The extent of positivity of the final excision margin in relationship to other relevant factors was evaluated as a predictor for local recurrence after breast conservation therapy (BCT). As part of an institutional practice policy for BCT in 509 stage I/II breast carcinomas, 105 cases had a final excision margin, which was positive. The median age for this cohort was 58 years and the median follow-up was 86 months. All positive margin patients received whole breast irradiation to 50 Gy-50.4 Gy followed by a boost to the tumor bed for an additional 20 Gy. The extent of positivity (EOP) of the excision margin was graded according to a four-point scale: focal, minimal, moderate, extensive. Cases were then analyzed for local failure according to EOP grade, histology, age, tumor size, total excision volume, re-excision, tamoxifen therapy, and chemotherapy. A focal or minimal EOP grade was found in 70% of specimens while an additional 26% were moderate or extensive. The incidence of invasive carcinoma with prominently associated DCIS was significantly greater in cases with an EOP grade of moderate/extensive. There were nine ipsilateral breast recurrences, eight of which could be evaluated for EOP grade. All recurrences were in or near the previous biopsy cavity. A Kaplan-Meier plot of freedom from local failure showed a significant (P = 0.008) difference between cases grouped by EOP grade of focal/minimal as compared to moderate/extensive. A Cox proportional hazards regression model found that the only variable significantly related at the P < or = 0.05 level to local failure was an EOP grade of moderate/extensive. For breast excision specimens with a positive final margin, an EOP grade of moderate/extensive is a predictor for local recurrence after BCT, which may be independent of other variables such as age or histology.