Achievable Accuracy and Reproducibility in Radiotherapy.

The use of Volumetric Modulated Arc Therapy (VMAT) is expected to increase, largely due to development of helical tomotherapy. VMAT minimises the occurrence of hotspots and irradiation of critical organs, providing more uniform dose while sparing critical organs. Two important characteristics of VMAT are its dynamic nature and dosimetric variability in radiation delivery. These present considerable challenge for clinical physicists as the implementation of the process contains a number of sources of uncertainty and thus require robust QA. This dissertation describes an evaluation of the currently achievable accuracy in VMAT delivery. Uncertainty in dose determination between primary standards is compared, as well as beam calibration of megavoltage photons at reference field sizes and for small field dosimetry associated with tomotherapy. Additional steps in the radiotherapy process including target volume delineation, evaluation of dose distribution and daily patient treatment (patient setup) have also been considered. The codes of practice currently employed in electron and photon radiotherapy beam dosimetry are well developed, in particular those of the International Atomic Energy Agency (IAEA TRS-398), the American Association of Physicists in Medicine (AAPM TG-51) and the Institute of Physics and Engineering in Medicine (IPEM 2003). For the photon and electron beams studied, the applied protocols show that all measured absorbed doses agree, with differences of less than 2. 0%. For the absorbed dose calibrations using tomotherapy, beam uncertainty in primary standard calibration has been estimated to be 1. 05%. An assessment of uncertainty in target volume delineation has also been considered. The combined uncertainty for the head & neck (H&N) study is 1. 67%, while for the prostate study the combined uncertainty is 5. 57%. Regarding actual beam delivery 15 Delivery QA plans for H&N patients were assessed for a tomotherapy system using Kodak X-Omat V film and an A1SL Ref F92722 ion chamber versus MapCheck2. Gamma matrix distribution was applied to evaluate the difference between measured and computed dose distributions. The combined uncertainty from H&N is 1. 11%, lower than that from the prostate study, which is 1. 16%. Generally, the combined uncertainties for the whole study in head and neck were found to be 2. 35% while for prostate cases, the combined uncertainty is 6. 29%.