|Year : 2017 | Volume
| Issue : 6 | Page : 936-942
Assessment of the scatter correction procedures in single photon emission computed tomography imaging using simulation and clinical study
Mehravar Rafati1, Hemmatollah Rouhani1, Ahmad Bitarafan-Rajabi2, Mahsa Noori-Asl3, Bagher Farhood4, Hadi Taleshi Ahangari5
1 Department of Medical Physics and Radiology, Faculty of Paramedicine, Kashan University of Medical Sciences, Kashan, Iran
2 Department of Nuclear Medicine, Cardiovascular Interventional Research Center, Rajaei Cardiovascular, Medical and Research Center, Iran University of Medical Sciences, Tehran, Iran
3 Department of Physics, Faculty of Sciences, University of Mohaghegh Ardabili, Ardabil, Iran
4 Department of Medical Physics and Biomedical Engineering, School of Medicine, Tehran University of Medical Sciences, Tehran, Iran
5 Department of Medical Physics, Faculty of Medicine, Semnan University of Medical Sciences, Semnan, Iran
|Date of Web Publication||13-Dec-2017|
Dr. Ahmad Bitarafan-Rajabi
Department of Nuclear Medicine, Cardiovascular Interventional Research Center, Rajaei Cardiovascular, Medical and Research Center, Iran University of Medical Sciences, Tehran
Source of Support: None, Conflict of Interest: None
Background: Compton-scattered photons transfer incorrect spatial information. These photons are detected in used photo-peak energy window. In this study, three scatter correction procedures including dual-energy window (DEW), three energy window (TEW), and new approach were evaluated, and then the best procedure based on simulation and clinical conditions introduced.
Materials and Methods: In this study, simulation projections and three-dimensional nonuniform rational B-spline–based Cardiac-Torso phantoms were produced by GEANT4 application for emission tomography simulation code. For clinical study, 2-day stress/rest myocardial perfusion imaging (MPI) protocol was performed with 99m Tc-sestamibi for 46 patients. Image quality parameters including contrast, signal-to-noise ratio (SNR), and relative noise of the background (RNB) were evaluated.
Results: The simulation results showed that contrast values for DEW, TEW, and new approach were (0.45 ± 0.07, 0.5 ± 0.08, and 0.63 ± 0.09), SNR values (4.74 ± 0.94, 5.58 ± 1.08, and 6.56 ± 1.24), and RNB values (0.33 ± 0.06, 0.33 ± 0.07, and 0.33 ± 0.05), respectively. In clinical study, the contrast values for DEW, TEW, and new approach were 0.53 ± 0.03, 0.57 ± 0.07, and 0.62 ± 0.04 in rest MPI and were 0.52 ± 0.04, 0.57 ± 0.06, and 0.6 ± 0.05 in stress MPI, respectively. Moreover, for the rest images, the SNR values were 7.65 ± 1.9, 9.08 ± 2.2, and 10.2 ± 1.75 and for stress images were 7.76 ± 1.99, 9.12 ± 2.25, and 10.17 ± 2.04, respectively. Finally, RNB values for rest and stress images were 0.12 ± 0.03, 0.13 ± 0.03, and 0.13 ± 0.03, respectively.
Conclusion: The simulation and the clinical studies showed that the new approach could be better performance than DEW, TEW methods, according to values of the contrast, and the SNR for scatter correction.
Keywords: Clinical study, GEANT4 application for emission tomography, scatter correction, single photon emission computed tomography
|How to cite this article:|
Rafati M, Rouhani H, Bitarafan-Rajabi A, Noori-Asl M, Farhood B, Ahangari HT. Assessment of the scatter correction procedures in single photon emission computed tomography imaging using simulation and clinical study. J Can Res Ther 2017;13:936-42
|How to cite this URL:|
Rafati M, Rouhani H, Bitarafan-Rajabi A, Noori-Asl M, Farhood B, Ahangari HT. Assessment of the scatter correction procedures in single photon emission computed tomography imaging using simulation and clinical study. J Can Res Ther [serial online] 2017 [cited 2019 May 19];13:936-42. Available from: http://www.cancerjournal.net/text.asp?2017/13/6/936/211659
| > Introduction|| |
Myocardial perfusion imaging by single photon emission computed tomography (SPECT) is a noninvasive method to diagnose coronary artery disease.,, In myocardial perfusion SPECT imaging, accuracy of quantification is affected by physical process including attenuation, scattering, partial volume effect, septal penetration, and the imaging characteristics including spatial and energy resolution, sensitivity, uniformity, as well as reconstruction methods.
