Abstract
Most of theoretical models encountered in radiobiology and radiotherapy use the absorbed energy dose (d) as the reference variable. By taking account of the nature of the elementary transformations which occur at molecular level in irradiated systems, it is shown that D is generally not a suitable variable, and that it is preferable to use a formalism in which the variables of radiation spectrum and intensity, and the time variable, are explicity separate. In parallel with the selection of suitable variables, a topological representation in the form of transformation graphs provides a clear schematization of the various processes and their respective roles. This representation, which includes physical, chemical and biological phenomena, clearly shows both the level of observations and the level of hypotheses. When searching for particular models one must first take into account the phenomenological requirements resulting from the observations which can be made, before introducing any complementary hypothesis. The basic models of target and hit theory generally do not satisfy these requirements. At it is easy to see, the formalism used is sufficiently general and can be utilized in fields different from the one described here.1.

This publication has 14 references indexed in Scilit: