Lee and Etzioni describe our best-guess estimate of background breast cancer incidence in the United States over three decades as prone to very small changes in the counterfactual incidence trend that would have significantly changed our result. They state that our final calendar year of estimate, 2008, illustrates how extrapolation of data over a long period of time can undermine the reliability of estimates of overdiagnosis based on excess incidence (EI).
However, our estimate was not based on a single year, but on the last 3 years of Surveillance, Epidemiology and End Results (SEER) data available at the time of our analysis, 2006–2008 . Secondly and more importantly, because the incidence rates of early- and late-stage breast cancer have been relatively stable since 2003 in the screened population of women over 40 years of age ( Fig 1 , based on SEER 13 regions, expanded from the 9 regions used for our prior report), the year(s) chosen since 2003 to compare to the prescreening era of 1976–1978 would lead to little variation in the estimated rate of breast cancer overdiagnosis.
Open full size image
Figure 1
Annual incidence of breast cancer in women over 40 years of age, 2003–2013, by early (ductal carcinoma in situ + localized) and late (regional + distant) stage at diagnosis, SEER 13 regions. Regressions are 2 o polynomials.
Moreover, because 5 more years of SEER data (2009–2013) have become available since our analysis, we can check the EI concern of Lee and Etzioni over an 11-year span since 2003, and not just 2006–2008. Our best-guess overdiagnosis rate estimate shows no trend toward an increase or decrease since 2003 ( Fig 2 , gray circles and zone) and little variability, with a mean rate during 2006–2013 of 30.2% (range, 28.8%–31.9%; 95% confidence interval, 28.3%–32.1%). The update also indicates that the annual number of women in the United States who have been overdiagnosed with breast cancer has continued to increase since our report but is slowing down as the best-guess estimate approaches 80,000 women per year ( Fig 2 , black triangles).
Open full size image
Figure 2
Variation in overdiagnosis estimates according to the calendar year since 2003, with 1976–1978 as prescreening era. Regressions are 2 o polynomials.
Our best-guess model assumed a background increase in breast cancer incidence of 0.25%/year rate that was based on the unscreened population under 40 years of age in the initial 9 SEER regions that represented 9% of the US population. The additional years available in SEER and its expanded population in 2000 to 18 regions representing 28% of the United States allowed us to check the validity of the assumption with greater accuracy. Figure 3 shows the incidence of breast cancer in the expanded unscreened population, in whom the rate increase during 2000–2013 was 0.08%/year. Thus, our assumption of the 0.25% background increase in older women of screening age was, if anything, conservative, and not a gross underestimate as criticized by others .
Open full size image
Figure 3
Annual incidence of all breast cancer (ductal carcinoma in situ, localized, regional, distant, unstaged) in women under 40 years of age, 2000–2013, SEER 18 regions. Regression is linear.
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Open full size image
Figure 4
Annual incidence of metastatic breast cancer (distant disease*) in women over 40 years of age, 1975–2013, SEER 9 regions. *SEER Historic Stage A classification.
Get Radiology Tree app to read full this article<
References
1. Lee C.I., Etzioni R.: Missteps in current estimates of cancer overdiagnosis. Acad Radiol 2017; 24: pp. 226-229.
2. Bleyer A., Welch H.G.: Effect of three decades of screening mammography on breast-cancer incidence. N Engl J Med 2012; 367: pp. 1998-2005.
3. Kopans D.B.: Arguments against mammography screening continue to be based on faulty science. Oncologist 2013; 19: pp. 107-112.
4. Harding C., Pompei F., Burmistrov D., et. al.: Breast cancer screening, incidence, and mortality across US counties. JAMA Intern Med 2015; 175: pp. 1483-1489. Published online July 6, 2015
5. Welch H.G., Prorok P.C., O’Malley A.J., et. al.: Breast-cancer tumor size, overdiagnosis, and mammography screening effectiveness. N Engl J Med 2016; 375: pp. 1438-1447.