Purpose : To measure the mutagenic effectiveness of low-filtered 30 kVp X-rays, mammography X-rays and conventional (200 kVp) X-rays in mammalian cells. Materials and methods : Two different cell lines and mutation assays were used. Exponentially growing SV40-transformed human fibroblasts were exposed to graded doses of mammography (29 kVp, tungsten anode, 50 μ m Rh filter) or conventional X-rays and the frequency of 6-thioguanine-resistent HPRT-deficient mutants was determined. Exponentially growing hamster A L cells, which contain a single human chromosome 11 conferring the expression of the human surface protein CD59, were subjected to magnetic cell separation (MACS) in order to remove spontaneous mutants before irradiation with low-filtered 30 kVp (tungsten anode, 0.5mm Al filter) or conventional X-rays. Fractions of radiation-induced CD59 - mutants were quantified by flow-cytometry after immunofluorescence labelling of CD59 proteins. Results : Mammography X-rays were more effective than conventional X-rays at inducing killing of human fibroblasts, whereas 30 kVp X-rays and conventional X-rays were about equally effective at killing A L cells. Mutant frequencies were linearly related to dose in both mutation assays. An RBE = 2.7 was calculated for the yield of HPRT mutants in human fibroblasts exposed to mammography relative to conventional X-rays and an RBE = 2.4 was obtained for the CD59 mutant frequency in A L cells irradiated with low-filtered 30 kVp relative to conventional X-rays. Conclusions : Both low-filtered 30 kVp and mammography X-rays are mutagenic in mammalian cells in vitro. It is unknown if and how the enhanced mutagenicity of mammography X-rays measured in human cells in vitro translates into breast cancer risk for predisposed women with an enhanced inherited risk for breast cancer. Although the ICRP guidelines attribute the same relative biological effectiveness to all radiations of low LET, including X- and γ-radiations of all energies for radiobiological protection purposes including the assessment of risks in general terms, they also state that 'for the estimation of the likely consequences of an exposure of a known population, it will sometimes be better to use absorbed dose and specific data relating to the relative biological effectiveness of the radiations concerned and the probability coefficients relating to the exposed population' (ICRP 1991: §32). This latter statement may apply for the population of familial predisposed women. We hope that the presented data on the enhanced mutagenicity of mammography X-rays may stimulate a re-evaluation of the risk assessment of mammography for familial predisposed women. In the meantime, one should be cautious and avoid early and frequent mammography exposure of predisposed women. Alternative examination methods should be applied for these women with an inherited increased risk for breast cancer.

Mutation induction in the HPRT gene of human fibroblasts after irradiation with mammography-like 29 kVp or 200 kVp x-rays shows radiohypersensitivity for doses smaller than approximately 0.5 Gy. Similarly, mutation induction in the CD 59 gene on human chromosome 11 in A(L) cells shows radiohypersensitivity for doses smaller than approximately 0.5 Gy after exposure to 200 kVp x-rays, but not after irradiation with low-filtered 30 kVp x-rays. The RBE values of 29 and 30 kVp x-rays relative to 200 kVp x-rays are strongly dose dependent. For neoplastic transformation of human hybrid (CGL1) cells after irradiation with 29 or 200 kVp x-rays or 60Co gamma rays a linear-quadratic dose relationship was observed with RBE values of approximately four and eight for mammography relative to 200 kVp x-rays and 60Co gamma rays, respectively.