It is known that the pineal gland has some antitumor activity. Melatonin, its most important hormone, has been shown to inhibit tumor growth in vivo and in vitro. Moreover, some investigations have demonstrated an altered melatonin secretion in cancer patients. Despite these interesting data, clinical trials have never been carried out to evaluate the effects of melatonin on human neoplasms. The aim of this study was to draw some preliminary conclusions on melatonin therapy in advanced human neoplasms. Nineteen patients suffering from advanced solid tumors, which did not respond to standard therapies, entered the study. Performance status (PS) was 20 or less in 9 cases, and more than 20 in the other 10. Melatonin was given intramuscularly at a daily dose of 20 mg at 3.00 p.m., followed by a maintenance period with lower doses in patients who had a remission, a stabilization of disease or an improvement in PS. Among patients with a PS higher than 20, a partial response was achieved in one case with cancer of the pancreas; moreover, 5 of 10 had stable disease, but the other 4 cases had a progression; an evident improvement of PS was obtained in 6 of the 10 cases. In contrast, among patients with a very poor PS, 7 of 9 died within the first 2 months of therapy. This preliminary study would suggest that melatonin may be of some value in treating cancer patients in whom standard antitumor therapies have failed, particularly in improving their PS and quality of life.