Hence, one factor affecting the image quality is the detection of the scattered photons. Scattering can occur in two ways: Compton scattering and coherent scattering. Compton scattering plays a main role in the nuclear medicine imaging. In addition, the probability of its occurrence in the energy range used in nuclear medicine is more than coherent scattering., Due to change of the direction, Compton-scattered photons transfer incorrect spatial information of their emitted location. Therefore, detection of them can lead to degrade the image contrast and overestimate the activity in each pixel of the image matrix. Decreasing the image contrast may cause uncertainty in important clinical details, particularly cold areas in the images.,,, As a result, the scattered photons have a significant effect on the interpretation and reporting the images as well as reducing the accuracy of diagnosis.
In SPECT imaging, a 20% window centered symmetrically on the photo-peak of 99mTc with energy of 140 keV is usually employed. Due to the low energy resolution of the scintillation, crystal of NaI (Tl) used in SPECT imaging systems. The scattered photons are detected in the used photo-peak energy window with the primary photons. Therefore, they contribute to the final reconstructed image. However, complete removal of the scattered photons with a limited energy resolution is not very practical.,,, Therefore, it is necessary to use a correction procedure for improving the final image quality and diagnostic accuracy. Finally, scatter correction procedures can increase the contrast-to-noise ratio, detection of the lesions, and quantification. Moreover, it is notable that use of the Monte Carlo (MC) simulation is a useful and essential way for defining the characteristics of the system that cannot be determined by examining. The simulations used in nuclear medicine are increasingly applied to develop new imaging tools and scatter correction procedures.,,
There are several studies that evaluated the scatter correction methods in SPECT imaging. Noori-Asl et al. investigated three scatter correction procedures (three energy window [TEW] method with trapezoidal and triangular approximations and a new approach) based on the estimation of the photo-peak scatter spectrum in SPECT imaging. Their findings showed that the new approach is a better approximation for estimation of the photo-peak scatter spectrum than the other two approximations. Furthermore, they concluded that the new approach improves the image contrast as compared to the two other methods. In another study, Asl et al. assessed six scatter correction procedures based on spectral analysis in SPECT imaging using the SIMIND simulation program. Their results demonstrated that the TEW procedure using triangular approximation due to ease of implementation, improvement in signal-to-noise ratio (SNR) and contrast, and the low noise level is proposed as most proper correction procedure.
In the present study, we evaluate the scattered correction procedures that have better performance than other procedures according to previous studies. Moreover, there is a possibility of their implementation on SPECT imaging systems in the real conditions (clinical trials). The evaluations are carried out in both simulation and clinical conditions. For this purpose, changes in the image quality parameters such as the contrast, SNR, and relative noise of the background (RNB) are investigated before and after applying the correction procedures.
| > Materials and Methods|| |
Scatter correction procedures
Two spectral procedures (Dual-Energy Window [DEW] and TEW using triangular approximation), which had the best performance in previous studies, along with our alternative procedure were evaluated in the simulation and clinical conditions. In each of the three scattered correction procedures, photo-peak and energy window were 140 and 126–154 keV, respectively.
Dual-energy window procedure
In DEW procedure, in addition to photo-peak range, a lower range was defined in the Compton spectrum of 99mTc from 92 to 125 keV.,, A number of photo-peak scattered photons (Spk) in specific pixel position of the image matrix (i, j) was obtained by multiplying k parameter (equal to 0.5) in a number of total photons of Compton range (Tc):
Three energy window procedure
In TEW procedure,, TEWs were used: (1) photo-peak range, (2) Wnw1 (narrow window at first photo-peak range from 123 to 129 keV, and (3) Wnw2 (narrow window at the latest of photo-peak range from 150 to 158 keV). With acceptance of two assumptions in higher ranges, there are only nonscattered photons; as numbers of theses photons are equal at both narrow windows:
As a result, triangle surface that its height is equal to scattered photons in the lower window (E1 = 126 keV) was used to estimate the scattered photons (Spk), according to the following equation:
In alternative procedure, in addition to accepting the above both assumptions, there is third assumption so that third narrow window (Wnw3 = 133–136 keV) in photo-peak range exists that scattered photons are same with Wnw1 range. Therefore:
In this procedure, approximation of the scattered photons in photo-peak window (Spk) obtains to calculate a trapezoidal surface, according to following formula:
Where Wd obtains by subtracting lower photo-peak energy range from center of third range (133 keV).
