I am deeply concerned about the recently published article by Dr. Archie Bleyer regarding screening mammography and the question of overdiagnosis . Dr. Bleyer offers a review of the literature and update of his prior work, but he has not updated his puzzling approach and conclusions. My concerns are many, and I will address only some of them here.
Dr. Bleyer’s review relies on modeling based on the author’s “best guess” about fundamental inputs such as the likely incidence rate of breast cancer in the absence of screening . The author estimated that the background growth in breast cancer incidence without screening would be 0.25% per year. Through careful analysis of multiple databases including the Surveillance, Epidemiology, and End Results database, the Connecticut Tumor Registry, and the United Kingdom female cancer registry data, Helvie et al. established that the annual percentage change in breast cancer incidence ranged from at least 0.8% up to 2.3% with a central estimate of 1.3% growth in breast cancer incidence per year in the absence of screening. This matches well with the median of seven models from the Cancer Intervention and Surveillance Modeling Network, which sets the increase in background incidence at 1.2% .
Although disagreements over assumptions may appear banal, in fact, accurate estimation of the annual percentage change in breast cancer incidence is crucial when attempting to estimate overdiagnosis. Dr. Bleyer’s marked underestimation of the upward trend in the incidence of breast cancer leads him to severely overestimate the difference between early- and late-stage cancers. As a result, he markedly overestimates overdiagnosis. Since the actual background incidence is significantly higher—just as Helvie et al. and the Cancer Intervention and Surveillance Modeling Network models predict—Dr. Bleyer assumes that the difference between reality and his best guess must be due to cancers that would only be detected by screening (i.e., overdiagnosis). Instead, the difference is due, in large part, to changes in the underlying risk of breast cancer in the population.
In addition to perpetuating his flawed modeling in this review, Dr. Bleyer also perpetuates several other discredited notions. Most notably, he rejects randomized controlled trials that show a significant decrease in breast cancer mortality because they do not also show a statistically significant decrease in all-cause mortality. This is simply a matter of resources. Tabár et al. demonstrated that to have sufficient power to demonstrate an improvement in all-cause mortality, a trial would require over 20 times as many subjects as the largest randomized controlled trial to date . Instead, Tabár et al. demonstrate that, among all women who develop breast cancer, there was a statistically significant 19% reduction in all-cause mortality for women who were invited to screening versus those in the control group.
Finally, Dr. Bleyer suggests that many breast cancers are not real and that, despite histologic evidence, we do not know “who really has cancer.” He again suggests that many invasive breast cancers develop and then regress spontaneously, although there are only a small number of anecdotal accounts of spontaneous regression . Also, he strongly recommends that screening regimens be matched to a patient’s personal risk profile, a strategy that sounds logical on the face of it but for which there are also no empirical data demonstrating that such a strategy would be effective. There are existing data that definitively demonstrate that screening mammography significantly reduces mortality due to breast cancer in women exposed to screening and reduces all-cause mortality among women who develop breast cancer. I would urge Dr. Bleyer and others who are particularly focused on the possibility of overdiagnosis to test alternatives that will save as many lives as the screening strategies that have already been proven.
References
1. Bleyer A.: Screening mammography: update and review of publications since our report in the New England Journal of Medicine on the magnitude of the problem in the United States. Acad Radiol 2015; 22: pp. 949-960.
2. Bleyer A., Welch G.: Effect of three decades of screening mammography on breast-cancer incidence. NEJM 2012; 367: pp. 1998-2005.
3. Helvie M.A., Chang J.T., Hendrick E., et. al.: Reduction in late-stage breast cancer incidence in the mammography era. Cancer 2014; 120: pp. 2649-2656.
4. Cronin K.A., Feuer E.J., Clarke L.D., et. al.: Impact of adjuvant therapy and mammography on U.S. mortality from 1975 to 2000: comparison of mortality results from the CISNET breast cancer base case analysis. J Natl Cancer Inst Mongr 2006; pp. 112-121.
5. Tabár L., Duffy S.W., Yen M.-F., et. al.: All-cause mortality among breast cancer patients in a screening trial: support for breast cancer mortality as an end point. J Med Screen 2002; 9: pp. 159-162.
6. Larsen S.U., Rose C.: [Spontaneous remission of breast cancer. A literature review]. Ugeskr Laeger 1999; 161: pp. 4001-4004.