PURPOSE: The MAGE-A3 protein is expressed in approximately 35% of patients with resectable non-small-cell lung cancer (NSCLC). Several immunization approaches against the MAGE-A3 antigen have shown few, but often long-lasting, clinical responses in patients with metastatic melanoma. PATIENTS AND METHODS: A double-blind, randomized, placebo-controlled phase II study was performed assessing clinical activity, immunologic response, and safety following immunization with recombinant MAGE-A3 protein combined with an immunostimulant (13 doses over 27 months) in completely resected MAGE-A3-positive stage IB to II NSCLC. The primary end point was disease-free interval (DFI). RESULTS: Patients were randomly assigned to either MAGE-A3 immunotherapeutic (n = 122) or placebo (n = 60). After a median postresection period of 44 months, recurrence was observed in 35% of patients in the MAGE-A3 arm and 43% in the placebo arm. No statistically significant improvement in DFI (hazard ratio [HR], 0.75, 95% CI, 0.46 to 1.23; two-sided P = .254), disease-free survival (DFS; HR, 0.76; 95% CI, 0.48 to 1.21; P = .248), or overall survival (HR, 0.81; 95% CI, 0.47 to 1.40; P = .454) was observed. Corresponding analysis after a median of 70 months of follow-up revealed a similar trend for DFI and DFS. All patients receiving the active treatment showed a humoral immune response to the MAGE-A3 antigen, although no correlation was observed with outcome. No significant toxicity was observed. CONCLUSION: In this early development study with a limited number of patients, postoperative MAGE-A3 immunization proved to be feasible with minimal toxicity. These results are being investigated further in a large phase III study.