Estimating RBEs at clinical doses from microdosimetric spectra.

To the Editor, Microdosimetry is perhaps the most useful tool for intercomparing the different types of radiation that are used in radiotherapy; the relative effects of different radiations are controlled by the different initial energy deposition patterns in cellular targets—the subject matter of microdosimetry. It is therefore appropriate that many authors have measured microdosimetric spectra in clinical beams, and have used these data in an attempt to predict clinically relevant relative biological effectiveness ~RBE!. It appears, however, that a particular formalism that has been used by a number of authors to calculate RBEs at clinical doses based on microdosimetric spectra is, in some important respects, incorrect. Here we discuss this issue, but point out that a correct application of microdosimetric theory does indeed allow predictions of RBE, both at low doses and at clinically relevant doses, based on measured microdosimetric spectra. Before commenting on the formalism that has appeared in the literature, we briefly review the fundamentals of microdosimetric theory as they apply to the predictions of RBE. The fundamental assumption is that relative biological effects are determined by the different energy deposition patterns that different radiations deposit in cellular targets. These energy deposition patterns are quantified through the stochastic quantity specific energy (z), defined as the energy per unit mass deposited in a particular cellular target. After exposure to a dose D of radiation type i , the probability density function of z in a set of cellular targets is given by the Poisson-weighted sum of the effects of different numbers of independent tracks passing through the targets: