P-285 Q-TWiST analysis for pembrolizumab plus chemotherapy versus chemotherapy as first-line treatment for patients with advanced esophageal cancer in the KEYNOTE-590 study

Abstract

In the phase 3 KEYNOTE-590 study (NCT03189719), first-line pembrolizumab plus chemotherapy provided statistically significant and clinically meaningful improvement in OS and PFS, maintained HRQoL, and was associated with a manageable safety profile versus chemotherapy in patients with locally advanced, unresectable or metastatic esophageal cancer or Siewert type 1 gastro-esophageal junction adenocarcinoma. The objective of this analysis was to assess Quality-adjusted Time Without Symptoms of disease progression or Toxicity of treatment (Q-TWiST) among patients receiving pembrolizumab + chemotherapy or chemotherapy in the KEYNOTE-590 study. Q-TWiST analysis divided survival time into 3 health states: time with grade 3+ toxicity before disease progression or death (TOX), time without symptoms or toxicities before progression (TWiST), and time from disease progression (REL). Q-TWiST was calculated as the restricted mean time spent in each health state, weighted by a health-state utility for that particular state, combined. Average utility weights for each health state were based on the pooled EQ-5D health utility score using base case utility weights of Utility[TWiST], 1.0; Utility[TOX], 0.5; and Utility[REL], 0.5. Treatment difference 95% CIs were generated using the non-parametric bootstrapping method. Data are reported for all randomized patients and patients with PD-L1 combined positive score (CPS) ≥10. Relative gains in Q-TWiST of ≥10% and ≥15% have been established as “clinically important” and “clearly clinically important,” respectively, in the literature. Data cutoff, July 2, 2020. In all patients at month 33, patients receiving pembrolizumab + chemotherapy had a 1.51 month (95% CI, 0.10-2.95) longer restricted mean time in TWiST (5.95 vs 4.44 months), 1.59 month (95% CI, 0.76-2.43) longer restricted mean time in TOX (3.71 vs 2.12 months), and a 0.01 month (95% CI, -1.93-1.85) shorter restricted mean time in REL (6.16 vs 6.18 months) versus chemotherapy. For the restricted mean Q-TWIST in all patients, the difference between the pembrolizumab + chemotherapy and chemotherapy arms favored the pembrolizumab + chemotherapy arm by 2.30 months (95% CI, 1.27-3.40; 10.89 vs 8.59 months) at month 33, representing a 18.1% relative Q-TWiST gain. In patients with PD-L1 CPS ≥10, those receiving pembrolizumab + chemotherapy had a 2.24 month (95% CI, 0.08-4.44) longer restricted mean time in TWiST (6.35 vs 4.11 months), 2.16 month (95% CI, 0.82-3.52) longer restricted mean time in TOX (4.39 vs 2.23 months), and a 0.19 month (95% CI, -2.44-2.99) longer restricted mean time in REL (5.98 vs 5.80 months) versus chemotherapy. For the restricted mean Q-TWIST in patients with PD-L1 CPS ≥10, the difference between the pembrolizumab + chemotherapy and chemotherapy arms favored the pembrolizumab + chemotherapy arm by 3.41 months (95% CI, 1.92-4.84; 11.54 vs 8.13 months) at month 33, representing a 28.1% relative Q-TWiST gain. Pembrolizumab + chemotherapy demonstrated clinically meaningful improvement in quality-adjusted survival based on Q-TWIST analyses compared with chemotherapy in patients with advanced esophageal cancer.