BACKGROUND Studies of human tumor-infiltrating lymphocytes (TILs) derived from patients with a variety of histologic types of cancer have demonstrated that cellular immune reactions against established malignancy exist in humans. OBJECTIVE We report the results of using autologous TILs plus high-dose bolus interleukin 2 (IL-2), with or without the concomitant administration of cyclophosphamide, in the treatment of 86 consecutive patients with metastatic melanoma. METHODS From May 1987 through December 1992, 86 patients (38 female and 48 male) with metastatic melanoma were treated (145 courses) with autologous TILs plus high-dose intravenous bolus IL-2 (720,000 IU/kg every 8 hours). TILs plus IL-2 were administered in two cycles separated by approximately 2 weeks. Two treatment cycles constituted one treatment course. Patients received a maximum of 15 doses of IL-2 per cycle given every 8 hours until grade 3 or 4 toxicity was reached that could not easily be reversed by standard supportive measures. All patients received concomitant medications to abrogate some of the side effects of IL-2 administration: acetaminophen (650 mg every 4 hours), indomethacin (50 mg every 8 hours), and ranitidine (150 mg every 12 hours). Fifty-seven of the 86 patients received a single intravenous dose of 25 mg/kg cyclophosphamide approximately 36 hours before receiving the first intravenous infusion of TILs plus IL-2. Six weeks after treatment, all known sites of disease were evaluated. RESULTS The overall objective response rate in these patients was 34% and was similar in patients receiving TILs and IL-2 alone (31%) or in conjunction with cyclophosphamide (35%). There was no significant difference in the objective response rate in patients whose therapy with high-dose IL-2 had failed (32%) compared with patients not previously treated with IL-2 (34%). The frequency of response to treatment was greater in those patients who were treated with TILs from younger cultures (P = .0001), TILs with shorter doubling times (P = .03), and TILs that exhibited higher lysis against autologous tumor targets (P = .0008). Patients who received TILs generated from subcutaneous tumor deposits had higher response rates (49%) compared with those receiving TILs from lymph nodes (17%; P = .006). There was one treatment-related death due to respiratory insufficiency. CONCLUSIONS Treatment with TILs and IL-2 with or without cyclophosphamide can result in objective responses in about one third of patients with metastatic melanoma. The side effects of treatment are transient in most patients, and this treatment can be safely administered. These results illustrate the potential value of immune lymphocytes for the treatment of patients with melanoma.