The Effect of Parallel-hole Collimator Material on Image and Functional Parameters in SPECT Imaging: A SIMIND Monte Carlo Study.

The collimator in single-photon emission computed tomography (SPECT) is a critical component of the imaging system and plays an impressive role in the imaging quality. In this study, the effect of the collimator material on the radioisotopic image and its functional parameters was studied. The simulating medical imaging nuclear detectors (SIMIND) Monte Carlo program was used to simulate a Siemens E.CAM SPECT (Siemens Medical Solutions, Erlangen, Germany) system equipped with a low-energy high-resolution (LEHR) collimator. The simulation and experimental data from the SPECT imaging modality using (99m)Tc were obtained on a point source and Jaszczak phantom. Seventeen high atomic number materials were considered as LEHR collimator materials. In order to determine the effect of the collimator material on the image and functional parameters, the energy resolution, spatial resolution, contrast, and collimator characteristics parameters such as septal penetration and scatter-to-primary ratio were investigated. Energy spectra profiles, full width at half maximums (FWHMs) (mm) of the point spread function (PSF) curves, system sensitivity, and contrast of cold spheres of the Jaszczak phantom for the simulated and experiment systems have acceptability superimposed. The results of FWHM and energy resolution for the 17 collimators showed that the collimator made of 98% lead and 2% antimony could provide the best FWHM and energy resolution, 7.68 mm and 9.87%, respectively. The LEHR collimator with 98% lead and 2% antimony offers the best resolution and contrast when compared to other high atomic number metals and alloys.

Introduction

Single-photon emission computed tomography (SPECT) is a nuclear medicine tomographic imaging modality. SPECT is a powerful tool for quantifying the three-dimensional distribution of a radiotracer in humans, especially in oncology. SPECT is considered to be the preferred modality in studying functional abnormalities between normal and cancer cells.[1234] In this process, the gamma rays from the administered radioisotope are encountered with photon attenuation and scattering. The gamma rays, before collision with the detector, encountered the collimator. The quality of the images in SPECT is seriously affected by the collimator's structure, properties of the detector, reconstruction algorithms, and also attenuation/scattering of gamma rays in the patient's body and the collimator.[5678] The collimation system is a critical component of a SPECT imaging system. Performance of a collimator was improved by increasing the absorption efficiency of the collimator material.[91011] Since nuclear medicine imaging deals with random phenomena such as radioactive decay and detection of photons, the Monte Carlo simulation is a suitable tool in nuclear medicine imaging. The purpose of the current study is to evaluate the effect of the parallel-hole collimator material for a low-energy high-resolution (LEHR) collimator on the image and system parameters.

Materials and Methods

In the present study, 17 materials were considered as LEHR collimator materials [Table 1]. These components and alloys have a high atomic number and linear attenuation coefficient, so they provide high-absorption efficiencies for gamma rays.[12] The Siemens E.CAM gamma camera equipped with two removable SE-LEHR (Siemens Medical Solutions, Erlangen, Germany) collimators was simulated by the simulating medical imaging nuclear detectors (SIMIND, Developed by Professor Michael Ljungberg, sweden) Monte Carlo simulation program version 5.0, which is well established for SPECT imaging.[5131415] A 1.5 mm 99mTc point source (3.7 MBq) and a Jaszczak Deluxe ECT phantom (model ECT/DLX/P) filled with water and 370 MBq activity of 99mTc were used for the experiment and simulated systems.[16] The energy resolution, spatial resolution, image contrast, and system sensitivity as a function of the collimator materials were determined according to the established and published quality control tests.[1718] Spatial and energy resolutions were determined by placing a 99mTc point source at the center of the field of view (FOV), at 10 cm from the collimator surface for both the systems and the planar images were acquired in matrices of 128 × 128 pixels, 0.27 mm pixel size, and 10 million counts for the systems. The sensitivities were determined by the ratio of the detected counts per second per unit activity (cps/MBq) for a point source placed at 25 cm from the detector surface and 10 million photons were detected. The Jaszczak Deluxe ECT phantom was also employed to calculate the image quality, and measurements were acquired experimentally and also by simulation. The phantom was positioned at 15 cm from the collimator surface and the projections along the axis of rotation were acquired in matrices of 128 × 128 pixels, 128 views, 360° clockwise rotation, 1.23 zoom factor, 0.39 mm pixel size, and 1 million counts/projection. The system setup for Jaszczak scanning is shown in Figure 1. The projections were obtained within the 15% energy window width centered on the 99mTc photopeak. The images were reconstructed by the filtered back projection (FBP) using a ramp combined with a Butterworth filter of order 5 and cutoff frequency of 0.25 cycles/cm. Spatial resolution (in mm) was determined by the smallest visible rods in the reconstructed images.

