Advances in molecular biology and immunology have renewed interest in the development of vaccines for the treatment or prevention of cancer. Research over the past 10 years has focused on the identification of suitable tumour antigens to use as targets for a variety of vaccine strategies. Carcinoembryonic antigen (CEA) was one of the first tumour antigens described, and is commonly expressed by a wide range of adenocarcinomas. Recent studies have identified several human-leukocyte-antigen-restricted epitopes (short peptides) within the CEA protein that can be recognised by human T lymphocytes (T cells). Although CEA-expressing tumour cells are generally weakly recognised by the immune system, several new strategies have been used to enhance immune responses against CEA. This includes using antibodies directed against CEA; inserting the CEA gene into recombinant viruses and bacteria as viral and bacterial vaccines; pulsing the CEA protein, peptides, DNA or RNA onto dendritic cells (specialised antigen-presenting cells); and combining CEA vaccines with cytokines or co-stimulatory molecules to increase vaccine effectiveness. Other factors that might be important in establishing systemic immunity against CEA are the dose, route, timing, and choice of vector and adjuvants for vaccine administration. Further research in understanding the fundamental processes involved in tumour-cell recognition by the immune system, better animal models, and improved clinical trial designs will help to define the full potential of CEA as a target for cancer vaccine development.