Image reconstruction of proton computed tomography modelled by Geant4 Monte Carlo toolkit

Abstract

Accuracy in the treatment planning of proton therapy depends on the accuracy of the information used to calculate the relative stopping power of tissues in the patient's body. This information is obtained from x-ray computed tomography images using a calibration curve to convert Hansfield numbers to relative stopping power values. Using x-ray computed tomography images leads to errors in estimating the proton range and the proton dose distribution in the treatment plan program. But applying the proton computed tomography eliminates this error and directly calculates the relative stopping power map of the tissues. In the present study, a modern proton computed tomography imaging system was simulated using the Monte Carlo Geant4 toolkit by tracing particle-to-particle trajectory. The purpose of this simulation was the improvement of density resolution of tissue without dose increment. The standard CIRS 062M phantom was irradiated with a 300 MeV proton beam. The energy, position, and direction of particle movement values before and after the phantom were stored in the root file by nuclear detectors. The image matrix phantom was reconstructed as a relative stopping power map using three radon analytical algorithms. The comparison was made regarding dose, density resolution, and RMSE concerning real phantom image data. The proposed algorithm improved the density resolution from 9.1% to 4.3% and RMSE from 26.43% to 6.81% by correcting the angles of the projections at the same dose level.

Keywords

• pct
• electron density phantom
• cirs062m
• reconstruction
• fbp