Objective. To evaluate pembrolizumab as adjuvant therapy for patients with resected, high-risk stage III melanoma.
Design. International randomized phase 3 trial.
Setting and participants. This multicenter international trial enrolled patients who had histologically confirmed cutaneous melanoma with regional lymph node metastasis (stage IIIA, IIIB or IIIC with no in-transit metastases). Patients had to have undergone a complete regional lymphadenectomy within 13 weeks before the start of treatment. Exclusion criteria were: ECOG performance status score > 1, autoimmune disease, current steroid use, and prior systemic therapy for melanoma. All tumor samples from melanoma-positive lymph nodes were required to be sent to the central lab for evaluation of programmed death ligand 1 (PD-L1) expression; PD-L1 positivity was defined as a tumor proportion score (TPS) ≥ 1%.
Intervention. Patients were randomized in a 1:1 fashion and stratified according to stage and geographic region. Local pharmacies were aware of trial-group assignments. Patients received either an intravenous infusion of pembrolizumab 200 mg or placebo every 3 weeks for a total of 18 doses or until disease recurrence or unacceptable toxicity occurred. If recurrence was detected, patients were able to cross over.
Main outcome measures. The primary outcome was recurrence-free survival (RFS) in the intention-to-treat population and in the subgroup of PD-L1–positive patients. Secondary endpoints included distant metastasis–free survival, overall survival (OS), safety, and quality of life.
Results. A total of 1019 patients were recruited from 123 centers in 23 countries: 514 were assigned to the pembrolizumab group and 505 were assigned to the placebo group. In the pembrolizumab group, 70 patients (13.8%) discontinued treatment because of an adverse event; in 66 patients of these patients the event was deemed drug-related. In the placebo group, 11 (2.2%) patients discontinued treatment due to an adverse event. Discontinuation due to disease recurrence was seen in 109 (21%) patients in the pembrolizumab group and 179 (35.7%) patients in the placebo group. The median duration of follow up was 15 months. In the overall intention-to-treat population, the 12-month RFS rate was 75.4% in the pembrolizumab group versus 61% in the placebo group (P < 0.001). At 18 months the RFS rates were 71.4% and 53.2%, respectively. The 18-month incidence of distant metastasis at recurrence was lower in the pembrolizumab group (16.7% vs. 29.7%, hazard ratio [HR] 0.53; 95% confidence interval 0.37 to 0.76). In those who were PD-L1–positive (n = 853), the 12-month RFS rate was 77.1% in the pembrolizumab group versus 62.6% in the placebo group. PD-L1 status had no impact on pembrolizumab efficacy. The benefit of pembrolizumab was noted across all subgroups, and no difference was seen in patients with stage IIIA, IIIB or IIIC disease. The benefit of pembrolizumab was similar in those with macroscopic or microscopic nodal metastasis. BRAF status did not influence RFS between the pembrolizumab and placebo groups.
Adverse events of grade 3 or higher were seen in 14.7% and 3.4% of the pembrolizumab and placebo groups, respectively. Immune-related adverse events of any grade were noted in 37% of patients in the pembrolizumab group. There was 1 pembrolizumab-related death secondary to myositis. Grades 3 or 4 immune-related events in the pembrolizumab group occurred at a low rate, including colitis (2% and 0.2%), hypophysitis (0.6% and 0%), and type 1 diabetes mellitus (1% and 0%).