT system
Mahsa Noori-Asl, Pooneh Sayyah-Koohi
Department of Physics, Faculty of Science, University of Mohaghegh Ardabili, Ardabil, Iran
| Date of Submission | 16-May-2022 |
| Date of Decision | 18-Sep-2022 |
| Date of Acceptance | 20-Sep-2022 |
| Date of Web Publication | 10-Jan-2023 |
Correspondence Address:
Dr. Mahsa Noori-Asl
Department of Physics, Faculty of Science, University of Mohaghegh Ardabili, Ardabil
Iran
 Source of Support: None, Conflict of Interest: None  | Check |
DOI: 10.4103/jmp.jmp_41_22
 Abstract | | |
Purpose: The aim of this study is to investigate the factors affecting the response of the pixelated CZT SPECT imaging systems and to compare the performance of these systems with the conventional SPECT imaging systems. Materials and Methods: By using the simulation technique, the effect of applied voltage, gap size between the anode pixels, and electron cloud mobility on the response of a pixelated CZT SPECT system are investigated. Then, the response of this system is compared with the conventional SPECT system in both single- and dual-radioisotope imaging. Results: The results of this study show that increasing the applied voltage, electron cloud mobility or decreasing the gap size, in the optimal range of these parameters obtained in this study, leads to reducing the lateral charge diffusion and consequently improving the probability of the complete charge collection by the target anode pixel. In dual-radioisotope imaging by the pixelated CZT SPECT system, although higher energy resolution results in better separation of photopeaks, the presence of a low-energy tail leads to an overestimation of counts in the low-energy photopeak. Conclusion: The use of the optimal values for the applied voltage, gap size, and electron cloud mobility strongly affect the performance of the pixelated CZT SPECT systems. In addition, the presence of a tail restricts the use of these systems for dual-radioisotope imaging and also, the use of the conventional methods for scatter correction.
Keywords: Energy spectrum, pixelated CZT SPECT imaging system, point source
How to cite this article:
Noori-Asl M, Sayyah-Koohi P. Investigation of response of the pixelated CZT SPECT imaging system and comparison with the conventional SPECT system. J Med Phys 2022;47:374-80 |
How to cite this URL:
Noori-Asl M, Sayyah-Koohi P. Investigation of response of the pixelated CZT SPECT imaging system and comparison with the conventional SPECT system. J Med Phys [serial online] 2022 [cited 2023 Mar 30];47:374-80. Available from: https://www.jmp.org.in/text.asp?2022/47/4/374/367422 |
| Introduction | |  |
Single-photon emission computed tomography (SPECT)[1],[2] is a nuclear medicine imaging technique that produces tomographic images from a three-dimensional (3D) distribution of the radiopharmaceutical injected into the patient's body.[3] The images obtained from this imaging system are strongly influenced by different factors such as photon attenuation and scattering,[1] patient motion,[4] detector response,[5] and collimator characteristics.[6] In order to achieve high-quality images for a more accurate diagnosis, the detectors used in nuclear medicine imaging systems must have some special properties such as high energy resolution, good photon conversion efficiency, high density, short decay time, short irradiation length, and good physical strength with a minimum afterglow.[7]
In a conventional SPECT imaging system, referred to as the “Anger Camera,” the sodium iodide (NaI) scintillation crystal is used as the detector material.[8] The indirect conversion process of the incident photon energy into an electrical signal through the emission of visible light in a scintillator leads to a statistical error in the measurement of the deposited energy.[9] The poor energy resolution of the detector causes the photons that have been scattered into the patient's body to be detected within the energy window used for imaging.[10] Due to changing the photon direction in the scattering event, these photons carry misleading information about their emission location. Therefore, detection of them leads to degrading the image quality.[11],[12] The use of a detector with a good energy resolution can minimize the unwanted contribution of these scattered photons in the final reconstructed image.[10]
The SPECT imaging systems based on cadmium–zinc–telluride (CZT) semiconductor detectors have been introduced in the first decade of the 21st century.[13] In a CZT detector, the interaction of the incident photon with the detector material leads to producing electron-hole pairs. The number of these pairs is initially proportional to the deposited energy of the incident photon. By applying an external voltage to the electrodes, the electrons and holes move in opposite directions toward the collection electrodes, inducing a charge that is amplified and processed to form an electrical signal.[14] Since, in the CZT semiconductor detector, the velocity of electrons is greater than the velocity of holes; therefore, the signal is extracted from the anode.[15] Each anode in the CZT imaging systems shows a pixel in the image at the corresponding position. Therefore, these detector systems are sometimes called “pixelated detectors.”[10] Due to the direct conversion of the energy and location of the incident photon into the electrical signal and no need for light tubes or photomultiplier tubes, the energy and spatial resolution of the CZT semiconductor detector is better than that of the NaI scintillation detector.[16]
