Identification of disulfidptosis in esophageal squamous cell carcinoma based on single-cell and bulk RNA-seq data to predict prognosis and treatment response

Abstract

Abstract Purpose Our study aims to identify the molecular subtypes of genes associated with disulfidptosis in ESCC, construct a scoring model to explore the differences in tumor growth behavior and find novel potential therapeutic targets. Methods Consensus cluster analysis was performed based on the GSE53625 dataset. The prognostic signature was constructed using univariate, multivariate, and Lasso-Cox regression analysis. The TCGA-ESCC dataset and single-cell RNA-seq data from the GSE160269 dataset was combined with trajectory analysis to analyze the prognostic signature. Additionally, the differences in tumor growth patterns, immune microenvironment, and cellular communication were explored, immunotherapy effects were predicted between high- and low-score groups, and potential therapeutic strategies were investigated to provide ideas for follow-up studies. Results We identified two distinct patterns of disulfidptosis expression with significant differences in overall survival. Then, we constructed the prognostic signature of disulfidptosis, and results showed patients with high score had worse prognosis. Univariate and multivariate Cox analysis demonstrated that the constructed prognostic signature was an independent prognostic factor and was validated in an independent validation set. The two subgroups differed in the proportion of immune cell infiltration and related signaling pathways in ESCC. The exploration of immunotherapy data confirmed our prognostic signature also had certain predictive power for immunotherapy. Regarding drug prediction, the results suggested the EGFR inhibitor had a stronger inhibitory effect on the low-score group. Conclusion This study provides a new prognostic signature for ESCC, explores new therapeutic targets, and provides new theoretical support for personalized treatment.