Magnetic resonance imaging of the breast is at this time superior to all other methods of breast cancer diagnosis. Within a few short years since its introduction, magnetic resonance mammography (MRM) can be labelled the "golden standard" for breast imaging and carcinoma detection. Nevertheless, in a certain percentage of high risk patients, there remains an unacceptable number of unclassifiable findings on MRM examination. We were able to demonstrate in 10% of our examined patients small (< 8 mm), occasionally solitary, but usually multiple lesions that exhibited with the use of contrast material an appearance characteristic of carcinoma. Yet in merely half of these cases, that were subsequently surgically examined, significant pathology was found. We assume that some false-negative findings are a matter of the lack of broad acceptance of this diagnostic method by surgeons and pathologists. In order to further the development of MRM in the diagnosis of breast cancer and to ascertain the relevance of these borderline findings, it is important to perfect the techniques of MRM-guided localization of lesions and MRM-guided fine needle biopsies. We except that this would also deepen our knowledge of the pathogenesis of breast carcinoma. MRM is today, however, the examination method of choice for young, high-risk patients, patients with solitary lesions of unknown dignity and patients with markedly fibrocystic breasts.