Cancer risk at low doses of ionizing radiation: artificial neural networks inference from atomic bomb survivors.

ABSTRACT: Cancer risk at low doses of ionizing radiation remains poorly defined because of ambiguity in the quantitative link to doses below 0.2 Sv in atomic bomb survivors in Hiroshima and Nagasaki arising from limitations in the statistical power and information available on overall radiation dose. To deal with these difficulties, a novel nonparametric statistics based on the 'integrate-and-fire' algorithm of artificial neural networks was developed and tested in cancer databases established by the Radiation Effects Research Foundation. The analysis revealed unique features at low doses that could not be accounted for by nominal exposure dose, including (i) the presence of a threshold that varied with organ, gender and age at exposure, and (ii) a small but significant bumping increase in cancer ri

SUBMITTER: Sasaki MS 

PROVIDER: S-EPMC4014156 | biostudies-literature | 2014 May

REPOSITORIES: biostudies-literature