Abstract 41P
Background
The efficacy of PD-1 inhibitors in pts with NSCLC may be related to the proportion of pts with negative (< 1%TPS/TPC) or high (≥ 50% TPS/TPC) PD-L1 expression. Studies included into NMA were quite heterogenous in proportion of PD-L1-negative patients and hyperexpressors. Systematic reviewand Bayesian individual patient data NMA (PROSPERO ID CRD42024521004) were conducted to assess modifying effect of PD-L1 expression on OS in pts with advanced nsNSCLC between chemotherapy (CT), immunotherapy (PD-1 inhibitors ± CTLA-4 inhibitors), targeted therapy and their combinations.
Methods
We assessed clinical heterogeneity and the risk of bias for studies included into NMA. Multilevel network meta-regression (ML NMR) with an interaction between treatment class and proportion of PDL1-negative pts was estimated.
Results
All studies in NMA were randomized with a low risk of bias in this domain (except for the Checkmate 227 study with an undetermined risk), in ML NMR randomization was respected by stratifying the baseline hazard by study. Hazard ratio for treatment effect modification by PD-L1 negative status is 0.92 (95% CrI, 0.51, 1.66), treatment rankings by OS are almost insensitive to the share of PD-L1-negative pts, and Prolgolimab+CT and Nivolumab+Bevacizumab+CT are the best options with the same OS profile for both PD-L1-negative and PD-L1-positive ones.
Conclusions
PD-L1 expression does not modify the effect of immunotherapy over CT ± Bevacizumab on the OS. Prologlimab+CT and Nivolumab+Bevacizumab+CT have benefit in the OS over all other treatment options irrespective to the proportion of pts with positive PD-1 expression.
Legal entity responsible for the study
The authors.
Funding
Has not received any funding.
Disclosure
All authors have declared no conflicts of interest.