(P123) National Trends in the Local Management of Early-Stage Paget Disease of the Breast

Mariam P. Korah, MD, Eugene Chung, MD, PhD, JD; Department of Radiation Oncology, USC Keck School of Medicine

Purpose and Objectives: Paget disease (PD) of the breast is a condition characterized by infiltration of the epidermis of the nipple with neoplastic cells with or without an underlying malignancy of the breast parenchyma. Due to the rarity of PD, the role of breast conserving surgery (BCS) and radiation therapy (RT) is not fully defined. The specific aims of this analysis are to study national patterns of care in the local management of PD and to determine breast cancer–specific survival (BCSS) by type of treatment in a large population-based cohort.

Materials and Methods: The Surveillance, Epidemiology, and End Results (SEER) database was queried for women ≥ 20 years of age diagnosed with Tis–T2 N0 M0 mammary PD who underwent definitive surgery +/– RT from 1998–2010, with minimum follow-up of 6 months. The cohort was stratified by type of treatment into four groups: BCS, BCS + RT, mastectomy (M), or M + RT. Clinical covariates were compared between the groups using the chi-square test. Cox multivariable regression analyses were performed to determine predictors of BCSS. Survival curves were calculated using the Kaplan-Meier method and compared using the log-rank test.

Results: The median follow-up time was 5 years (range: 0.5–12.9 yr). The study cohort comprised 1,509 women, most of whom were ≥ 50 years (79%). The distribution of underlying histologic subtypes was as follows: PD with infiltrating ductal carcinoma (IDC, 43%), PD with ductal carcinoma in situ (DCIS, 46%), and PD with no underlying tumor (11%). High-grade histology was identified in 48% of patients, with estrogen receptor and progesterone receptor positivity in 27% and 18% of patients, respectively. Allocation to treatment groups was as follows: BCS (n = 200), BCS + RT (n = 216), M (n = 1,046), and M + RT (n = 47). Mastectomy rates were highest among patients with PD-IDC (87%). Rates of mastectomy were 74% in the early study period (1998–2004) vs 71% in the latter study period (2005–2010) (P = .12). Overall, lymph node sampling or dissection was performed in 72% of patients: 92% of those with PD-IDC, 61% of those with PD-DCIS, and 44% of those with PD and no demonstrable tumor. BCSS at 5 and 8 years was 93% and 91% for PD-IDC, 98% and 96% for PD-DCIS, and 95% and 95% for PD without underlying tumor, respectively. BCSS was higher among patients in the BCS + RT group (94% at 8 years) compared with those who received BCS alone (91% at 8 years) (P = .54). No difference in BCSS was noted between the BCS + RT and M-alone groups (94% at 8 years) (P = .98). After adjusting for patient and tumor characteristics, no differences were observed in BCSS based on type of surgery (P = .61). Patients who required postmastectomy RT had poorer pretreatment tumor characteristics and inferior BCSS (83% at 8 years).

Conclusions: This population-based analysis shows that mastectomy with lymph node assessment is the most commonly employed approach for local disease management in early-stage PD. Rates of mastectomy have remained relatively constant throughout the study era. Despite conferring comparable rates of BCSS to mastectomy, BCS + RT appears to be underutilized in the management of PD of the breast.