Impact of sensitive volume dimensions on microdosimetric measurements of radiation quality for proton therapy

Introduction Macroscopic dosimetric quantities (dose, LET) are ill suited for describing the effect of clinical proton beams at sub-cellular level. Microdosimetry provides a stochastic description of microscopic interactions, suitable for treatment planning. Despite increasing advancements, poor standardisation and lack of code of practice hinder clinical adoption: detector sizes and geometries vary across implementations, making cross-comparison difficult. Approach The impact of detector dimensions on microdosimetric means is analysed via Monte Carlo simulations at several points along a proton Bragg curve, taking care of uncoupling dimensions parallel (thickness) and transversal (width) to the beam. Influences of different path length definitions are concurrently analysed. Main results Increasing the thickness introduces an averaging effect that reduces dose-weighted means by as much as a 50% (for a 10µm detector thickness, compared to 1µm), but has a small impact on the frequency-weighted mean. Increasing the detector width captures a higher fraction of secondary electrons, and thus causes an increase of 10-20% in every microdosimetric mean. For non-spherical sensitive volumes, a mean path length scored via Monte Carlo represents the best solution, but the detector thickness constitutes a good approximation. Significance The results provide a range of recommended dimensions for suitable sensitive volume geometries for microdosimetric measurements in proton beams, along with a quantitative estimate of the expected microdosimetric means variation within such range. Suggestions for standardising the critical parameter of the mean chord length are likewise given. This paves the way to further systematic standardisation for future development and use in clinics.