GEANT4 simulation of exit energy in proton medical imaging

• Published in: Radiation physics and chemistry (Oxford, England : 1993) Vol. 167; p. 108338
• Main Authors: Silva, R.C.L., Denyak, V., Schelin, H.R., Hoff, G., Paschuk, S.A., Setti, J.A.P.
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01.02.2020; Elsevier BV
• Subjects: Absorbers; Computed tomography; Computer simulation; Energy; Energy distribution; Energy spectra; Exit energy spectrum; GEANT4; Imaging techniques; Medical imaging; NMR; Nuclear magnetic resonance; Proton medical imaging; Protons; GEANT4; Proton medical imaging; Exit energy spectrum
• ISSN: 0969-806X; 1879-0895
• DOI: 10.1016/j.radphyschem.2019.05.028

Imaging techniques using protons as incident particles are currently being actively developed as a substitute for X-ray computed tomography and nuclear magnetic resonance methods in proton therapy. In proton imaging, density resolution depends on the position of the maximum of the exit energy distribution and the width of this distribution which influences uncertainty in the definition of the exit energy. In recent years, many research groups have used the GEANT4 toolkit to simulate proton imaging devices, without, however, detailed analysis of the reproducibility of the physical quantities simulated by the different models in GEANT4 for the case of thick absorbers. In the present report, we performed a careful quantitative comparison of the experimental and simulated parameters of the exit energy spectra that are important for proton imaging. The difference in exit energy between the BERT and BIC models does not exceed 0.5%, while between EMY and EMV models it is always much larger and reaches 2.5% and is accompanied by a significant difference in spectrum shape. The differences between exit energies and widths of the simulated and measured distributions do not exceed 15% and depend on the exit energy (absorber thickness). •Difference in exit energy between the BERT and BIC models does not exceed 0.5%.•Such difference between EMY and EMV models reaches 2.5%.•Differences between the simulated and measured exit energies do not exceed 15%.•Differences in distribution width (simulated and experimental) do not exceed 12%.