Early BMI increase as a predictive biomarker for immune checkpoint inhibitor ±chemotherapy efficacyin advanced NSCLC: Integrating clinical, genomic, and circulating proteomic data

Background: Baseline body mass index (BMI) has emerged as a prognostic factor for immune checkpoint inhibitor (ICI) in advanced NSCLC. However, with chemo-immunotherapy becoming standard of care, whether early BMI changes predict efficacy and differ by treatment regimen remains unclear. Identifying early predictive changes could inform timely interventions to improve treatment efficacy. Methods: This retrospective study included patients with advanced NSCLC treated with ICI +/- chemotherapy. Baseline and longitudinal BMI were collected. Baseline covariates included age, sex, histology, smoking status, ECOG PS, TMB, PD-L1 TPS, and therapy line. Multivariable logistic and Cox models assessed early BMI changes at 3, 6, and 9 weeks and objective response rate (ORR), progression-free survival (PFS), and overall survival (OS), adjusting for baseline BMI and covariates. Time-dependent analyses verified temporal associations. To understand molecular profiles associated with baseline BMI (≥25 vs. <25 kg/m²), we assessed circulating proteins (Olink, 2700 proteins) in 141 patients and genomic alterations (OncoPanel) in 728 patients, adjusting for age, sex, and smoking. Results: Among 1,110 patients (303 chemo-immunotherapy, 797 ICI monotherapy), early BMI increase independently predicted improved outcomes in both regimens, with stronger effects in chemo-immunotherapy. In chemo-immunotherapy, 1% BMI increase at 3 weeks predicted higher ORR (OR 1.48, 95% CI 1.15-2.03), PFS (HR 0.78, 95% CI 0.71-0.86), and OS (HR 0.77, 95% CI 0.69-0.86). Predictive value decreased over time (3, 6, 9 weeks), particularly in chemo-immunotherapy. Older patients with smoking history, lower TMB receiving chemo-immunotherapy were more likely to lose weight within 3 weeks. Genomically, KRAS, KMT2A, RASA1, FANCA, CTNNB1, and ATM mutations were enriched in higher BMI patients, while EGFR, PMS1, FH, TP53, and BAP1 mutations were more common in lower BMI patients. Circulating proteins (SSC4D, LEP, CDHR2 in higher BMI; FSHB, GHRL in lower BMI) showed sex-specific patterns and metabolic regulatory roles. Conclusions: Early BMI increase predicts improved ICI outcomes in advanced NSCLC, with weight monitoring offering opportunities to optimize efficacy. Baseline BMI-associated genomic and proteomic profiles provide complementary insights: circulating proteins establish chronic metabolic selective environments, while tumor mutations represent evolutionary outcomes.