Characterizing the performance of a proton tomography system for x-ray CT cross-calibration

Speaker: Elena Fogazzi
Institution: University of Trento and INFN TIFPA, Trento, Italy

Abstract

Purpose: INFN research projects have recently studied the feasibility of proton Computed Tomography (pCT) as a possible clinical method to cross-calibrate x-ray CT (xCT), aiming at improving the dose computation accuracy in proton treatment plans. Herein, we report on the performance of our pCT apparatus.

Methods: The pCT system was tested in June 2022 at the Trento Proton Therapy Center with a proton beam at 211 MeV kinetic energy, and with a custom-built phantom made of five different cylindrical inserts, with density in the range [0.7-2.2] g/cm3 (Fig. 1) to probe the performance parameters, i.e. spatial resolution, accuracy and noise spectrum. A filtered backprojection algorithm, taking into account the protons' most likely path, allowed reconstructing the phantoms' Relative Stopping Power (RSP) 3D maps. The cylinders' edges were employed to investigate the spatial resolution through the calculation of the Edge Spread Function. The pCT RSP values of the inserts were instead compared with multilayer ionization chamber measurements to estimate the system accuracy .

Results: According to preliminary analysis, a resolution of about 0.65 lp/mm was estimated. Direct measurements of RSP values of the phantom inserts showed a mean absolute percentage error of 0.5%. Finally, statistical analysis suggested a minimum obtainable noise magnitude of about 0.005 for RSP.

Conclusions: The obtained performances allow designing a first clinical application for our pCT system. To this purpose, the RSP maps of a set of stabilized, heterogeneous, biological phantoms have also been acquired and will be compared voxel-by-voxel with the corresponding xCT images, obtaining a cross-calibrated xCT calibration curve i.e. a SPR-HUs look-up table. Once validated, this procedure might be offered to proton therapy centers not equipped with a pCT system, representing a new and direct calibration method to be adopted in treatment planning.

Resources