Author: Gisell Ruiz Boiset, Paulo ROBERTO Costa, Renan Felix dos Santos, Gustavo Freire Pereira da Silva, Alessandra Tomal 👨🔬
Affiliation: University of São Paulo (USP), Institute of Physics, Universidade Estadual de Campinas. Instituto de Física Gleb Wataghin 🌍
Purpose: Evaluate the feasibility of a bench-top PCCT using Geant4 Monte Carlo simulations. The study applied the Geant4 framework on determining the electronic density of materials as a function of CT number at two energy ranges.
Methods: In this study, the Geant4 toolkit3 (release 11.1.2) was used. A focal spot-to-image detector distance of 1045 mm was adopted. A 50 kV X-ray spectrum measured by a CdTe spectrometer was incorporated. The virtual photon-counting detector consisted of a 5122 matrix with pixel size of 0.025 mm2 and an ideal efficiency was modeled. Low and high energy acquisition ranges with an energy threshold of 30 keV were adopted and CT images were reconstructed using FBP. Two virtual cylindrical phantoms were designed: a calibration phantom (6 mm radius) with internal cylindrical inserts (1 mm radius) of polypropylene, nylon, silicon dioxide, polycarbonate and Teflon; and a test phantom with 3 mm paraffin and 2 mm polyethylene inserts, all materials immersed in water. The axis of these phantoms was positioned at 225 mm from the focal spot. ROIs were selected and the HUs measured for the reconstructed images from low (HUlow) and high (HUhigh) energy ranges. A dual-energy subtracted quantity was estimated as DHU = (1 + α)HUhigh - αHUlow, where α is a weighting factor. A linear regression between the electronic densities, ρe, and DHU, fitted based on the calibration phantom, was used to estimate the ρe of the test phantom materials.
Results: The linear fit resulted in α=1.6 (r2=0.94). Using test phantom, the ρe were estimated with accuracies 0.0%, 0.4%, 4.2%, and 8.6%, respectively, to air, water, polyethylene, and paraffin.
Conclusion: Based on MC simulations, the proposed experimental model demonstrated to be a viable alternative for characterizing materials based on their electronic densities. The test phantom's composition was determined with inaccuracies lower than 9%.