Pemetrexed and platinum with or without pembrolizumab for tyrosine kinase inhibitor (TKI)-resistant, EGFR-mutant, metastatic nonsquamous NSCLC: Phase 3 KEYNOTE-789 study.

LBA9000Background: EGFR TKIs are standard 1L therapy for metastatic NSCLC with sensitizing EGFR mutations; however, most patients (pts) ultimately experience PD. We report the protocol-specified final analysis (FA) from the randomized, double-blind, phase 3 KEYNOTE-789 study of pemetrexed (pem) and platinum-based chemotherapy (chemo) with or without pembrolizumab (pembro) as subsequent therapy for pts with TKI-resistant, EGFR-mutant, metastatic nonsquamous NSCLC (NCT03515837). Methods: Adults with histologically or cytologically confirmed stage IV nonsquamous NSCLC, ECOG PS of 0 or 1, documented DEL19 or L858R EGFR mutation, and progression after EGFR TKI treatment were enrolled. Pts were randomized 1:1 to 35 cycles of pembro 200 mg Q3W or placebo (pbo) Q3W plus 4 cycles of pem and carboplatin or cisplatin Q3W followed by maintenance pem. Randomization was stratified by PD-L1 TPS (<50% vs 50%), prior osimertinib (yes vs no), and region (East Asia vs not East Asia). Dual primary endpoints were PFS per RECIST v1.1 by blinded independent central review (BICR) and OS. ORR and DOR per RECIST v1.1 by BICR and safety were secondary endpoints. Final PFS testing was completed at the second interim analysis (IA2; data cutoff, Dec 3, 2021); all other endpoints were assessed at FA (data cutoff, Jan 17, 2023). Efficacy boundaries based on actual observed events were 1-sided P = 0.0117 for PFS (IA2) and P = 0.0118 for OS (FA). Results: 492 pts were randomized to pembro + chemo (n = 245) or pbo + chemo (n = 247). At IA2, median PFS (95% CI) was 5.6 (5.5-5.8) mo with pembro + chemo vs 5.5 (5.4-5.6) mo with pbo + chemo; HR 0.80 (95% CI, 0.65-0.97); P = 0.0122; and the results did not reach statistical significance. Median (range) time from randomization to data cutoff at FA (Jan 17, 2023) was 42.0 (29.5-53.9) mo. At FA, median OS (95% CI) was 15.9 (13.7-18.8) vs 14.7 (12.7-17.1) mo. While the HR for OS (0.84 [95% CI, 0.69-1.02]; P = 0.0362) favored pembro + chemo vs pbo + chemo, it did not reach statistical significance. OS rates at 12-mo were 61.6% vs 59.4% and at 24-mo were 30.6% vs 26.4%. HR for OS was similar in PD-L1 TPS >= 50% (HR, 0.84) and TPS <50% groups (HR, 0.85). ORR (95% CI) in ITT was 29.0% (23.4%-35.1%) with pembro + chemo vs 27.1% (21.7%-33.1%) with pbo + chemo. Median DOR was 6.3 (2.3 to 40.8+) mo vs 5.6 (1.8+ to 40.6+) mo. Grade 3 treatment-related AEs occurred in 43.7% of pts in pembro + chemo arm and 38.6% in pbo + chemo arm; grade 5 AEs occurred in 0.4% vs 0.8%. Grade >= 3 immune-mediated AEs and infusion reactions occurred in 4.5% of pts in the pembro + chemo arm and 2.0% in the pbo + chemo arm; 0.4% vs 0% had grade 5 events. Conclusions: In the KEYNOTE-789 study, addition of pembro to chemo in pts with TKI-resistant, EGFR-mutant, metastatic nonsquamous NSCLC did not significantly prolong PFS and OS in comparison to pbo + chemo. AEs were manageable in both arms, and no new safety signals were identified. Clinical trial information: NCT03820986.