Introduction: While a positive margin after an attempt at breast conservation therapy (BCT) is a reason for concern, there is more controversy regarding close margins. When re-excisions are performed, there is often no residual disease in the new specimen, calling into question the need for the procedure. We sought to examine the incidence of residual disease after re-excision for close margins and to identify predictive factors that may better select patients for re-excision. Methods: Our IRB-approved prospective breast cancer database was queried for all breast cancer patients who underwent a re-excision lumpectomy for either close or positive margins after an attempt at BCT. Close margins are defined as 2 mm for invasive carcinoma and 3 mm for DCIS. Clinicopathologic features were correlated with the presence of residual disease in the re-excision specimen. Results: Three hundred three patients (32%) underwent re-operation for either close (173) or positive (130) margins. Overall, 33% had residual disease identified, 42% of DCIS patients and 29% of patients with invasive disease, nearly identical to patients with positive margins. For patients with DCIS, only younger age was significantly related to residual disease. For patients with invasive cancer, only multifocality was significantly associated with residual disease ). However, patients without multifocality still had a substantial risk of residual disease. Discussion: The presence of residual disease appears equal between re-excisions for close and positive margins. No subset of patients with either DCIS or invasive cancer could be identified with a substantially lower risk of residual disease.

Bias in referral patterns and variations in multi-disciplinary management may impact breast conservation therapy (BCT) rates between hospitals. Retrospective studies of BCT rates are limited by their inability to differentiate indicated mastectomies versus those chosen by the patient. Our prospective breast cancer data base was queried for patients with invasive breast cancer who underwent surgical therapy at the University of Michigan over a 3-year period. Demographics, stage and histology were recorded along with the reason mastectomy was performed, categorized as "by need" (contraindication to BCT) or "by choice." Multivariate analysis was used to identify factors significantly associated with mastectomy by choice. BCT was associated with tumor size, histology and nodal status, but not older age, either by choice or by need. Of the 34% of patients initially felt to be poor candidates for BCT, it was absolutely contraindicated in 44%, while 56% were thought to have a tumor-to-breast size ratio too large for successful BCT. Of this latter group, 80% underwent neo-adjuvant chemotherapy in an attempt to downstage the primary tumor and perform BCT, which was successful in over half the patients. For the patients initially thought to be good candidates for BCT, only 15% chose to undergo mastectomy, while 5% eventually required mastectomy due to failed attempts to achieve negative margins. Overall, the BCT rate was 63%, however without the use of neo-adjuvant chemotherapy, the BCT rate would have been only 53%. At a tertiary referral center, BCT rates are driven more by contraindications than patient choice, and may be heavily skewed towards mastectomy due to referral patterns. In addition to tumor factors such as stage and histology, BCT rate can be dramatically impacted by neo-adjuvant chemotherapy or genetic counseling. Examining BCT rates alone as a measure of quality, therefore, is not an appropriate standard across institutions serving diverse populations.