Abstract CT188: 5-year survival and other long-term outcomes from KEYNOTE-006 study of pembrolizumab (pembro) for ipilimumab (ipi)-naive advanced melanoma

Introduction: Pembro significantly improved OS vs ipi in advanced melanoma in the KEYNOTE-006 study (NCT01866319). We present outcomes from the 5-year follow-up of this study−the longest to date in a randomized phase 3 trial of pembro in advanced cancer. Long-term follow-up for pts completing 2 years of pembro and data for pts treated with a second course (2nd course) of pembro are also reported. Methods: Eligible pts (N=834) were randomly assigned (1:1:1) to pembro 10 mg/kg Q2W or Q3W for up to 2 years or ipi 3 mg/kg Q3W for 4 doses. Eligible pts who completed pembro or stopped for CR and then progressed could receive an additional 12 mo of pembro if they met inclusion criteria. End points included OS and ORR per irRC by investigator review. Pembro completion was defined as discontinuation with CR, PR, or SD after ≥94 weeks of pembro. Results were pooled from the 2 pembro arms. Results: At data cutoff (Dec 3, 2018), median follow-up of surviving pts was 57.7 mo (range, 0.03-62.1). Median OS (95% CI) was 32.7 mo (24.5-41.6) in the combined pembro arms (n=556) and 15.9 mo (13.3-22.0) in the ipi arm (n=278) (HR, 0.73). Five-year OS rates (95% CI) were estimated to be 38.7% (34.2-43.1) and 31.0% (25.3-36.9), respectively. For pts receiving first-line pembro, median OS (95% CI) was 38.7 mo (27.3-50.7) in the combined pembro arms (n=368) and 17.1 mo (13.8-26.2) in the ipi arm (n=181) (HR, 0.73); 5-year OS rates (95% CI) were estimated to be 43.2% (38.0-48.3) and 33.0% (25.8-40.3), respectively. A total of 103 (18.5%) pts completed 2 years of pembro; median survival follow-up from pembro completion was 34.5 mo; OS rate at 36 mo was 93.8%. Of the 103 pts, 76 were progression-free and 27 had PD. Median time from pembro end to progression was 16.8 mo (range, 0.99-33.9). Thirteen pts received 2nd course pembro; best overall response (BOR) in 1st course was 6 CR, 6 PR, and 1 SD. Median duration of 2nd-course pembro was 9.7 mo; BOR on 2nd course was 3 CR, 4 PR, 3 SD, and 1 PD (response assessment was pending for 2 pts). All 3 pts with 2nd-course CR and 2 of 4 with PR had ongoing response; the remaining 2 pts who had 2nd-course PR subsequently progressed. Four pts discontinued 2nd-course treatment before 12 mo (2 due to PD, 1 due to G2 interstitial pneumonia, and 1 due to physician decision). Six pts had grade1/2 TRAEs during 2nd-course pembro; there were no grade 3/4 TRAEs or deaths. Conclusions: Pembro continued to show improved OS vs ipi in 5-year follow-up of pts w