Development and Validation of Ischemic Events Related Signature After Carotid Endarterectomy

Abstract

Background: Ischemic events after carotid endarterectomy (CEA) in carotid artery stenosis patients are unforeseeable and alarming. Therefore, we aimed to establish a novel model to prevent recurrent ischemic events after CEA.Methods: Ninety-eight peripheral blood mononuclear cell samples were collected from carotid artery stenosis patients. Based on weighted gene co-expression network analysis, we performed whole transcriptome correlation analysis and extracted the key module related to ischemic events. The biological functions of the 292 genes in the key module were annotated via GO and KEGG enrichment analysis, and the protein-protein interaction (PPI) network was constructed via the STRING database and Cytoscape software. The enrolled samples were divided into train (n = 66), validation (n = 28), and total sets (n = 94). In the train set, the random forest algorithm was used to identify critical genes for predicting ischemic events after CEA, and further dimension reduction was performed by LASSO logistic regression. A diagnosis model was established in the train set and verified in the validation and total sets. Furthermore, fifty peripheral venous blood samples from patients with carotid stenosis in our hospital were used as an independent cohort to validation the model by RT-qPCR. Meanwhile, GSEA, ssGSEA, CIBERSORT, and MCP-counter were used to enrichment analysis in high- and low-risk groups, which were divided by the median risk score.Results: We established an eight-gene model consisting of PLSCR1, ECRP, CASP5, SPTSSA, MSRB1, BCL6, FBP1, and LST1. The ROC-AUCs and PR-AUCs of the train, validation, total, and independent cohort were 0.891 and 0.725, 0.826 and 0.364, 0.869 and 0.654, 0.792 and 0.372, respectively. GSEA, ssGSEA, CIBERSORT, and MCP-counter analyses further revealed that high-risk patients presented enhanced immune signatures, which indicated that immunotherapy may improve clinical outcomes in these patients.Conclusion: An eight-gene model with high accuracy for predicting ischemic events after CEA was constructed. This model might be a promising tool to facilitate the clinical management and postoperative surveillance of carotid artery stenosis patients.

T cell-mediated tumor killing patterns in head and neck squamous cell carcinoma identify novel molecular subtypes, with prognosis and therapeutic implications

“…Firstly, to screen for differentially expressed genes (DEGs) in tumor and normal samples, the ‘ edgeR’ R package was utilized. P values <0.05 and abs (log2 fold change) >1is defined as the significance threshold [ 19 ]. Secondly, univariate Cox regression analysis was used to screen genes with independent prognosis significance for HNSCC patients (P <0.05).…”

A machine learning-derived gene signature for assessing rupture risk and circulatory immunopathologic landscape in patients with intracranial aneurysms

“…Weighted gene co-expression network analysis (WGCNA), an advanced bioinformatics method capable of identifying gene interrelationships and gene-phenotype relationships, was utilized to recognize gene co-expression modules intimately related to IA rupture. Similar to previous studies (21,22), top 5,000 genes selected according to the median absolute deviation were employed to construct co-expression networks based on weighted correlation adjacency matrix and cluster analysis via the WGCNA R package (23). An appropriate soft threshold β was calculated via the "pickSoftThreshold" function to meet the criteria for scale-free network and construct a weighted adjacency matrix.…”

Identification of biomarkers and analysis of infiltrated immune cells in stable and ruptured abdominal aortic aneurysms

“…The gene set files used in this study were downloaded from the Molecular Signatures Database version 7.5.1 4 . The enrichment scores of GO and KEGG pathways terms in each group were calculated using the gene set enrichment analysis (GSEA) software (version 4.2.3) (22), and terms enriched in the AAA or aortic rupture group were identified. A nominal p-value of < 0.05 and false-discovery rate q value of < 0.25 were considered as significantly enriched in the AAA or aortic rupture group.…”