急性髓系白血病中RUNX1突变基因相关预后模型的构建

    Construction of a prognostic model associated with RUNX1 mutations for acute myeloid leukemia

    • 摘要:
      目的寻找RUNX1突变型急性髓系白血病(AML)和未突变型AML的差异基因并用来构建预后模型。
      方法从RUNX1突变组与未突变组中筛选AML中的差异基因,通过单因素Cox对差异基因进行筛选并构建多因素Cox回归预测模型。根据模型得到病人的风险评分并通过生存曲线和ROC曲线予以评估。
      结果从RUNX1突变组与未突变组中筛选得到89个差异基因,其中30个基因上调,59个基因下调。全部的差异基因通过单因素Cox回归筛选,得到38个基因构建多因素Cox回归模型。基于双向逐步回归法进一步筛选得到10个基因,包括BIK、APP、MLLT3、C10orf10、PLXNC1、FHL1、CST3、TGLL1、HOXA5、KIAAO125,使用该10个基因构建预后基因模型,风险评分公式为:风险评分=-0.100×(BIK)+0.215×(APP)+-0.232×(MLLT3)+0.112×(C10orf10)+0.160×(PLXNC1)+0.113×(FHL1)+-0.167×(CST3)+-0.152×(IGLL1)+0.164×(HOXA5)+0.084×(KIAA0125);其中BIK、MLLT3、CST3和IGLL1的风险比小于1,而其他6个基因的风险比大于1。生存分析结果表明高风险评分组的总体生存率明显低于低风险评分组(P < 0.01)。ROC曲线对未来1、3和5年的总体生存率预测的AUC分别为0.709、0.769和0.771。
      结论成功构建突变型AML差异基因预后模型,其中BIK、MLLT3、CST3和IGLL1可能是AML预后保护因素,APP、C10orf10、PLXNC1、FHL1、HOXA5和KIAAO125可能是AML预后的危险因素。

       

      Abstract:
      ObjectiveTo identify the differentially expressed genes in RUNX1 mutant acute myeloid leukemia(AML) and non-mutant AML and construct a prognostic model derived from RUNX1 mutations.
      MethodsDifferentially expressed genes in AML were screened from the RUNX1 mutated AML group and non-mutated AML group by univariate Cox, and a multivariate Cox regression prognostic model was constructed.The risk score of the patients was obtained according to the model, which was evaluated by the survival curve and the ROC curve.
      ResultsA total of 89 differentially expressed genes were screened from the RUNX1 mutant and non-mutant groups, 30 of which were up-regulated and 59 of which were down-regulated.All the differentially expressed genes were screened by univariate Cox regression, and 38 genes were obtained to construct multivariate Cox regression model.Based on the bidirectional stepwise regression method, 10 genes including BIK, APP, MLLT3, C10orf10, PLXNC1, FHL1, CST3, TGLL1, HOXA5 and KIAAO125 were further screened to construct the prognostic gene model.The 10 genes were used to construct a prognostic gene model, and the risk scoring formula was as follows.Risk score=-0.100×(BIK)+ 0.215×(APP)+ -0.232×(MLLT3)+ 0.112×(C10orf10)+ 0.160×(PLXNC1)+ 0.113×(FHL1)+ -0.167×(CST3)+ -0.152×(IGLL1)+ 0.164×(HOXA5)+ 0.084×(KIAA0125).The risk scores of BIK, MLLT3, CST3 and IGLL1 were less than 1, while the risk scores of the other six genes were greater than 1.Survival analysis of the high and low risk rating groups showed that the overall survival rate of high risk rating group was significantly lower than that of low risk rating group(P < 0.01).The AUC predicted by ROC curve for overall survival in 1, 3 and 5 years were 0.709, 0.769 and 0.771, respectively.
      ConclusionsThe prognostic model of mutant AML differentially genes was successfully constructed.The BIK, MLLT3 CST3 and IGLL1 may be the prognostic protective factors for AML.APP, C10orf10, PLXNC1, FHL1, HOXA5 and KIAAO125 may be the risk factors for prognosis of AML.

       

    /

    返回文章
    返回