Evaluation of proton path estimators for spatial resolution modification in images obtained by Proton computed tomography

Abstract

Proton computed tomography (pCT) can reduce proton therapy uncertainty by measuring Relative Stopping Power (RSP) directly. The spatial resolution of the pCT images decreases due to the multi colomb scattering (MCS) of protons inside the phantom. This reduction of image quality can be compensated by using the most probable proton path in the reconstruction algorithm. In this study, a pCT system was simulated by particle-to-particle tracking of protons using the Geant4 toolkit. This simulation improves the spatial resolution of images obtained from applying different estimators of the proton path, including straight line path (SLP), cubic spline path (CSP), and most likely path (MLP). The Catphan528 phantom was irradiated with 200MeV protons. The energy, position, and direction of the particle were recorded before and after the phantom. The RSP image matrix was modified by weigthing factors obtaind using SLP, CSP, and MLP path esitimators and image was reconstructed using FBP. Spatial resolution and root mean square errors (RMSE) were compared to phantom image data. According to the results, the MLP method is less error-prone and more accurate than other methods in resolving spatial resolutions. For 100,000 protons, with image resolution ranging from 1 to 0.1 mm, the spatial resolution increased from 3 to 9 line pairs/cm, while the RMSE increased from 8.11% to 14.97%.

Keywords

• proton computed tomography
• image reconstruction
• monte carlo simulation
• proton path estimator