According to the Swedish Radiation Safety Authority's Regulations in Radiotherapy (SSMFS 2008:33) are Swedish clinics required to verify the individual dose to patients, prior the first time a new treatment is given. To fulfil this requirement, current clinics perform in-vivo dosimetry in three dimensional conformal radiation therapy (3DCRT), which means that the entrance of each patient field is measured with a diode in a specific point and compared to the calculated dose in the same point.
Volumetric Modulated Arc Therapy (VMAT) is an advanced treatment technique where radiation is delivered by a linear accelerator with a continuous rotating gantry combined with dynamic collimator movements and varying dose rates. Due to the complexity of VMAT, the previous in vivo dosimetry procedure is not applicable. Current practice in Sweden is to perform pre-treatment quality assurance (QA). The delivered dose is compared with the planned dose based on measurements with different methods, for instance: point dose verification and/or dose distribution verification performed in a homogenous phantom. One disadvantage is the difficulty to interpret deviations between predicted and measured doses, if the deviation is caused by failures of the accelerator performance or the theoretical patient dose distributions. Another weakness is that the recalculations are performed for a homogenous phantom and not for the real patient. Furthermore, there are a resolution limitations caused by the distance between the detectors in dose verification phantoms.
In a previous work (SSM report 2017:13) the authors of this project developed a Monte Carlo (MC) patient-specific pre-treatment QA system. The system allows the calculation and evaluation of dose distributions for VMAT plans in patient geometry based on Computed Tomography (CT) images. The QA system was tested by performing retrospective analysis of radiation therapy plans created for treatment of cancer. An advanced software tool was developed for generation of MC compliant voxel phantoms from patient DICOM data. A stand-alone analysis module was also developed to perform quantitative calculation of deviations between planned and MC dose distributions.
Objectives of the project
The aim of this project was to improve the quality assurance of VMAT planned patient dose distributions for some specific diagnosis, e.g. cervix uterus- and head-neck treatments, using the MC system developedin the previous work. In addition, a pilot study of ionization chamber response in dynamic fields has been performed.
The MC system developed previously has been extended by including a new TrueBeam accelerator model. The functionality of the stand-alone analysis module has been improved. Deviations between clinical- and MC calculated dose distributions have been investigated for 70 patient plans by comparison of Dose Volume Histogram (DVH) parameters and by more detailed 3D analysis. Diagnose specific tolerance criteria for treatment of four cancer sites: prostate, thorax, gynaecological and head-neck, are suggested and evaluation procedure recommended for practical implementation. Furthermore, the response of an ionization chamber in a dynamic MLC field has been studied and compared to that in a static field.