Abstract P3-07-08: Don't mix apples and oranges: Not all non-screen-detected breast cancers belong to the same category in a cancer registry dataset

Background: In population-based datasets breast cancers are often classified as screen-detected or not screen-detected (or symptoms-detected) without further information, resulting in substantial heterogeneity in the latter group. In fact, this group includes symptoms-detected and fortuitously detected cancers,as well as ‘interval breast cancers’ (i.e. detected between screenings), all of which have different morphological characteristics than screen-detected cancers. Objective: To compare the demographic features, tumour characteristics and survival of women with not-screen detected breast cancer according to whether or not they reported previous mammography screening. Methods: The study population comprised1696 women aged 50-69 years oldwith non-screen detected invasive breast cancer recorded at the Geneva cancer registry between 1990 and 2007. According to the information on the variable "previous mammography screening" the women were classified as those never having had a screening test and those having had one. We comparedtumour characteristics and prognostic factors through chi square test, and calculated 5-year breast cancer specific survival and multivariate Cox proportional hazard models to assess independent determinants of mortality. Results:Among the non-screen detected breast cancers there were 904 womenwho never had a screening mammography and 792 who declared to have had at least one. Women previously screened were more often 55-59 years old, diagnosed after 1996, of high social class, treated in the private sector and had a positive family history. These women had smaller tumours (mean 21.6 mm vs. 27.2, p Five year breast cancer-specific survival was 83% among never screened women (95% Confidence Intervals [95%CI] 81-86) and 91% among those previously screened (95% CI: 89-93; log rank test=20.62, p Conclusions:Within the category of non-screen detected breast cancers co-exist at least two distinct groups of patients, with different tumours and, potentially, treatments and prognoses. Comparing women with screen-detected breast cancer against this heterogeneous group of women could produce biased results. Recording information about screening use before the breast cancer diagnosis would help to correctly classify these women. Citation Format: Elisabetta Rapiti, Massimo Usel, Robin Schaffar, Hyma Schubert, Christine Bouchardy. Don9t mix apples and oranges: Not all non-screen-detected breast cancers belong to the same category in a cancer registry dataset [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 P3-07-08.