A substantial proportion of patients with breast cancer are treated with the antioestrogen tamoxifen. As with other endocrine therapies, clinical experience has shown that some tumours in which growth is initially attenuated by tamoxifen treatment become resistant to continued drug treatment and resume growth. The mechanisms underlying the development of tamoxifen resistance have yet to be described but represent an important focus of research with the aim of defining what other therapies might be effective following tamoxifen treatment. Secondly, an understanding of tamoxifen resistance might suggest means to develop more effective agents for primary treatment of the disease. The development of pure antioestrogens, for example ICI 164,384 and ICI 182,780, which differ pharmacologically from tamoxifen in being entirely free of oestrogen partial-agonist activity, together with cell and animal models of tamoxifen resistant human breast cancer, has revealed one mechanism which might be of considerable clinical significance. Pure antioestrogens were shown to inhibit the proliferation of a greater proportion of tumor cells than tamoxifen in vitro, a differential effect that was attributed to the oestrogenic activity of tamoxifen. Subsequently, cell culture studies have shown that breast cancer cell lines selected for resistance to tamoxifen can still remain sensitive to the growth inhibitory action of pure antioestrogens. Similarly, the growth of human breast tumours in nude mice, which is initially attenuated by tamoxifen but then resumes, can be inhibited by pure antioestrogens. Both types of experiment are consistent with the view that tamoxifen resistance in these model systems is due to the oestrogenic action of tamoxifen. Thus, it can be predicted that in some patients whose tumours recur during tamoxifen therapy, a further response to pure antioestrogen treatment might occur. Studies to examine this hypothesis are currently being undertaken with ICI 182,780. One mechanism which might account for the experimental observations is an intrinsic heterogeneity amongst breast tumour cells in their response to tamoxifen, i.e. that there are at least two different populations of cells; one population which responds to tamoxifen as an antioestrogen and one which "reads" tamoxifen as an oestrogen. The growth advantage thus conferred on the latter population would lead to its predominance. If this is what actually happens in a proportion of human tumours, it can be argued that primary treatment of the tumour with a pure antioestrogen, rather than tamoxifen, would be preferred since a more complete and longer-lasting response would be predicted. Recent comparative studies with human breast tumours grown in nude mice support these predictions.