AbstractThe COBRA (COre Biopsy after RAdiological localization) study showed that in a controlled study setting, stereotactic large core needle biopsy (LCNB) is as reliable for diagnosing nonpalpable breast lesions as open surgical biopsy. In the present study, we evaluated the diagnostic performance of stereotactic LCNB in routine clinical practice. Between February 2000 and June 2002, data on all patients (n = 955) with nonpalpable breast lesions referred for LCNB were collected. High risk underestimate rate, ductal carcinoma in situ (DCIS) underestimate rate and sensitivity rate after 2 years were calculated. The usefulness of the COBRA guidelines in clinical practice was determined. Nine hundred five successful biopsies were performed in 874 patients. Of the high‐risk lesions, 27% were found to be carcinomas on open biopsy, which is comparable to the results of the COBRA study (23%). The DCIS underestimate rate (28%) was higher than found in the COBRA study (17%). No carcinomas were missed after a follow‐up period of 2 years. Ninety‐six percent of patients were treated according to the COBRA guidelines. The diagnostic performance of stereotactic LCNB in patients with nonpalpable breast lesions seems to be comparable in a controlled study setting and routine clinical practice. © 2007 Wiley‐Liss, Inc.