Mammographic screening before age 50 years in the UK: comparison of the radiation risks with the mortality benefits.
Berrington de González A., Reeves G.
Mammographic screening before age 50 years is less effective than at older ages and the associated radiation risks are higher. We estimated how many breast cancer deaths could be caused and how many could be prevented by a decade of annual two-view mammographic screening starting at ages 20, 30 and 40 years, respectively, in the UK; for all women, and for women with first-degree relatives affected with breast cancer. We extrapolated from a radiation risk model to estimate the number of radiation-induced breast cancer deaths, and used results from randomised trials, which suggest a reduction in breast cancer mortality of 10-20% in women invited to screening before age 50 years, to estimate the number of deaths that could be prevented. The net change in breast cancer deaths was defined as the number of radiation-induced deaths minus the number of prevented deaths. For all women, assuming a reduction in mortality from screening of 20%, a decade of annual screening was estimated to induce more deaths than it prevents if started at age 20 years and at age 30 years (net increase = 0.86 and 0.37 breast cancer deaths, respectively, per 1000 women screened). The corresponding estimate for screening starting at age 40 years was a net decrease of 0.46 deaths/1000 women screened and a zero net change assuming a 10% mortality reduction. Results for women with first-degree relatives with breast cancer were generally in the same direction but, because their background incidence rates are higher, the net increases or decreases were greater. In conclusion, our estimates suggest that a decade of annual two-view mammographic screening before age 40 years would result in a net increase in breast cancer deaths, and that starting at age 40 years could result in a material net decrease only if breast cancer mortality is reduced by about 20% or more in women screened. Although these calculations were based on a number of uncertain parameters, in general, the conclusions were not altered when these parameters were varied within a feasible range.