Pan-Immune-Inflammation Value as an Independent Prognostic Marker in Patients with Brain Metastases

Jiacheng Li^1,2,3Menghan Liu⁴Yingtong Liu¹Zheran Liu^2,3Jiazhen Liu^1,2,3Yuping Xie^5*

• ¹Chengdu University of Traditional Chinese Medicine, Chengdu, China
• ²West China Hospital of Sichuan University Department of Biotherapy, Chengdu, China
• ³Sichuan Provincial Key Laboratory of Nuclear Physics and Medical Research, Sichuan University, Chengdu, China
• ⁴Sichuan Academy of Medical Sciences and Sichuan People's Hospital Department of Nephrology, Chengdu, China
• ⁵Department of Oncology, West China Fourth Hospital, Sichuan University, Chengdu, China

Background: Systemic inflammation and immune dysregulation are recognized as key determinants of cancer progression and survival. The pan-immune-inflammation value (PIV), a hematologic-based composite biomarker, may reflect the host's immune-inflammatory status. Its prognostic significance in brain metastases (BM), however, remains undefined. Methods: In a single-center retrospective cohort, 3,856 consecutive patients with radiologically confirmed BM diagnosed between 2013 and 2021 were included. PIV was calculated as neutrophil count × platelet count × monocyte count, divided by lymphocyte count. All blood cell counts were recorded in units of 10^9 cells per liter. Complete blood counts were taken within 7 days before the start of treatment. The optimal PIV cut-off, derived using maximally selected log-rank statistics, defined low and high PIV groups. OS was analyzed using multivariable Cox models adjusted for age, performance status, number of BM and extracranial metastases. A PIV-augmented GPA nomogram was developed and internally validated with bootstrap resampling. Time-dependent concordance indices, calibration and integrated discrimination improvement (IDI) were used to assess model performance. Subgroup and sensitivity analyses examined robustness across systemic and local treatment modalities, primary tumor types, sex and alternative PIV parameterizations. Results: The PIV cut-off separated 1,570 patients with low PIV and 2,286 with high PIV. High PIV was associated with worse OS and remained independently prognostic (hazard ratio 1.40; 95% This is a provisional file, not the final typeset article confidence interval 1.29–1.52; p < 0.001), with consistent effects across treatment modalities, primary tumor types and sex. Alternative cut-offs and modeling PIV as a continuous variable (per 1-standard-deviation increase) produced effect estimates similar to the primary analysis. Adding PIV to the GPA modestly improved discrimination and increased IDI by 0.010 (95% confidence interval 0.006–0.015; p < 0.001); the PIV+GPA nomogram showed good 1-year calibration. Conclusions: PIV is an independent prognostic factor for OS in BM patients. Incorporating this marker into the Graded Prognostic Assessment modestly improves risk stratification and supports an accessible nomogram for individualized survival prediction. External prospective validation, including longitudinal assessment of the pan-immune-inflammation value and integration with molecular and imaging markers, is needed before routine clinical implementation.