Establishment and Validation of a Novel Nutritional-Immune-Inflammatory Score Model for Predicting Survival Prognosis in Hepatocellular Carcinoma Patients Treated with PD-1 Inhibitors.

Abstract

Purpose: Immune checkpoint inhibitors, particularly PD-1 inhibitors, are widely used in hepatocellular carcinoma therapy, many received PD-1 inhibitors beyond first-line, but heterogeneous treatment responses require reliable biomarkers. The interaction of immune function, nutritional status, and inflammatory responses affects tumor progression and survival, yet their prognostic value in PD-1 inhibitor-treated HCC patients remains unclear. This study developed a novel nutritional-immune-inflammatory score (NIIS) to evaluate its prognostic value in HCC patients receiving PD-1 inhibitors.

Patients and methods: We analyzed 355 HCC patients treated with PD-1 inhibitors (training: n=249; validation: n=106), the cohort included 18.6% Child-Pugh B patients. Fourteen nutritional, immune, and inflammatory biomarkers were evaluated. Prognostic indicators were selected via univariate and LASSO Cox regression. The NIIS was constructed and validated for OS prediction. A nomogram integrating the NIIS with clinical variables was developed and validated based on calibration curves, AUC, and DCA, and compared with the BCLC staging system. The primary outcome assessed was OS from the initiation of PD-1 inhibitor therapy in HCC patients.

Results: The NIIS (ALRI, APRI, PALBI, AAPR) showed strong prognostic stratification. High-risk patients had shorter OS (training: P =1.764×10^-8; verification: P=2.775×10^-6). Higher NIIS were significantly associated with advanced tumor stage, poor liver function grade, multiple and larger tumors, tumor thrombus, vascular invasion, and elevated AFP levels (P<0.05). Multivariate Cox analysis confirmed the NIIS as an independent prognostic factor for OS (training: HR=1.565, 95% CI: 1.273-1.925; verification: HR=1.341, 95% CI: 1.065-1.687). A nomogram integrating the NIIS with clinical variables was constructed for individualized prognosis prediction, demonstrating superior predictive performance compared to the conventional BCLC staging system.

Conclusion: The NIIS and nomogram provide a clinically useful tool for risk stratification in HCC immunotherapy, this model outperforming conventional staging systems and may optimize patient selection for PD-1 inhibitor therapy. Prospective multicenter studies are warranted to validate its generalizability.