A Five-Gene Risk Score Model for Predicting the Prognosis of Multiple Myeloma Patients Based on Gene Expression Profiles

Multiple myeloma is a heterogeneous plasma cell malignancy that remains incurable because of the tendency of relapse for most patients. Survival outcomes may vary widely due to patient and disease variables, therefore a more accurate prognostic model is needed to establish to improve prognostic precision and guide clinical therapy. Here, we developed a risk score model based on myeloma gene expression profiles from 3 independent datasets: GSE6477, GSE13591, GSE24080. In this model, highly survival-associated five genes, including EPAS1, ERC2, PRC1, CSGALNACT1, and CCND1, are selected by using the least absolute shrinkage and selection operator (Lasso) regression, univariate, and multivariate Cox regression. At last, we analyzed three validation datasets (including GSE2658, GSE136337 and MMRF) to examine the prognostic efficacy of this model by dividing patients into high-risk and low-risk groups based on the median risk score. The results indicated that the survival of patients in low-risk group was greatly prolonged compared with their counterparts in high-risk group. Therefore, the five-gene risk score model could increase accuracy of risk stratification and provide effective prediction for the prognosis of patients and instruction for individualized clinical treatment.