In this study, we innovatively established the ICI profile of ESCC by mining GEO and TCGA databases and discovered a robust sixteen-gene prognostic signature from the ICI differences. Specifically, our novelty lay in using ICI profile, differential expression gene identification, univariate Cox models, and Lasso regression in the model training phase.
The adoption of an independent TCGA dataset, Kaplan-Meier analysis, Cox regression, ROC curve, IAUC, and IBS in the validation process, moreover, highlighted our innovativeness. Most importantly, we compared our signature with published research to prove ours' superiority.
At the end of the study, we discovered important mechanisms related to gene signature through GSEA, immune gene correlation analysis, and immune infiltration analysis and speculated that resting mast cells might potentially help the predictive ability of the signature.
Esophageal squamous cell carcinoma (ESCC) is a life-threatening thoracic tumor with a poor prognosis. The tumor microenvironment (TME) mainly comprises tumor cells and tumor-infiltrating immune cells mixed with stromal components. The latest research has displayed that tumor immune cell infiltration (ICI) is closely connected with the ESCC patients' clinical prognosis. This study was designed to construct a gene signature based on the ICI of ESCC to predict prognosis.
Based on the selection criteria we set, the eligible ESCC cases from the GSE53625 and TCGA-ESCA datasets were chosen for the training cohort and the validation cohort, respectively. Unsupervised clustering detailed grouped ESCC cases of the training cohort based on the ICI profile. We determined the differential expression genes (DEGs) between the ICI clusters, and, subsequently, we adopted the univariate Cox analysis to recognize DEGs with prognostic potential. These screened DEGs underwent a Lasso regression, which then generated a gene signature. The harvested signature's predictive ability was further examined by the Kaplan-Meier analysis, Cox analysis, ROC, IAUC, and IBS. More importantly, we listed similar studies in the most recent year and compared theirs with ours. We performed the functional annotation, immune relevant signature correlation analysis, and immune infiltrating analysis to thoroughly understand the functional mechanism of the signature and the immune cells’ roles in the gene signature's predicting capacity.
A sixteen-gene signature (ARSD, BCAT1, BIK, CLDN11, DLEU7-AS1, GGH, IGFBP2, LINC01037, LINC01446, LINC01497, M1AP, PCSK2, PCSK5, PPP2R2A, TIGD7, and TMSB4X) was generated from the Lasso model. We then confirmed the signature as having solid and stable prognostic capacity by several statistical methods. We revealed the superiority of our signature after comparing it to our predecessors, and the GSEA uncovered the specifically mechanism of action related to the gene signature. Two immune relevant signatures, including GZMA and LAG3 were identified associating with our signature. The immune-infiltrating analysis identified crucial roles of resting mast cells, which potentially support the sixteen-gene signature's prognosis ability.
We discovered a robust sixteen-gene signature that can accurately predict ESCC prognosis. The immune relevant signatures, GZMA and LAG3, and resting mast cells infiltrating were closely linked to the sixteen-gene signature's ability.Le texte complet de cet article est disponible en PDF.
Keywords : Esophageal squamous cell carcinoma, Immune infiltration, Tumor microenvironment, Gene signature, Biomarker