Abstract Background: Though current guidelines classify breast tumors with ≥1% estrogen receptor (ER) positivity by immunohistochemistry as ER positive, “borderline” tumors expressing ≥1 – <10% positivity have biological similarities to ER negative tumors in prior studies. We sought to describe molecular features of ER-borderline breast cancer in a diverse cohort of incident cases and identify characteristics associated with high-risk borderline tumors. Methods: We used the Carolina Breast Cancer Study, a large population-based study that oversampled young and black women in 44 counties of North Carolina, to study characteristics of ER-borderline cases compared to other ER-positive and ER-negative tumors. Gene expression of ESR1, PAM50 subtype, and risk of recurrence score (ROR-PT, calculated from subtype, proliferation level, and tumor size; categorized as low, medium, high) were quantified using RNA counting methods, with finite mixture modeling to establish a cutoff for ESR1-high vs. -low tumors. The relative frequency differences of clinical and genomic features of borderline vs. positive/negative tumors were estimated using linear regression, adjusted for age and race. Recurrence risk was evaluated by ER status, with and without receipt of endocrine therapy using Kaplan-Meier curves and Cox proportional hazards models. Results: Of 2,859 eligible patients, 8% (n=217) were ER borderline, 26% (n=757) were ER negative and 66% were ER positive. Compared to other ER positive tumors (which were 5% basal-like, 89% luminal, 8% high ROR-PT, 16% ESR1-low), borderlines (40% basal-like, 42% luminal, 26% high ROR-PT, 71% ESR1-low) had a higher relative frequency of basal-like subtype (+37.7%, 95% CI 27.1, 48.4), high ROR-PT (+52.4%, 95% CI 36.8, 68.0), and ESR1-low status (+26.3%, 95% CI 20.2, 32.5). In log-rank analysis, recurrence risk in borderlines was not significantly different from ER negatives. However, after adjusting for patient and tumor factors, recurrence risk in borderlines treated without endocrine therapy was the highest of all subtypes (HR 2.9, 95% CI 1.6, 5.0) and persisted after further adjusting for ROR-PT. Results are summarized in Table 1. Frequency of high ROR-PT was 22% among white women with borderline tumors but was significantly more common among black women with borderline tumors (+26.0%, 95% CI 7.1, 45.0). Among borderlines, high grade was associated with basal-like subtype (+43.4%, 95% CI 18.7, 68.0; reference = non-basal subtype, 52% high grade). Table 1:Recurrence-free survival by ER subtype and endocrine therapy receipt Hazard Ratio (HR) Model 1HR Model 2ER PositiveREFREFER Borderline, no endocrine therapy2.9 (1.6, 5.0)2.3 (1.1, 4.9)ER Borderline, endocrine therapy1.5 (0.7, 3.2)1.5 (0.6, 3.7)ER Negative1.8 (1.3, 2.6)1.5 (0.9, 2.4)Adjusted for (1) age, race, stage, grade, year of diagnosis, (2) variables in (1) and ROR-PT Conclusion: Consistent with prior studies, borderline tumors demonstrated characteristics that were intermediate between other ER positive and negative tumors, with heterogeneity of genomic features. Given the higher recurrence risk among borderlines, identifying aggressive borderline tumors is important. Black women and those with high tumor grade may benefit from further genomic assessment to guide treatment decisions. Citation Format: Benefield HC, Allott EH, Sun X, Reeder-Hayes KE, Troester MA. Clinical and genomic characteristics of borderline ER-positive breast cancers [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P1-09-05.
Read full abstract