Gate Simulation Of A Clinical Pet Using The Computing Grid

Abstract

Nowadays, Nuclear Medicine becomes a potential research field of a growing importance at the universities. This fact can be explained, on one hand, by the increasing number of purchases in medical imaging devices by the hospitals, and, in other hand, by the number of PhD students and researchers becoming interested to medical studies. A Positron Emission Tomography (PET) system is a functional medical imaging technique which provides 3D images of the living processes inside the body relying on radioisotopes usage. The physics of PET systems is based on the detection in coincidence of the two 511 keV ?-rays, produced by an electron-positron annihilation, and emitted in opposite directions, as dictated by the conservation of energy and momentum physics laws. The radioactive nuclei used as sources of emission of positrons for PET systems are mainly 11C, 13N, 15O, and 18F, which are produced in cyclotrons, and decay with half-lives of 20.3 min, 9.97 min, 124 sec, and 110 min, respectively. These radioisotopes can be incorporated in a wide variety of radiopharmaceuticals that are inhaled or injected, leading to a medical diagnosis based on images obtained from a PET system. The PET scanners consists mainly of a large number of detector crystals arranged in a ring which surround the patient organ (or phantom in simulations) where the radioisotope tracer (e.g.: 18F-FDG) is inoculated. The final 3D image, representing the distribution of the radiotracer in the organ (or the phantom), is obtained by processing the signals delivered by the detectors of the scanner (when the ?-rays emitted from the source interact with the crystals) and using image reconstruction algorithms. This allows measuring important body functions, such as blood flow, oxygen use, and glucose metabolism, to help doctors evaluate how well organs and tissues are functioning and to diagnose and determine the severity of or treat a variety of diseases. The simulation of a real experiment using a GATE-modeled clinical Positron Emission Tomography (PET) scanner, namely PHILIPS Allegro, has been carried out using a computing Grid infrastructure. In order to reduce the computing time, the PET simulation tasks are split into several jobs submitted to the Grid to run simultaneously. The splitting technique and merging the outputs are discussed. Results of the simulation are presented and good agreements are observed with experimental data. Keywords—Grid Computing; Monte Carlo simulation; GATE; Positron Emission Tomography; splitting