Introducing Biological Microdosimetry for Ionising Radiation

Microdosimetry is important for radiation protection, for understanding mechanisms of radiation action, and for radiation risk assessment. This article introduces a generic, Monte Carlo based approach to biological microdosimetry for ionising radiation. Our Monte Carlo analyses are carried out with a widely used Crystal Ball software. The approach to biological microdosimetry presented relates to quantal biological effects data (e.g. cell survival, mutagenesis, neoplastic transformation) for which there is an initial linear segment to the dose-response curve. The macroscopic dose data considered were selected such that is could be presumed that the vast majority of cells at risk have radiation dose delivered to their critical target. For cell killing, neoplastic transformation, and mutagenesis, the critical biological target for radiation is presumed to be DNA. Our approach to biological microdosimetry does not require detailed information about the mass, volume, and shape of the critical biological target. Further, one does not have to know what formal distribution function applies to the microdose distribution. However, formal distributions are required for the biological data used to derive the non-parametric microdose distributions. Here, we use the binomial distribution to characterise the variability in the number of cells affected by a fixed macroscopic dose. Assuming this variability to arise from variability in the microscopic dose to the critical biological target, a non-parametric microdose distribution is generated by the standard Monte Carlo method. The non-parametric distribution is then fitted using a set of formal distributions (beta, exponential, extreme value, gamma, logistic, log-normal, normal, Pareto, triangular, uniform, and Weibull). The best fit is then evaluated based on statistical criteria (chi-square test). To demonstrate the application of biological microdosimetry, the standard Monte Carlo method is used with radiobiological data for cell survival after exposure in vitro to alpha radiation in order to generate presumed distributions of the microscopic dose to the critical target (DNA). Results are presented indicating that for a low macroscopic alpha radiation dose to a large group of C3H 10T1/2 cells, the microdose distribution over critical targets of individual cells that receive radiation microdoses may be adequately characterised using a log-normal or gamma distribution.