Head and neck squamous cell carcinoma (
HNSCC) is one of the most aggressive
malignancies with complex phenotypic, etiological, biological, and clinical heterogeneities. Previous studies have proposed different clinically relevant subtypes of
HNSCC, but little is known about its corresponding prognosis or suitable treatment strategy. Here, we identified 101 core genes from three prognostic pathways, including
mTORC1 signaling, unfold
protein response, and UV response UP, in 124 pairs of
tumor and matched normal tissues of
HNSCC. Moreover, we identified three robust subtypes associated with distinct molecular characteristics and clinical outcomes using consensus clustering based on the gene expression profiles of 944
HNSCC patients from four independent datasets. We then integrated the genomic information of The
Cancer Genome Atlas (TCGA)
HNSCC cohort to comprehensively evaluate the molecular features of different subtypes and screen for potentially effective therapeutic agents. Cluster 1 had more arrested oncogenic signaling, the highest immune cell infiltration, the highest
immunotherapy and chemotherapeutic responsiveness, and the best prognosis. By contrast, Cluster 3 showed more activated oncogenic signaling, the lowest immune cell infiltration, the lowest
immunotherapy and
chemotherapy responsiveness, and the worst prognosis. Our findings corroborate the molecular diversity of
HNSCC tumors and provide a novel classification strategy that may guide for prognosis and treatment allocation.