Objectives: IO agents that block the programmed cell death protein 1 (PD-1) and programmed death-ligand 1 (PD-L1) have been approved in monotherapy or combination settings for the treatment of mNSCLC. The objectives of this work were to: (1) Develop MBMAs based on objective response rate (ORR), median progression free survival (mPFS), median overall survival (mOS), and Grade ≥3 treatment-related adverse events (TRAEs Gr3+); and (2) Draw a comparative treatment landscape and evaluate the tradeoffs between clinical benefit and risk amongst approved agents.
Methods: MBMA of ORR, mPFS, mOS and TRAEs Gr3+ were conducted using a database containing results from 50, 54, 56 and 47 randomized controlled and single-arm studies (published 2015-2022) with approved PD-(L)1 agents representing 15,486, 18,095, 18,700 and 19,583 mNSCLC patients, respectively. Treatments were grouped into 3 categories – IO monotherapy, Chemotherapy (CT) and IO+CT. Models were developed to quantify drivers of variability on outcomes across trials (i.e., treatment, covariates, random effects). Tumor cells PD-L1 expression ( < 1%: Negative; 1-49%: Low; >=50%: High), line of therapy, ECOG PS, histology, geography, sex, age, and smoking status were evaluated as potential covariates.
Results: PD-L1 expression and line of treatment were identified as significant covariates for IO monotherapy, while histology and race were significant for both IO mono- and IO+CT combo-therapy. Low PD-L1 expression was associated with a lower ORR (ratio [95%]: 0.33 [0.28, 0.40]), mPFS (ratio [95%]: 0.61 [0.51, 0.73]) and mOS (ratio [95%]: 0.73 [0.62, 0.86]) than high PD-L1 expression for IO monotherapy. Regardless of PD-L1 expression, treatment-experienced populations had lower ORR (ratio [95% CI]: 0.64 [0.52, 0.78]), mPFS (ratio [95% CI]: 0.653 [0.57, 0.74]) and mOS (ratio [95% CI]: 0.780 [0.70, 0.86]) than treatment-naïve populations for IO monotherapy. Populations with squamous histology were associated with lower mPFS (ratio [95% CI]: 0.84 [0.77, 0.91]) and mOS (ratio [95% CI]: 0.73 [0.68, 0.79]) than non-squamous populations. Asian majority populations had higher mOS than non-Asian majority populations (ratio [95% CI]: 1.26 [1.15, 1.39]). Irrespective of PD-L1 expression, similar efficacy was observed across IO within both mono- and combination-therapy after adjusting for covariates. No clinically meaningful covariates were identified for TRAEs Gr3+. Regardless of PD-L1 expression, fewer patients experienced TRAEs Gr3+ with IO alone (14%) compared to CT alone (39%) or IO+CT (48%). In patients with high PD-L1 expression, mOS was comparable for IO with or without CT. In patients with low PD-L1 expression, mOS was higher for IO+CT (16 months) than IO (12 months).
Conclusions: These results are consistent across all the efficacy and safety endpoints tested and in line with treatment guidelines for mNSCLC patients, highlighting the potential utility of this work.
