High-dose intravenous interleukin-2 (IL-2) therapy (14 doses/course, 2 courses/cycle) for metastatic melanoma or kidney cancer induces infrequent, although major responses. In this trial, we evaluated a new schedule (dose of 600,000 IU/kg, 8 h between doses, 5 doses/course, 4 courses at weekly intervals/cycle) of high-dose IL-2, in which we inserted more planned breaks while maintaining high cumulative dose delivery, and investigated the relationship between dendritic cells (DC) and response to treatment. Target dose delivery was attained: median IL-2 cumulative dose per patient was 11.4 and 10.8 million units/kg (cycles 1 and 2, respectively). Major responses were observed in patients with kidney cancer (n=20; 3 complete and 2 partial responses) and melanoma (n=16; 1 partial response). Adverse events appeared comparable with those typically associated with high-dose IL-2. From this data set, we introduce the hypothesis-generating observation that patients who had more favorable outcomes had high pretreatment DC-to-myeloid-derived suppressor cell (MDSC) ratios, similar to the ratio observed in healthy individuals. However, even in patients with the most favorable outcome, after treatment, there were IL-2-induced changes in the DC-to-MDSC ratio, specifically increases in MDSCs. This modified IL-2 schedule is a feasible option, with a more uniform dose delivery over the treatment cycle, a similar toxicity profile, and observed complete, durable response in patients with renal cancer. Pretreatment assessment of DC phenotypic or maturational status may be a starting point to predicting response to high-dose IL-2 cytokine immunotherapy in patients with melanoma and kidney cancer.