Chemoprevention is a recently introduced and rapidly growing area of oncology that is identifying agents with a potentially preventive role in cancer. Several clinical trials have recently shown the feasibility of this approach in reducing the risk of major human cancers. In the USA, a large trial that demonstrated a reduction of approximately 50% in the risk of developing breast cancer led to Food and Drug Administration (FDA) approval of tamoxifen as a preventive agent in women at increased risk. Although the results could not be reproduced in two smaller European trials, further investigations into this agent are clearly warranted. Raloxifene, another selective oestrogen receptor modulator which has reduced the risk of breast cancer in a trial in women with osteoporosis, is being compared with tamoxifen in a large primary prevention trial in at-risk women. Retinoids are a group of compounds that have proved especially effective in reducing the occurrence of second primary tumours in subjects with skin, head and neck or liver cancer. Fenretinide, a synthetic retinoic acid derivative, has recently been shown to decrease the occurrence of a second breast malignancy in premenopausal women. Results with non-steroidal anti-inflammatory drugs (NSAIDs) have proved consistently encouraging in epidemiological studies in lowering the incidence of colorectal cancer. Clinical trials with selective cyclo-oxygenase inhibitors potentially devoid of gastrointestinal (GI) toxicity are currently underway in at-risk subjects. Calcium and selenium have also received much attention as chemopreventive agents. Originally investigated against skin cancer, selenium showed efficacy in reducing prostate, lung and colon cancer incidence. Similarly, vitamin E was effective in reducing prostate cancer incidence and mortality in a lung cancer prevention trial in heavy smokers. The challenges of conducting well-designed and unequivocal chemoprevention trials are considerable, but advances in techniques of identification of at-risk subjects and establishing surrogate endpoint biomarkers should contribute greatly to future studies. Current knowledge suggests that a pharmacological approach to preventing cancer, using natural or synthetic agents, could become an important way forward.