A Study on Determination of an Optimized Detector for Single Photon Emission Computed Tomography.

The detector is a critical component of the single photon emission computed tomography (SPECT) imaging system for giving accurate information from the exact pattern of radionuclide distribution in the target organ. The SIMIND Monte Carlo program was utilized for the simulation of a Siemen's dual head variable angle SPECT imaging system with a low energy high resolution (LEHR) collimator. The Planar and SPECT scans for a (99m)Tc point source and a Jaszczak Phantom with the both experiment and simulated systems were prepared and after verification and validation of the simulated system, the similar scans of the phantoms were compared (from the point of view of the images' quality), namely, the simulated system with the detectors including bismuth germanate (BGO), yttrium aluminum garnet (YAG:Ce), Cerium-doped yttrium aluminum garnet (YAG:Ce), yttrium aluminum perovslite (YAP:Ce), lutetium aluminum garnet (LuAG:Ce), cerium activated lanthanum bromide (LaBr3), cadmium zinc telluride (CZT), and sodium iodide activated with thallium [NaI(Tl)]. The parameters of full width at half maximum (FWHM), energy and special resolution, sensitivity, and also the comparison of images' quality by the structural similarity (SSIM) algorithm with the Zhou Wang and Rouse/Hemami methods were analyzed. FWHMs for the crystals were calculated at 13.895, 14.321, 14.310, 14.322, 14.184, and 14.312 keV and the related energy resolutions obtained 9.854, 10.229, 10.221, 10.230, 10.131, and 10.223 %, respectively. Finally, SSIM indexes for comparison of the phantom images were calculated at 0.22172, 0.16326, 0.18135, 0.17301, 0.18412, and 0.20433 as compared to NaI(Tl). The results showed that BGO and LuAG: Ce crystals have high sensitivity and resolution, and better image quality as compared to other scintillation crystals.

Introduction

Single photon emission computed tomography (SPECT) imaging studies the distribution of an administered radiotracer in vivo. The images are strongly affected by the attenuation, scattering, and response of the detector. The conventional detector is mainly made from sodium iodide activated by thallium [NaI(Tl)] in nuclear medicine imaging. In SPECT, the used radioisotopes mainly decay by gamma irradiation, with the energy ranging from 50 KeV to 511 KeV. On the other hand, the favorite detector should have some characteristics such as high energy resolution, appropriate photon conversion efficiency, high density, short decay time, short radiation length as well as appropriate physical strength, with least afterglow, in order to obtain images with the best quality for a proper diagnosis.[1] Bismuth germanate (BGO), due to its better sensitivity and higher density[2] with a low afterglow and also as a nonhygroscopic crystal was suggested as a proper candidate for radioisotopic imaging.[3] BGO crystal is commonly used in positron emission tomography (PET) scanners for its high photo fraction that is needed in PET imaging.[4] However, yttrium aluminum garnet (YAG: Ce) and Cerium-doped yttrium aluminum perovskite (YAP: Ce) have the benefit of an average density but YAP: Ce also has energy resolution better than YAG: Ce.[5] On the other hand, YAG: Ce has the ability to detect x-rays and low energy gamma ray photons. Also, cerium-doped lutetium aluminum garnet (LuAG: Ce) has higher density than YAG: Ce and faster decay time than BGO crystal.[6] Cerium activated lanthanum bromide (LaBr3) has a good energy resolution and low decay time with a high temperature stability also at room temperature.[567] A new crystal, cadmium zinc telluride (CZT) with a high density, very short decay time, and excellent energy resolution were recently suggested in the optical and nuclear medicine imaging systems. CZT crystal significantly reduces imaging time owing to its very short decay time, and it also administered the dosage of the patients in nuclear medicine imaging.[8] This study planned to compare the abovementioned crystals for an optimized SPECT imaging [Table 1].

