S4-4: Clarifying the Risk of Breast Cancer in Women with Atypical Breast Lesions.

Abstract

Abstract Background: Women diagnosed with atypical ductal hyperplasia (ADH), atypical lobular hyperplasia (ALH), lobular carcinoma in situ (LCIS), and borderline ADH/DCIS are at increased risk for breast cancer, but the precise degree of risk varies widely in the literature. Information from prior studies is limited by grouping ADH and ALH together and by small cohort sizes. Objectives: To identify women with a pathologic diagnosis of ADH, ALH, LCIS, and borderline ADH/DCIS using Natural Language Processing. To evaluate breast cancer risk based on atypia type. Methods: Using Natural Language Processing, we reviewed all electronically available pathology reports from Massachusetts General Hospital, Brigham and Women's Hospital, and Newton-Wellesley Hospital (members of Partners HealthCare System) from 1987–2010. We identified all women with a diagnosis of ADH, ALH, LCIS, and borderline ADH/DCIS with no prior or concurrent diagnosis of breast cancer. We determined the incidence of subsequent invasive and noninvasive breast cancer, the side of cancer diagnosis compared to original atypia side, and the time to cancer diagnosis for each atypia type. Results: We reviewed 76,333 path reports in 42,950 unique individuals and identified 3049 women who were diagnosed with atypical breast lesions over this 14-year period; 1233 (40.4%) had ADH, 851 (27.9%) had ALH, 595 (19.5%) had LCIS, and 370 (12.1%) had borderline ADH/DCIS. The mean age for atypia diagnosis was 51 years (range: 18–93). At a mean follow-up of 66 months, cancer occurred in 7.0% of women with ADH, 11.3% of women with ALH, 11.1% of women with LCIS, and 8.4% of women with borderline ADH/DCIS. The median time to breast cancer diagnosis was 48 months with ADH, 50 months with ALH, 47 months with LCIS, and 60 months with borderline ADH/DCIS. Significantly more ipsilateral cancers developed than contralateral cancers for all types of atypia combined (p=0.027). The development of invasive versus noninvasive breast cancer was not significantly affected by atypia type. Subsequent cancers were DCIS in 121 patients (43.4%) and invasive in 158 patients (56.6%). Kaplan Meier curves for time to cancer diagnosis based on atypia type were created. The curves for ADH and borderline ADH/DCIS were similar and significantly different than the curves for ALH and LCIS (p<0.001). The estimated 5 and 10-year breast cancer risks for each atypia type are presented in Table 1. Conclusion: A diagnosis of ADH, ALH, LCIS, or borderline ADH/DCIS increases a woman's risk of invasive and noninvasive breast cancer in either breast. The breast cancer risk at 5 and 10 years is significantly higher in those with ALH or LCIS compared to those with ADH or borderline ADH/DCIS, but there is little difference in risk between ADH and borderline ADH/DCIS or between LCIS and ALH. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr S4-4.