To perform a double-blind randomized trial with American Joint Commission on Cancer (AJCC) stage III melanoma patients for the following reasons: (1) to confirm our previous finding that patients with antibodies against the melanoma differentiation antigen GM2 have an improved prognosis, and (2) to demonstrate clinical benefit from GM2 antibody induction.

One hundred twenty-two patients with AJCC stage III melanoma who were free of disease after surgery were randomized: 58 to receive treatment with the GM2/BCG vaccine, and 64 to receive treatment with bacille Calmette-Guèrin (BCG) alone. All patients were pretreated with low-dose cyclophosphamide (Cy).

GM2 antibody was detected in 50 of 58 patients treated with GM2/BCG and seven of 64 patients treated with BCG alone. With a minimum follow-up period of 51 months, there was a highly significant increase in the disease-free interval (P = .004) and a 17% increase in overall survival (P = .02) in these 57 antibody-positive patients, confirming our earlier experience. Exclusion of all patients with preexisting GM2 antibodies (one in the GM2/BCG group and five in the BCG group) from statistical analysis resulted in a 23% increase in disease-free interval (P = .02) and a 14% increase in overall survival (P = .15) at 51 months for patients treated with the GM2/BCG vaccine. However, when all patients in the two treatment groups were compared as randomized, these increases were 18% for disease-free interval and 11% for survival in the GM2/BCG treatment group, with neither result showing statistical significance.

(1) Vaccination with GM2/BCG induced immunoglobulin M (IgM) antibodies in most patients. (2) GM2 antibody production was associated with a prolonged disease-free interval and survival. (3) Comparison of the two arms of this trial as randomized fails to show a statistically significant improvement in disease-free interval or survival for patients treated with GM2/BCG vaccines.