The results reported by JA Hanley and colleagues (1) are logically implausible. The study found a non-significant 9% (95% CI : 20% reduction to 4% increase) reduction in breast cancer mortality in half of Ireland where the mammography screening programme was introduced in 2000 (region 1) than in the other half of Ireland where the programme was introduced in 2007 (region 2). Despite the fact that this result was statistically non-significant, authors made a great deal of speculations by which they concluded that the real influence of screening on breast cancer mortality reductions is greater than 9%. For supporting their conclusions, authors show in Table 1 of the article data suggesting a decline in the incidence rates of stage 2-4 cancers in 2000-07 in region 1 while in region 2, incidence rates would have been increasing. Authors restricted their age-specific analyses according to time from invitation to screening but it is likely that in the absence of screening, similar trends would be observed in women in age groups not eligible for screening. In contrast with the reported results, the data collected by the Irish cancer registry shows that from 1994 to 2012, the incidence rates of stage 2-4 breast cancers has remained stable at around 80 cases per 100,000 women per year (2). If rates of stage 2-4 cancers in region 2 from 2000 to 2008 were actually on the increase, it means that rates before 2000 were increasing in all the country. The stability of rates from 1994 to 2000 does not agree with this hypothesis and the results displayed in Table 1 are probably fortuitous. In addition, there is no formal statistical testing of results displayed in Table 1 and Figure 4 of the article, which adds to the suspicion of fortuitous finding. Finally, results displayed in Figure 4 would reflect the finding of prevalent cases of stage 2-4 cancers in 2000-03 followed by a transient compensatory decrease in 2004-09 due to the advances in detection in 2000-03, with a progressive return to the stable incidence rates that prevailed from 1994 to 2012. Authors surmise that after 2009, the incidence rate differences between the two regions would vanish because of declines in the incidence of stage 2-4 cancers in region 2 where screening started in 2007. Again, this hypothesis is implausible because if true, the incidence rates of stage 2-4 cancers after 2007 reported by the Irish cancer registry should decrease, which never happened. We conclude that the study by Hanley and colleagues does not provide evidence that mammography screening can reduce the incidence rate of advanced breast cancers and the incidence of breast cancer deaths in populations.

1. Hanley JA, Hannigan A, O'Brien KM. Mortality reductions due to mammography screening: Contemporary population-based data. PLoS One 2017;12(12):e0188947.
2. National Cancer Registry Ireland. Breast Cancer. Cancer Trends;29(March 2016).



Prof. Philippe Autier, MD, MPH, PhD
University of Strathclyde Institute of Global Public Health at iPRI
International Prevention Research Institute (iPRI,www.i-pri.org)
Espace Européen d’Ecully, Bâtiment G
Allée Claude Debussy
69130 Ecully ouest Lyon
France
Philippe.autier@i-pri.org