We previously demonstrated that the antigen-specific immune complexes captured by the monoclonal antibody MAb JSI in a sandwich enzyme-linked immunosorbent assay were associated with recurrent melanoma. To determine the potential use of antigen-specific immune complex analysis in predicting the development of recurrent melanoma, we prospectively analyzed serum obtained from patients with melanoma following surgical treatment. Two hundred fifty-three patients have been followed up for a median of 25 months (range, 17 to 29 months). Seventy-seven patients (30%) have developed recurrent melanoma. Antigen-specific immune complexes correlated with the stage of disease at time of entry into the study. The absence of antigen-specific immune complexes in postoperative serum samples is predictive of a disease-free status. Long-term follow-up will define the false-positive rate of antigen-specific immune complex analysis. Continued refinement of this approach should lead to clinically useful methodology to monitor human melanoma.