Differential Benign Breast Disease Co-Occurrence with Cancer in Caucasian and African American Women.

Abstract

Abstract Background: Breast Cancers (BCs) in Caucasian (CA) and African American (AA) women have different characteristics. Recently, there have been reports, mostly focusing on Caucasians that Benign Breast Diseases (BBDs) may be risk factors for BCs. There are also a few reports of different co-occurring patterns of BBDs with BCs between the two populations. In the Clinical Breast Care Project (CBCP) more than 4,000 subjects have been enrolled following HIPAA-compliant, IRB-approved protocols, with more than 35,000 biospecimens collected. For patients requiring biopsies, an expanded pathology report for research is completed. The occurrences of any of the 131 pathologic conditions, including 66 BBDs, are recorded. All the results are reviewed by the same pathologist. The CBCP provides a uniformed pathology dataset for a comprehensive characterization of the association of BBDs with BCs.Method: The diagnoses were made from potentially multiple biopsies obtained over a short period of time, mostly within 60 days. CBCP subjects with BBD diagnosis and ethnicity information from their most recent visit were selected, giving a dataset totaling 1479 CA and 484 AA women. We studied the association between BBDs and BCs (including in situ, invasive, and malignant NOS) in these two groups using the chi-square test.Results: In both populations 6 BBDs are positively associated with BCs, including atypical ductal hyperplasia (ADH) (4% vs 15% for AA p<3.8E-05, meaning 4% cancer-free cases with ADH compared to 15% cancer cases co-occurring with it; 7% vs 16% for CA p<1.9E-08), microcalcifications (27% vs 56% for AA p<1.2E-09, 35% vs 53% for CA p<8.0E-12), multiple (peripheral) papillomas (6% vs 16% for AA p<0.0005, 4% vs 8% for CA p<0.0002), columnar cell hyperplasia (9% vs 21% for AA p<0.0003, 6% vs 18% for CA p<3.5E-12), and moderate intraductal hyperplasia (IDH) (14% vs 28% for AA p<0.0002, 14% vs 18% for CA p<0.03). On the contrary, 3 BBDs are negatively associated with BCs including fibroadenoma (30% vs 10% for AA p<4.9E-07; 21% vs 7% for CA p<1.0E-13), and mild IDH (10% vs 2% for AA p<0.003; 13% vs 7% for CA p<0.0002). Eight BBDs differentially co-occur with BCs between AA and CA. BCs in AA are associated with sclerosing adenosis (13% vs 31%, p<4.5E-06), fibrocystic changes (45% vs 61%, p<0.001), fibroadenomatoid nodule (6% vs 12%, p<0.042), and cysts (42% vs 57%, p<0.004). BCs in CA are positively associated with columnar cell hyperplasia with atypia (3% vs 8% p<3.6E-05), atypical lobular hyperplasia (1% vs 5%, p<6.0E-05), and radiation changes (0% vs 1.4% p<0.004), but negatively associated with duct ectasia (10% vs 2%, p<7.9E-10).Discussion Our findings are in accordance with established BC risk factors such as ADH and moderate IDH for both ethnic groups. We further found that 8 BBDs differentially co-occur with BCs between AA and CA groups, and interestingly the 4 significant BBDs detected in the AA group typically show >10% increased co-occurrence rate in cancer cases compared to cancer-free cases, whereas the 4 BBDs detected in the CA group almost always show a <10% co-occurrence rate overall. More research is undergoing to understand the implications of this distinct co-occurrence pattern. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 3066.