《Table 3 Multivariate Cox regression analysis conducted in TCGA for overall survival and progression

  提示：宽带有限、当前游客访问压缩模式

本系列图表出处文件名：随高清版一同展现

《基于预后相关基因构建的分析模型可识别高风险胶质母细胞瘤（英文）》

1. 获取 高清版本忘记账户？点击这里登录

1. 下载图表忘记账户？点击这里登录

Previously，several known molecular features have been verified to be associated with survival times of GBM patients.Hence，the relationship of our risk score system with the previously described molecular classifications and clinicopathologic features was analyzed（Table 4，5）.Here，we found almost all the tumors in the high-risk group were IDH wild type.With TCGA classifications，we found that most tumors of the mesenchymal subtype were clustered in the high-risk group while most tumors of the proneural and neural subtypes fell in the low-risk group.We also rearranged the GBM samples in two datasets according to their risk scores，and the hierarchical clustering showed that 17 survival-related genes exhibited distinct expression patterns between the low-risk and high-risk groups（Figure7）.The high-risk group expressed higher levels of risky genes and lower levels of protective genes as compared to the low-risk group.Furthermore，we performed statistical analysis（t test）for the distribution of risk scores among various molecular subtypes.The mesenchymal subgroup showed higher risk scores in comparison with the other TCGA classifications（classical，proneural and neural subtypes），and the IDH wild type group had higher risk scores than IDH mutant group（Figure 7）.These results indicated that our risk scoring system was associated with clinical characteristics and high risk scores positively correlated with risky clinicopathologic features.