Early dynamic immune-inflammatory and nutritional recovery signature is associated with durable benefit and early progression in advanced non-small cell lung cancer receiving first-line PD-1/PD-L1 inhibitor-based therapy: a real-world retrospective cohort study.

Background: Durable benefit from first-line PD-1/PD-L1 inhibitor-based therapy remains heterogeneous in advanced non-small cell lung cancer (NSCLC). Baseline biomarkers, including PD-L1 tumor proportion score, tumor mutational burden, and static inflammatory indices, only partly capture this heterogeneity and may not reflect host response. We evaluated whether a cycle 2 immune-inflammatory and nutritional recovery signature from blood markers was associated with durable clinical benefit (DCB), early progression, and survival. Methods: This single-center retrospective cohort included 420 patients with advanced NSCLC treated with first-line PD-1/PD-L1 inhibitor-based therapy at Chengdu Third People's Hospital between January 2019 and September 2025, with data cutoff on March 31, 2026. The primary analytic cohort comprised 409 patients with a rule-derived cycle 2 dynamic recovery signature: favorable recovery, intermediate, or persistent failure. DCB at 6 months and early progression within 3 months were the primary endpoints. Associations were estimated using Firth-type penalized logistic regression and Cox models adjusted for clinical covariates and baseline NLR, PNI, and ALI. Model performance was assessed using discrimination, calibration, decision curves, and bootstrap optimism correction. Results: Among 409 classifiable patients, 141 (34.5%), 115 (28.1%), and 153 (37.4%) had favorable recovery, intermediate, and persistent failure signatures. DCB occurred in 61.7, 43.5, and 33.3%, and early progression occurred in 11.3, 18.3, and 23.5%, respectively. Compared with favorable recovery, persistent failure was associated with lower DCB (adjusted odds ratio [OR], 0.34; 95% CI, 0.21-0.56), higher early progression (adjusted OR, 2.15; 95% CI, 1.13-4.07), and shorter progression-free survival (adjusted hazard ratio [HR], 3.34; 95% CI, 2.40-4.64). Its adjusted association with overall survival was weaker (HR, 1.22; 95% CI, 0.86-1.72). The clinical plus dynamic signature model had optimism-corrected AUCs of 0.619 for DCB and 0.562 for early progression. The composite signature did not outperform individual continuous dynamic biomarkers. Conclusion: A rule-derived cycle 2 immune-inflammatory and nutritional recovery signature was independently associated with DCB and progression-free survival, but discrimination was modest, early-progression performance was weak, and no robust adjusted overall survival association was observed. It should be interpreted as an exploratory risk-stratification summary, not a clinically ready prediction tool. External validation and prospective evaluation are required before clinical use.