Background: Breast cancer mortality has been declining in many western countries, including Switzerland,since the late 1980s. This has largely been credited to mammography screening and improved treatment. However, mortality trends are also decreasing among unscreened women, though most breast cancer deaths still occur in that population.
Objective: The objective of this study was to analyse trends in,and factors affecting the survival of, women whose breast cancer was not detected by screening in the Geneva female population from 1990 to 2007. In Geneva an organized screening programme started in 2001 while opportunistic screening has existed since the beginning of the 1990s.
Methods: The study population comprised1696 women aged 50-69 years oldwith invasive breast cancer that was not detected by screening and that was recorded at the population-based Geneva cancer registry. We studied tumour characteristics and prognostic factors across 6 time periods through chi square and trend tests. To assess whether breast cancer specific survival had improved over time, we calculated 5-year specific survival and performed multivariate Cox proportional hazard models to assess independent determinants of mortality.
Results:Median age of the women at diagnosis was 59 years. During the 18 year study periodthere was a decrease in the proportion of diagnoses among women of low social class (25.8% in 1990-92 vs 17.3% in 2005-07, p=0.001). No change in the distribution of stage at diagnosisor hormone receptor status was observed, while between the first and last period there was an increase in cancers with lobular morphology (6.8% vs. 19.0%, p<0.001), and a decrease in poorly or undifferentiated cancers (36.7% vs 28.7%, p<0.001). The use of breast conserving surgery with radiotherapy (BCS), chemotherapy and endocrine therapy increased significantly along the period (41.2% to 58.8%, p<0.001; 41.6% to 52.8, p=0.030; and 34% to 69.9%, p<0.001, respectively).
Five year breast cancer-specific survival was 82% in 1990-92 (95% Confidence Intervals [95%CI] 77-87) and 89% in 2005-07 (95% CI: 83-92; log rank test=8.68, p=0.122). In the Cox multivariate model there was a trend towards improved survival, but it was only statistically significant when comparing the period 2005-07 to the first period (Hazard Ratio 0.47, 95%CI: 0.22-0.98). Increasing age, stage, and grade, hormone receptor–negative disease, and not receiving BCS or endocrine therapy were all independently associated with a worse breast cancer–specific survival.
Conclusions:We observed an improvement in survival only in recent years among women whose breast cancers were not detected by screening; this appears to be associated with improved treatment. This suggests that the breast cancer mortality reduction observed in Switzerland since the late 1980s is not likely attributable to changes in treatment before 2005, but rather to the generalization of screening.
Citation Format: Elisabetta Rapiti, Thomas Agoritsas, Massimo Usel, Robin Schaffar, Hyma Schubert, Christine Bouchardy. Trends and determinants of breast cancer survival among unscreened women: A population-based study [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P3-07-09.