BACKGROUND. The trials of bacille Calmette-Guerin (BCG) as adjuvant therapy in malignant melanoma conducted over the preceding 2 decades have presented conflicting claims of efficacy.

OBJECTIVE. Determination of the role of BCG immunotherapy in malignant melanoma.

METHODS. Critical analysis of randomized clinical trials of stage I and II melanoma and all reported trials of intralesional and oral BCG in stage III melanoma was conducted. A literature search used the Medline data base (1966–1992);bibliographic reviews of relevant texts and pertinent articles.

RESULTS. No significant benefit of BCG as postsurgical adjuvant therapy in stage I malignant melanoma was observed. Although two of seven trials in stage II melanoma demonstrated benefit with the addition of BCG, the trial with the greatest power in this series detected no difference in outcomes. In stage III malignant melanoma, there was no significant benefit with addition of BCG to various chemotherapeutic regimens. Oral BCG monotherapy produced complete responses in 6%, partial responses in 1%, and extended survival in 7% of patients. Objective responses were observed primarily in patients with intracutaneous non-visceral metastases. Pooled analysis of 15 non-controlled trials of intralesional BCG injections revealed complete responses in 19%, partial responses in 26%, and extended survival in 13% of patients with stage III melanoma. Objective responses to intralesional BCG were more likely in patients with solely cutaneous metastases, particularly intradermal lesions.

CONCLUSION. Pooled analysis of non-placebo controlled trials of intralesional BCG for stage III malignant melanoma supports a trend to enhanced survival in patients with cutaneous non-visceral metastases.