PURPOSE: To determine the usefulness of ultrasound-guided core biopsy for the diagnosis of non-palpable beast lesions. MATERIALS AND METHODS: Between April 1996 and December 2000, 932 lesions in 901 patients were the object of ultrasound-guided core biopsy. Of these, 440 non-palpable lesions ranging in size from 0.3 to 3.0 (average, 0.9)cm, and found in 428 patients (all women aged, on average, 43.9 years), were included in this study. The pathologic results of core biopsy were compared with the available surgical data, and clinical and radiologic follow-up data were also reviewed. A 16-gauge needle was used in 197 lesions, and a 14-gauge neadle in the other 243. RESULTS: At core biopsy, 53 lesions were diagnosed as invasive carcinoma, and 45 of these were excised. Forthfour were confirmed as invasive carcinoma, and in one case there was no residual tumor. Seven lesions, diagnosed as ductal carcinoma in situ at core biopsy, were surgically removed, and the final diagnosis was ductal carcinoma in four cases and invasive carcinoma in two. Two of four cases initially diagnosed as atypical ductal hyperplasia were finally diagnosed as invasive carcinoma after surgery. Six lesions diagnosed at core biopsy asbenign were later found to be malignant (false-negative rate, 8.3%). Radiologic imaging suggested that all six lesions-for two of which, a 14-gauge needle was used, and for four, a 16-gauge needle-were malignant. The false-negative rate was 5.1% and 12%, respectively, whithout statistical significance (p=0.26). CONCLUSION: Ultrasound-guided core needle biopsy for non-palpable breast lesions is useful and can replace surgical excision. To avoid false-negative assessment, however, strict radiologic-histopathologic correlation is required.