Baseline and Longitudinal Pan-Immune-Inflammation Value Compared with Other Inflammatory Indices in Patients Receiving Immune Checkpoint Inhibitors: A Retrospective Cohort Study.

Abstract

Objectives: The pan-immune-inflammation value (PIV), calculated from neutrophil, platelet, monocyte, and lymphocyte counts, is a peripheral blood biomarker of systemic immune-inflammatory status. We evaluated the prognostic value of baseline PIV and its trajectory over the first six months of treatment for overall survival (OS) and progression-free survival (PFS) in patients receiving immune checkpoint inhibitors (ICIs) and benchmarked PIV against the neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and systemic immune-inflammation index (SII).

Methods: We retrospectively analyzed 205 patients treated at a single tertiary center for non-small-cell lung cancer, renal cell carcinoma, melanoma, or other solid tumors who received an ICI (90.2% nivolumab). PIV was measured at baseline, month 3, and month 6. The outcome-independent median value (587.7) defined the PIV-low and PIV-high groups. Cox regression (per doubling of log2-transformed PIV), Kaplan-Meier analysis, 3- and 6-month landmark models, time-updated Cox models, and restricted cubic splines were used, with adjustment for age, sex, and tumor type. Discrimination was assessed using Harrell's C-index, 1000-resample bootstrap internal validation, and time-dependent area under the curve (AUC).

Results: The median follow-up was 50.6 months; 168 deaths and 190 progression-or-death events occurred. Each doubling of baseline PIV was independently associated with worse OS (adjusted HR, 1.24; 95% confidence interval [CI], 1.14-1.36; p<0.001) and PFS (HR, 1.18; 95% CI, 1.09-1.28; p<0.001). Median OS was 11.96 vs. 5.62 months, and median PFS was 6.21 vs. 3.45 months, in the PIV-low vs. PIV-high groups (log-rank p<0.001 for both). Month-3 and month-6 absolute PIV values remained prognostic in landmark analyses, whereas the relative change in PIV from baseline to month 3 was not associated with either outcome. The association was approximately log-linear (nonlinearity p>0.05 for OS and PFS). PIV, NLR, PLR, and SII showed similar discrimination (adjusted C-index for OS: 0.650, 0.647, 0.633, and 0.652, respectively); PIV had the lowest Akaike information criterion (AIC) and Bayesian information criterion (BIC) values among the four indices.

Conclusion: Baseline and on-treatment absolute PIV values, but not early relative changes in PIV, are independently associated with survival outcomes in patients receiving ICIs, with prognostic performance comparable to that of established inflammatory indices. PIV may serve as a simple, repeatable biomarker to support risk stratification during ICI therapy.