This study compared pre-operative staging with MR mammography (MRM) and positron emission tomography (PET) in patients with clinically suspected breast cancer according to the Breast Imaging Reporting and Data System, category 5. A total of 43 patients with breast cancer were examined. MRM included both T(2) weighted turbo spin echo sequences and T(1) weighted gradient echo sequences (three-dimensional fast low angle shot) before and after application of gadolinium-DPTA. All patients then underwent examination with a modern full-ring PET scanner following injection of fluorodeoxyglucose. We evaluated the efficacy of these methods in the diagnosis of primary tumour, contralateral carcinomas, bifocal, trifocal or multifocal disease, as well as non-invasive cancer portions and tumour size. Determination of patients' N-status was only attempted using PET. All findings were validated by histological examination. MRM was slightly superior to PET in several areas, such as in the respective methods' sensitivity and specificity. Sensitivities for MRM and PET were: 100% vs 93.0% in diagnosis of the primary tumour; 100% vs 100% in diagnosis of contralateral carcinomas; and 95.2% vs 92.5% in diagnosis of bifocal, trifocal or multifocal disease. Specificities for MRM and PET were: 100% vs 97.5% in diagnosis of contralateral carcinomas; and 96.8% vs 90.3% in diagnosis of bifocal, trifocal or multifocal disease. Non-invasive cancer portions and tumour sizes were equally well determined with both methods. The sensitivity of PET for detection of lymph node involvement was 80% and specificity 95%. MRM and PET were superior to conventional methods in nearly all areas studied; the findings of one or both of the methods impacted positively on patients' surgical treatment in 12.5-15% of cases. Pre-operative MRM and/or PET can have a positive influence on surgical treatment planning. Therefore, it appears useful to perform pre-operative staging with MRM or PET in these patients.