Mammography use, incidence, and survival of primary breast lymphoma.

Abstract

e13087 Background: Breast lymphoma is rare and, unlike in carcinoma, surgery is not the primary treatment modality, so early detection may not contribute to survival. Caon et al. ( Clin Breast Cancer 2012) reported that 85% of aggressive lymphomas are diagnosed by a palpable mass, while 56% of indolent subtypes, by mammography. Our objective was to compare incidence, clinical features, and survival of breast lymphomas according to prevalence of screening mammography. Methods: From the United States SEER registry, we selected primary breast lymphomas diagnosed in 2000-2013 among women ≥40 years old, distinguishing DLBCL, marginal zone (MZL), follicular (FL), and other histologies. Mammography use in county of residence was expressed as % of women ≥40 years old with a mammogram in past 2 years, estimated from the National Cancer Institute’s small-area models. We calculated incidence rates (IR) for each county, and compared them in a Poisson model (with 95% confidence intervals, CI). Head and neck (H&N) lymphomas were used as sensitivity controls. Overall survival (OS, from diagnosis) was compared in a Cox model adjusting for age, race, and median income in the county. Results: We identified 1662 cases of breast lymphoma, with a median age of 69 years. The most common subtypes were DLBCL (39%), MZL (26%), and FL (16%). Stage distribution was: 1 in 62%, 2 in 17%, and 3 or 4 in 21%. Incidence of breast lymphoma increased by 24% for every 10% increase in mammography use (IR ratio, 1.24, 95%CI, 1.15-1.34, P<.0001). This relationship was absent for controls with H&N lymphoma (IR ratio, 1.02, 95%CI 0.97-1.07, P=.43). In breast lymphoma, there was no difference in histologies ( P=.19), stage ( P=.83), or OS by mammography prevalence (hazard ratio 1.01 for every 10% increase, 95%CI, 0.87-1.18, P=.87). OS at 5 years was 60% for DLBCL, 77% for MZL, and 84% for FL. Conclusions: Prevalent mammography use is associated with increased incidence of breast lymphoma, without evident change in stage distribution or survival. Further research will be needed to investigate responsible causal pathways, or potential overdiagnosis. [Table: see text]