Design and analysis of epidemiological studies of excess cancer among children exposed to Chernobyl radionuclides.

Within the last decade, a substantial amount of attention has been devoted to etiological research on the association between exposure to fallout radionuclides from the Chernobyl accident and radiation-induced late effects (cancer) among children. A majority of the studies completed to date have been of the descriptive type, which only correlate average population exposure with average rate of cancer incidence as a function of calendar period. Since individual dosimetry is not performed in descriptive studies, it is unclear whether exposure precedes the development of cancer and a final decision cannot be made regarding the association between radiation exposure and cancer. This paper reviews the background epidemiology and outlines an analytical study design that is needed to clarify the unclear association between Chernobyl fallout exposure and childhood cancer. We discuss the essential elements of an analytical case-control design such as genetic predisposition, vital statistics, sample size and power determinations, ascertainment of cases and controls, and phenomenological dose modeling to establish individual doses. Examples such as cytogenetic biodosimetry, medical radiation dosimetry, and cytogenetic characterization of leukemia to minimize exposure and diagnostic misclassification are provided. We recommend the analytical methods described in this paper for studying the role of Chernobyl radionuclides and development of childhood cancer.