[en] In Japan, a cohort study of radiation workers has been conducted since 1990. The cohort population consisted of about 176,000 workers (mostly males) who had been registered in the centralized radiation dose registry system and engaged in various radiation works at nuclear facilities. Statistical analyses were performed mainly on the 2,934 deaths, of which 1,191 were cancer cases, detected among 119,000 male subjects during the prospective follow-up. The standardized mortality ratio showed that for any cancers mortality was not different between this population and Japanese general population. By the trend test, though significantly increasing trend in accord with increasing doses was not observed for both cancer in all sites and leukemia, it was highly significant for esophagus cancer and external causes of deaths. Results of the questionnaire survey study of lifestyle of radiation workers suggested that increasing trend of these diseases was at least partly due to the influence of some confounding factors. As a result of reviewing published studies, including the present work, trend of mortality from cancer in all sites with increasing doses seems still unclear, whereas for leukemia it appears to stay flat under 100 mSv but rapidly rise up in the doses higher than this as if fitting to either a linear-quadratic or threshold models. (author)