Affiliation: Gillings School of Global Public Health, Department of Epidemiology
Purpose: Symptomatic cancers generally have poor prognosis compared to screen-detected cancers and likelihood of screen detection may vary as a function of biological subtype or imaging characteristics of the breast cancer. The aims of this study were to study the association between breast cancer subtype and 1) mode of detection and 2) radiologic/ imaging features. Methods: In the first aim, we identified 1497 women diagnosed with primary invasive breast cancer from a linked data set between the Carolina Breast Cancer Study and the Carolina Mammography Registry. Among recently-screened (within 24 months) women (n=370, 25%), 45% of cancers were screen-detected (N=165), and 55% were interval-detected (N=205). Interval cancer was evaluated in association with clinical and genomic characteristics. In the second aim, 412 women with mammograms within 2 years before to 30 days after diagnosis were identified and associations between subtype and radiologic features were assessed.
Results: Interval cancer was associated with large tumors (>2 cm) (OR=2.3; 95% C.I.: 1.5, 3.7), positive nodal status (OR=1.8; 95% C.I.: 1.1, 2.8), and triple negative cancer (OR=2.5; 95% C.I.: 1.1, 5.5). Associations between interval detection and genomic characteristics were strong, and suggested that the vast majority of screen-detected cancers were indolent (96% were low risk of recurrence; 71% were Luminal A). Both young (<50) and African-American women showed higher relative frequency of masses and lower frequency of calcifications compared to older (≥ 50) and White women. Masses were less frequent among interval-detected vs. screen-detected women (46% vs. 33%, p=0.04). Relative to Luminal A breast cancers (42% presenting as masses), PAM50 Basal-like and HER2-enriched subtypes were more likely to present as masses (59% and 72%, respectively). Few Basal-like and ROR-PT high cancers presented with calcifications (n=4/49 Basal-like and n=3/30 ROR-PT high).
Conclusions: Underlying cancer biology plays a role in screen detection; some interval cancers arise from aggressive tumor biology and distinct molecular and genomic subtypes of breast cancer present with distinct mammographic features. Results of this research add to our understanding of mammographic screening limitations and helps prioritize research questions in the context of evolving radiologic practices.