microRNA标志物在早期诊断室间隔缺损中的价值*

Value of microRNA markers in early diagnosis of VSD*

摘要:

目的 评价血浆microRNA（miRNA）在室间隔缺损（VSD）早期诊断中的临床应用价值以及作为VSD分子诊断标志物的可行性，并建立临床早期诊断模型。方法 采用以济南市儿童医院、泰安市妇幼保健院为基础的病例-对照研究方法开展本次研究。研究分为VSD组85例和对照组80例，分3个阶段进行检测：首先选取VSD组及其匹配对照组各3例，应用miRNA全基因组表达谱芯片进行初筛；然后扩大样本量（20例VSD组vs 15例对照组），对选定的8个有差异的miRNA采用实时荧光定量聚合酶链反应（qRT-PCR）进行芯片结果的验证；最后进一步扩大样本（62例VSD组vs 62例对照组）进行qRT-PCR检测，筛选先天性心血管病早期诊断的生物标志物，建立多指标的Logistic回归模型，并应用MedCalc软件进行受试者工作特征（ROC）曲线分析评价诊断模型的价值。结果 芯片杂交初筛结果显示，与对照组比较，有36个表达有差异的miRNA（P?<0.05），两阶段大样本qRT-PCR验证，检测到5种血浆miRNA（let-7e-5p、miR-155-5p、miR-222-3p、miR-433-3p和miR-487b-5p）在VSD中的表达差异有统计学意义（P?<0.05），可以选择作为具有VSD早期诊断意义的生物标志物。5种血浆miRNA的ROC曲线下面积（AUC）分别在0.678～0.832，其中miR-222-3p的ROC曲线下面积最大为0.832。建立多指标Logistic回归分析模型，5种血浆miRNAs联合模型可提高VSD的诊断效率，ROC曲线下面积达到0.955，敏感性和特异性分别为83.87%和95.16%；3种血浆miRNA（let-7e-5p、miR-155-5p、miR-222-3p）联合诊断模型，ROC曲线下面积达到0.910，敏感性和特异性分别为82.30%和90.30%，接近5个miRNA的诊断效率。结论 5种血浆miRNA作为VSD早期诊断标志物，具有一定的可信性。将其整合后可建立起多指标的联合诊断模型。联合3种血浆miRNA构建诊断模型可提高VSD的诊断效率，临床应用价值更大。

Abstract:

Objective To evaluate the clinical value of plasma microRNA in the early diagnosis of ventricular septal defect (VSD) and its feasibility as a molecular diagnostic marker for VSD, and to establish a clinical early diagnosis model. Methods This study was conducted using a Jinan Children's Hospital and Taian Maternal and Child Health Hospital-based case-control study. A total of 85 cases were collected for the VSD group and 80 cases for the control group. The test was performed in three stages. At first, 3 cases of ventricular septal defect (VSD) and 3 matched controls were selected, and the microRNA whole genome expression profile chip was used for primary screening. Then, the sample size was expanded (20 VSD cases VS 15 controls). The selected 8 differential microRNAs were verified by real-time quantitative PCR (qRT-PCR). Finally, further expanding samples (62 VSD cases VS 62 controls) were selected for qRT-PCR to screen biomarkers for early diagnosis of congenital heart disease and to establish a multi-index logistic regression model, and using MedCalc software for ROC (Receiver-Operator Characteristic curve)curve to analyze the value of the diagnostic model. Results Compared with control group, the results of microarray screening showed that there were 36 differential-expressed microRNAs (P?