In-silico comparison of X-ray and proton computed tomography for proton therapy dose simulation with a full Monte Carlo treatment planning
Résumé
Proton therapy treatment efficiency directly relies on the proton range derived from tissue-relative stopping power (RSP) maps. In clinical practice, RSP is obtained from the conversion of X-ray CT Hounsfield units. This calibration step introduces uncertainties on the dose distribution. Proton computed tomography (pCT) has the potential to improve the accuracy of the proton therapy treatment by a direct derivation of RSP maps. This study aims to quantify the dosimetric accuracy of pCT. A full Monte Carlo proton therapy treatment planning has been developed. Both X-ray and proton CT images of a human computional phantom are simulated with an equivalent 2mGy dose. RSP maps are derived from both images and used to estimate the dose distribution. RSP maps derived from pCT are significantly more accurate and less sensitive to common X-ray artefacts. Proton CT based treatment enables a more precise proton range estimate of about 3-4%. This Monte Carlo method can be applied to any clinical case in order to estimate the expected margins reduction of a proton therapy treatment planning based on pCT.