Technical Report No. 7-88
Comparison of the dose-response relationships for chromosome aberration frequencies between the T65D and DS86 dosimetries
Preston DL, McConney ME, Awa AA, Ohtaki K, Itoh M, Honda TEditor’s note: A publication based on this report was published in New Dosimetry at Hiroshima and Nagasaki and Its Implications for Risk Estimates. Bethesda, Md, USA, NCRP, 1988. pp 185-202. (Proceedings of the 23rd Annual Meeting of National Council on Radiation Protection and Measurements, Washington DC,1987)
Summary
Cytogenetic data, derived from cultured lymphocytes of atomic bomb survivors and controls in the ABCC-RERF Adult Health Study cohort, have been analyzed to determine differences in the dose-response relationships for chromosome aberrations between the T65D and DS86 dose estimates and to assess differences between Hiroshima and Nagasaki. The data consist of blood samples collected between 1968 and 1980 from 1,245 individuals (788 from Hiroshima and 457 from Nagasaki), who have a DS86 kerma estimate of less than 4 Gy (678 in Hiroshima and 381 in Nagasaki) or who were not in the cities at the time of the bombings (110 in Hiroshima and 76 in Nagasaki). For each person, the number of cells with at least one structural chromosome aberration was related to dose using binomial regression models. The variability of the observed proportion of aberrant cells as a function of estimated dose is much greater than one would expect under a binomial model, consequently all models have been adjusted for extra-binomial variability.
For a linear dose-response model, the average percentage of cells with at least one chromosome aberration increases less rapidly with dose in Nagasaki than in Hiroshima. The magnitude of the intercity difference in the percentage of cells with aberrations per gray is less for DS86 than for T65D, though the difference is statistically significant for both kerma and bone marrow dose with either dosimetry. The percentage of cells with aberrations per gray for DS86 kerma estimates is about 60% greater than the corresponding T65D slope. For marrow dose, however, there is very little change in slopes between the two dosimetries. Analyses to test nonlinearity in the dose-response function indicate significant departures (p < 0.001) from linearity, using both dosimetries for both kerma and marrow dose. Comparison of aberration dose-response functions for the old and new dosimetries depends heavily upon the relative biological effectiveness (RBE) of neutrons; however, it is difficult to estimate separate gamma ray and neutron response functions and, hence, RBE from these data. Therefore, comparative results are presented for a range of RBE relationships under various linear (L) and linear-quadratic linear (LQ-L) models. As an illustrative result, if one assumes an LQ-L model similar to models reported in the cytogenetic literature, with a limiting RBE of 20 at zero dose, the DS86 slope (the percentage of cells with aberrations per sievert) is 120% greater than the corresponding T65D value.
Cytogenetic data, derived from cultured lymphocytes of atomic bomb survivors and controls in the ABCC-RERF Adult Health Study cohort, have been analyzed to determine differences in the dose-response relationships for chromosome aberrations between the T65D and DS86 dose estimates and to assess differences between Hiroshima and Nagasaki. The data consist of blood samples collected between 1968 and 1980 from 1,245 individuals (788 from Hiroshima and 457 from Nagasaki), who have a DS86 kerma estimate of less than 4 Gy (678 in Hiroshima and 381 in Nagasaki) or who were not in the cities at the time of the bombings (110 in Hiroshima and 76 in Nagasaki). For each person, the number of cells with at least one structural chromosome aberration was related to dose using binomial regression models. The variability of the observed proportion of aberrant cells as a function of estimated dose is much greater than one would expect under a binomial model, consequently all models have been adjusted for extra-binomial variability.
For a linear dose-response model, the average percentage of cells with at least one chromosome aberration increases less rapidly with dose in Nagasaki than in Hiroshima. The magnitude of the intercity difference in the percentage of cells with aberrations per gray is less for DS86 than for T65D, though the difference is statistically significant for both kerma and bone marrow dose with either dosimetry. The percentage of cells with aberrations per gray for DS86 kerma estimates is about 60% greater than the corresponding T65D slope. For marrow dose, however, there is very little change in slopes between the two dosimetries. Analyses to test nonlinearity in the dose-response function indicate significant departures (p < 0.001) from linearity, using both dosimetries for both kerma and marrow dose. Comparison of aberration dose-response functions for the old and new dosimetries depends heavily upon the relative biological effectiveness (RBE) of neutrons; however, it is difficult to estimate separate gamma ray and neutron response functions and, hence, RBE from these data. Therefore, comparative results are presented for a range of RBE relationships under various linear (L) and linear-quadratic linear (LQ-L) models. As an illustrative result, if one assumes an LQ-L model similar to models reported in the cytogenetic literature, with a limiting RBE of 20 at zero dose, the DS86 slope (the percentage of cells with aberrations per sievert) is 120% greater than the corresponding T65D value.