Background: ASCO’s Choosing Wisely initiative recommends reducing certain low value services, including advanced imaging in early-stage (ES) breast cancer (BC). Few have explored baseline frequency of, or documented clinical reasons for, advanced imaging in ES BC. We analyzed data at an NCI comprehensive cancer center to characterize pt scenarios in which advanced imaging was ordered. Methods: We analyzed all pts with ES BC from January 2010 to June 2012 at the Duke Cancer Institute (DCI). We used the Choosing Wisely criteria for ES BC (clinical Stage IIb or less) and advanced imaging (nuclear bone scan, computed tomography (CT), positron emission tomography (PET), and brain magnetic resonance imaging). We then searched an administrative database for advanced imaging performed in the Duke Health System within 60 days after diagnosis. Three independent reviewers abstracted the medical charts of 40% of cases, randomly selected, to explore reasons for image ordering. Descriptive statistics and the chi-square test were performed. Results: Of 1,143 ES BC cases identified, 20.6% (235 pts) had at least one advanced imaging procedure performed. Imaging type varied widely (41% CT, 22% PET, and 36% bone scans, p<0.001). Pts with advanced imaging were more likely hormone receptor negative, HER2-negative, younger (age<50), and with higher stage (Stage IIb vs. Stage IIa or less) disease (all p<0.001). Of the 95 abstracted cases with imaging, 62% (59 pts) were obtained for further staging, 17% to evaluate a major concurrent disease (6 oncologic, 10 non-oncologic), and 19% for exam or history findings worrisome for distant disease. 15% (9/59 pts) of images ordered for staging were abnormal. Overall, 45% (43/95 pts) of advanced imaging ordered were for a major concurrent disease, worrisome findings, or staging which ultimately revealed an abnormality. Conclusions: Advanced imaging is ordered in a minority of ES BC cases. When ordered, almost half of images aid in clarifying a complex clinical decision. Our data highlight the benefit of quality measures in reducing low value services, while also acknowledging that allowances for measure non-adherence are needed to promote patient-centered cancer care.