Image Optimization in Single Photon Emission Computed Tomography by Hardware Modifications with Monte Carlo Simulation

Document Type: Original Paper

Authors

1 Professor, Medical Physics Research Center, Faculty of Medicine, Mashhad University of Medical Sciences, Mashhad, Iran

2 Assistant Professor, Medical Physics Dept., Faculty of Medicine, Tabriz University of Medical Sciences, Tabriz, Iran

3 Assistant Professor, Nuclear Medicine Dept., Imam Reza Hospital, Mashhad University of Medical Sciences, Mashhad, Iran

4 Associate Professor, Nuclear Medicine Dept., Imam Reza Hospital, Mashhad University of Medical Sciences, Mashhad, Iran

5 Professor, Medical Radiation Physics Dept., Clinical Sciences-Lund, Lund University, Lund, Sweden

Abstract

Introduction: In Single Photon Emission Computed Tomography (SPECT), the projection data used for image reconstruction are distorted by several factors, including attenuation and scattering of gamma rays, collimator structure, data acquisition method, organ motion, and washout of radiopharmaceuticals. All these make reconstruction of a quantitative SPECT image very difficult. Simulation of a SPECT system is a convenient method to assess the impact of these factors on the image quality.
Materials and Methods: The SIMIND Monte Carlo program was employed to simulate a Siemens E.CAM SPECT system. Verification of the simulation was performed by comparing the performance parameters of the system. The verified system was used for SPECT simulations of homogenous and inhomogeneous voxelized phantoms in conjugation with hardware modifications. The resulting data were compared with those obtained from the simulated system without any modifications. Image quality was assessed by comparing the Structural SIMularity index (SSIM), contrast, and resolution of images.
Results: The energy spectra acquired from both simulated and real SPECT systems demonstrated similar energy peak regions. The resulting full-widths-at-half-maximums were 13.92 keV for the simulation and 13.58 keV for experimental data, corresponding to energy resolutions of 9.95% and 9.61%, and with calculated sensitivities of 85.39 and 85.11 cps/MBq, respectively. Better performance parameters were obtained with a hardware-modified system constructed using a 0.944 cm thickness NaI(Tl) crystal covered by a layer of 0.24 cm aluminum, a  slat of 4.5 cm Pyrex as a backscattering medium, and a parallel hole collimator of Pb-Sb alloy with 2.405 cm thickness.
Conclusion: The modeling of a Siemens E.CAM SPECT system was performed with the SIMIND Monte Carlo code. Results obtained with the code are in good agreement with experimental results. The findings demonstrate that the proposed hardware modifications in the system appear to be suitable for further improvement of the performance parameters of the system, indicating that future investigations can be conducted on using the system for supplementary studies on image improvement in the field of nuclear medicine.

Keywords

Main Subjects