Comparative evaluation of two commercial PET scanners, ECAT EXACT HR+ and Biograph 2, using GATE
Introduction
Monte Carlo simulations are used in nuclear medicine to model positron emission tomography (PET) or single photon emission tomography (SPECT) imaging systems in order to develop and assess tomographic reconstruction algorithms and correction methods for improved image quantification [1]. For these purposes several simulation codes have been used in the past. Geant4 application for tomographic emission (GATE) is a generic Monte Carlo simulation platform based on a general-purpose code GEANT4 [2] and designed to answer the specific needs of PET/SPECT applications [3], [4]. GATE includes specific modules necessary to perform realistic simulations, including modules managing time and time-dependent processes (detector and source movements, radioactive decay, and dynamic acquisitions), complex source distributions and easy description of scanner geometry. The ability to synchronize all time-dependent components allows a coherent description of the acquisition process and is one of the most innovative features of this package. GATE is an open source software and its development and validation is carried out by members of the OpenGATE collaboration.
The purpose of this study is to validate a Monte Carlo model for the simulation of the commercial available SIEMENS PET scanners of ECAT EXACT HR+ and the PET/CT Biograph 2 using GATE simulation package (version 2.2.0). Both scanners were chosen because of their wide area of applications throughout the last years. Their comparative evaluation on a simulation level is going to contribute to our better understanding of the influence of each one of their basic operational parameters on the efficiency and sensitivity of these systems. Furthermore, the comparative presentation of both the simulation and experimental data can trigger a thorough analysis in order to optimize the simulation parameter values at GATE and achieve tolerable and satisfying agreement between simulation and experimental output. This simulation code will be used in a second phase in order to study scatter phenomena and motion artifacts. The simulation results will be used to optimize image reconstruction algorithms, with emphasis on dynamic PET studies.
Section snippets
Description of geometry
The scanner's physical and technical specifications were obtained from Siemens medical and CTI innovations data sheets. Direct measurements at the place of the installation were performed as well.
The ECAT EXACT HR+ PET scanner (Fig. 1) is located at Hammersmith Imanet Hospital at London. It consists of 4 block rings of 72 detector blocks each. Each block is constructed of an 8×8 BGO crystal array. The dimensions of each crystal element are 4.05 mm×4.39 mm×30 mm. The detector ring diameter is 82.4
Results
The National Electrical Manufacturers Association (NEMA) performance protocols describe a series of explicitly defined experiments that have a significant recognition at the field of performance measurements for both PET and SPECT systems [8], [9]. Therefore we had designed our simulation experiments according to these protocols. We have used ROOT data analysis framework [10] as an output data format. The simulated values produced for each scanner about the performance parameters of spatial
Discussion and conclusion
The spatial resolution values, obtained using the simulated GATE model, are within 9% of the experimental values. Furthermore we observe that the results of the simulation are consistently indicating improved spatial resolution in comparison to the respective experimental measurements. These discrepancies are observed due to the absence of modeling within GATE the light shielding both within and between the detector blocks, the inherent limitations of the resolution of the photomultiplier
Acknowledgments
The authors would like to thank all the GATE users and developer community for their valuable suggestions and insight they have provided us during this validation study.
References (12)
Geant-4 a simulation toolkit
Nucl. Instr. and Meth. A
(2003)- et al.
ROOT-An object oriented data analysis framework
Nucl. Instr. and Meth. A
(1997) - et al.
Monte Carlo Simulations in PET and SPECT
Q. J. Nucl. Med.
(2002) GATE: a simulation toolkit for PET and SPECT
Phys. Med. Biol.
(2004)- I. Buvat, et al., Unified description and validation of Monte Carlo simulators in PET, Nuclear Science Symposium...
- S. Jan, et al., Monte Carlo Simulation for the ECAT EXACT HR+System Using GATE, IEEE Trans. Nucl. Sci. NS-52(3)...
Cited by (43)
Development and validation of an accurate GATE model for NeuroPET scanner
2017, Physica MedicaCitation Excerpt :However, changing this value does not significantly affect the results, as the NeuroPET/CT and small-bore PET systems cannot use time-of-flight information with the current hardware technologies. In addition, we found that a quantum efficiency (QE) of 80% as a free and varied scaling factor [25,26] matches the results of the experimental values. This parameter is applied to individual events within the blocks in the digitizer chain [26].
Detective quantum efficiency (DQE) in PET scanners: A simulation study
2017, Applied Radiation and IsotopesCitation Excerpt :Within the context of the model, the principal imaging parameters i.e. MTF, NNPS and DQE were estimated by the simulation of thin layer chromatography (TLC) plane sources filled with 18F-FDG using a software model based on the GATE Monte Carlo package. The GATE package, which was developed by the Open-GATE collaboration, is an open-source extension of the Geant4 Monte Carlo toolkit (Strul et al., 2003; Jan et al., 2004) GATE was used in combination with the STIR image reconstruction software (Karakatsanis et al., 2006; Thielemans et al., 2006, 2012; Tsoumpas et al., 2011, Karakatsanis et al., 2016a) to obtain the plane source reconstructed images. To date, various GATE studies have been published on implementing STIR reconstruction for large commercial PET scanners (Karakatsanis et al., 2006; Gonias et al., 2007; Tsoumpas et al., 2011; Zagni et al., 2013).
Validation of a Monte Carlo simulation of the Inveon PET scanner using GATE
2016, Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors and Associated EquipmentCitation Excerpt :Using an easy-to-learn macro mechanism to configurate simple or highly sophisticated experimental settings, GATE now plays a key role in the design of new medical imaging devices, in the optimization of acquisition protocols and in the development and assessment of image reconstruction algorithms and correction techniques [11–13]. Several validation studies of GATE models for clinical PET and pre-clinical small animal PET have recently been completed [14–19]. For the clinical PET scanner, Gonias et al. [14] validated a GATE model for the simulation of Siemens PET Biograph™ 6 scanner, and compared the results with experimental data, obtained in accordance with the National Electrical Manufacturers Association (NEMA) NU 2-2001 performance measurement protocol (National Electrical Manufacturers Association 2001).
Monte-Carlo simulations of clinically realistic respiratory gated <sup>18</sup>F-FDG PET: Application to lesion detectability and volume measurements
2015, Computer Methods and Programs in BiomedicineCitation Excerpt :Geant4 Application for Emission Tomography (GATE) [10] was used to perform the simulations. The SIEMENS Biograph Sensation 16 Hi-Rez PET scanner corresponding to that used in the Department of Nuclear Medicine of the Henri Becquerel Center and Rouen University Hospital was modeled [11–13]. Its physical characteristics are given in Table 1.
Modeling and simulation of 4D PET-CT and PET-MR images
2013, PET ClinicsCitation Excerpt :Among these, GATE,54 an open source software package (www.opengatecollaboration.org), is the most used. It can accommodate different scanner geometries and has been validated for numerous PET and SPECT scanners.55–57 Although GATE has the additional advantage of allowing modeling of time-dependent phenomena under realistic acquisition conditions, its main limitation is the extremely high computational demands for simulations without compromising statistical accuracy.