The prognostic significance of neuroendocrine differentiation in prostatic malignancy is controversial, but the results of recent studies with markers such as chromogranin A and neurone-specific enolase suggest that neuroendocrine differentiation, as reflected by increased tissue expression or blood concentrations of these neuroendocrine secretory products, is associated with a poor prognosis, tumour progression, and androgen independence. As all malignant neuroendocrine cells are devoid of androgen receptors and the expression of neuroendocrine cells is not suppressed by androgen ablation, clonal propagation of androgen receptor-negative neuroendocrine cells may have an important role in the development of androgen-independent prostatic carcinoma. This has significant implications for the treatment of prostate cancer, because several of the hormones that are secreted by neuroendocrine differentiated, malignant prostatic cells are potential candidates for use in drug treatment. A limited number of hormones have been tested in this context, in particular somatostatin, bombesin, and serotonin. As there is currently no successful treatment for differentiated prostate cancer, new therapeutic procedures and trials need to be developed to test drugs based on Endocrine-Related Cancer (1999) 6 503-519 neuroendocrine hormones or their antagonists.