Comprehensive validation of a Monte Carlo kV-CBCT model using OSL and spectral measurements

International audiencePurpose or Objective: The use of Cone-Beam computed Tomography (CBCT) is progressively increasing in radiotherapy treatments, but additional doses induced are not well quantified and could amplify the risk for patients to develop a second cancer. There is a need, expressed by the medical physics community, to develop tools to estimate, report and potentially help reducing CBCT doses. We hence developed a Monte Carlo (MC) model for the XVI kV-CBCT system. The dosimetric and geometric accuracy of the simulated beams was evaluated by comparisons withmeasurements in a water tank, and x-ray energy spectra acquisitions. Before clinical use, the model requires an evaluation in anthropomorphic phantoms in which were inserted OSL NanoDots (Landauer). The purpose of the present study is to develop an accurate dosimetric protocoltaking into account for OSL energy dependence in keV energy range.Material and Methods: A MC model of the XVI was developed using the PENELOPE code. The dosimetric and geometrical evaluation of the beam MC models was performed by comparing simulations with lateral and depth-dose profiles measured using a PTW Farmer-type chamber, and on-axis energy spectra measured with a CdTe detector. These comparisons were performed at 120, 100 and 80 kVp, and for different filtration/collimation couples. For OSL measurements, the first step was to perform, in different beam qualities, in-air cross-calibrations with a PTW Farmer type chamber. At this energy range, OSL exhibit strong energy dependence, so the signal needs to be corrected for the spectral variations between calibration and measurement conditions. Thus, to ensure accurate dose measurements, a correction method was developed using calculated spectra. The dosimetric protocol was validated by performing dose profiles with OSL inserted in a PMMA tube submerged in water. Preliminary comparisons with XVI model were made with acquisitions in a home-made heterogeneous phantom consisting of a water tank equipped with PMMA, bone and lung equivalent inserts.Results: Experimental and simulated lateral and depth-dose profiles, and energy spectra, are in excellent agreement (Fig 1A).These results validate that the MC model accurately reproduces the dosimetric and geometric properties of the XVI beams. The uncorrected OSL profiles in the PMMA tube over-estimate by 15 % the calculated doses. However, energy corrected measurements are matching the simulations and the differences not exceed 7.5 % (Fig 1B and 1C). Table 1 presents doses measured at different points in the heterogeneous phantom and discrepancies not exceed 11.3 %.Conclusion: The dosimetric protocol developed for OSL allows accurate measurements of imaging doses, and will be then used to validate the dose calculation tool in pre-clinical conditions. Preliminary results obtained in the home-made phantom highlight the accuracy of XVI MC model. Further validations are on-going in anthropomorphic phantoms