科研之友
微信
新浪微博
Facebook
分享链接
摘要
Multiple myeloma (MM) is a kind of aggressive tumor prevalent with high heterogeneity. Abnormal expression of certain genes may lead to the occurrence and development of MM. Nowadays, it is not commonly seen in clinical research to predict the prognostic circumstances of patients with MM by multiple gene expression profiling method. Identification of potential genes in prognostic process could be beneficial for clinical management of MM. Therefore, we aimed to build a risk fraction model to screen out the prognostic indicator for clinical outcome of MM. Microarray data were downloaded from the Genome Expression Omnibus (GEO) datasets with accession numbers GSE24080 and GSE57317. A total of 279 samples were selected out randomly. Besides, a risk formula was constructed and verified in the dataset. Time-dependent receiver operating characteristic (ROC) curve was applied in evaluating the accurate prognostic conditions of patients. Finally, a ten genes model in the training dataset was found to be closely related to the survival condition of MM patients. Patients with MM were divided into two groups, high-risk and low-risk, by the expression of these ten genes, and significant statistical difference was found between the two groups. Furthermore, the result of multivariate cox regression and stratified analysis indicated that this modelwas independent of other clinical phenotypes. ROC curves also showed its feasibility to predict the survival status of MM patients. Our results demonstrated that the fraction risk model constructed by the selected ten genes could be used to predict survival status of multiple myeloma patients, which could also help in improvement of prognostic and therapeutic tool of MM.
