Antigen presenting cells are inactive within tumor tissue because of local immunosuppression. Tumor infiltrating lymphocyte signal activation transducing mechanisms are also seriously impaired. Administration of granulocyte macrophage-colony stimulating factor may lead to antigen-presenting cell recovery and interleukin-2 may restore local tumor infiltrating lymphocyte activation. Moreover, interleukin-2 increases the systemic lymphocyte population, an event which seems to correlate with a better prognosis. The present phase I-II study was carried out to examine whether intralesional injection of granulocyte macrophage-colony stimulating factor followed by subcutaneous interleukin-2 would induce a clinical response in advanced, pretreated and elderly melanoma patients.Fourteen patients over 60 years of age received intralesional granulocyte macrophage-colony stimulating factor (150 micrograms per lesion on day 1), generally divided between the two largest cutaneous lesions, followed by perilesional subcutaneous interleukin-2 (3.000.000/IU) for 5 days (3 to 7) every 3 weeks. All patients received 6 courses of treatment unless progression occurred. Clinical evaluation of the treated cutaneous lesions was assessed at the baseline and before every cycle. Distant lesions were checked every two cycles.Four clinical responses (2 partial responses and 2 minimal responses) (28.5%), which also involved lesions that had not been directly treated, and seven cases of stable disease were observed. The response duration for partial response and minimal response was 9, 4, 4 and 2.5+ months, respectively. Stable disease (50%) recorded in the 7 patients was short term, 3-6 months. Three patients rapidly progressed after 2, 2, and 1 therapy cycles, respectively. The patient who reached the best partial response had a fairly high absolute lymphocyte count (1600 to 2400/mm3). The second one, who reached a complete remission after subsequent locoregional chemotherapy and hyperthermia, had a low absolute lymphocyte count which had doubled, however, by the end of treatment. Blood lymphocyte values in the other patients were too varied to allow any correlation with clinical response. Therapy was well tolerated and only mild fever was observed, with the exception of one patient who had grade 3 fever, with muscle pain and arthralgia.Considering the very low toxicity observed, this treatment might be indicated in elderly patients for whom systemic therapy is no longer a viable option. Improved scheduling and timing could result from further studies.