Prognostic Value of Inflammatory Biomarkers as Continuous Variables in Metastatic Renal Cell Carcinoma: Linearity Assessment, Proportional Hazards Assumption Testing, and Time‐Varying Effects

OBJECTIVES: To evaluate inflammatory biomarkers (neutrophil-to-lymphocyte ratio [NLR], C-reactive protein [CRP]) as continuous variables and assess linearity and the proportional hazards (PH) assumption, including time-varying effects, in metastatic renal cell carcinoma (mRCC) treated with first-line immune checkpoint inhibitor (ICI)-based combinations. METHODS: This multicenter retrospective study included 151 patients with mRCC. Log-transformed NLR and CRP were tested for linearity (quadratic terms) and PH (Schoenfeld residuals). Multivariable Cox regression was adjusted for International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) risk factor count and treatment regimen. Time-varying effects were assessed using marker × log(time) interactions and 12-month landmark analyses; incremental discrimination beyond IMDC was evaluated using Harrell's concordance index. RESULTS: Log-transformed NLR and CRP showed no statistically significant evidence of nonlinearity (p = 0.211 and p = 0.055, respectively). NLR showed evidence of PH violation for all outcomes, whereas CRP did not. On multivariable analysis, NLR remained associated with overall survival (OS; adjusted hazard ratio [HR] 2.10, 95% confidence interval 1.26-3.49, p = 0.005) and cancer-specific survival (adjusted HR 2.04, 1.20-3.47, p = 0.009). Landmark analysis suggested a stronger OS association during 0-12 months (HR 2.61, 1.49-4.59) than after 12 months (0.93, 0.37-2.34). Adding NLR to IMDC modestly improved OS discrimination (ΔC-index +0.038, bootstrap p = 0.052). CONCLUSIONS: Log-transformed inflammatory markers showed no statistically significant evidence of nonlinearity. NLR showed exploratory evidence of time-varying prognostic associations, with stronger discrimination during earlier follow-up.