Degree Name

Doctor of Philosophy


University of Wollongong. Department of Engineering Physics


This thesis examines skin doses in the currently experimental field of Magnetic Resonance Image guided Radiotherapy (MRIgRT). This modality promises realtime MRI quality soft tissue resolution whilst x-ray beam treatment is occurring.The potential advantage of the high image resolution offered is the improved tracking of organ and tumour motion. This in turn will allow a more conformal treatment of the tumour tissue. The ultimate improvement is then the reduction of dose to healthy tissue whilst dose escalation to tumour cells. In turn this would lead toimproved survival rates of radiotherapy patients.

At present this modality is still in an experimental stage however at least 2 prototype machines currently exist. Preliminary studies in the literature describe potential limitations of MRIgRT being related to excessive skin doses caused by the magnetic fields. Until this thesis no studies existed which predict the skin doses at a 10 micron resolution. This is particularly important in obtaining accurate estimates of the dose at the skin structure depths. This resolution reflects the biologically significant structures of the skin, hence the values are of direct clinical interest.

In this thesis the skin dose from 6 MV photon beams has been studied both experimentally and theoretically. The effects of transverse magnetic fields on perturbing this dose has also been studied via Geant4 Monte Carlo simulations. The skin dose arrises from the photon beam and it is altered by the effect of the magnetic field, in particular, the influence on the secondary electrons produced by the x-rays. Ion chamber and Attix chamber detector measurements have been used to confirm the accuracy of the Geant4 Monte Carlo Model.

A detailed Monte Carlo model has been used which utilises high resolution voxel scoring. This enables the extraction of the 70 μm skin doses, the epidermal, and the dermal skin doses in the entry and exit regions of various phantoms. Beam field sizes of 5x5, 10x10, 15x15, and 20x20 cm2 have been simulated, exposed to magnetic fields ranging between 0 and 3 T. The effect of entry and exit surface orientation with respect to the beam central axis is also investigated over the range of +75◦ to-75◦. The skin doses have been calculated at the beam central axis, and over the entire entry and exit surfaces of the phantoms used.

In the beam entry region the role of air-generated lepton contamination has been investigated. In transverse magnetic fields of 6 0.4 T some lepton contamination is not fully removed by the magnetic field. Such contamination should be included in simulations to produce accurate skin dose calculations. The use of helium air-bags on the entry surface has also been simulated, however it has been shown not to be of benefit due to the the bolus effect of the helium bag wall material against the entry surface. There is a considerable entry skin dose increase predicted for strong magnetic fields (1-3 T) and large positive surface angles. This becomes > 100% Dmax for 1 T at around 60◦, and for 3 T this occurs at 0◦. At negative surface angles the magnetic field can actually lead to skin dose reductions relative to the case of no magnetic field.

In the exit region the effect of the Electron Return Effect (ERE) has been examined. Opposite to the entry side, significant skin dose increases are predicted at negative surface angles. At -45◦ and 0.6-3 T, the skin dose is around 100% greater than at 0 T. For a 20 cm thick phantom these are skin doses of 70% of Dmax. Also,the lower magnetic fields lead to the greatest increases in exit side skin dose. The exit skin dose increases can however be reduced significantly with the use of 1 cm thick exit bolus (water layer). Simulations performed show how this exit bolus almost completely eliminates the negative effects of the magnetic fields.

Overall, the results of this thesis quantify the skin doses expected in 6 MV transverse field MRIgRT. Despite many scenarios that lead to skin dose increases on both the entry and exit sides, there are also many techniques to counter or avoid these. Further high resolution Monte Carlo simulations will most definitely play anintegral part in the continued characterisation of the skin doses in MRIgRT.



Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.