GATE Monte Carlo simulation of GE Discovery 600 and a uniformity phantom

GATE (Geant4 Application Tomography Emission) Monte Carlo simulations have been successful in the application of emission tomography for precise modeling of various physical processes. Most previous studies on Monte Carlo simulations have only involved performance assessments using virtual phantoms. Although that allows the performance of simulated positron emission tomography (PET) to be evaluated, it does not reflect the reality of practical conditions. This restriction causes substantial drawbacks in GATE simulations of real situations. To overcome the described limitation and to provide a method to enable simulation research relevant to clinically important issues, we conducted a GATE simulation using real data from a scanner rather than a virtual phantom and evaluated the scanner is performance. For that purpose, the system and the geometry of a commercial GE PET/ CT (computed tomography) scanner, BGO-based Discovery 600 (D600), was developed for the first time. The performance of the modeled PET system was evaluated by using the National Electrical Manufacturers Association NEMA NU 2-2007 protocols and results were compared with those of the reference data. The sensitivity, scatter fraction, noise-equivalent count rate (NECR), and resolution were estimated by using the protocol of the NEMA NU2-2007. Sensitivities were 9.01 cps/kBq at 0 cm and 9.43 cps/kBq at 10 cm. Scatter fractions were 39.5%. The NECR peak was 89.7 kcps @ 14.7 kBq/cc. Resolutions were 4.8 mm in the transaxial plane and 5.9 mm in the axial plane at 1 cm, and 6.2 mm in the transaxial plane and 6.4 mm in the axial plane at 10 cm. The resolutions exceeded the limited value provided by the manufacturer. The uniformity phantom was simulated using the CT and the PET data. The output data in a ROOT format were converted and then reconstructed by using the C program and STIR (Software for Tomographic Image Reconstruction). The reconstructed images of the simulated uniformity phantom data had comparable quality even though improvement in the quality is still required. In conclusion, we have demonstrated a successful simulation of a PET system by using scanned data. In future studies, the parameters that alter the imaging conditions, such as patient movement and physiological change, need to be studied.