OBJECTIVE: This investigation compares the frequency of histologic underestimation of breast carcinoma that occurs when a large-core needle biopsy reveals atypical ductal hyperplasia or ductal carcinoma in situ with the automated 14-gauge needle, the 14-gauge directional vacuum-assisted biopsy device, and the 11-gauge directional vacuum-assisted biopsy device. SUBJECTS AND METHODS: Evaluation of 428 large-core needle biopsies yielding atypical ductal hyperplasia (139 lesions) or ductal carcinoma in situ (289 lesions) was performed. The results of subsequent surgical excision were retrospectively compared with the needle biopsy results. RESULTS: For lesions initially diagnosed as ductal carcinoma in situ, underestimation of invasive ductal carcinoma was significantly less frequent using the 11-gauge directional vacuum-assisted biopsy device when compared with the automated 14-gauge needle (10% versus 21%, p < 0.05) but was not significantly less frequent when compared with the 14-gauge directional vacuum-assisted device (10% versus 17%, p > 0.1). For lesions diagnosed initially as atypical ductal hyperplasia, underestimation of ductal carcinoma in situ and invasive ductal carcinoma was significantly less frequent using the 11-gauge directional vacuum-assisted biopsy device compared with the 14-gauge directional vacuum-assisted device (19% versus 39%, p = 0. 025) and with the automated 14-gauge needle (19% versus 44%, p = 0. 01). CONCLUSION: The frequency of histologic underestimation of breast carcinoma in lesions initially diagnosed as atypical ductal hyperplasia or ductal carcinoma in situ using large-core needle biopsy is substantially lower with the 11-gauge directional vacuum-assisted device than with the automated 14-gauge needle and with the 14-gauge directional vacuum-assisted device.