Biological dosimetry of chronic radiation exposure in interventional radiologists using dicentric assay and a Bayesian framework.

Cytogenetic studies in populations chronically exposed to mutagens are challenging due to a number of methodological constraints. The aim of the study was to demonstrate the applicability of dicentric assay, combined with advanced data interpretation, for radiation biodosimetry in chronically exposed medical staff. Dicentric frequencies were measured in cultured blood lymphocytes from 12 interventional radiologists with 5-42 years of professional exposure and 14 unexposed controls. The data processing methodology included a mathematical "unfolding" of observed aberration yields, a robust linear dose-response coefficient derived from in vitro low-dose calibration curve for dicentrics, and a Bayesian-type statistical framework for calculating posterior probability densities for "true" aberration yields and corresponding dose estimates. The mean dicentric frequency in radiologists significantly exceeded the control level, but individual yields showed no dependence on the years of service (r = 0.268). After "unfolding", a strong positive correlation emerged (r = 0.843). "Unfolded" dicentric yields were converted into cumulative dose estimates of 49 - 595 mGy. Bayesian posterior probability densities of "true" aberration yields corresponded to the most probable annual doses of 3.7 - 27.9 mGy. Confidence intervals of annual doses, which are "more probable to be true than non-true", ranged from 0.8 to 39.3 mGy, and their upper limits exceeded 20 mGy in eleven cases out of twelve. The proposed approach substantially enhances the accuracy and interpretative power of cytogenetic biodosimetry of chronic irradiation. The resulting dose estimates can be used for assessment of radiation-associated health risks and support individual radiation protection measures for occupationally exposed individuals.