To have a realistic simulation of the patient's anatomy with clinical information, a 3D nonuniform rational B-spline – based Cardiac-Torso (NCAT) phantom was used.,, In this study, activity distribution in various parts of the body was similar to these alternative by Segars and Tsui. GEANT4 application for emission tomography (GATE) simulation code is adequately flexible in modeling of advanced tools in nuclear medicine, particularly SPECT. It is ready to use in voxelized phantoms such as NCAT. For simulation of the SPECT scanner in GATE, existence of a scripting through macro language is essential. All projections (un-scatter and scatter) were obtained from GATE version 7.0 (international Open GATE collaboration (version 7.0)) program on a relatively fast computer. The scanner characteristics include dual head SPECT with size of 5.905 cm × 61.1 cm × 46.5 cm. Collimator dimensions of low energy–high resolution were 2.405 cm × 59.1 cm × 44.5 cm with hexagon holes that their radiuses are 0.111 cm. The dimensions of NaI (Tl) crystal used were 0.59 cm × 59.1 cm × 44.5 cm. Energy resolution of 20% was set for 99mTC at energy of 140 keV.
This study was performed on 46 patients (22 males and 24 females) with cardiac scan by SPECT scanner. For all patients who their cardiac SPECT images were evaluated clinically, imaging protocol of 2-day 99mTc-sestamibi stress/rest was used. In stress phase, 20 mci of technetium-labeled was injected through intravenous (IV) into patients, and imaging was performed 20 min late. Rest phase that was performed 1 day after stress phase scan. In this phase, 20 mci of technetium-labeled was also injected through IV into patients, and imaging was performed about 45–60 min after injection. Imaging was started from right anterior oblique angle and continued to left posterior oblique angle so that 180° was covered completely. Image acquisition was performed by dual head SPECT/CT Symbia T2 Siemens with low energy–high resolution collimator that its photo-peak was 140 keV. Furthermore, its energy resolution was 20% and its matrix size was 64 × 64 in 32 views.
For evaluating the performance of scatter correction procedures in reconstructed projections of simulation and clinical conditions by filtered back projection method in MATLAB program with version 7.4 and using Hann filter, three evaluation criteria were used which are defined separately the following:
In cardiac image, contrast of reconstructed image is defined as the ratio of subtraction of the mean counts in cardiac cavity between two peaks (N̄c) from the mean counts of two peaks (N̄p), divided by the mean of two peaks:
Relative noise of the background
To calculate the RNB, standard deviation (SD) in the cavity of cardiac is divided into its mean counts:
SNR is defined as the numerator of contrast equation divide by the numerator of RNB equation:
In this study, to analyze simulated and clinical data, SPSS version 11.5 (SPSS Inc., Chicago, IL, USA) was used. Clinical samples were collected with IR.KAUMS.REC.1394.177 ethical code and patients consciously participated in this study.
| > Results|| |
By analyzing the simulation data related to 64 slices, the following results were obtained for each of three parameters of contrast, RNB and SNR using before and after scatter correction procedures [Table 1] and [Figure 1].
|Table 1: Mean, standard deviation, and coefficient of variation related to the simulated image quality parameters (contrast, relative noise of the background, and signal-to-noise ratio) of 64 slices for before and after scatter correction procedures|
Click here to view
|Figure 1: Plots related to quality parameters of the simulated images for before and after scatter correction procedures. Parts of a, b, and c are related to contrast, relative noise of the background, and signal-to-noise ratio, respectively|
Click here to view
ANOVA test was used to assess the results of quality parameters between four groups (before and after scatter correction procedures). The results showed that there was a significant difference between contrast results of four groups (P < 0.05). The findings showed that there was no significant difference between RNB of four groups (P > 0.05). Furthermore, the results demonstrated that there was a significant difference between SNRs of four groups (P < 0.05). According to the listed values in [Table 1], alternative, TEW, and DEW scatter correction procedures had the highest values for contrast and SNR, respectively. For RNB levels, there was not much difference between the three scatter correction procedures [Table 2] and [Figure 2].
