50 Background: Current methods for intraoperative assessment of lumpectomy margins are limited, and a meaningful proportion of patients require re-excision to achieve acceptable margins. There is little available information regarding the relationship of mammographic breast density (BD) and positive margin rate. Methods: The MarginProbe device uses radiofrequency spectroscopy to evaluate the margins of lumpectomy specimens. The current study utilized data from the MarginProbe Pivotal Trial (Dune Medical Devices, Israel, NCT00749931). This randomized clinical trial compared adjunctive intraoperative use of the MarginProbe device with surgeons’ standard approach to lumpectomy surgery. For the current analysis, data was compiled from the period prior to patient randomization and device use. Variables of interest included BD, patient and tumor characteristics, and the margin status of the main lumpectomy specimen (prior to device use in the device arm). For the purpose of this analysis, a positive margin was considered tumor on ink. Statistical analysis was performed with univariate and multivariate analysis, and linear/logistic regression. Results: A total of 664 patients were enrolled in the trial. 450 patients had preoperative breast density information available, and formed the basis for this analysis. As expected, higher BD was associated with younger age, lower BMI and smaller breast and specimen volume. Increased BD was also associated with increased use of preoperative MRI imaging (odds ratio 2.2, p<0.0001). Higher BD was also associated with a significant increase in main lumpectomy specimen positive margin rate (Table). The odds ratio was 1.46 per change in density category (p=0.011). BD remained significantly associated with positive margins after controlling for age, BMI and breast volume. Conclusions: Higher BD is an independent risk factor for positive margins in main lumpectomy specimens, suggesting that adjunctive methods for intraoperative margin assessment may be particularly helpful in these patients. [Table: see text]