Table 1

The percentage of metals and alloys that were used as the collimator material for a SPECT imaging system

Figure 1

The system setup for Jaszczak scanning by the Siemens E.CAM gamma camera equipped with two removable SE-LEHR collimators

Results and Discussion

Figure 2 shows the normalized energy spectra from the experiment and simulated gamma cameras for a 99mTc point source scanning. Energy spectra profiles of the systems have acceptability superimposed.

Figure 2

Energy spectra for a 99mTc point source scanning at 10 cm from the collimator surface with a SPECT system. Related SIMIND simulated spectra (dashed) and experimental energy spectra (solid) are presented

The calculated energy resolutions for the 140 keV photopeak within 15% windows obtained for simulation and experiment were 10.14% and 10.13%, respectively. The full width at half maximums (FWHMs) of point spread function (PSF) curves for the 10 cm source to collimator distance were determined to be 7.74 mm for both the systems; also, the sensitivities were calculated to be 74.76 cps/MBq and 74.90 cps/MBq, respectively. The qualities of the images were compared in terms of the contrast and spatial resolution obtained by images of the cold spheres and cold rods in the Jaszczak phantom.[19] The results of image contrast for the phantom obtained from simulation and also by experiment were calculated by Equation 1 and the data are shown in Table 2.

Table 2

Results for calculated contrast of cold spheres of Jaszczak phantom from reconstructed SPECT acquisitions obtained experimentally and by SIMIND simulation

Spatial resolution was obtained by viewing the smallest visible rods for the reconstructed SPECT images of the Jaszczak phantom. The smallest cold rods observed were 9.5 mm in diameter for both the systems. Verification of the simulated system was previously described in detail by Bahreyni Toossi et al.[15]

where, Msp and Mcy correspond to the minimum pixel value in cold spheres and the maximum pixel value in the phantom cylinder, respectively.

Afterward, following a change in the collimator materials, the functional parameters such as sensitivity, energy resolution, spatial resolution or rather the FWHMs of PSF curves were measured for the 17 collimators [Table 3]. The collimator characteristics parameters such as septal penetration and the ratio of the scattered-to-primary photons in the photopeak energy windows were investigated and also the imaging parameters were evaluated by comparing the contrast of Jaszczak phantom cold sphere with 19.1 mm in diameter [Table 4]. We have chosen the 19.1 mm sphere because the contrast of the sphere has shown most contrast similarity between the simulation and the experiment systems.

Table 3

The results of functional parameters for a 99mTc point source scanning by simulated SPECT with different collimator materials

Table 4

The results of collimator characteristics parameters for a 99mTc point source and contrast of 19.1 mm cold sphere of Jaszczak phantom, which were scanned by simulated SPECT with different collimator materials

Effect of the collimator materials on the image contrast of the phantom is shown in Figure 3. According to Table 4, platinum has, on the one hand, the least penetration but on the other hand, the highest scatter-to-primary ratio in the energy windows. There was also a reduction in the sensitivity for the collimator material according to Table 3. According to Tables 3 and 4, the best spatial resolution, energy resolution, and contrast were obtained by the collimator made of 98% lead and 2% antimony [lead antimony (1)] with the best FWHM and energy resolution of 7.68 mm and 9.87%, respectively. It appears that the alloy, as the collimator material, could provide the best image quality in SPECT imaging. It should also be noted that there is always a trade-off between the image resolution and sensitivity in nuclear medicine imaging modality, and the alloy showed a proper spatial resolution from the point of view of the image quality and not sensitivity.

Figure 3

The reconstructed images of cold spheres of simulated Jaszczak phantom filled with 370 MBq 99mTc that were sorted according to the decreased contrast. The phantom consisting of six cold spheres (the cold rod diameters were 4.8 mm, 6.4 mm, 7.9 mm, 9.5 mm, 11.1 mm, and 12.7 mm). The acquisition parameters were 128 × 128 matrix, 128 projections, 360° clockwise rotation, 1.23 zoom factor, 3.9 mm pixel size, and 1 million counts per projection using a dual-head camera (the Siemens E.CAM gamma camera equipped with two removable SE-LEHR collimators). Images were reconstructed by the FBP with a Butterworth filter of order 5 and cutoff frequency of 0.25 cycles/cm. (a) The best contrast that is related to lead antimony (1) collimator (b) The good contrast achieved by Wood's metal collimator (c) The moderate contrast achieved by lead antimony (2), and finally (d) The poor contrast achieved by gold/copper collimator

Conclusion

We have used the SIMIND Monte Carlo simulation program to investigate the effect of collimator material on the image and functional parameters in SPECT imaging. The results of our study showed that the LEHR collimator made of 98% lead and 2% antimony offers the best resolution and contrast compared to other high atomic number metals and alloys.