Between 1995 and 1996, 96 consecutive patients with nonpalpable breast lesions detected by abnormal mammogram underwent core needle biopsies under stereotactic guidance at Johns Hopkins Bayview Medical Center. The cytologic diagnoses rendered on touch imprints and the histopathologic diagnoses made on the core needle biopsies were retrospectively reviewed to ascertain the accuracy and the validity of this procedure. These imprints were made of the cores as on-site evaluation in order to assist the radiologist. Separate diagnoses were rendered on the imprints and the cores. Follow-up excisional/open biopsies were then correlated to determine the usefulness of making touch imprints from the core needle biopsies toward the overall management of a patient with an abnormal mammogram. The core needle biopsies were obtained with a 14-gauge needle and biopsy gun. The cytologic diagnoses rendered on touch imprints and histopathologic diagnoses on core needle biopsies were compared and the concordance rate was determined. The subsequent surgical follow-up was analyzed and correlated with the imprint cytology and the core needle biopsy interpretation to ascertain the pathologic outcome. A total of 100 cases were reviewed: 4 patients had two lesions biopsied. Eighty-four cases showed complete cytohistologic correlation, that is, a high concordance rate. These 84 cases were divided into four categories: malignant (24), atypical (2), benign (55), and unsatisfactory (3). Of these concordant cases, there were 34 subsequent excisional biopsies and all except one confirmed the core needle biopsy diagnosis. The exception had a benign touch imprint and core but the excisional biopsy showed multifocal ductal carcinoma in situ. Of the 100 cases, 16 were nonconcordant (the cytologic diagnosis fell into a different category from the histologic diagnosis). Seven of these nonconcordant cases resulted in excisional biopsies and all but one showed the core diagnosis to be correct. The only exception was a case with atypical cytology and a benign core biopsy with the follow-up excision showing atypia. The rest of the nonconcordant cases with atypical imprints and benign cores had no follow-up surgery, showing that the clinicians are inclined to depend on the core biopsy diagnosis. While our study demonstrates the accuracy and concordance of cytologic touch imprints, the surgical follow-up data reveal that there does not appear to be any additive value to rendering a separate diagnosis on touch imprints of core needle biopsies.