On the other hand, due to the charge carrier trapping and recombination in the CZT semiconductor detector, the electrons and holes produced at the location of photon interaction cannot be collected by the electrodes completely. Therefore, it is important to consider the incomplete charge collection because of its direct effect on measuring the incident gamma-photon energy. In addition, another factor that can affect incomplete charge collection is the diffusion of the electron charge cloud [Figure 1]. Due to the diffusion phenomenon, the radius of the electron cloud groves when moving toward the pixelated anode. Near the surface of the pixelated anode, the radius of this cloud may be greater than the size of the target anode pixel, causing a part of the charge to enter the gap region (without considerable electric field) or even the neighboring pixels (”charge sharing”). The charge sharing in the pixelated CZT detector leads to recording the primary photons outside of the photopeak area and creating a tail in the low-energy range of the spectrum of the radioisotope used in imaging.[14] A simplified model for the performance characterization of a pixellated CZT imaging system is presented in reference.[17] | Figure 1: Illustration of an array of anode pixels together with the charge diffusion and charge sharing for four different typical gamma-ray interaction locations (a-d). Circles with a larger radius indicate a larger charge diffusion
Click here to view |
In the present study, by using the simulation technique, in the first stage, we evaluate the effect of three factors, including the applied voltage, gap size, and electron cloud mobility (CM) in imaging with a pixelated CZT SPECT system. This evaluation is performed by investigating the corresponding changes in the energy spectra obtained from the simulation of a technetium-99m (99mTc) point source in the air. In the next stage, to compare the response of the pixelated CZT SPECT system with the conventional SPECT system based on the NaI scintillation detector, the spectral characteristics of these two systems in single-radioisotope imaging with 99mTc and dual-radioisotope imaging with 99mTc and iodine-123 (123I) are investigated.
| Materials and Methods | | |
In this study, the SIMIND (SIMulation of Imaging Nuclear Detectors) Monte Carlo program[18],[19] is used to simulate the pixelated CZT SPECT imaging system and to produce the energy spectra. The simulation is performed by considering the contribution of both charge carriers (electron and hole), and by using the combined energy resolution model.[14] The default values of parameters used in this study are: applied voltage 600 V, electron mobility 5 × 103 V/cm2, hole mobility 4 × 102 V/cm2, gap size between anodic pixels 0.11, detector size 0.246 cm, exponential decay constant 0.12. The imaging is performed by setting a 20% energy window centered at the photopeak energies of the used radioisotopes (140 keV for 99mTc and 159 keV for 123I).
This study is performed in two stages. In the first stage, we use the energy spectrum obtained from the simulation of a 99mTc point source in the air for evaluation of the system response. In this stage, by changing the applied voltage, gap size between the anode pixels, and electron cloud mobility in three separate steps, the corresponding changes in the response of the pixelated CZT SPECT imaging system are investigated. These changes include changes in full-width at half-maximum (FWHM), peak counts (maximum counts in the spectrum), photopeak area (The integral of counts between the lower and upper energy window channels), and Compton area (the integral of counts from the first channel to the lower energy window channel).
In the second stage, to compare the response of the pixelated CZT SPECT system with the conventional NaI SPECT system, the simulation of the point source is performed in two following situations:
- The point source in the air (without attenuating medium),
- The point source inside an attenuating medium.
The attenuating medium is a horizontal cylindrical water-filled phantom (diameter 22 cm and height 20 cm) without activity. The point source is located in the center and on the axis of this cylindrical phantom. Simulation is performed by acquiring 64 projections in a 360° rotation of SPECT camera. This comparison is performed using the energy spectra obtained from two systems in single-radioisotope (99mTc) imaging and dual-radioisotope imaging (99mTc/123I simultaneous imaging).
| Results | | |
Investigation of the response of the pixelated CZT SPECT system to the change of parameters
Change in the applied voltage
In the first step, to investigate the effect of the applied voltage in the response of the pixelated CZT SPECT system, the voltage was changed in the range of 100 V to 1200 V with 100 V steps. [Figure 2] shows the changes in the energy spectrum obtained from the simulation of a 99mTc point source with increasing the applied voltage in this range. Furthermore, the diagrams in [Figure 3] show the corresponding changes in the FWHM and peak counts of the energy spectra. | Figure 2: (a) Illustration of the changes in the energy spectra obtained from the simulation of a 99mTc point source with increasing the applied voltage. (b) For better comparison, the spectra are shown in the energy range of 120–160 keV.
