Standard breast conservation therapy consists of lumpectomy and whole breast radiation with boost. The effectiveness of radiation in cases of positive margins is controversial. Two potential treatment-related factors are time to initiation of radiation and localization of the boost. Here, we examine long-term outcomes of positive margin cases treated with an upfront interstitial brachytherapy boost. This particular treatment arrangement may reduce those treatment-related effects on local control. Historically, a low dose interstitial implant was a common boost technique. One approach administered the boost before external beam. A review of 521 cases treated at our institution was conducted. Patients were selected for this study if they received an upfront brachytherapy boost and had close or positive margins. Forty-four breast cancers were identified. Median follow-up of survivors was 11.3 years (8.1-21.7). Implant was performed at lumpectomy (12) or axillary dissection (32). Margin statuses were: focal carcinoma at ink (37), multifocal carcinoma at ink (1), carcinoma <1 mm from ink (2), and DCIS at ink (4). Median tumor size was 1.2 cm (0.5-3.5 cm). Ten patients had nodal involvement. Total median dose was 60 Gy (58.6-65.3 Gy). Median boost dose was 15 Gy (12-20.3 Gy). Lumpectomy to boost interval was median of 3 weeks (0-10.8). No ipsilateral breast recurrences or second primaries were identified. Four developed contralateral breast cancer. Eleven are deceased, four from breast cancer-all from metastasis. 12-year Kaplan-Meier estimates were: overall survival 78 ± 7 %, cause specific survival 93 ± 4 %, and recurrence-free survival 82 ± 6 %. Univariate analysis identified nodal disease as significant for cause specific survival (log rank p = 0.005). No ipsilateral breast recurrences were found. Early administered radiation and accurate boost localization were identified as suspected treatment-related factors for local recurrence. When these two treatment-related factors are minimized, long-term local control rates do not suffer.