Uncertainty of stochastic parametric approach to bone marrow dosimetry of Sr.

The objective of this study is to evaluate the uncertainties of the dosimetric modeling of active marrow (AM) exposure from bone-seeking Sr. The stochastic parametric skeletal dosimetry (SPSD) model was specifically developed to study the long-term effects resulting from chronic Sr exposure in populations of the radioactively contaminated territories of the Southern Urals region of the Russian Federation. The method permits the evaluation of the dose factors ( and , which convert the radionuclide activity concentration in trabecular () and cortical () bone volumes into dose rate in the , and their uncertainties. The sources of uncertainty can be subdivided into inherent uncertainties related to the individual variability of the simulated objects and introduced uncertainties related to model simplifications. Inherent uncertainty components are the individual variability of bone chemical composition, bone density, bone micro- and macro-architecture as well as distribution within the skeleton. The introduced uncertainties may result from the stylization of bone segment geometry, assumption of uniform cortical thickness, restriction of bone geometry and the selection of the applied voxel resolution. The inherent uncertainty depends on a number of factors of influence. Foremost, it is the result of variability of distribution within the skeleton. Another important factor is the variability of bone micro- and macro-architecture. The inherent uncertainty of skeletal-average dose factors was found to be about 40-50%. The introduced uncertainty associated with the SPSD model approach does not exceed 16% and mainly depends on the error of bone-shape stylization. The overall inherent and introduced uncertainties of and are below 55% and 63%, respectively. The results obtained will be incorporated into the stochastic version of the Techa River Dosimetry System (TRDS-2016MC) that provides multiple realizations of the annual doses for each cohort member to obtain both a central estimate of the individual dose and information on the dose uncertainty.