|Table 2: Demonstrates mean, standard deviation, and coefficient of variation related to the rest and stress image quality parameters (contrast, relative noise of the background, and signal-to-noise ratio) for before and after scatter correction procedures|
Click here to view
|Figure 2: Plots related to quality parameters of rest and stress images for before and after scatter correction procedures. Parts of a, b, and c are related to contrast, relative noise of the background, and signal-to-noise ratio, respectively|
Click here to view
Paired t-test was used to assess the results of image quality parameters in the two groups of rest and stress for each of the scatter correction procedures. The results demonstrated that there is no significant difference between contrast values of all cases in the two groups of rest and stress for each of the scatter correction procedures (P > 0.05). Then, ANOVA test showed that there is a significant difference between different scatter correction procedures in each of the two groups for all cases (P < 0.05). The finding showed that there is no significant difference between RNB values of all cases in the two groups of rest and stress for each of the scatter correction procedures (P > 0.05). Then, ANOVA test showed that there is no significant difference between different scatter correction procedures in each of two groups for all cases (P > 0.05). Furthermore, the results demonstrated that there was no significant difference between SNR values of all cases in the two groups of rest and stress for each of the scatter correction procedures (P > 0.05). Then, ANOVA statistic test shows that there is a significant difference for different correction procedures in each of two groups for all cases (P < 0.05) [Figure 3] and [Figure 4].
|Figure 3: A slice from reconstructed image of the nonuniform rational B-spline – based Cardiac-Torso phantom without correction (a), with scatter corrections procedures of dual-energy window (b), three energy window (c), and alternative (d)|
Click here to view
|Figure 4: A slice from reconstructed clinical image without correction (a), with scatter corrections procedures of dual-energy window (b), three energy window (c), and alternative (d)|
Click here to view
| > Discussion|| |
For evaluating and improving of scatter correction procedures, MC simulations are very valuable because only the simulations are able to distinguish nonscatter and scatter photons. Among MC simulation codes, GATE code is very flexible and accurate in simulation. This code considers all conditions of SPECT imaging such as transmission of photons in the collimator, crystal, shielding of the head, table of scan, and phantom. In addition, it has predictability of tracking scattered photons.,
To the best of our knowledge, the scatter correction procedures of DEW, TEW, and alternative were not evaluated in clinical conditions, so in this study, mentioned scatter correction procedures were evaluated in simulation and clinical conditions. The aim of this study was the introduction of the best scatter correction procedure in SPECT imaging technique using GATE simulation and clinical conditions. Image quality parameters evaluated for comparison of different scatter correction procedures were contrast, RNB, and SNR levels. Contrast in SPECT imaging is a factor that contributes to the diagnosis, and it should increase to improve image quality; as increasing the contrast image can happen to use scatter correction procedures. Furthermore, because of removing a few number of primary photons when using the scatter correction procedures, increasing the relative noise in SPECT images can be expected. On the other hand, SNR includes both parameters of contrast and relative noise, and it can represent values related to remove the primary and scattered photons.
In this study, contrast range for cardiac SPECT images in the clinical situation without using scatter correction methods, for rest and stress phases were 0.413–0.5685 and 0.3988–0.5773, respectively. After applying the scatter correction procedures, contrast range in rest phase for the procedures of DEW, TEW, and alternative was 0.4467–0.6003, 0.5036–0.6526, and 0.5328–0.679, respectively, as well as for stress phase was 0.4298–0.5953, 0.4851–0.6426, and 0.5164–0.678, respectively. Jaszczak et al. performed a study on phantoms and patients to evaluate DEW correction procedure. In their study, the factor K was considered equal to 0.5, and evaluation parameters of image were contrast, SNR, and SD. By comparing the results, contrast and SNR values in our study were more than Jaszczak study. Asl et al. evaluated six scatter correction procedures in SPECT imaging using SIMIND MC simulation. Their results demonstrated that the TEW procedure using triangular approximation is alternative as most proper scatter correction procedure, due to ease of implementation, uniform performance, and good improvement of the image contrast. To compare the contrast values in DEW and TEW with triangular approximation procedures related to this study and results of our study, the findings showed that difference between values of contrast in these two correction procedures for our study was more than their study so had lower SD. In other study by Noori-Asl et al., three scatter correction procedures (trapezoidal approximation, triangular approximation, and their alternative procedure) in SPECT imaging were investigated. Their finding showed that the alternative procedure has better scatter fraction in contrast of cold background and hot-sphere image rather than two other procedures and but with a slightly more noise.
