Breast cancer remains the commonest cause of death from cancer in women in most of the Western world. There is considerable evidence that breast cancer risk is influenced by environmental factors and can therefore potentially be modified. In this paper we describe evidence suggesting a relationship of lipid peroxidation to breast cancer risk, and propose that the method used to generate this information might usefully be applied to other disease states, and make some suggestions for further work. We have compared the urinary excretion of the mutagen malonaldehyde (MDA) in premenopausal women at different risks for breast cancer as determined by the appearance of the breast parenchyma on mammography. MDA was measured in 24-h urine samples from both groups and excretion in 30 women with mammographic dysplasia (high risk) was found to be approximately double that of 16 women without these radiological changes (p less than 0.02). These results suggest that mammographic dysplasia may be associated with lipid peroxidation. Further study of environmental factors associated with states that precede the development of breast and other cancers may lead to the identification of factors that can be modified and that may prevent the development of malignant disease.