An approach to the interpretation of clinical data on the tumour control probability-dose relationship.

A conventional statistical model allows predicting the sterilisation rate as a function of dose. However, the computation requires data on biological parameters (proportion of clonogenic cells, survival per fraction, multiplication rate) which are inaccessible for human tumours. The curative dose 50% (TCD50) can be used as a synthesis of these parameters and its significance for the response-dose relationship of a population of tumours of uniform radiosensitivity is discussed. The slope of the dose-control curve provides vital information regarding the variation in radiocurability of the various individual tumours. The model allows the analysis of the clinical data and the separation of tumour subsets with different radiation responsiveness. It provides an evaluation of the benefit which could be obtained from the identification of the subsets and a guidance for the clinical, pathological and biological studies which relate to this identification. The change in the response-dose relationships with the tumour size cannot (usually) be explained by the cell number increase alone. Other possible factors of reduced radiocurability are discussed.