弥漫性大B细胞淋巴瘤预后相关microRNA筛选及其预后价值

目的 识别与弥漫性大B细胞淋巴瘤(diffuse large B-cell lymphoma ，DLBCL)预后相关的microRNA，并结合临床数据评估其预后价值。方法 从GEO (gene expression omnibus)数据库中获取116例DLBCL患者的MicroRNA表达谱数据与临床数据。采用lasso - logistic回归模型识别与DLBCL预后相关的miRNAs，卡方和Fisher确切概率检验筛选与DLBCL患者总生存期(overall survival,OS)相关的临床指标，再利用logistic逐步回归法筛选出预测因子并构建其单因素及多因素Cox比例风险模型。采用时依ROC (time - dependent receiver operating characteristic)曲线评价各模型的预测能力，各模型预测准确性用一致性统计量C评价。结果 Hsa - miR - 23a (HR = 0.20, 95%CI: 0.10～0.39, P＜0.001)和hsa - miR - 566 (HR = 6.69, 95%CI: 2.40～18.65, P＜0.001)与DLBCL患者的OS相关。单因素Cox回归模型与多因素Cox回归模型相比，C统计量差异具有统计学意义(IPI与IPI + hsa - miR - 23a + hsa - miR - 566: Z = 36.2, P＜0.001, hsa - miR - 23a + hsa - miR - 566与IPI +hsa - miR - 23a + hsa - miR - 566: Z = 7.0, P＜ 0.001)。结论 Hsa - miR - 23a高表达是DLBCL患者预后的保护因素，hsa - miR - 566高表达是DLBCL患者预后的危险因素，IPI、Hsa - miR - 23a和hsa - miR - 566这3个指标联合建模可提高DLBCL患者5年生存预测的准确性。

Prognostic-related microRNA of diffuse large B-cell lymphoma Screening and prognostic value

Objective To explore the expression of microRNAs in diffuse large B-cell lymphoma(DLBCL) and combine clinical indicators to evaluate their prognostic effects and value. Methods MicroRNA expression profile and clinical data of 116 patients with DLBCL were obtained from the GEO(gene expression omnibus) database. Lasso-logistic regression was used to identify the DLBCL-related microRNAs, and logistic stepwise regression was used to screen the miRNAs and clinical indicators which related to overall survival(OS) of DLBCL patients, and then establish the univariate and multivariate Cox proportional risk model on this basis. The ROC (time-dependent receiver operating characteristic) curve is applied to evaluate the prediction ability of each model, and the prediction accuracy of each model is evaluated with the consistency statistic-C. Results Hsa-mir-23a (HR=0.20, 95%CI: 0.10-0.39, P<0.001) and hsa-mir-566 (HR=6.69, 95%CI: 2.40-18.65, P=P<0.001) were significantly correlated with the OS. Compared with the multivariate Cox regression model, C statistics was statistically significantly different(IPI and IPI + hsa-mir-23a + hsa-mir-566: Z=36.2, P<0.001, hsa-mir-23a + hsa-mir-566 and IPI + hsa-mir-23a + hsa-mir-566: Z=7.0, P<0.001). Conclusion The high expression of hsa-mir-23a is a protective factor for the prognosis of patients with DLBCL, and the high expression of hsa-mir-566 is a risk factor for the prognosis of DLBCL patients. The prediction of 5-year survival accuracy of DLBCL patients may be improved by Combining IPI, hsa-mir-23a and hsa-mir-566.