Cutaneous melanoma prognostic risk marker screening based on glycolysis gene expression profiles
10.3760/cma.j.cn114453-20210108-00008
- VernacularTitle:基于糖酵解基因表达谱筛选皮肤黑色素瘤预后风险标志物
- Author:
Qiaoli SHANG
1
;
Leren HE
Author Information
1. 中国医学科学院北京协和医学院整形外科医院整形七科 100144
- Keywords:
Melanoma;
Prognosis;
Glycolysis;
Cox regression analysis;
Proportional hazards models;
The Cancer Genome Atlas
- From:
Chinese Journal of Plastic Surgery
2021;37(8):875-883
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To establish a prognosis model of cutaneous melanoma through bioinformatics analysis of glycolysis-related genes, therefore enhancingthe precision of classification and treatment.Methods:The mRNA expression profiles and clinical data of melanoma patients (470 cases in total, including 290 males and 180 females) in The Cancer Genome Atlas (TCGA) database were obtained, data of normal skin tissue samples (812 cases, including 467 males and 345 females) in GTEx database were used as controls. Gene set enrichment analysis (GSEA) was used to extract differentially expressed genes in the glycolysis pathway that significantly associated with melanoma. The mRNAs significantly correlated with melanoma prognosis were screened by Cox regression analysis, and the prognostic risk score (RS) model of melanoma patients was constructed. According to the median RS value of the risk scoring model, 450 melanoma patients who were eligible for survival analysis were divided into the high-risk group and the low-risk group, and the difference in survival time and survival rate between the two groups was analyzed. In addition, ROC curves were drawn to evaluate the predictive value of RS value on the mortality risk of melanoma patients at 10 and 20 years.Results:GSEA analysis showed that the glycolysis-related REACTOME_GLYCOLYSIS pathway was significantly enriched in melanoma samples, with a standard differentiation score (NES) of 1.74 and a false discovery rate (FDR) of 0.016. By univariate and multivariate Cox regression analysis of 117 differentially expressed genes (71 up-regulated and 46 down-regulated) in this pathway, 9 mRNA molecules closely related to the prognosis of cutaneous melanoma were found out, including four protective factors (STAT3, MLXIPL, IDUA and AURKA) and five risk factors (KIF20A, PGM1, GYS1, ALG1 and HDAC4). According to the regression coefficient of multivariate Cox regression analysis and mRNA expression level of each gene, the prognostic risk scoring model was established: RS = 0.31×KIF20A + 0.19× PGM1-0.27 ×STAT3 + 0.24× GYS1-0.25 × MLXIPL-0.19 ×IDUA + 0.33× ALG1-0.28 ×AURKA + 0.26×HDAC4. The median RS value of 0.83 was used as the cut-off value, and the patients were divided into the high-risk group (225 cases) and the low-risk group (225 cases). The overall survival time of the high-risk group was (6.72±0.55) years, which was significantly lower than that of the low-risk group (13.60±1.03) years ( P<0.001). The 10-year and 20-year survival rates in the high-risk group were significantly lower than those in the low-risk group (24.1% vs. 50.9%; 0 vs. 25.2%; P<0.001); the area under the ROC curve of RS value was 0.730 and 0.749 for 10-year and 20-year survival prediction of melanoma patients, respectively. Multivariate Cox regression analysis showed that RS value was an independent prognostic factor of melanoma. Conclusions:The risk-score model could effectively group the melanoma patients according to the prognosis, and had high degree of credibility. This model may be a new method to guide the accurate targeting treatment of melanoma.
