The purpose of this study was to evaluate the correlation between histologic tumor grade obtained at breast core needle biopsy (CNB) with that obtained from the excised breast specimen. A total of 899 records from all patients with breast cancer from January 1999 to January 2003 were reviewed. Of these, 341 lesions had CNB for diagnosis and had pathology results prospectively graded for both the core and excised specimens. Patient age, lesion type and size, imaging method for biopsy, number of cores, core pathology, surgery performed, final diagnosis, nuclear grade, and combined grade were recorded. In 286 lesions (83.9%), nuclear grade was reported for both core and excised specimens. Of these, 219 (76.6%) were in agreement. Of the 67 that differed, 66 (98.5%) differed by one grade; 46 (68.7%) were upgraded. The groups did not differ in average tumor size, tumor pathology, or number of cores. In 288 lesions (84.5%), combined grade was reported for both core and excised specimens. Of these, 214 (74.3%) were in agreement. Of the 74 that differed, 71 (95.9%) differed by one grade and 50 (67.6%) were upgraded. Again, there were no significant differences in tumor size, type, or number of cores. Overall, nuclear grade and combined grade obtained at CNB each showed agreement with the corresponding excised specimen grading approximately 75% of the time. Grading obtained by CNB, if changed by pathologic evaluation of the excised tumor, is most commonly upgraded. Differences, when they occurred, could not be attributed to tumor size, type, or the number of cores obtained.