A pyroptosis score-based model for prognosis prediction of esophageal squamous cell carcinoma

Abstract

Abstract Background: Emerging evidences have demonstrated the dual effect of pyroptosis in cancer progression, pro- or anti-tumorigenic roles in differential cancer types. This study aimed to identify prognostic molecules in esophageal squamous cell carcinoma (ESCC) based on the bioinformatic analysis of pyroptosis-related genes (PRGs) expression profiles. Method:Firstly, the pyroptosis gene sets were collected from the MSigDB database, and the mRNA levels were detected in ESCC and normal tissue samples from TCGA. Consensus cluster analysis was conducted to classify patients into different clusters that varied significantly in overall survival times. A pyroptosis score (PYS)- mediate model was generated and the property was assessed in the validation set. Results: Here, we found sixteen PRGs were dysregulated in ESCC samples. The subtyping of ESCC based on differential expressed PRGs could classify patients into three clusters. A prognostic gene signature comprising two genes (HLA-E and TNFSF10) in the three clusters was established, and samples could be classified to two PYS groups. The PYS model displayed excellent performance in prognostic prediction both in the training and validation set. Patients in high PYS group implied universal inflammatory cell infiltration and the better response of paclitaxel and methotrexate (P < 0.05). Conclusion: A PYS-based risk model comprising two genes that could predict the prognosis of ESCC patients. Our finding demonstrated the correlation between pyroptosis, survival outcome, immunity, and chemotherapy response in ESCC.