Pembrolizumab or Placebo Plus Etoposide and Platinum as First-Line Therapy for Extensive-Stage Small-Cell Lung Cancer: Randomized, Double-Blind, Phase III KEYNOTE-604 Study

Abstract

Purpose: Pembrolizumab monotherapy has shown antitumor activity in patients with small-cell lung cancer (SCLC). The randomized, double-blind, phase III KEYNOTE-604 study compared pembrolizumab plus etoposide and platinum (EP) with placebo plus EP for patients with previously untreated extensive-stage (ES) SCLC.

Methods: Eligible patients were randomly assigned 1:1 to pembrolizumab 200 mg once every 3 weeks or saline placebo for up to 35 cycles plus 4 cycles of EP. Primary end points were progression-free survival (PFS; RECIST version 1.1, blinded central review) and overall survival (OS) in the intention-to-treat population. Objective response rate (ORR) and duration of response were secondary end points. Prespecified efficacy boundaries were one-sided P = .0048 for PFS and .0128 for OS.

Results: Of the 453 participants, 228 were randomly assigned to pembrolizumab plus EP and 225 to placebo plus EP. Pembrolizumab plus EP significantly improved PFS (hazard ratio [HR], 0.75; 95% CI, 0.61 to 0.91; P = .0023). Twelve-month PFS estimates were 13.6% with pembrolizumab plus EP and 3.1% with placebo plus EP. Although pembrolizumab plus EP prolonged OS, the significance threshold was not met (HR, 0.80; 95% CI, 0.64 to 0.98; P = .0164). Twenty-four-month OS estimates were 22.5% and 11.2%, respectively. ORR was 70.6% in the pembrolizumab plus EP group and 61.8% in the placebo plus EP group; the estimated proportion of responders remaining in response at 12 months was 19.3% and 3.3%, respectively. In the pembrolizumab plus EP and placebo plus EP groups, respectively, any-cause adverse events were grade 3-4 in 76.7% and 74.9%, grade 5 in 6.3% and 5.4%, and led to discontinuation of any drug in 14.8% and 6.3%.

Conclusion: Pembrolizumab plus EP significantly improved PFS compared with placebo plus EP as first-line therapy for patients with ES-SCLC. No unexpected toxicities were seen with pembrolizumab plus EP. These data support the benefit of pembrolizumab in ES-SCLC.

Trial registration: ClinicalTrials.gov NCT03066778.

Figures

FIG 1.

CONSORT diagram. (a) The participant allocated to the pembrolizumab plus etoposide and platinum (EP) group who received placebo plus EP in error was included in the placebo plus EP as-treated population. ITT, intention to treat.

FIG 2.

Progression-free survival (PFS) assessed per RECIST version 1.1 by blinded, independent central review in the intention-to-treat population at the second interim analysis. (A) Kaplan-Meier estimates of PFS. (B) Forest plot of PFS in subgroups. The protocol-specified final analysis of PFS occurred at the second interim analysis. In (B), analysis for the overall population is based on a Cox regression model with treatment as a covariate stratified by platinum chemotherapy, Eastern Cooperative Oncology Group (ECOG) performance status, and lactate dehydrogenase (LDH) concentration; for subgroups, analyses are based on an unstratified Cox regression model with treatment as a covariate. +, no PD at the last disease assessment; CPS, combined positive score; EP, etoposide and platinum; HR, hazard ratio; PD-L1, programmed death ligand 1; ULN, upper limit of normal.

FIG 3.

Overall survival (OS) in the intention-to-treat population at final analysis. (A) Kaplan-Meier estimates of OS. (B) Forest plot of OS in subgroups. In (B), analysis for the overall population is based on a Cox regression model with treatment as a covariate stratified by platinum chemotherapy, Eastern Cooperative Oncology Group (ECOG) performance status, and lactate dehydrogenase (LDH) concentration; for subgroups, analyses are based on an unstratified Cox regression model with treatment as a covariate. CPS, combined positive score; EP, etoposide and platinum; HR, hazard ratio; PD-L1, programmed death ligand 1; ULN, upper limit of normal.

FIG 4.

Kaplan-Meier estimates of duration of response assessed per RECIST version 1.1 by blinded, independent central review at final analysis in participants who experienced complete or partial response. EP, etoposide and platinum.

FIG A1.

Multiplicity diagram for type I error control. The initial α allocated to each hypothesis is represented in the ovals; the weights for reallocation from each hypothesis to the others are represented in the boxes on the lines that connect the hypotheses. If the progression-free survival (PFS) test is significant, the overall survival (OS) hypothesis may be tested at α = .025. If the OS test is significant, the PFS hypothesis may be tested at α = .025. If both PFS and OS are significant, objective response rate (ORR) may be tested at α = .025. The actual boundaries were updated on the basis of the number of events observed and the α spent at previous analyses using the Lan-DeMets O’Brien-Fleming spending function.

FIG A2.

Progression-free survival (PFS) assessed per RECIST version 1.1 by blinded, independent central review in the intention-to-treat population at the final analysis. (A) Kaplan-Meier estimates of PFS. (B) Forest plot of PFS in subgroups. In (B), analysis for the overall population is based on a Cox regression model with treatment as a covariate stratified by platinum chemotherapy, Eastern Cooperative Oncology Group (ECOG) performance status, and lactate dehydrogenase (LDH) concentration; for subgroups, analyses are based on an unstratified Cox regression model with treatment as a covariate. CPS, combined positive score; EP, etoposide and platinum; HR, hazard ratio; PD-L1, programmed death ligand 1; ULN, upper limit of normal.

FIG A3.

Kaplan-Meier estimates of overall survival (OS) in a post hoc analysis of the as-treated population at final analysis. EP, etoposide and platinum; HR, hazard ratio.

FIG A4.

Risk difference between treatment groups for adverse events of any cause in the as-treated population. (A) Adverse events of any grade with incidence ≥ 10%. (B) Adverse events of grade 3-5 with incidence ≥ 2%. EP, etoposide and platinum; pembro, pembrolizumab.