Pembrolizumab (P) was a common first-line therapeutic option for advanced non-small cell lung cancer (aNSCLC) with a PD-L1 tumor percentage score (TPS) of more than or equal to 50% either used alone or in conjunction with platinum-based chemotherapy (PCT). There were no predictive biomarkers available to help with therapy selection. In the computerized databases of 5 Israeli cancer centers, 423 patients with EGFR/ALK/ROS1-wild-type PD-L1 TPS of more than or equal to 50% aNSCLC received P (n=302) or PCT (n=121) as a first-line treatment. Overall survival (OS, months [mo]) was compared with blood biomarkers (BB: NLR, dNLR, PLR, SII, LIPI, ALI) to produce a prediction score. mOS was 17.2mo (95% CI, 13.2-36.5) and 21.3mo (95% CI, 14.8-NR) in groups P and PCT, respectively, in the propensity score matching analysis (n=236; 118 patients in each group matched for age, sex, and ECOG PS). In univariate and multivariate COX regression analysis, NLR, dNLR, PLR, LIPI, and ALI were substantially linked with OS in group P (P<.05) However, none of the BB in group P (group A) PCT showed a significant association. Age of more than or equal to 65, female sex, never-smoking status, adenocarcinoma histology, and dNLR of more than or equal to 3 was used to create a prediction score (each attribute earning one point). OS was considerably longer with PCT compared with P: mOS NR (95% CI, 15.3-NR) and 8.7mo (95% CI, 5.8-13.7) (P=.0005) in patients with prediction score 3-5, but not in patients with predictive score 0-2 (P =.61). The suggested dNLR-based Score appeared to predict OS with P and PCT, despite the constraints of the retrospective analysis.