Click here to view |
 | Figure 3: Diagrams of (a) FWHM and (b) peak counts related to the energy spectra shown in Figure 2, with respect to the applied voltage. FWHW: Full-width at half-maximum
Click here to view |
From these figures, by increasing the value of the applied voltage, the widths of the peaks decrease, and the peaks become narrower. In addition, the maximum counts in the energy spectra (peak counts) and the peak heights increase. This is because by increasing the applied voltage, the electrons are more directed toward the target anode pixel and their lateral diffusion to the adjacent pixels (charge sharing) reduces. In the higher voltage range (600–1200 V), due to the almost complete collection of the charge by the target pixel, the energy spectra get closer to each other.
Change in the gap size
In the second step, the changes in the 99mTc energy spectrum by increasing the distance between the anode pixels at a constant voltage = 600 V were investigated. For this purpose, the gap size between the anode pixels was changed in the range of 0 to 0.5 with steps of 0.05. [Figure 4] shows the energy spectra obtained from the simulation for the different values of the gap size in the range of interest. From this figure, when there is no gap between the anode pixels (the anode is almost a continuous plate), the obtained energy spectrum has the narrowest width and the highest peak at 140 keV, with the lowest tail in the low-energy range of the spectrum. By increasing the gap size, the peak heights in the energy spectra decrease, and in contrast, the peak widths and also, low-energy tails increase. As the gap size increases further, the peak heights tend to zero, and a continuum appears in the low-energy range of 0 to 60 keV. For a better comparison, the bar graphs of the changes in the Compton and photopeak areas are shown in [Figure 5]. | Figure 4: Illustration of the changes in 99mTc energy spectrum by increasing the gap size between the anode pixels
Click here to view |
 | Figure 5: The bar graphs of (a) photopeak area and (b) compton area related to the energy spectra shown in Figure 4, with respect to the gap size
Click here to view |
The gradual reduction of the peak heights can be explained by the decrease in the size of the anode pixels as a result of the increase in the size of gaps. In fact, by decreasing the pixel size, the electron charge collected from a single gamma event by the target anode pixel decreases and a part of the charge enters to the space between the pixels and is lost. The incomplete charge accumulation by the target anode pixel leads to increasing the area under the continuum in the low-energy region of the spectrum (Compton area) and decreasing the area under the photopeak area.
Change in the electron cloud mobility
In the third step, the effect of the electron cloud mobility on the energy spectra obtained from the simulation of the 99mTc point source in a constant voltage = 600 V with a gap size = 0.246 cm was investigated. For this purpose, the electron cloud mobility in the range of 0.05 to 10 was changed. The results of this simulation are shown in [Figure 6] and [Figure 7]. From [Figure 6], although in the low range of changes (0 to 1), by increasing the electron cloud mobility, obvious differences are seen in the obtained energy spectra; in the higher range (1 to 10), these differences are small. This is due to increasing the electron cloud mobility, the charge diffusion decreases, and therefore, the possibility of the electron cloud collection by the target anode pixel increases. This is also clear from the diagrams of changes in the FWHM and peak counts shown in [Figure 7]. From this figure, in the high range of the electron cloud mobility (CM > 1), the values obtained for both diagrams are almost constant. | Figure 6: Illustration of the changes in the energy spectra obtained from the 99mTc point source simulation by increasing the electron CM. (For better comparison, the spectra are shown in the energy range of 120–160 keV in right). CM: Cloud mobility
Click here to view |
 | Figure 7: Diagrams of (a) FWHM and (b) peak counts related to the energy spectra shown in Figure 6, with respect to the electron cloud mobility. FWHW: Full-width at half-maximum
Click here to view |
Comparison of the response of the pixelated CZT SPECT imaging systems
Single-radioisotope (99mTc) imaging
Results for a point source in the air
The energy spectra obtained from the simulation of a 99mTc point source in the air for both the conventional SPECT system based on the NaI scintillation detector and the pixelated CZT SPECT system are shown in [Figure 8]a. For better comparison, the two spectra were normalized to the peak heights. Since the simulation of the point source was performed in the air without considering the attenuating medium, it was expected that there would be no counts in the low-energy region of the energy spectra, as is clearly the case with the spectrum obtained from the conventional NaI SPECT system. However, in the case of the spectrum obtained from the pixelated CZT SPECT system, due to the two phenomena of the electron cloud diffusion (charge sharing) and the charge carrier recombination inside the detector, a tail in the low-energy region of the spectrum is present. | Figure 8: The energy spectra obtained from the 99mTc point-source simulation (a) in the air and (b) in the attenuating medium for both the conventional NaI SPECT system and the pixelated CZT SPECT system. CZT: Cadmium–zinc–telluride, SPECT: Single-photon emission computed tomography
Click here to view |
Results for a point source inside an attenuating medium
In the presence of an attenuating medium, due to the occurrence of Compton scattering for a number of photons, as they pass through the medium, a continuum was expected to appear in the low-energy region of the spectra obtained from both systems. From [Figure 8]b, in the case of the pixelated CZT SPECT system, due to the two-mentioned phenomena, this continuum covers a significant part of the energy spectrum compared to the spectrum obtained from the conventional NaI SPECT system.
