Abstract
Background: Head and neck squamous cell carcinoma's tumor immune microenvironment (TIME) plays an important role in tumorigenesis and progression, but its clinical significance remains unclear. Therefore, the TIME needs to be better understood in order to improve the response of diagnosis and therapy. Methods: The gene expression and clinical data of 569 HNSCC patients were obtained from The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO). Immune-related genes (IRGs) from the ImmPort database were used for immunotyping of HNSCC patients, and independent GEO datasets were used for subtype verification and comprehensive molecular identification. Results: The patients were divided into three subtypes (C1, C2, and C3) related to different gene expression profiles. The three subtypes showed widely different patterns in tumor genetic distortion, immune cell composition, cytokine profile, and so on, verifying that the immune-enhanced C2 subtype was associated with better prognosis. In addition, the stroma-deficient C1 subtype may be more efficient for the immune response than the C3 subtype. Furthermore, using WGCNA on the IRGs of those three subtypes, we found two C2-positive gene modules closely related to infection- and immune-associated pathways in the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database, and the two modules had 22 common pathways. Conclusion: This study improves the power for prognosis prediction and develops new therapeutic strategies to stratify HNSCC patients into clinically significant groups through TIME-related prognostic signature.