1. Progressive and Prognostic Performance of an Extracellular Matrix-Receptor Interaction Signature in Gastric Cancer
- Author
-
Liping Chen, Muqing He, Yuting Mao, Zijing Hu, and Xiangchou Yang
- Subjects
Male ,0301 basic medicine ,Oncology ,Medicine (General) ,medicine.medical_specialty ,Article Subject ,Angiogenesis ,Clinical Biochemistry ,Kaplan-Meier Estimate ,medicine.disease_cause ,Disease-Free Survival ,Extracellular matrix ,Correlation ,03 medical and health sciences ,R5-920 ,0302 clinical medicine ,Text mining ,Stomach Neoplasms ,Internal medicine ,Biomarkers, Tumor ,Genetics ,medicine ,Humans ,Gene Regulatory Networks ,Molecular Biology ,Gene ,Neoplasm Staging ,Framingham Risk Score ,Cell adhesion molecule ,business.industry ,Gene Expression Profiling ,Biochemistry (medical) ,General Medicine ,Prognosis ,Extracellular Matrix ,Gene Expression Regulation, Neoplastic ,030104 developmental biology ,030220 oncology & carcinogenesis ,Disease Progression ,Female ,Carcinogenesis ,business ,Research Article - Abstract
The role of an extracellular matrix- (ECM-) receptor interaction signature has not been fully clarified in gastric cancer. This study performed comprehensive analyses on the differentially expressed ECM-related genes, clinicopathologic features, and prognostic application in gastric cancer. The differentially expressed genes between tumorous and matched normal tissues in The Cancer Genome Atlas (TCGA) and validation cohorts were identified by a paired t -test. Consensus clusters were built to find the correlation between clinicopathologic features and subclusters. Then, the least absolute shrinkage and selection operator (lasso) method was used to construct a risk score model. Correlation analyses were made to reveal the relation between risk score-stratified subgroups and clinicopathologic features or significant signatures. In TCGA (26 pairs) and validation cohort (134 pairs), 25 ECM-related genes were significantly highly expressed and 11 genes were downexpressed in gastric cancer. ECM-based subclusters were slightly related to clinicopathologic features. We constructed a risk score model = 0.081 ∗ log 2 CD 36 + 0.043 ∗ log 2 COL 5 A 2 + 0.001 ∗ log 2 ITGB 5 + 0.039 ∗ log 2 SDC 2 + 0.135 ∗ log 2 SV 2 B + 0.012 ∗ log 2 THBS 1 + 0.068 ∗ log 2 VTN + 0.023 ∗ log 2 VWF . The risk score model could well predict the outcome of patients with gastric cancer in both training ( n = 351 , HR: 1.807, 95% CI: 1.292-2.528, P = 0.00046 ) and validation ( n = 300 , HR: 1.866, 95% CI: 1.347-2.584, P = 0.00014 ) cohorts. Besides, risk score-based subgroups were associated with angiogenesis, cell adhesion molecules, complement and coagulation cascades, TGF-beta signaling, and mismatch repair-relevant signatures ( P < 0.0001 ). By univariate (1.845, 95% CI: 1.382-2.462, P < 0.001 ) and multivariate (1.756, 95% CI: 1.284-2.402, P < 0.001 ) analyses, we regarded the risk score as an independent risk factor in gastric cancer. Our findings revealed that ECM compositions became accomplices in the tumorigenesis, progression, and poor survival of gastric cancer.
- Published
- 2020
- Full Text
- View/download PDF