RNB ranges for cardiac SPECT images in the clinical situation without using scatter correction procedures, for rest and stress phases were 0.0482–0.2529 and 0.0634–0.1866, respectively. After applying the scatter correction procedures, RNB ranges in rest phase for the procedures of DEW, TEW, and alternative were 0.0553–0.2546, 0.0717–0.2457, and 0.0779–0.2719, respectively, as well as for stress phase were 0.0908–0.1824, 0.1018–0.2017, and 0.0977–0.228, respectively. The findings showed that RBN levels increase after using the scatter correction procedures to improve the quality image, but increasing the RNB values were no statistically significant; as the RNB values on images after applying the various correction procedures were approximately same to images without scatter correction. The results of our study related to RNB were consistent with other studies. The results of Jaszczak et al. related to noise level showed that pure noise levels increase for corrected images and our results were inconsistent with their results. Furthermore, results of Noori-Asl et al. showed that using alternative scatter correction procedure, noise level in corrected images increases slightly, whereas the results of our study indicated that values of noise for three scatter correction procedures are nearly same and there was no significant difference between them.
SNR ranges for cardiac SPECT images in the clinical situation without using scatter correction procedures, for rest and stress phases were 4.0439–13.6153 and 4.2722–13.8805, respectively. After applying the scatter correction procedures, SNR ranges in rest phase for the procedures of DEW, TEW, and alternative were 4.9978–15.5144, 6.4564–18.5802, and 9.0284–18.8563, respectively, as well as for stress phase were 5.3149–15.9765, 6.0391–17.2349, and 8.542–19.8261, respectively. The results showed that after applying the DEW scatter correction procedure, SNR values in our study were more than Jaszczak study of Jaszczak et al. Furthermore, Noori-Asl et al. showed that after applying the alternative scatter correction procedure, image contrast slightly improves relative to the uncorrected image. In other study by Asl et al., their results demonstrated that the TEW scatter correction procedure using triangular approximation increases image contrast than the uncorrected image. The results of our study were inconsistent with the findings of Noori-Asl et al.
Our findings showed that there is good consistent between clinical and simulation results. Staelens et al. validated the results of GATE simulation based on a comparison between simulation and experimental data from a standard gamma camera. Overall, their results showed very good agreement between the simulation and clinical data, and GATE simulation was able to simulate a gamma camera. Lazaro et al. performed validation of the GATE simulation for modeling a CsI (Tl) scintillation camera. Their results showed a good agreement between experimental and simulated data, as experimental spatial resolutions were anticipated with an error <100 μm. Furthermore, they showed that difference between the simulated and experimental system sensitivities for various source-to-collimator distances was within 2% as well as for sources located in water, experimental, and simulated scatter fractions in a 98–182 keV energy window differed by <2%. Experimental and simulated energy spectra had good agreement between 40 and 180 keV. Finally, their results demonstrated the flexibility and ability of GATE code to simulate original detector designs. In another study, Assié et al. carried out validation of the GATE code for indium-111 imaging. Their results showed that there is a good agreement between simulated and experimental energy spectra. Clinical and simulation spatial resolution had difference <2% in the air, whereas the values of simulated sensitivity were within 4% of the experimental values. Finally, they suggested that GATE code is capable for accurate simulation of SPECT acquisitions. In 2015, Lee et al. simulated Siemens Symbia SPECT/CT imaging device with 131I radioisotope by GATE code. Their simulation and experimental results demonstrated a same pattern with a difference of <15% in spatial resolution values with increasing distance. For the future studies, we will evaluate the effect of scatter, attenuation, and motion correction individually and simultaneously on simulation and clinical images.
| > Conclusion|| |
In this study, the highest value of contrast for simulation and clinical images was related to the alternative scatter correction procedure, as there is a significant difference between this procedure with other correction procedures and before correction images. RNB values for three scatter correction procedures were approximately same, but the SNR values related to the alternative procedure was more than other procedures and before correction images.
The results of the GATE simulation and patients showed that the alternative scatter correction procedure can introduce as the best procedure to improve of the scattering in SPECT images. It is reasonable, due to less removal of primary photons by the alternative scatter correction procedure compared to other scatter correction procedures and its more similarity to photo-peak true curve. However, the alternative scatter correction procedure is somewhat more difficult in practice because it needs to four energy ranges.
It is necessary to thanks and gratitude from members of the Nuclear Medicine Department of Rajaei Cardiovascular, Medical and Research Center, for their cooperation in this study.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| > References|| |
Anagnostopoulos C, Harbinson M, Kelion A, Kundley K, Loong CY, Notghi A, et al.
