Analysis of generalized models of radiogenic risk. Part 2. UNSCEAR models

Analyzed generalized risk models are presented in UNSCEAR 2006. In the paper is analyzed generalized models of absolute and relative risks from tables 45-58 the report of UNSCEAR 2006. Presented the mathematical formula that generalizes these models together with unificated constants tables for all of the locations and risk indicators. As the main criterion of reliability of the models considered the compliance of the corresponding background models of morbidity and mortality (EAR/ERR), with the background data for the composite population ICRP, Japanese, and Russian populations. A fundamental disadvantage of non-linear models UNSCEAR consist that they are giving the models background morbidity and mortality that are dose-dependent. Background incidence and mortality models for linear risk models in quantitative, and to some cancers and qualitative terms, are not consistent with statistical data. The sets of models UNSCEAR and ICRP, both fitted to the same LSS data, in general poor agree with each other. The results of the comparative analysis of the two sets of models do not allow giving clear preference to one of them. For estimates of radiogenic cancer of the thyroid is preferable to apply the UNSCEAR model, while for cancers of the colon and lung more accurate estimates can be obtained by ICRP models. The choice of model should performed separately for each cancer localization with regard to the objectives of risk assessment and characteristics of the solved problems.