The Question of Radiation Causation of Cancer in Hanford Workers

Abstract

An independent study of the Hanford worker death sample of Mancuso confirms the association of radiation, accumulated at rates below the permissible rate, with cancer induction (p < 0.05). The calculated doubling dose, for cancers overall (mean age at exposure of approx. 45 yr) is 43.5 rad, which is consistent with and confirmatory of the value 33.7 rad reported for cancers overall by Mancuso, Stewart and Kneale (in both instances these doubling doses are for males). However, the confidence limits on the doubling dose are quite broad, and hence the study can neither rule out a much lower doubling dose nor can it rule out a much higher doubling dose that would be consistent with those reported for persons receiving radiation at higher dose rates and with higher accumulated doses. This study does not concur with the conclusion of Mancuso et al., that certain tissues are unusually radio-sensitive for cancer induction. An alternative explanation is provided for the apparently low doubling doses for such entities as myeloma. It will require one or more series of comparable size to this Hanford death series to resolve the question of differential radiosensitivity of tissues. The trends suggest, in accord with earlier experience with higher total dose irradiations, that persons young at the time of radiation exposure show lower doubling doses than do persons older at the time of exposure. This differs from the conclusions of Mancuso et al. It is concluded that the Hanford data for females are too sparse to justify any conclusions concerning the relative sensitivity of males and females for radiation induced cancer. The largest part of the association of radiation with cancer in this study arises from those persons who received a cumulative dose of 10 rad or more in the course of employment. A much larger data base would be required to draw any conclusions concerning radiation carcinogenesis at doses below 10 rad, cumulative. Indeed, a much larger data base, of the same type as that provided by Mancuso, would aid materially in resolving several residual questions.