The diagnosis of atypical ductal hyperplasia (ADH) at needle core breast biopsy (NCB) is typically regarded as an indication for surgical excision. Although ADH is an intermediate risk nonobligate precursor lesion, the rationale for further therapy is the result of a reported high prevalence of a concomitant more advanced lesion (typically ductal carcinoma in situ) as the index lesion. To assess whether certain histopathologic features of ADH in NCB are predictive of open biopsy outcomes, the authors correlated the extent and pattern of ADH in 47 core biopsies (11-or 14-gauge) with the subsequent surgical specimen. Extent of ADH on NCB was ascertained by determining the number of large ducts and/or terminal duct-lobular units affected, with involvement of one large duct or one terminal duct-lobular unit representing a single focus, involvement of one duct and one terminal duct-lobular unit as two foci, and so on. Of the 47 cases, ADH was restricted to < or =2 foci in 24 cases (51.1%), confined to 3 foci in 8 cases (17.0%), and involved > or =4 foci in 15 cases (31.9%). The corresponding histopathologic findings at excision were benign lesions without atypia (n = 14), focal residual ADH (n = 13), atypical lobular hyperplasia (n = 3), ductal carcinoma in situ (n = 15), and invasive mammary carcinoma (n = 2). When the number of foci of involvement by ADH on NCB (based on an average of 11.6 cores per case) was correlated with the open biopsy results, all cases of ADH limited to < or =2 foci had no worse lesion on excision, whereas ADH present in > or =4 foci was found to be a strong predictor of a more advanced lesion on excision (p <0.0001, chi2). When histologic pattern was evaluated, all cases of pure micropapillary ADH on NCB showed pure micropapillary ductal carcinoma in situ on excision.