Abstract PR-12: Is the screening-related increase of breast cancer mainly caused by lesions that would undergo spontaneous regression if left untreated? Free

In Norway mammography screening is associated with a 50% long-term increase of the incidence of invasive breast cancer. Based on an analysis of incidence data from the Norwegian screening program, we have recently argued that the majority of these extra cancers would have undergone spontaneous regression in the absence of screening (Zahl PH, Mæhlen J, Welch HG. The natural history of breast cancer detected by screening mammography. Archives of Internal Medicine 2008; Nov 24th). At mammography a large number of ductal carcinoma in-situ (DCIS) are detected and treated as low grade cancer. The objective of the present study is to repeat this analysis on a data set containing both invasive cancer and DCIS.

The data from the Norwegian Cancer Registry include invasive cancer and DCIS from the last 16 years for women of four counties in which the age group 50-69 years (n =166 000) was been invited to biennial mammography screening from 1996. We compared the 6-year cumulative incidence in two age matched and partly overlapping cohorts. The first cohort (the test group) included women born 1932-46 invited biennially during the 6-year observation period 1996-2001. The second cohort (the control group) included women born 1928-42 and invited once at the end of the 6-year observation period 1992-97. Note that for both cohorts the age span was 50-64 years at the start of the respective observation periods.

The 6-year cumulative incidence of invasive cancer and DCIS combined was 28 % higher in the first cohort than in the second cohort. (2206 vs. 1739 per 100 000; RR = 1.27 (95% CI: 1.20 - 1.34). A sensitivity analysis of the potential effect of HRT showed that at most 2% of the 27% difference could be explained by differences in the HRT use. We also tested if low sensitivity could explain the difference - it could not explain the difference (P<0.001).

We have shown that the screening related incidence increase in the first four years of the 6-year observation period in the first cohort is not compensated for by a corresponding incidence increase when the previously un-screened women in the second cohort were invited in year 5-6. Increasing use of HRT and low sensitivity in the screening did not explain the difference. We propose that most of the screening-related incidence increase of invasive cancers and DCIS is due to tumors that in the absence of screening would have undergone spontaneous regression.

Citation Information: Cancer Prev Res 2008;1(7 Suppl):PR-12.