Back to search

University of Patras

Simulation of image formation in nuclear medicine imaging systems using Monte Carlo methods

Abstract

dc:description

The overall purpose of this study was to propose a novel method for the complete image quality characterization and optimization of Positron Emission Tomography (PET) scanners with Monte-Carlo (MC) methods. Α model was developed using the Monte Carlo package of Geant4 Application for Tomographic Emission (GATE) and the software for tomographic image reconstruction (STIR) with cluster computing to obtain reconstructed images. The PET scanner used in this study was the General Electric Discovery-ST (US). The GATE model was validated by comparing results obtained in accordance with the National Electrical Manufacturers Association NEMA-NU-2-2001 protocol (Bettinardi et al 2004, Mawlawi et al 2004). Validation images were reconstructed with the commonly used 2D Filtered Back Projection (FBP2D) and the Kinahan and Rogers FPB3DRP Reprojection Algorithms. Image quality, in terms of the Modulation Transfer Function (MTF), was initially assessed with a novel plane source consisting of a Thin Layer Chromatography (TLC) plate, simulated by a layer of Silica gel on Aluminium foil substrates immersed in Fluorodeoxyglucose (18F-FDG) bath solution (44.4MBq). MTF was assessed from the evaluated STIR digital reconstructed images of the plane source. MTF was also assessed a) in three dimensions, in lines passing through the central axis of the PET scanner, by placing the plane source only horizontally and vertically (5x10cm), b) by scanning 18cm of the transaxial Field Of View (FOV) through the simulation of a horizontal large plane source (18x10cm) and c) by evaluating reconstructed point source images. Furthermore, the study aimed to compare the proposed method with the more traditional technique based on a line source. The complete image quality characterization was assessed in terms of the Detective Quantum Efficiency (DQE) by estimating the MTF and the Normalized Noise Power Spectrum (NNPS) of the 18F-FDG TLC plane source (1 MBq). MTFs curves were estimated from transverse reconstructed images of the plane source, whereas the NNPS data were estimated from the corresponding coronal images. Images were reconstructed by the Maximum Likelihood Estimation (MLE)-OSMAPOSL reprojection algorithm by using various subsets (3 to 21) and iterations (1 to 20). Additionally, the influence of different scintillating crystals on PET scanner’s image quality, in terms of the MTF, the NNPS and the DQE, was also investigated. In this case, OSMAPOSL imageAbstract .29reconstruction was assessed by using 15 subsets and 3 iterations. The simulated spatial resolution in terms of Full Width at Half Maximum (FWHM) agreed with published data of Mawlawi et al (2004), within less than 3.29% in 2D and less than 2.51% in 3D with published data of others, respectively. The 2D values for the sensitivity, the scatter fraction and the count-rate were found to agree within less than 0.46%, 4.59% and 7.86%, respectively with corresponding published values. Accordingly, the corresponding 3D values were found to agree to less than 1.62%, 2.85% and 9.13%, respectively with Mawlawi et al (2004) published values. Sensitivity, which was also estimated in the absence of attenuation material by simulating an ideal source, showed differences between the extrapolated and the ideal source values (with and without attenuation) ranging in 2D from 0.04% to 0.82% (radial location R=0cm) in 2D and from 0.52% to 0.67% in 3D mode (radial locations R=10cm). By using this model, sensitivity can be obtained in a more simplified procedure. The simulated noise equivalent count rate was found to be 94.31kcps in 2D and 66.9kcps in 3D at 70 and 15kBq/mL respectively, compared to 94.08kcps in 2D and 70.88kcps in 3D at 54.6kBq/mL and 14kBq/mL respectively, from the published by others values. The simulated image quality was found in excellent agreement with these published values. The MTFs obtained using the FBP2D were in close agreement to those obtained by the FBP3DRP, whereas the MTFs of the OSMAPOSL show, in all cases, that higher frequencies are preserved than in the case of the FBP. FBP reconstructed images obtained from the large horizontal plane source showed that the MTF was found to degrade gradually as we move towards the edge of the FOV. The MTFs of the FBP images along the vertical direction were slightly lower than the corresponding horizontal ones. In addition, the plane source method is less prone to noise than the conventional line source method (sd=0.0031 and 0.0203, respectively). In the case of DQE investigation, MTF values assessed from the evaluated STIR digital reconstructed images, of the TLC based plane source, were found to improve as the number of iterations increased up to 12 and remain almost constant thereafter. Furthermore, variation in the number of subsets didn’t show any effect on the MTF, for equal number of iterations. The noise levels, in terms of the NNPS, in the reconstructed PET image, were found to decrease with the corresponding increase of both the number of iterations and subsets. DQE values were influenced by both MTF and NNPS and were found to increase with the corresponding increase in both number of iterations and subsets.Georgios Karpetas PhD Thesis30Finally, the PET scanner configuration, incorporating LuAP crystals, provided the optimum MTF values in both 2D and 3DFBP whereas the corresponding configuration with BGO crystals was found with the higher MTF values after OSMAPOSL. The scanner incorporating BGO crystals were also found with the lowest noise levels and the highest DQE values after all image reconstruction algorithms. In conclusion, our study showed that the imaging performance of PET scanners can be fully characterized, further improved and simplified by investigation of the imaging chain components through MC methods. The simulated PET evaluation method, based on a TLC plane source, requires materials commonly found in a clinical environment and can be experimentally implemented and used in clinical practice. In this study it was used for the image quality assessment and optimization, but it can be also useful in research for the further development of PET and SPECT (Single Photon Emission Tomography) scanners though GATE simulations. Reconstructed images by STIR can be also used to obtain radiopharmaceutical distribution of images and direct dose maps, quite useful to nuclear medicine practitioners.

Degree

thesis:*
Grantor dc:publisher
University of Patras
Year dc:date
2013

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Karpetas, George
  • Καρπέτας, Γεώργιος

Subjects

dc:subject × 13

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/29322

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Karpetas, George; Καρπέτας, Γεώργιος. Simulation of image formation in nuclear medicine imaging systems using Monte Carlo methods. University of Patras, 2013. http://hdl.handle.net/10442/hedi/29322