Abstract

The current DS86 dosimetry system for the Japanese bomb survivors indicates that neutron doses were so low that they prevent the direct derivation of any useful estimates of neutron risk. However, the large body of thermal neutron activation measurements carried out over many years in Hiroshima and Nagasaki appear to indicate that current DS86 neutron doses may have been significantly underestimated in Hiroshima. An earlier companion paper has provided an update of neutron activation measurements. While a large body of data appears to support a significant increase, there is ongoing debate and review regarding its validity. However, as yet, there are no detailed, peer-reviewed, published refutations of the neutron activation data which appear to support an increase in neutron doses. In this paper, we consider the impact of possible future revisions in the DS86 dosimetry on radiation risk estimates. We consider the extreme range of possibilities from maintaining the existing DS86 values, to changes in neutron doses in accord with the majority of existing neutron activation data. We have used the latest cancer incidence data and cancer mortality data for the A-bomb survivors, and neutron doses have been modified using a neutron revision factor (NRF) in line with the latest thermal neutron activation measurements in Hiroshima. In contrast to previous analyses, a nonlinear relationship between log(NRF) and slant range has been used which better represents the data beyond slant ranges of ~1 km. The impact on the evaluation of neutron relative biological effectiveness (RBE) and gamma radiation risk estimates has been assessed. While DS86 neutron doses are too low to allow any useful direct evaluation of neutron risk or neutron RBE, it becomes possible to derive more meaningful values if neutron doses are increased in Hiroshima in line with the broad range of thermal neutron activation measurements. The uncertainties are smallest for the cancer incidence data. The best estimates of neutron RBE give upper 95% confidence limits of about 6 for all solid tumours for the incidence data and about 28 for the mortality data. The uncertainties in neutron RBE for leukaemia incidence are larger, and estimation at doses below about 0.1 Gy is not possible. There is no significant change in the excess relative risk for gamma radiation for all solid tumours taken together, compared with the current DS86 dosimetry. The results preclude neutron RBE values significantly greater than the current ICRP radiation weighting factors, which range between 5 and 20, depending on energy. Whether or not the Japanese bomb survivors can indeed form the basis for useful, directly determined neutron risks clearly depends on the veracity of existing neutron activation data. This is currently the subject of careful international scrutiny and the outcome is eagerly awaited.