In the commentary article “Mammogram Trials” published recently in Academic Radiology , the scientific evidence supporting screening mammography is questioned . Contrary to the article’s plural title, a single randomized control trial is presented, the Canadian National Breast Screening Study, which is the only one trial that failed to demonstrate a reduced mortality from breast cancer . Thus, the article’s discussion of screening mammography trials is far from balanced.
Multiple randomized control trials (New York Health Insurance Plan, Malmö I and II, Swedish Two-county [Kopparberg and Östergötland], Stockholm, Gothenburg, UK Age trial, Edinburgh) have been published and are reviewed by Smith et al. in Reference . To debate the utility of screening mammography without considering all of the data introduces considerable bias. Pooled analysis of eight randomized control trials that include the Canadian National Breast Screening Study, despite its methodological and image quality shortcomings, still demonstrates a 20% reduction in breast cancer mortality for women invited to screening mammography (95% confidence intervals 14–27%) . This figure is actually an underestimate of the benefit achieved in clinical practice because there are (1) women in the group invited to screening who chose not to undergo screening (termed noncompliance) and (2) women in the control group who chose to undergo screening (termed contamination).
Data from service screening studies, which reflect women who are actually screened not just invited to screening as in a randomized control trial, provide a closer estimate of the actual benefit of screening mammography when used in clinical practice. European and Canadian service screening programs report roughly a 40% reduction in breast cancer mortality in screened women compared to unscreened women . This figure is approximately double the benefit shown by randomized control trials.
I therefore disagree with Jha and Ware’s opinion that “the benefits (of screening mammography) are so small at a population level.” To put these numbers into perspective, analysis by Hendrick and Helvie published in the American Journal of Roentgenology showed that approximately 6500 additional women in the United States would die from breast cancer each year if those between 40 and 49 years go unscreened and those 50–74 years old are only screened biennially . As a physician scientist, I am deeply distressed by the implication that a single study yielding negative results could be used to influence healthcare policy regarding screening mammography at the disregard of the majority of data demonstrating a clinical benefit. Thousands of women’s lives should not be put at risk because of data “cherry picking.”
References
1. Jha S., Ware J.B.: Mammogram trials. Acad Radiol 2015; 22: pp. 973-975.
2. Miller A.B., Wall C., Baines C.J., et. al.: Twenty five year follow-up for breast cancer incidence and mortality for of the Canadian National Breast Screening Study: a randomized screening trial. BMJ 2014; 348: pp. g366.
3. Smith R.A., Duffy S.W., Gabe R., et. al.: The randomized trials of breast cancer screening: what have we learned?. Radiol Clin North Am 2004; 42: pp. 793-806. v
4. Broeders M., Moss S., Nystrom L., et. al.: The impact of mammographic screening on breast cancer mortality in Europe: a review of observational studies. J Med Screen 2012; 19: pp. 14-25.
5. Coldman A., Phillips N., Wilson C., et. al.: Pan-Canadian study of mammography screening and mortality from breast cancer. J Natl Cancer Inst 2014; 106: pii: dju261
6. Hendrick R.E., Helvie M.A.: United States Preventive Services Task Force screening mammography recommendations: science ignored. AJR Am J Roentgenol 2011; 196: pp. W112-W116.