Pembrolizumab plus ipilimumab or placebo in previously untreated metastatic NSCLC with PD-L1 tumor proportion score >= 50%: KEYNOTE-598 3-year follow-up

Abstract

Background: At protocol-specified interim analysis 1 (IA1; data cutoff Sep 1, 2020) of the phase 3 KEYNOTE-598 study (NCT03302234), adding ipilimumab (ipi) to pembrolizumab (pembro) did not improve OS (HR, 1.08) or PFS (HR, 1.06) vs placebo (pbo) + pembro in patients (pts) with previously untreated stage IV NSCLC with PD-L1 tumor proportion score (TPS) 50%. Per external DMC recommendation at IA1, ipi and pbo were discontinued and pembro monotherapy continued in both arms. We report 13 additional mo follow-up from IA1 and outcomes of pts who completed 35 cycles of pembro and pts who started second-course pembro. Methods: Eligible pts were randomized 1:1 to ipi 1 mg/kg or pbo Q6W for up to 18 doses (stratified by ECOG PS [0 vs 1], region [East Asia vs non-East Asia], and histology [squamous vs nonsquamous]). All pts received pembro 200 mg Q3W for up to 35 cycles. Dual primary endpoints were OS and PFS per RECIST v1.1 by BICR. Results: 568 pts were randomized to pembro + ipi vs pembro + pbo (n ¼ 284 each). Median time from randomization to data cutoff (Oct 1, 2021) was 33.6 (range, 25.4- 44.6) mo. After discontinuing ipi/pbo for all pts, median OS and PFS remained similar between groups (Table). Grade 3-5 treatment-related AEs occurred in 99/282 (35.1%) vs 57/281 (20.3%) pts in the pembro + ipi vs pembro + pbo arms, respectively. Among 52 pts vs 71 pts initially treated with pembro + ipi vs pembro + pbo who completed 35 cycles of pembro, ORR was 88.5% vs 87.3%, respectively; OS and PFS rates 1 y from completing pembro were 86.3% vs 87.6% and 72.8% vs 83.5%, respectively. 11 pts started second-course pembro