Integrated analyses identify potential prognostic markers for uveal melanoma.

Uveal melanoma (UM) is the most common primary intraocular malignant tumor in adults, which has a high rate of metastases and can induce vision loss and even death to the patients. To identify suitable prognostic markers of UM for the early detection or prognosis prediction would be an essential step toward successful management of the disease. Herein, we extracted the mRNA expression data along with the clinical information from The Cancer Genome Atlas (TCGA) database. A total of eight co-expression modules were constructed by 5,000 genes based on the weighted gene co-expression network analysis (WGCNA). We found the blue and yellow modules were significantly associated with clinical stage. The Cox regression analyses found the blue, yellow, green and brown modules were significantly associated with overall survival (OS), while the blue, yellow, brown, green and pink modules were significantly associated with recurrence-free survival (RFS). Furthermore, the hallmark pathway enrichment analyses found the genes encompassed in the blue, yellow, and brown modules were significantly enriched in critical pathways involved in tumorigenesis and progression process, such as EMT and KRAS pathways. The hub-genes in these three modules were visualized by Cytoscape software and further validated by an external Gene Expression Omnibus (GEO) dataset. Besides, the OS and RFS predicting signatures were constructed based on the validated hub-genes according to the LASSO Cox regression model. The UM patients were assigned to low-/high-risk population. The survival analyses indicated high-risk patients mostly had bad OS/RFS rate compared with the low-risk population. The receiver operating characteristic (ROC) curve proved the stability and superiority of the two signatures. To sum up, our findings provide a framework of co-expression network of UM and identify a series of biomarkers, which will benefit from improving the prognosis prediction of UM patients.

Copyright © 2019. Published by Elsevier Ltd.