Stereotactic core needle biopsy (SCNB) allows specific histopathologic diagnoses to be made without surgery and has been demonstrated to be an accurate, cost-effective method of diagnosing breast disease, particularly nonpalpable lesions. However, recent studies have concluded that the diagnosis of atypical ductal hyperplasia (ADH) by means of SCNB has resulted in nearly equal odds that a coexisting malignant lesion will be missed. Furthermore, others have concluded that SCNB diagnosed as DCIS cannot reliably indicate the absence of tumor invasion in surgical excision. Between 1993 and 1998, 1,221 consecutive SCNB of mammographically identified lesions were performed using a 14-gauge automated device with an average of 5.3 cores obtained per lesion. ADH was identified in 19 (1.6%) lesions and DCIS in 89 (7.3%). Surgical biopsy was performed in 89 of these patients and histopathologic results from SCNB and surgical biopsies were reviewed and correlated. In 12 cases of ADH diagnosed by SCNB, surgical biopsy showed ADH in 8 (67%) cases and DCIS in the other 4 (33%) cases. In 77 cases of DCIS diagnosed by SCNB, a surgical biopsy showed DCIS in 55 (71%) cases, 6 more cases (8%) had DCIS with focal microinvasion, and 15 (19%) had invasive ductal carcinoma. In one case no residual tumor was found at surgery. In the author's patient population, the diagnosis of ADH at SCNB indicates high probability of DCIS or residual ADH in the surgical biopsy. The diagnosis of DCIS at SCNB is confirmed in the majority of surgical biopsies; however, a significant number of cases may show microinvasion or invasive carcinoma.