Fast-Track of F-18 Positron paths simulations

• DOI: 10.48550/arxiv.2403.06307
• Publication Date: 2024-03-10
• Authors: Youness Mellak; Konstantinos Chat...; Alexandre Bousse; Catherine Chez-Le...; Dimitris Visvikis; Julien Bert
• Keywords: FOS: Physical sciences; Medical Physics (physics.med-ph); Physics - Medical Physics

In recent years, the use of Monte Carlo (MC) simulations in domain Medical Physics has become a state-of-the-art technology that consumes lots computational resources for accurate prediction particle interactions. The generative adversarial network (GAN) been recently proposed as an alternative to improve efficiency and extending applications tools both medical imaging therapeutic applications. This study introduces new approach simulate positron paths originating from Fluorine 18 (18 F) isotopes through utilization GANs. methodology developed pure conditional transformer least squares (LS)-GAN model, designed generate paths, track their interaction within surrounding material. Conditioning factors include pre-determined number interactions, initial momentum emitted positrons, derived emission spectrum F. By leveraging these conditions, model aims quickly accurately electromagnetic interactions paths. Results were compared outcome produced with Geant4 Application Tomography Emission (GATE) MC toolkit. Less than 10 % difference was observed calculation mean maximum length path 1-D point spread function (PSF) three different materials (Water, Bone, Lung).