Abstract Background: The National Health Service Breast Screening Programme (NHSBSP) invites women between 50 and 70 years for mammography every three years, with abnormal imaging leading to further assessment. Core needle biopsy (CNB) is the standard for obtaining a definitive pre-operative diagnosis. However a proportion of CNBs will be reported as showing indeterminate pathology, with uncertain malignant potential. This necessitates surgical diagnostic excision biopsy, with attendant potential morbidity, and highlights the importance of pre-operative diagnosis. This study aimed to categorise the histological lesions prompting B3 classification arising within a screening programme, and to quantify the definitive surgical outcomes following excision biopsy. Material and Methods: Data was collected retrospectively from January 2000 — December 2010, including demographic details, radiological presentation, relevant clinical findings, results of preoperative investigations, reports of diagnostic excision biopsy and subsequent surgical management. Patients were included if they had a B3 CNB result and subsequent excision biopsy, following attendance for screening mammography. Primary outcome was classification of B3 lesions and positive predictive values (PPV) for malignancy. Secondary outcomes were subsequent surgical procedures. Results: There were 239 B3 lesions, representing 7.1% of all CNB carried out within the screening programme. Mean patient age was 55.6 years, and median lesion size was 14.5 mm (range: 4–25mm). Eighty seven lesions (36.4%) were malignant on excision: 28 (11.7%) invasive and 52 (21.8%) in situ. Lesion specific PPVs for malignancy were as follows: phyllodes tumour 50.0%, atypical ductal hyperplasia 40.0%, columnar cell lesion 40.0%, papillary lesion 26.7%, lobular neoplasia in situ 26.7%, and radial scar/complex sclerosing lesions (RS/CSL) 25.9%. The PPV of RS/CSL with atypia was 65.2%, and without 25.8%. Of 87 malignant cases, 55 (63.2%) went on to have a further surgery. 27 (50.9%) had re-excision of surgical margins, 10 (18.2%) had a mastectomy, 8 (14.5%) had a wide local excision, 2 had an axillary biopsy and one patient had a quadrantectomy. 18 patients had two further operations, the second was most commonly an axillary clearance, and 3 patients had three further operations. Discussion: B3 lesions are heterogeneous in nature, and continue to mandate surgical excision, demonstrated by a PPV of 36.4%, the highest reported figure in the literature to date. B3 CNB represents a significant surgical workload, demonstrated by the fact that 63.2% cases go on to have further surgery, and that 8.8% will require more than one further operation. Currently there is no pre-operative investigation that reliably predicts the likelihood of B3 associated malignancy. It has been demonstrated that the use of large volume vacuum assisted biopsy can reduce the rate of under diagnosis of both in situ and invasive malignancy, and this would clearly reduce the requirement for repeated surgery reported in this study. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P5-09-04.