Definitions of recurrence and toxicity categories, and follow-up visit windows, were see more provided by the ASBrS and its independent scientific advisory committee to BSI. Management and analysis of the data at BSI occurs only through in-depth discussions between statisticians at BSI and the ASBrS. For the purposes of this analysis, negative margins were defined as greater than or equal to 2 mm between all inked margins and the tumor. Close margins were defined as less than 2 mm of space to an inked margin, and positive margins were defined as “tumor on ink” (focal or otherwise).

No central pathology was performed and margin classifications were based on reporting from the treating institution. An IBTR was defined as the reappearance of breast cancer in the treated breast before development of a distant metastasis and was required to be confirmed pathologically (12). A true recurrence/marginal miss (TR/MM) was defined as a recurrence of the treated cancer within or immediately adjacent to the primary tumor site. An elsewhere failure (EF) was defined as an IBTR several centimeters from the primary site. Investigators were also asked to classify regional failures as axillary, supraclavicular, or internal mammary in location. Overall survival NVP-LDE225 purchase in this

study reflected all deaths, cancer related or otherwise, whereas cause-specific survival was based on deaths attributed only to breast cancer. For this analysis, follow-up was complete by December 2011. All time intervals were calculated from the date of MammoSite RT system explantation. Differences in clinical, pathologic, and treatment-related variables among negative-margin and close-margin, positive-margin,

and close/positive-margin patients were performed via the pairwise Wilcoxon rank sum test and pairwise χ2 tests. Differences in clinical outcomes were analyzed using the log-rank test. Kaplan–Meier Sinomenine tests were used to calculate clinical outcomes. Univariate analysis of IBTR was performed for negative-margin and close/positive-margin patients; within each group, the analysis was repeated for invasive and ductal carcinoma in situ (DCIS) cases separately. All tests were two sided and declared statistically significant if the p-value was less than or equal to 0.05. Version 8.0 or higher of the SAS (Cary, NC) statistical software package was used to provide all statistical analyses. A total of 1440 patients with 1449 treated breasts were analyzed including 1326 (91.5%) with negative margins, 110 (7.6%) with close margins, and 13 (0.9%) with positive margins. Median follow-up was 58.5 months for margin-negative patients, 64.5 months for women with close margins, and 63.1 months for women with positive margins.