Dual-radioisotope (99mTc/123I) imaging
[Figure 9] shows the energy spectra obtained from the simulation of the 99mTc and 123I point sources together with the energy spectra of simultaneous imaging of these two radioisotopes for both the pixelated CZT SPECT system and the conventional NaI SPECT system. From this figure, due to the lower widths of the photopeaks in the energy spectra obtained from the pixelated CZT SPECT system, the overlap of the two photopeaks is less than that for the conventional NaI SPECT system, resulting in better distinction and separation of the two photopeaks. On the other hand, in the case of the pixelated CZT SPECT system, the spectrum obtained from 99mTc/123I simultaneous imaging [dashed line diagram in [Figure 9]] shows an overestimation at 140 keV photopeak of the 99mTc radioisotope. The reason for this is the addition of the low-energy tail of the 123I spectrum to the 99mTc peak counts. | Figure 9: The 99mTc and 123I energy spectra together with the spectrum obtained from the simultaneous imaging for (a) the conventional NaI SPECT system and (b) the pixelated CZT SPECT system. CZT: Cadmium–zinc–telluride, SPECT: Single-photon emission computed tomography, NaI: Sodium iodide
Click here to view |
This simulation was performed for a point source in the air. Certainly, the problem would be further complicated by considering the attenuating medium that leads to the addition of a Compton continuum in the spectrum of both radioisotopes.
| Discussion | | |
In this study, in the first stage, the response of the pixelated CZT SPECT imaging system to the changes in the applied voltage, gap size, and electron cloud mobility was investigated in three separate steps. In the second stage, the differences in the response of the pixelated CZT SPECT system and the conventional NaI SPECT system in both single-and dual-radioisotope imaging were compared.
The results obtained from the first stage showed that by increasing the applied voltage, the number of counts recorded in the photopeak energy window and, consequently, the photopeak sensitivity and efficiency increase, which is a result of the reduced charge sharing between the adjacent anode pixels. Another result is that, at the higher voltages, from 600 V to 1200 V, the differences in the photopeak widths and heights of the resulting energy spectra decrease, which indicates approaching a state in which the electron cloud is almost completely collected by the target pixel. Hence, the voltage of 600 V at the beginning of this range was chosen as the desired operating voltage in the next steps.
On the other hand, a gradual increase in the gap size, which means a gradual decrease in the size of the anode pixels, leads to a decrease in the photopeak height and an increase in the low-energy tail of the spectrum. This is due to reducing the possibility of the complete collection of the electron cloud generated in a single event by the target anode pixel and also, the entrance of more electron cloud into the gaps. Finally, in the gap = 0.5, the probability of the complete charge collection is reduced to zero, and the photopeak in the spectrum at 140 keV disappears.
In the third step, it was shown that by keeping constant the other parameters affecting the charge collection, the increase of the electron cloud mobility, due to reducing the charge sharing, leads to an increase in the height of the photopeak in the spectrum. This increase continues until the value of the charge cloud mobility equals 1. After that, the observed changes become less and less, which indicates reaching to a constant final state.
The results obtained from the second stage showed that, due to the presence of the charge diffusion and trapping of the charge carriers in the pixelated CZT SPECT system, the spectrum obtained from this system has a low-energy tail, even in the absence of an attenuating medium. This means spreading the primary (unscattered) events throughout the energy spectrum, while, in the case of the conventional NaI SPECT system, these events are limited to the photopeak area. This makes it difficult to use the usual scatter correction methods based on energy windows for the pixelated CZT SPECT system because the contributions of the charge diffusion and trapping are always added to the contribution of the scattered counts in the low-energy region of the spectrum. Therefore, the spatial distribution and the number of counts in this region can no longer be used to estimate the scattered counts included in the photopeak window.
