Degree Name

Master of Science (Research)


School of Engineering Physics


Introduction With the adoption of technologies such as stereotactic radiosurgery in the treatment of cancer, there is an increasing trend towards smaller field sizes where the importance of accurate penumbral measurements is critical. Small segments are also common in intensity modulated radiation therapy deliveries; hence accurate dose assessment at the edge of multi-leaf collimated segmented fields is also paramount.

Clinically used detectors have significant detector volumes that contribute to measurement of wider penumbral dose profiles than the beam produces. This overestimate of penumbral width in turn has an impact on the radiotherapy treatment planning modelled dose distributions used for patient treatment. This is because the penumbra broadening in the dose profile affects the source size parameter used in radiotherapy treatment planning system. In this thesis, the extent of penumbral broadening was quantified and methods to produce data with effectively zero detector volumes were investigated. This data was used to calculate a source size for the computer model to best match the measured data.

Methods Data was measured for a 6 MV beam (Varian Clinac 600C) using a diamond detector, a pinpoint detector, and a 0.125 cc ionisation chamber. Extrapolation and deconvolution techniques were used to calculate zero detector volume data. The extrapolation technique was studied in detail and a new verification technique, which involved R2 and dose differences, was developed to calculate the fit and errors associated with the extrapolation method. The amount of penumbral broadening and source size overestimation in Pinnacle decreased with decreasing detector diameter. Results In this study, penumbral broadening of up to +1.8 mm (80%-20% penumbra) due to the detector volume effect was found to occur across both large and small field sizes and this resulted in overestimations in the source size parameter in the Pinnacle radiotherapy treatment planning system by +1.2 mm for the 0.125cc ionisation chamber (from the zero detector source size of 0.9 mm).

The effect of source size overestimation in Pinnacle was studied by the calculation of dose distributions with the virtual zero detector dataset and the 0.125 cc ionisation chamber dataset.

The point in the middle of the field had minimal change but there were changes in the dose distribution which were due to a summation of penumbral perturbations of each beam.It was found that for large field sizes (~10×10 cm2) the summed doses in the treatment region were underestimated by approximately 0.5%. For small field sizes (1×1 cm2) summed dose in the treatment region was overestimated by approximately 3.5% while over the whole region there was an overestimation of approximately 11%. For the case of a 3DCRT prostate plan, changes in dose were underestimated by to 1% for volumes typical of the PTV and overestimated by up to +1.5% for volumes typical of organs at risk.

Equations were derived that produced agreeable links between the detector volume and the penumbral width as well as the penumbral width and the source size parameter in Pinnacle. The coefficients required in these equations were calculated from datasets obtained from the measurement of dose profiles by physical detectors and the calculation of dose profiles in the treatment planning system respectively. The use of these equations could be used to estimate and/or correct for the detector volume effect on the source size parameter in the treatment planning system with a minimum of beam measurement time. However, further investigations are required to verify this over a wide range of conditions such as beam energy and collimator design.

The 1D dose profiles measured with different detectors were analysed in terms of intersection point and inflection point. The results indicated that there were significant deviations of both these points from a normalised dose of 50% with small field sizes. There was an overestimate of the radiation field size (50%) by 0.8 mm measured with the 0.125cc ionisation chamber at the field size of 1×1 cm2 but at other field sizes measured the radiation field size was within ±0.2 mm. The intersection point determined the spatial location of overestimation and underestimation of point and summed dose. The overall summed dose was found to be unaffected by the detector volume effect at a field size of 2.3x2.3 cm2, which was similar to the minimum field size for lateral electron equilibrium (2.6x2.6 cm2).

Conclusions- The results of a survey of different radiotherapy institutions indicated that approximately half of measurements done for use in modelling the Pinnacle radiation treatment planning system involved the use of ionisation chambers (approximately 0.1 cm3). In this study, it was demonstrated that (1) the detector volume effect is significant as matching the model to broad penumbra overestimates the virtual source size parameter by the order of +1 mm in Pinnacle; (2) that the effect on dose distributions for single fields in the penumbra are the dose may be different by 1-10% compared with zero detector profile matched data (3) that corrections to the detector volume can be made with a new single detector technique combined with a predictive equation. This makes the correction more feasible with consideration to time constraints.

02Whole.pdf (4585 kB)