Techniques for the increased utilization of dose response variability in proton therapy

Particle therapy is a form of radiation therapy in which protons and heavier ions are used, as opposed to photons in conventional radiation therapy. The biological effectiveness of particles compared to photons is often quantified as relative biological effectiveness (RBE). In clinical practice, protons are assumed to be 10% more efficient than photons, despite the fact that RBE is known to vary. On the other hand, variable RBE models can be used to describe the RBE at a given position as a function of a few parameters, such as the linear energy transfer (LET) of the beam. Questions of accuracy and validation have prevented the clinical introduction of variable RBE models. In this thesis, we tried to develop a variable RBE model for protons and carbon ions, and then apply it in a proton planning study. We started with developing a beam model for protons. It was based on measured data at the Skandion Clinic in Uppsala, Sweden. It is capable of describing the spatial, angular and energy distributions of a proton beam at a certain position in a treatment room. This, coupled with a particle transport engine, allows for accurate study of the physical properties of a clinical beam. Prior to developing our RBE model, we studied a number of publications containing proton in vitro cell survival data. It was found that the particle beams used included heavy secondary particle contamination and thus this need not be accounted for separately in a proton RBE model based on this data. Taking this into account, the subsequent RBE model did not provided increased accuracy compared to the considered proton RBE models. For carbon ions, accuracy was increased. Coupled with a treatment planning system, treatment plans taking into account RBE variability can thus be made with this RBE model. Finally, we applied the nanoCluE RBE model in a proton dose painting planning study, where the tumor target is given a heterogeneous dose based on an estimated radio sensitivity map of the tumor such that more resistant areas are given higher doses. Variable RBE was not beneficial in increasing the control probability of the tumor, but it did help in decreasing doses to nearby, healthy tissue