Finally, the results obtained from the simultaneous 99mTc/123I dual-radioisotope imaging showed that although the better energy resolution of the pixelated CZT SPECT system leads to less overlap of the two photopeaks, the presence of a low-energy tail in the energy spectra obtained from this system leads to an overestimation of counts in the 140 keV peak.
| Conclusion | | |
The results obtained from this study showed that the use of the optimal values for the applied voltage, gap size between the anode pixels, and electron cloud mobility strongly affect the performance of the pixelated CZT SPECT imaging systems. On the other hand, the presence of the low-energy tail from the charge trapping, recombination of the charge carriers, and charge diffusion (charge sharing), in the energy spectra obtained from the pixelated CZT SPECT imaging systems, restricts the use of these systems for dual-radioisotope imaging and also, the use of the conventional approximations for scatter correction of images.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
| References | | |
| 1. | Groch MW, Erwin WD. SPECT in the year 2000: Basic principles. J Nucl Med Technol 2000;28:233-44.  |
| 2. | Groch MW, Erwin WD. Single-photon emission computed tomography in the year 2001: Instrumentation and quality control. J Nucl Med Technol 2001;29:12-8. |
| 3. | Kapucu OL, Nobili F, Varrone A, Booij J, Vander Borght T, Någren K, et al. EANM procedure guideline for brain perfusion SPECT using 99mTc-labelled radiopharmaceuticals, version 2. Eur J Nucl Med Mol Imaging 2009;36:2093-102. |
| 4. | Chen QS, Franken PR, Defrise M, Jonckheer MH, Deconinc F. Detection and correction of patient motion in SPECT imaging. J Nucl Med Technol 1993;21:198-205. |
| 5. | Khoshakhlagh M, Islamian JP, Abedi M, Mahmoudian B, Mardanshahi AR. A Study on determination of an optimized detector for single photon emission computed tomography. World J Nucl Med 2016;15:12-7. [ PUBMED] [Full text] |
| 6. | Razavi SH, Kalantari F, Bagheri M, Namiranian N, Nafisi-Moghadam R, Mardanshahi A, Emami-Ardekani A, Ardekani MS, Razavi-Ratki SK, Characterization of low, medium and high energy collimators for common isotopes in nuclear medicine: A Monte Carlo study. Iran J Nucl Med 2017;25:100-4. |
| 7. | Liu B, Shi C. Development of medical scintillator. Chin Sci Bull 2002;47:1057-63. |
| 8. | Peterson TE, Furenlid LR. SPECT detectors: The anger camera and beyond. Phys Med Biol 2011;56:R145-82. |
| 9. | Moszynski M, Nassalski A, Syntfeld-Kazuch A, Swiderski L, Szczesniak T. Energy resolution of scintillation detectors – New observations. IEEE Trans Nucl Sci 2008;55:10628. |
| 10. | Mattsson E. The Alcyone CZT SPECT Camera. Evaluation of Performance Using Phantom Measurement and Monte Carlo simulation. Master of Science Thesis, Lund University; 2013. |
| 11. | Asl MN, Sadremomtaz A, Bitarafan-Rajabi A. Evaluation of six scatter correction methods based on spectral analysis in (99m) Tc SPECT imaging using SIMIND Monte Carlo simulation. J Med Phys 2013;38:189-97. [ PUBMED] [Full text] |
| 12. | Noori-Asl M, Sadremomtaz A, Bitarafan-Rajabi A. Evaluation of three scatter correction methods based on estimation of photopeak scatter spectrum in SPECT imaging: A simulation study. Phys Med 2014;30:947-53. |
| 13. | Kincl V, Drozdová A, Vašina J, Panovský R, Kamínek M. Cadmium-zinc-telluride SPECT scanners – New perspectives in nuclear cardiology. Cor Vasa 2015;57:214-8. |
| 14. | Pretorius H, Liu C, Fan P, Peterson M, Ljungberg M. Monte Carlo simulations of the GE discovery Alcyone CZT SPECT systems. IEEE Trans Nucl Sci 2015;62:832-9. |
| 15. | Knoll GF. Radiation Detection and Measurement. New York: John Wiley & Sons; 2010. |
| 16. | Iniewski K. CZT detector technology for medical imaging. J Instrum 2014;9:C11001. |
| 17. | Guerra P, Santos A, Darambara DG. Development of a simplified simulation model for performance characterization of a pixellated CdZnTe multimodality imaging system. Phys Med Biol 2008;53:1099-113. |
| 18. | Ljungberg M, Strand SE. A Monte Carlo program simulating scintillation camera imaging. Comp Meth Prog Biomed 1989;29:257-72. |
| 19. | Ljungberg M, Strand SE, King MA. Monte Carlo Calculations in Nuclear Medicine Applications in Diagnostic Imaging. London: IOP Publishing; 1998. |
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8], [Figure 9]
|
