The purpose of this study is to assess the positive predictive value (PPV) for malignancy of core needle biopsy (CNB) demonstrating lobular neoplasia (LN). From 3920 CNBs, 35 (0.89%) LNs (14 atypical lobular hyperplasia - ALH - and 21 lobular carcinoma in situ - LCIS) were identified. Twenty-eight patients underwent surgical excision and seven radiologic follow-up. We describe the imaging findings and excision histology outcomes. We report the PPV for malignancy based on excision histology (n=28) and on excision or follow-up (n=35), and according to the histologic type, biopsy probe and guidance, lesion diameter, and BI-RADS category. PPV for malignancy (based on excision histology) was 46.4% (13/28) and PPV (based on excision or follow-up) was 37.1% (13/35). The overall rate of malignancy for LN was 37.1% (13/35), with a PPV for malignancy of ALH and LCIS of 7.1% (1/14) and 57.1% (12/21), respectively (p=0.003). Estimates of the PPV for malignancy were: stereotactic-guided vacuum-assisted biopsy (22.7%) versus ultrasound-guided automated CNB (61.5%), p=0.053; lesions<20mm (31.2%) versus lesions>20mm (100%), p=0.043; lesions classified as BI-RADS 3 (16.7%) versus BI-RADS 4 or 5 (41.4%), p=0.377. Underestimation of malignancy was therefore more likely in cases of LCIS, US-guided CNB, and lesions that were large and suspicious on imaging. Nevertheless, the absence of these features does not spare the need for surgical excision in lobular neoplasia diagnosed on CNB.