Abstract
The current understanding of the prognostic significance of natural killer (NK) cells and their tumor microenvironment (TME) in hepatocellular carcinoma (HCC) is limited. Thus, we screened for NK-cell-related genes by single-cell transcriptome data analysis and developed an NK-cell-related gene signature (NKRGS) using multi-regression analyses. Patients in the Cancer Genome Atlas cohort were stratified into high- and low-risk groups according to their median NKRGS risk scores. Overall survival between the risk groups was estimated using the Kaplan-Meier method, and a NKRGS-based nomogram was constructed. Immune infiltration profiles were compared between the risk groups. The NKRGS risk model suggests significantly worse prognoses in patients with high NKRGS risk (p < 0.05). The NKRGS-based nomogram showed good prognostic performance. The immune infiltration analysis revealed that the high-NKRGS-risk patients had significantly lower immune cell infiltration levels (p < 0.05) and were more likely to be in an immunosuppressive state. The enrichment analysis revealed that immune-related and tumor metabolism pathways highly correlated with the prognostic gene signature. In this study, a novel NKRGS was developed to stratify the prognosis of HCC patients. An immunosuppressive TME coincided with the high NKRGS risk among the HCC patients. The higher KLRB1 and DUSP10 expression levels correlated with the patients' favorable survival.