基于TCGA数据的黑色素瘤预后相关miRNA模型

A miRNA model to predict prognosis of melanoma based on TCGA data

目的对黑色素瘤患者进行精准的分组以提高治疗的针对性,建立与患者预后密切相关的mi RNAs风险评分模型。方法通过收集TCGA数据库中黑色素瘤患者的mi RNA表达谱和临床资料相关数据,并采用单因素和多因素COX比例风险回归模型进行分析并构建风险评分模型。结果通过单因素和多因素COX比例风险回归分析筛选出6个与黑色素瘤预后密切相关的mi RNA分子,分别为hsa-mi R-7702,hsa-mi R-4442,hsa-mi R-4444-2,hsa-mi R-4461,hsa-mi R-203b和hsa-mi R-1910,经多因素COX比例风险回归分析后建立风险评分值的风险评分公式:Risk-Score=(0.245 2×mi R-4461)-(0.101 8×mi R-7702)-(0.197 5×mi R-4442)+(0.108 3×mi R-203b)+(0.133 3×mi R-1910)-(0.173 3×mi R-4444-2)。结论风险评分模型可以有效的根据患者的预后将黑色素瘤患者进行分组,具有较高的可信度,或可以成为指导黑色素瘤精准治疗的新方法。

Objective To establish a miRNAs Risk-Score model related to the prognosis of melanoma for accurately grouping the pa- tients to improve the targeting of treatment. Methods We downloaded the miRNA expression profile and clinical data of melanoma from TCGA （The Cancer Genome Atlas）. The univariate and multivariate COX proportional hazards regression models were used to an- alyze the data and created the Risk-Score model. Results Six miRNAs （hsa-miR-7702, hsa-miR-4442, hsa-miR-4444-2, hsa-miR- 4461, hsa-miR-203b and hsa-miR-1910） closely associated with the prognosis of melanoma were screened. The melanoma of Risk- Score formula created by multivariate COX proportional hazards regression analysis as follows： （0.245 2×miR-4461 ）-（0.101 8×miR- 7702）-（0.197 5×miR-4442）＋（0.108 3×miR-203b）＋（0.133 3×miR-1910）-（0.173 3×miR-4444-2）. Conclusion The Risk-Score model eould 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.