Impact on overall survival after adjusting for treatment crossover in first-line metastatic NSCLC trials for pembrolizumab plus chemotherapy.

Abstract

8596 Background: High crossover rates were observed in the chemotherapy alone arm in KEYNOTE(KN)-189 and KEYNOTE-407, attributable to protocol-permitted crossover to pembrolizumab monotherapy and subsequent anti-PD1/PD-L1 therapies per clinical practice. The objective was to assess the impact on overall survival (OS) by adjusting for treatment crossover. Methods: The analysis used pooled data from KN-189 global (NCT02578680; cut-off date: 08MAR2022) and Japan extension (NCT03950674; cut-off date: 07FEB2023), and pooled data from KN-407 global (NCT02775435; cut-off date: 23FEB2022) and China extension (NCT03875092; cut-off date: 10FEB2023). The survival time of participants (pts) in the chemotherapy alone arm who crossed over to pembrolizumab or subsequent anti-PD1/PD-L1 therapy was adjusted using a Two-Stage Estimation model (TSE) and Rank-Preserving Structural Failure Time model (RPSFT). Subsequently, a stratified Cox regression model was used to estimate the treatment effect of pembrolizumab+chemotherapy (P+C) relative to chemotherapy. Results: Overall, 646 and 669 patients were included in the analysis for KN-189 & KN-407 with a median follow-up of 64.6 months (range, 59.3-82.8) and 56.7 months (range, 50.1-69.3), respectively. In KN-189, adjusting OS for 56.3% pts who crossed over from the chemotherapy arm resulted in a larger treatment effect estimate for P+C relative to chemotherapy than observed in the unadjusted Intention-to-Treat (ITT) analysis. Similarly, for KN-407, after adjusting OS for 53.7% pts who crossed over, compared with chemotherapy alone, P+C showed more pronounced benefit than in the unadjusted ITT approach. The greater treatment effect for P+C was consistently demonstrated in both studies regardless of the PD-L1 TPS cutoff values. Conclusions: After adjusting for treatment crossover, pembrolizumab + chemotherapy demonstrated greater magnitude of OS benefit versus chemotherapy alone in both trials. TSE and RPSFT analyses support the robustness of these findings showing similar direction and magnitude of treatment effects across PD-L1 TPS cut-offs. Clinical trial information: NCT02578680 ; NCT03950674 ; NCT02775435 ; NCT03875092 . KN-189 KN-407 Population N Unadjusted ITT TSE RPSFT N Unadjusted ITT TSE RPSFT All comers 431 vs 215 0.60 (0.50; 0.72) 0.44 (0.33, 0.58) 0.46 (0.35, 0.61) 332 vs 337 0.65 (0.55, 0.76) 0.47 (0.35, 0.63) 0.52 (0.40, 0.67) TPS <1% 140 vs 66 0.51 (0.37, 0.70) 0.41 (0.27, 0.63) 0.42 (0.28, 0.63) 115 vs 121 0.75 (0.56, 0.99) 0.59 (0.35, 0.98) 0.64 (0.42, 0.98) TPS ≥1% 267 vs 132 0.65 (0.52, 0.82) 0.48 (0.32, 0.71) 0.50 (0.34, 0.73) 207 vs 209 0.60 (0.48, 0.74) 0.40 (0.28, 0.59) 0.45 (0.32, 0.63) TPS 1-49% 132 vs 61 0.66 (0.47, 0.92) 0.48 (0.27, 0.86) 0.54 (0.33, 0.88) 114 vs 124 0.59 (0.45, 0.79) 0.38 (0.22, 0.64) 0.45 (0.30, 0.70) TPS ≥ 50% 135 vs 71 0.66 (0.47, 0.93) 0.48 (0.26, 0.87) 0.48 (0.26, 0.87) 93 vs 85 0.63 (0.45, 0.88) 0.44 (0.24, 0.79) 0.46 (0.26, 0.81)