The exploitation of biological differences between normal and malignant cells is a logical approach to novel treatments for breast cancer. The potential targets for such therapy include the products of proto-oncogenes and oncogenes, inhibition of growth factor receptor signalling and the immunological exploitation of antigenic differences between normal and malignant cells. Monoclonal antibody technology was heralded as a potential 'magic bullet' for cancer therapy following its discovery in the mid-1970s, but it is only in the past few years that such technology has entered mainstream clinical practice. The humanised murine monoclonal antibody to HER2 (trastuzumab) has significant anti-tumour activity but with minimal toxicity, and has been licensed for use in patients with advanced breast cancer. A different approach has been the use of enzyme inhibitors to interfere with the signalling pathways downstream of growth factor receptors (e.g. farnesyl transferase inhibitors). It is likely that effective targets for such therapies will be identified in the next few years. There have been significant advances in our understanding of human immunology which have coincided with the identification of so-called tumour-associated antigens (TAA). These developments have resulted in a resurgence of interest in tumour immunotherapy. Peptides derived from these TAAs have been used to generate tumour-specific immune responses. An alternative strategy has been to immunise patients using viral vectors and plasmid cDNA encoding the TAA. In some studies, notably those in patients with advanced melanoma, significant clinical responses have been observed. Cell-based strategies including autologous tumour cell vaccines, allogeneic tumour cell vaccines and dendritic cell vaccines have been used, and significant responses have been reported in several studies. Few of these methods have so far been applied to breast cancer, but the possible benefits and drawbacks of such an approach will be discussed.