The association of POSTN with postoperative recurrence risk in early-stage lung adenocarcinoma: From gene networks to cellular functions
Abstract
Purpose Periostin (POSTN) has been identified as a biomarker highly correlated with the risk of recurrence of lung adenocarcinoma (LUAD). Increasing the understanding of POSTN’s role in LUAD recurrence may facilitate the development of more effective clinical interventions. Methods Firstly, Weighted Gene Co-expression Network Analysis (WGCNA) analysis was used to identify key genes related to LUAD disease recurrence. Subsequently, the diagnostic performance of these genes was evaluated using Receiver Operating Characteristic (ROC) curves and validated using external datasets. The most effective genes were then analyzed using single- and multivariate Cox regression, linear regression, and stratified analysis to identify their clinical effects. Threshold effect analysis was used to identify the optimal threshold for POSTN expression, and an in vitro experiment was conducted to evaluate the effect of POSTN on tumor cell growth and invasion. Results WGCNA analysis identified four genes strongly associated with LUAD recurrence. POSTN demonstrated good prediction performance for disease recurrence in both The Cancer Genome Atlas (TCGA) and GSE31210 datasets (Area Under the Curve (AUC)=0.693 and 0.743, respectively). Furthermore, POSTN was identified as an independent risk factor for disease recurrence in univariate and multivariate Cox regression analysis, and its expression level was found to be linearly related to disease recurrence in different populations. Threshold effect analysis indicated that a POSTN expression level of 12.5 was associated with a > 50% risk of recurrence within 1 year. In vitro experiments demonstrated that POSTN silencing could inhibit tumor cell proliferation and invasion. Conclusion POSTN is positively correlated with the risk of early LUAD recurrence within 1 year after treatment, indicating its potential as a biomarker for predicting recurrence.
Article Details
Authors (5)
Youde Xiao
Xuelian Lin
Hui Gong
Juan Hu
Dan Wang