Abstract
OBJECTIVE: Esophageal cancer (EC) ranks among the most prevalent malignancies globally and represents a significant and growing public health burden. This study aimed to construct a prognostic model leveraging anoikis-related genes (ARGs) to predict patient survival and elucidate the immunological microenvironment in EC. The findings are anticipated to enhance prognostic accuracy and inform therapeutic strategies, ultimately improving patient outcomes and treatment efficacy. METHODS: A comprehensive analysis was conducted using 11 control samples and 159 EC samples obtained from The Cancer Genome Atlas (TCGA) database, alongside associated clinical features. A total of 794 ARGs were curated from GeneCards database. Functional enrichment analyses of EC-related differentially expressed ARGs were performed using Gene Ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG). Prognostic differential ARGs associated with EC were identified through univariate Cox regression analysis, while LASSO regression was employed to minimize overfitting and construct a robust risk prognostic model. The EC cohort was stratified into training and testing groups for model development and verification. Model performance was evaluated through risk curves, survival curves, time-dependent receiver operating characteristic (ROC) curves, ROC curves for the riskscore and clinical features, and independent prognostic analysis. A nomogram with high predictive accuracy was also developed to estimate the prognosis of EC patients. To assess the impact of the risk prognosis model on the immune microenvironment of EC, analyses included tumor microenvironment analysis, single-sample gene set enrichment analysis (ssGSEA), immune cell infiltration correlation analysis, and differential analysis of immune checkpoint expression. Drug sensitivity profiling was conducted to identify potential therapeutic agents for EC. Finally, the expression of selected ARGs was validated at the mRNA level in EC cell lines using real-time quantitative PCR (RT-qPCR). RESULTS: The ARG-based risk prognostic model was constructed incorporating four high-risk ARGs (CDK1, IL17A, FOXC2, and OLFM3) and two low-risk ARGs (PIP5K1C and MAPK1). This model demonstrated strong predictive accuracy for the survival outcomes of EC patients. Immune correlation analyses revealed that the high-risk group exhibited significantly lower immunological scores compared to the low-risk group. Notably, immune cells such as macrophages and mast cells were markedly downregulated in the high-risk group. Additionally, key immunological functions, including APC co-inhibition, parainflammation, Type I IFN Response, and Type II IFN Response, were significantly suppressed in the high-risk group. Eight immune checkpoint-related genes (TNFRSF25, TNFRSF14, CD70, TNFSF15, TMIGD2, CD160, TNFSF18, and HHLA2) displayed distinct expression differences between high- and low-risk groups. The nomogram developed from this model demonstrated high efficacy in predicting EC patient prognosis. Furthermore, six potential therapeutic agents for EC were identified: BIRB.0796, Camptothecin, CHIR.99021, Methotrexate, PF.4708671, and Vorinostat. Finally, the mRNA expression levels of ARGs were validated using RT-qPCR in EC cell lines. Compared to normal esophageal epithelial cells (NE-2), CDK1 and MAPK1 were significantly upregulated in two EC cell lines (KYSE-30 and KYSE-180). CONCLUSION: This study provides valuable insights into the prognostic outcomes and immune microenvironment of EC through the analysis of ARGs. Furthermore, several potential therapeutic agents for EC were identified, offering promising avenues for treatment. These findings hold significant potential for enhancing the survival outcomes of EC patients and provide meaningful guidance for clinical decision-making in managing this malignancy.