An ensemble dose-response model for monoenergetic ions and doses relevant for space radiation carcinogenesis

For future missions to the moon or Mars, exposure to space radiation is a significant hazard requiring accurate risk assessment tools to guide and evaluate mitigation strategies. NASA has developed a cancer risk assessment model to perform such projections, but uncertainties associated with radiation quality and dose-rate effects are substantial. Quality factors (QFs) are used to describe the enhanced effectiveness of radiation types found in space for inducing a health effect (carcinogenesis, in this case) compared to terrestrial radiation. Space radiation QFs are developed on the basis of dose-response models used to describe ground-based experimental data for ion and gamma exposures over a range of doses. Two challenges arise when attempting to develop such models. First, data has been collected for only a sparse subset of the particle types and energies comprising the highly energetic mixed-field space environment. Second, ground-based animal carcinogenesis studies are limited in number, necessitating the use of surrogate biological endpoints. Robust methods are needed to describe available data and make predictions where no data exist. In this work, an ensemble dose-response model is presented as a first step towards a new foundation for improving QFs used in space radiation risk assessment. The ensemble is based on a generalized dose-response model framework including terms representing targeted effects, non-targeted effects, and cell-killing/sterilization. Available options for each of these key dose-response components are collected to yield 378 ensemble members. The members are weighted according to a performance-based statistical metric, determined by comparing each member to experimental data. Three relevant global experimental datasets (i.e., each covering a range of particle types and doses) are considered and used to demonstrate results. The ensemble dose-response model is then used to calculate radiation quality metrics such as relative biological effectiveness (RBE) and relative effects ratio (RER). Experimental statistical uncertainties are propagated through RBE/RER calculations. It is shown that the ensemble dose-response model provides an objective and robust foundation upon which space radiation QFs and associated uncertainties can be constructed.