Identification of aneuploidy-related gene signature to predict survival in head and neck squamous cell carcinomas

Background: To parse the characteristics of aneuploidy related riskscore (ARS) model in head and neck squamous cell carcinomas (HNSC) and their predictive ability on patient prognosis.Methods: Molecular subtyping of HNSC specimens was clustered by Copy Number Variation (CNV) data from The Cancer Genome Atlas (TCGA) dataset applying consistent clustering, followed by immune condition evaluation, differentially expressed genes (DEGs) analysis and DEGs function annotation. Weighted gene co expression network analysis (WGCNA), protein-protein interaction, Univariate Cox regression analysis, least absolute shrinkage and selection operator (LASSO) and stepwise multivariate Cox regression analysis were implemented to construct an ARS model. A nomogram for clinic practice was designed by rms package. Immunotherapy evaluation and drug sensitivity prediction were also carried out.Results: We stratified HNSC patients into three different molecular subgroups, with the best prognosis in C1 cluster among 3 clusters. C1 cluster displayed greatest immune infiltration status. The most DEGs between C1 and C2 groups, mainly enriched in cell cycle and immune function. We constructed a nine-gene ARS model (ICOS, IL21R, CCR7, SELL, CYTIP, ZAP70, CCR4, S1PR4 and CD79A) that effectively differentiates between high and low-risk patients. Patients in low ARS group showed a higher sensitivity to immunotherapy. A nomogram built by integrating ARS and clinic-pathological characteristics helped predict clinic survival benefit. Drug sensitivity evaluation found that 4/9 inhibitor drugs (MK-8776, AZD5438, PD-0332991, PHA-665752) acted on the cell cycle. Conclusions: We classified 3 molecular subtypes for HNSC patients and established an ARS prognostic model, which offered a prospective direction for prognosis in HNSC.