BACKGROUND

Therapies targeting immune molecules have rapidly been adopted and advanced the treatment of hepatocellular carcinoma (HCC). Nonetheless, no studies have reported a systematic analysis between immunological profiles and clinical significance in HCC.

METHODS

We comprehensively investigated immune patterns and systematically correlated 22 types of both adaptive and innate immune cells with genomic characteristics and clinical outcomes based on 370 HCC patients from The Cancer Genome Atlas (TCGA) database through a metagene approach (known as CIBERSORT). Based on the coupled with integrated high-dimensional bioinformatics analysis, we further independently validated six immune subsets (CD4 T cells, CD8 T cells, CD20 B cells, CD14 monocytes, CD56 NK cells, and CD68 macrophages), and shortlisted three (CD4 T cells, CD8 T cells, and CD56 NK cells) of which to investigate their association with clinical outcomes in two independent Zhongshan cohorts of HCC patients (n = 258 and n = 178). Patient prognosis was further evaluated by Kaplan-Meier analysis and univariate and multivariate regression analysis.

RESULTS

By using the CIBERSORT method, the immunome landscape of HCC was constructed based on integrated transcriptomics analysis and multiplexed sequential immunohistochemistry. Further, the patients were categorized into four immune subgroups featured with distinct clinical outcomes. Strikingly, significant inter-tumoral and intra-tumoral immune heterogeneity was further identified according to the in-depth interrogation of the immune landscape.

CONCLUSION

This work represents a potential useful resource for the immunoscore establishment for prognostic prediction in HCC patients.