Assessing the Potential Clinical Impact of Variable Biological Effectiveness in Proton Radiotherapy

It has long been known that proton radiotherapy has an increased biological effectiveness compared to traditional x-ray radiotherapy. This arises from the clustered nature of DNA damage produced by the energy deposition of protons along their tracks in medium. This effect is currently quantified in clinical settings by assigning protons a relative biological effectiveness (RBE) value of 1.1 corresponding to 10% increased effectiveness compared to photon radiation. Numerous studies have shown, however, that the RBE value of protons is variable and can deviate substantially from 1.1, but experimental data on RBE and clinical evidence of its variability remains limited. The potential for using the variable RBE of proton radiation to improve clinical treatment plans has been theorized, but it is accepted that more experimental in vitro and in vivo data are needed before clinical adaptation of these techniques may occur. Nevertheless, it will be important to identify strategies in which the variable nature of proton RBE may be used to inform treatment planning. The goal of this work is thus to investigate if the assumption of a constant proton RBE has an adverse effect in current clinical applications and if the variable biological effectiveness of protons can be quantified from clinical data. First, results from high-resolution experiments quantifying proton RBE are compared to multiple models for calculating RBE. A new model is then proposed which can more accurately reproduce the experimental results. These models are implemented in a Monte Carlo-based dose calculation system and their output is compared for a cohort of pediatric patients treated for brain tumors with proton radiotherapy who subsequently presented with post-treatment image changes identified on magnetic resonance imaging. One RBE model is identified as the best candidate for further study; however, results of volumetric analyses of RBE-weighted dose prove inconclusive in correlating with image changes. A model is developed that can describe the probability of voxel-level image changes (signifying normal tissue damage) based on proton dose and linear energy transfer. The model constitutes the first clinical evidence for the variable biological effectiveness of protons and holds promise for the improvement of proton therapy treatment planning