Re: Effectiveness of and overdiagnosis from mammography screening in the Netherlands: population based study
As scientists we have an obligation, both to the scientific community and to the public, to interpret data responsibly. Ionizing radiation is a recognized carcinogen on which an enormous number of studies have been conducted. There is only indirect evidence that x rays at the dose levels used in mammography could induce breast cancer, but to be prudent risks are predicted by extrapolating downward from higher doses using the linear, no-threshold (LNT) model.
Daniel Corcos, in response to the article by Autier et al. (1), suggests that the excess cancers inferred by Autier’s analysis are radiation induced, citing his own yet unpublished work and that of others. Corcos’ work is not available for scrutiny, but his reference to the 2004 publication of Hooker et al. (2) does not support his suggestion. Indeed Hooker observed that departure (of mutations in a mouse model) from the LNT model for the dose range of mammography (a few milligrays) results in reduced risk, contradicting Corcos’ idea. In the same issue of Radiation Research Preston et al. (3) , re-examining atomic bomb dosimetry observed that estimated risks for radiation induction of solid tumours should be reduced by about 8% from previously accepted values.
In the Autier article itself, strong inferences are made, based on cancer data for which the means of detection was not available, as discussed in the response from Tabar. Much more relevant direct evidence is available from individual level data, where it is known who received screening and who did not and which cancers were detected by screening. Analyses of such data have demonstrated a clear mortality benefit from mammography screening.
Making such apparently definitive statements using incomplete data or incorrectly interpreting the results of others only serves to weaken the credibility of scientific research.
Martin J. Yaffe, PhD, C.M.
1. Autier P, Boniol M, Koechlin A, et al. Effectiveness of and overdiagnosis from mammography screening in the Netherlands: population based study. BMJ 2017;359:j5224
2. Hooker AM, Bhat M, Day TK, et al. The linear no-threshold model does not hold for low-dose ionizing radiation. Radiat Res 2004;162(4):447-52.
3. Preston DL, Pierce DA, Shimizu Y, Cullings HM, Fujita S, Funamoto S, Kodama K. Effect of Recent Changes in Atomic Bomb Survivor Dosimetry on Cancer Mortality Risk Estimates. Radiat. Res. 2004;162(4):377-389.
Competing interests: No competing interests