Classification of chronic radiation sickness cases using neural networks and classification trees.

Chronic radiation sickness is a deterministic radiation health effect observed among the Mayak Production Association workers in Russia. In this study, unsupervised neural networks were used to cluster hematological measurements in a subset (n = 88) of the Mayak Production Association population while excluding from the analysis the radiation dose and the historical clinical diagnosis. Clusters of observations that had lower average leukocyte and thrombocyte counts were labeled "affected" and those having higher average blood cell counts were labeled "unaffected." The class (cluster) membership for each individual was used subsequently as a dependent variable in a classification tree model in order to identify significant features of the underlying classification model. After re-classification of cases using this method, the results showed a better data separation between the blood cell counts for affected vs. unaffected groups compared to those based on historical classification, and a greater difference between group means for differential blood counts was observed than for the historical diagnosis. The reclassification of diagnostic groups changed the group mean radiation doses. The geometric means (and 95% CL) of cumulative radiation dose equivalent from external exposures, based on the historical diagnosis, are 0.31 (0.0035, 3.4) vs. 1.7 (0.0007, 18) Sv. After clustering and classification tree analyses, the group geometric means were 0.78 (0.0014, 8.6) vs. 1.5 (0.0007, 17) and 0.82 (0.0013, 9.0) vs. 1.4 (0.0008, 16) Sv, using (respectively) whole blood cell counts or differential counts as the independent variables. The approach presented here is useful as a diagnostic aid for both retrospective analyses and in the event of future radiation accidents.

Deterministic effects from occupational radiation exposures in a cohort of Mayak PA workers: data base description.

Project 2.3 of the Joint Coordinating Committee on Radiation Effects Research (JCCRER) is a study of deterministic health effects among a cohort of Russia nuclear workers. The preliminary study population includes a stratified random sample of 221 radiation workers who were employed in a cohort of 8,055 workers at the Mayak PA facilities for at least one year during the period from 1948 to 1958. High annual doses, approaching 1 Gy per year from external and internal radiation sources, were reported for a significant proportion of the workers in this cohort. The present data set includes 96 cases of chronic radiation sickness (CRS), 14 cases of acute radiation syndrome (ARS) and 13 cases of plutonium pneumosclerosis (PPn). The remainder of the sample consists of "uninjured workers" who had no known history of radiation illness or injury; however, the uninjured workers are not "controls" for radiation exposure. The data base is currently being expanded to 600 individuals sampled from the cohort of workers from 1948 to 1958 to allow a more complete analysis of the deterministic health effects and comparisons with existing health effect models. The final data base will be used with state-of-the-art modeling techniques to determine threshold doses and dose-response relationships for key clinical diagnostic variables.