Abstract
Associations between HLA class I antigen expression and the efficacy of a melanoma vaccine (Melacine; Corixa Corp.) were initially described in stage IV melanoma. Similar associations were observed in S9035, a phase III adjuvant trial evaluating Melacine for 2 years compared with observation in patients with stage II melanoma. This report provides long-term results. The effects of treatment on relapse-free survival (RFS) and overall survival (OS) were evaluated, and prespecified analyses investigated associations between treatment and HLA expression. Multivariable analyses were adjusted for tumor thickness, ulceration and site, method of nodal staging, and sex. P = 0.01 was considered statistically significant in subset analyses to account for multiple comparisons. For the entire study population of 689 patients, there were no significant differences in RFS or OS by treatment arm. HLA serotyping was performed on 553 (80%) patients (vaccine, 294; observation, 259). Among the subpopulation with HLA-A2 and/or HLA-Cw3 serotype, vaccine arm patients (n = 178) had marginally improved RFS (adjusted P = 0.02) and significantly improved OS compared with observation arm patients (n = 145), with 10-year OS of 75% and 63%, respectively [hazard ratio (HR), 0.62; 99% confidence interval (CI), 0.37-1.02; P = 0.01]. There was no impact of HLA-A2 and/or HLA-Cw3 expression on observation arm patients. An analysis of mature data from S9035 indicates a significant OS benefit from adjuvant vaccine therapy for patients with HLA-A2- and/or HLA-Cw3-expressing melanoma. The possibility of interactions between HLA type and outcome should be considered in future immunotherapy trials. Further investigations of melanoma-associated antigens present in Melacine and presented by HLA-A2 and HLA-Cw3 may be warranted.