The phase 1 trial (KEYNOTE-001) testing the efficacy of pembrolizumab demonstrated an ORR of 33% in the total population of patients treated and an ORR of 45% in those who were treatment-naive. Additionally, the median OS was 23 months for the total population and 31 months for treatment-naive patients, with only 14% of patients experiencing a grade 3 or 4 adverse event.22 The KEYNOTE-002 phase 2 trial compared 2 different pembrolizumab doses (2 mg/kg and 10 mg/kg every 3 weeks) to investigator’s choice chemotherapy (paclitaxel plus carboplatin, paclitaxel, carboplatin, dacarbazine, or oral temozolomide) in 540 patients with advanced melanoma with documented progression on ipilimumab with or without prior progression on molecularly targeted therapy if positive for a BRAF V600 mutation. The final analysis demonstrated significantly improved ORR with pembrolizumab (22% at 2 mg/kg vs 26% at 10 mg/kg vs 4% chemotherapy) and significantly improved 24-month PFS (16% vs 22% vs 0.6%, respectively). There was a nonstatistically significant improvement in median OS (13.4 months vs 14.7 months vs 10 months), although 55% of the patients initially assigned to the chemotherapy arm crossed over and received pembrolizumab after documentation of progressive disease.23,24
Because PD-1 inhibition improved efficacy with less toxicity than chemotherapy when studied in progressive disease, subsequent studies focused on PD-1 inhibition in the frontline setting. CheckMate 066 was a phase 3 trial comparing nivolumab to dacarbazine as first-line therapy for 418 patients with untreated metastatic melanoma who did not have a BRAF mutation. For the primary end point of 1-year OS, nivolumab was superior to dacarbazine (72.9% vs 42.1%; hazard ratio [HR], 0.42; P < 0.001). Treatment with nivolumab also resulted in superior ORR (40% vs 14%) and PFS (5.1 months vs 2.2 months). Additionally, nivolumab therapy had a lower rate of grade 3 or 4 toxicity compared to dacarbazine (11.7% vs 17.6%).25
The KEYNOTE-006 trial compared 2 separate dosing schedules of pembrolizumab (10 mg/kg every 2 weeks versus every 3 weeks) to ipilimumab (3 mg/kg every 3 weeks for 4 cycles) in a 1:1:1 ratio in 834 patients with metastatic melanoma who had received up to 1 prior systemic therapy, but no prior CTLA-4 or PD-1 inhibitors. The first published data reported statistically significant outcomes for the co-primary end points of 6-month PFS (47.3% for pembrolizumab every 2 weeks vs 46.4% for pembrolizumab every 3 weeks vs 26.5% for ipilimumab; HR, 0.58 for both pembrolizumab groups compared to ipilimumab; P < 0.001) and 12-month OS (74.1% vs 68.4% vs 58.2%) with pembrolizumab compared to ipilimumab. Compared to ipilimumab, pembrolizumab every 2 weeks had a hazard ratio of 0.63 (P = 0.0005) and pembrolizumab every 3 weeks had a hazard ratio of 0.69 (P = 0.0036). The pembrolizumab groups was also had lower rates of grade 3 to 5 toxicity (13.3% vs 10.1% vs 19.9%).5 Updated outcomes demonstrated improved ORR compared to the first analysis (37% vs 36% vs 13%), and improved OS (median OS, not reached for the pembrolizumab groups vs 16.0 months for the ipilimumab group; HR, 0.68, P = 0.0009 for pembrolizumab every 2 weeks versus HR 0.68, P = 0.0008 for pembrolizumab every 3 weeks).26 In addition, 24-month OS was 55% in both pembrolizumab groups compared to 43% in the ipilimumab group. Grade 3 or 4 toxicity occurred less frequently with pembrolizumab (17% vs 17% vs 20%).
Further analysis from the KEYNOTE-006 trial data demonstrated improved ORR, PFS, and OS with pembrolizumab compared to ipilimumab in tumors positive for PD-L1 expression. For PD-L1-negative tumors, response rate was higher, and PFS and OS rates were similar with pembrolizumab compared to ipilimumab. Given that pembrolizumab was associated with similar survival outcomes in PD-L1-negative tumors and with less toxicity than ipilimumab, the superiority of PD-L1 inhibitors over ipilimumab was further supported, regardless of tumor PD-L1 status.27
In sum, PD-1 inhibition should be considered the first-line immunotherapy in advanced melanoma, either alone or in combination with ipilimumab, as discussed in the following section. There is no longer a role for ipilimumab monotherapy in the first-line setting, based on evidence from direct comparison to single-agent PD-1 inhibition in clinical trials that demonstrated superior efficacy and less serious toxicity with PD-1 inhibitors.5,26 The finding that ORR and OS outcomes with single-agent PD-1 inhibitors are higher in treatment-naive patients compared to those receiving prior therapies also supports this approach.22