Identifying head and neck squamous cell carcinoma

Although several biomarkers can predict human papillomavirus (HPV) status in head and neck squamous cell carcinoma (HNSC), their accuracy is low. We assessed the HPV status of HNSC using a Scissor algorithm in conjunction with bulk sequencing and single-cell sequencing data. A Scissor scoring system was constructed for predicting HPV status using the gene set variation analysis algorithm. The biological functions and signaling pathways influencing the Scissor scoring model were described using correlation analysis and functional annotation. We used the Scissor algorithm to identify 1070 HPV-positive cells in single-cell sequencing data based on HPV status and then screened out genes that were significantly expressed in the Scissor-positive cells. We then constructed a Scissor score model using the 43 genes specifically expressed in Scissor-positive cells. The model had high accuracy in predicting HPV status in HNSC patients (area under curve in the discovery set: 86.6%, area under curve in the validation set: 96.1% and 79.8%). We also found that patients with higher Scissor scores had better outcomes, higher tumor mutational burden and tumor neoantigen burden, higher levels of immune checkpoint infiltration, and more cytotoxic T cell infiltration, and better outcomes with radiotherapy and immunotherapy. We constructed a predictive model that can accurately predict HPV status in patients with HNSC with greater accuracy than most current biomarkers.