Outcome of Women with Ductal Carcinoma In Situ (DCIS) Treated with Breast-Conserving Surgery Alone: A Case-Control Study of 225 Patients from the Cancer Research Network
L Collins, S Schnitt, N Achacoso, R Haque, L Nekhlyudov, S Fletcher, C Quesenberry, L Habel. Beth Israel Deaconess Med Ctr, Boston; Kaiser Permanente Northern California, San Francisco; Kaiser Permanente Southern California, Los Angeles; Harvard Pilgrim Health Care, Boston
Background: Radiation therapy (RT) following breast-conserving surgery (BCS) is now the standard treatment for most women with DCIS. However, the identification of a subset of patients who can safely be spared RT and be adequately treated by BCS alone is an important clinical goal.
Design: Among 3,037 women with DCIS treated with breast-conserving therapy between 1990-2001 at three health maintenance organizations in the Cancer Research Network, 1,298 received BCS alone. Within this group, we performed a case-control study of 225 women (97 cases;128 controls) to assess clinical and pathologic factors associated with local recurrence (LR). Histologic sections were reviewed by pathologists blinded to case-control status. Immunostains for various biomarkers were performed on paraffin sections. Adjusted relative risks (RRs) for LR were calculated for clinical and pathologic factors.
Results: Among the women treated with BCS alone, the 5-yr cumulative LR rate was 16%. Detection of DCIS as a mass or other clinical symptom was associated with an increased risk of LR (RR=2.4; 95% CI 1.1-5.6). Among the pathologic features analyzed, LR was not significantly related to nuclear grade, architectural pattern, or comedo necrosis. Patients with a single low power field (LPF) of DCIS had a significantly lower LR risk compared to those with >10 LPFs (RR=0.3; 95% CI 0.1-0.9). In addition, those with positive or uncertain surgical margins had over a three-fold increase in LR risk compared to those with negative margins (RR=3.6; 95% CI 1.4-9.4). Close margins (<1 mm) were associated with an increased LR risk of borderline significance (RR=2.6; 95% CI 1.0-6.6). None of the biomarkers studied (ER, PR, HER2, Ki67, VEGF, p53 and COX2) were significantly related to risk of LR. In addition, using combinations of ER, PR and HER2 status as surrgoates for molecular phenotype (luminal A or B, HER2 and basal types), we found no signifcant associations between molecular subtype of DCIS and risk of LR.
Conclusions: Among this population of women with DCIS treated with BCS alone the LR rate was unacceptably high (3%/year). The only pathologic features significantly related to LR were larger lesion size and positive or uncertain margins of excision.
Tuesday, March 10, 2009 9:30 AM
Poster Session III # 28, Tuesday Morning