To determine the reliability of stereotactic large core needle biopsy (SLCNB) in the diagnostic work-up of suspected carcinomatous non-palpable breast lesions.Prospective.Between 1 February 2000 and 6 June 2002, data were collected on all patients with a non-palpable breast lesion who were scheduled to undergo a SLCNB. SLCNB procedures were performed in 4 centres and patients were referred from 40 hospitals in the Netherlands. Histological diagnosis at SLCNB and subsequent therapy were recorded. Results were compared with results of the preceding 'Core biopsy after radiological localisation' (COBRA) study, where each large core needle biopsy was followed by an excision biopsy. Follow-up data were also obtained from the Dutch National Automated Pathology Archive (PALGA). The adherence to the COBRA guidelines was also evaluated.Nine hundred and fifty-five patients were included with 995 lesions, and 905 biopsy procedures were completed in 874 patients (mean age: 59 years; range: 23-86). Of the high-risk lesions, 27% were found to be carcinomas upon excision biopsy, which is comparable to the results of the COBRA study (23%). Twenty-eight percent of the ductal in-situ carcinomas were found to be invasive, which is higher than in the COBRA study (17%). No malignancies were found during the limited follow-up period (mean: 20.0 months; range: 5.8-34.0). Ninenty-six percent of the patients were treated according to the COBRA guidelines.No malignancies were missed during the limited follow-up period, and therefore, to date, the sensitivity of SLCNB seems to be comparable to that estimated in the study setting.