Procedure guidelines for radionuclide myocardial perfusion imaging. Heart 2004;90 Suppl 1:i1-10.
Niu X, Yang Y, Jin M, Wernick MN, King MA. Effects of motion, attenuation, and scatter corrections on gated cardiac SPECT reconstruction. Med Phys 2011;38:6571-84.
Yokota S, Mouden M, Ottervanger JP, Engbers E, Knollema S, Timmer JR, et al.
Prognostic value of normal stress-only myocardial perfusion imaging: A comparison between conventional and CZT-based SPECT. Eur J Nucl Med Mol Imaging 2016;43:296-301.
Buvat I, Rodriguez-Villafuerte M, Todd-Pokropek A, Benali H, Di Paola R. Comparative assessment of nine scatter correction methods based on spectral analysis using Monte Carlo simulations. J Nucl Med 1995;36:1476-88.
Staelens S, Strul D, Santin G, Vandenberghe S, Koole M, D'Asseler Y, et al.
Monte Carlo simulations of a scintillation camera using GATE: Validation and application modelling. Phys Med Biol 2003;48:3021-42.
Zaidi H. Quantitative Analysis in Nuclear Medicine Imaging. New York: Springer Science Business Media; 2006.
Cherry SR, Sorenson JA, Phelps ME. Physics in Nuclear Medicine. 4th
ed. Philadelphila: Elsevier Health Sciences; 2012.
Saha GB. Physics and Radiobiology of Nuclear Medicine. 4th
ed. New York, NY, USA: Springer; 2012.
Ljungberg M, King MA, Hademenos GJ, Strand SE. Comparison of four scatter correction methods using Monte Carlo simulated source distributions. Energy 1994;150:200.
Ghaly M, Links JM, Frey E. Optimization of energy window and evaluation of scatter compensation methods in myocardial perfusion SPECT using the ideal observer with and without model mismatch and an anthropomorphic model observer. J Med Imaging (Bellingham) 2015;2. pii: 015502.
Beekman FJ, Kamphuis C, Frey EC. Scatter compensation methods in 3D iterative SPECT reconstruction: A simulation study. Phys Med Biol 1997;42:1619-32.
Kadrmas DJ, Frey EC, Tsui BM. Application of reconstruction-based scatter compensation to thallium-201 SPECT: Implementations for reduced reconstructed image noise. IEEE Trans Med Imaging 1998;17:325-33.
Zaidi H, Sgouros G, editors. Therapeutic Applications of Monte Carlo Calculations in Nuclear Medicine. Taylor & Francis: CRC Press; 2002.
Imbert L, Galbrun E, Odille F, Poussier S, Noel A, Wolf D, et al.
Assessment of a Monte-Carlo simulation of SPECT recordings from a new-generation heart-centric semiconductor camera: From point sources to human images. Phys Med Biol 2015;60:1007-18.
Lazaro D, Buvat I, Loudos G, Strul D, Santin G, Giokaris N, et al.
Validation of the GATE Monte Carlo simulation platform for modelling a CsI(Tl) scintillation camera dedicated to small-animal imaging. Phys Med Biol 2004;49:271-85.
Lee YS, Kim JS, Kim KM, Lim SM, Kim HJ. Determination of energy windows for the triple energy window scatter correction method in I-131 on a Siemens SYMBIA gamma camera: A GATE simulation study. J Instrum 2015;10:P01004.
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.
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.
] [Full text]
Jaszczak RJ, Floyd CE Jr., Coleman RE. Scatter compensation techniques for SPECT. IEEE Trans Nucl Sci 1985;32:786-93.
Jaszczak RJ, Greer KL, Floyd CE Jr., Harris CC, Coleman RE. Improved SPECT quantification using compensation for scattered photons. J Nucl Med 1984;25:893-900.
Segars WP, Lalush DS, Tsui BM. A realistic spline-based dynamic heart phantom. IEEE Trans Nucl Sci 1999;46:503-6.
Segars WP, Tsui BM. Study of the efficacy of respiratory gating in myocardial SPECT using the new 4-D NCAT phantom. IEEE Trans Nucl Sci 2002;49:675-9.
Xu XG, Eckerman KF. Handbook of Anatomical Models for Radiation Dosimetry. Taylor & Francis: CRC Press; 2009.
Assié K, Gardin I, Véra P, Buvat I. Validation of the Monte Carlo simulator GATE for indium-111 imaging. Phys Med Biol 2005;50:3113-25.
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2]