Table 1

Certain quantities of the crystals may be used as nuclear medicine detectors

Materials and Methods

Monte Carlo simulation

The SIMIND Monte Carlo program was used for simulating Siemen's dual head variable angle SPECT imaging system with seven detectors, including BGO, YAG:Ce, YAP:Ce, LuAG:Ce, LaBr3, CZT, and NaI(Tl), a 99mTc point source, and also a Jaszczak Phantom. The program is freely downloadable from the site: www.radfys.lu.se/simind. The SIMIND program mainly consists of two programs: CHANGE that defines the system and scanning parameters and SIMIND that performs the actual simulation. The program can also simulate a nonuniform attenuation from voxel-based phantoms and includes several types of variance reduction techniques.[9]

Phantom studies

A 99mTc point source and an acrylic cylindrical Jaszczak Deluxe Phantom[10] were used for both the experiment and the simulation studies. The 99mTc point source has been considered to have an activity of 37 MBq, with an isotropic source emitting 140 keV gamma photons. Moreover, the acrylic cylindrical Jaszczak Deluxe Phantom, with the spherical inserts measuring 9.5 mm, 12.7 mm, 15.9 mm, 19.1 mm, 25.4 mm, and 31.0 mm in diameter and the inserts of rods measuring 4.8 mm, 6.4 mm, 7.9 mm, 9.5 mm, 11.1 mm, and 12.7 mm diameter were filled with a uniform solution of 370 MBq 99mTc. As the SIMIND simulated SPECT projections are noise-free, for realization the noise was added according to the administered dose.[11] The phantom is used for routine quality assurance and control, including system spatial resolution, contrast, and uniformity.[9] The experiment and also the simulated SPECT of the phantom were generated. The acquisition parameters were matrix of 128 × 128, 128 views, 1.23 zoom factor, and 3.9 mm pixel size, and the images were reconstructed with the filtered back projection reconstruction using a Butterworth filter with an order of 5 and in cutoff frequency of 0.25. The qualities of the produced images were compared in terms of image contrast and spatial resolution.[9101112]

Quantitative image evaluation

The structural similarity (SSIM) algorithm has two methods for a quantitative image evaluation, namely Zhou Wang and Rouse/Hemami. The SSIM metric calculates various windows of an image. The measure between two images x and y of common size N × N is:

Where μ is the average of x and μ is the average of y, is the variance of x and is the variance of y. σ is the covariance of x and y. C1= (k1L)2 and C2= (k2L)2 are two variables to stabilize the division with a weak denominator. L is the dynamic range of the pixel values (L = 255 for 8 bits/pixel gray scale images) and finally, k1 = 0.01 and K2 = 0.03 by default.[1314]

Results

Point source simulation

Figure 1 shows the energy spectra produced by the experiment and simulated gamma cameras from a 99mTc point source. Some minor differences may be observed between the simulated and the experimental energy spectra; the experimental spectrum presents a wider peak that may be explained by the superimposition of the energy peaks with the x-ray energy of 99mTc.[15] Also, Figure 2 shows energy spectra for the related 99mTc point source scans by the simulated system with the detectors of BGO, YAG: Ce, YAP: Ce, LuAG: Ce, LaBr3, and CZT crystals. Contributions of the Compton and photoelectric interactions in the whole spectra (simulated) and in the selected window, and also some functional parameters are given in Table 2.

Figure 1

Energy Spectra of the simulated (solid) and experiment (dashed) 99mTc point source scanning with NaI(Tl) detector

Figure 2

Energy spectra from simulation of a 99mTc point source scanning with BGO, YAG:Ce, YAP:Ce, CZT, LuAG:Ce, and LaBr3 crystals

Table 2

Results of some functional parameters for the simulation of a 99mTc point source scanning with BGO, YAG:Ce, YAP:Ce, CZT, LuAG:Ce, and LaBr3 crystals

Jaszczak Phantom simulation

In this study, we simulated SPECT scanning of the hot and cold Jaszczak Phantoms and compared the related experiment and simulated reconstruction images with the SSIM algorithm by the Zhou Wang and Rouse/Hemami methods. The methods gave suitable information on comparing the images from the point of view of image contrast, luminance, and structure as well as provided a multiscale-SSIM (MS-SSIM) similarity index [Table 3]. The same comparisons were done for the images after changing the crystals of the simulated SPECT system with CZT, LaBr3, LuAG: Ce, YAG: Ce, YAP: Ce, and BGO [Table 4]. The quality of the resultant images in Figures 3 and 4 have shown the acquisition of pie slice of the simulated hot and cold Deluxe Phantoms by SPECT with the seven detectors.

Table 3

The data on comparing cold Jaszczak Phantom pie slice images with SSIM algorithm, and the Zhou Wang and Rouse/Hemami methods for the experimented and simulated reconstructed SPECT images of the phantom

Table 4

The data on comparison of pie slice images from a hot Jaszczak Phantom scanning with BGO, YAG: Ce, YAP: Ce, CZT, LuAG: Ce, and LaBr3 crystals by SSIM algorithm and the Zhou Wang and Rouse/Hemami methods

Figure 3

Images from the simulated SPECT of cold Jaszczak Phantom acquisition, consisting of 148 rods with 10 mCi 99mTc by BGO, YAG:Ce, YAP:Ce, CZT, LuAG:Ce, and LaBr3 crystals

Figure 4

Images from the simulated SPECT of hot Jaszczak Phantoavm acquisition, consisting of 148 rods with 10 mCi 99mTc by BGO, YAG:Ce, YAP:Ce, CZT, LuAG:Ce, and LaBr3 crystals

Discussion

SIMIND as a Monte Carlo program has benefits in the simulation of the hardware of SPECT. The effects of scintillator crystal on the performance parameters, sensitivity, and efficiency as well as the results of quantitative and qualitative analysis of reconstructed SPECT images have been studied in the present research. We simulated the SPECT system with NaI (Tl) and six different crystals, including CZT, LaBr3, LuAG: Ce, YAG: Ce, YAP: Ce, and BGO. Figure 1 shows the energy spectra of the simulated and experiment 99mTc point sources scanned with NaI(Tl) detector. Angela et al. (2011) studied the LaBr3 crystal in vivo for the administration of radioiodine (I-131) in the thyroid gland to obtain acceptable results and consequently enunciated it as an appropriate detector. The properties that make the LaBr3:Ce scintillator detector attractive for different applications based on γ-ray spectrometry include very suitable energy resolution and very fast light decay, enabling high count rate applications, high temperature stability, high gamma detection efficiency, stability at room temperature, and promising technology for manufacturing the crystal at larger sizes.[7] Furthermore, Liu B et al. reviewed the BGO crystal that has occupied more than 50% of the PET market.[1] In this regard, Derenzo et al. compared the NaI (Tl) crystal with the BGO crystal and found the latter to be more sensitive mainly due to its higher density.[16] Our results [Table 2] also showed that BGO crystals have higher sensitivity, 91.8050 cps/MBq with full width at half maximum (FWHM) of 13.895 KeV, and peak/Compton 7.4274 with a lower energy resolution 9.8545 (%) than other crystals. In addition, Lo Meo Sergio et al. studied some nuclear medicine crystals and have determined that LuAG: Ce with energy resolution of 6.7 and 70 ns decay time is better than BGO.[17] Also, Chewpraditkul W et al. demonstrated that LuAG: Ce has better detection rate than YAG:Ce from the point of view of higher density (6.76 g/cm3) and higher atomic number (58.9).[6] In our study, we used SSIM algorithm for comparing the simulated with the experimental Jaszczak Phantom reconstructed images. Table 3 shows the MS-SSIM numbers of 0.98279 and 0.98793 for the hot Jaszczak Phantom images with the Zhou Wang and Rouse/Hemami methods, respectively. Meanwhile, we changed the crystal material during the simulation of Jaszczak Phantom and studied images’ quality in six scintillator crystals, including BGO, YAG: Ce, YAP: Ce, LuAG: Ce, LaBr3, and CZT. The related SSIM algorithms were obtained with MS-SSIM numbers 0.22172, 0.16326, 0.18135, 0.17301, 0.18412, and 0.20433 by the Zhou Wang method and 0.21097, 0.15060, 0.16929, 0.16219, 0.17253, and 0.19323 by the Rouse/Hemami method for a quantity study of the simulation images of the Jaszczak Phantom. However, these results show the LuAG: Ce detector with some more different MS-SSIM numbers and also with some difference in quality between the Jaszczak Phantom images and the other crystals.

Conclusion

The poor resolution of SPECT has impaired its use in clinical practice.[14] It is well known that the best detector with a high fast process can provide beneficial effect for better diagnosis and reduce the time of scans. The results of our current study showed that the BGO crystal is more sensitive than other crystals [Table 2]. Moreover, the Zhou Wang and the Rouse/Hemami methods for hot Jaszczak Phantom images showed that LuaG: Ce and CZT crystals have good images for the smallest visible rods of the hot Jaszczak Phantom [Figure 4] and also that LuAG: Ce crystals provide better images, determined with SSIM algorithms [Table 4]. These data might encourage the planning of studies in more detail and allow users to have a better investigation compared to other equipments commonly utilized in nuclear medicine.