Prediction of invasive placenta through cell-free RNA analysis

为了开发预测侵袭性胎盘的策略，我们已经收集了超过660份血液样本。然而，由于在这些病例中没有人并发侵袭性胎盘，我们无法评估侵袭性胎盘的预测标记物。因此，我们用来预测先兆子痫。这项研究的目的是使用一组无细胞或细胞的mRNA标记物来预测妊娠15-20周的一系列患者的先兆子痫的发生。分析了62名在验血时无症状的先兆子痫患者和310名对照的数据。采用判别分析进行多变量分析。在无细胞RNA的分析中，单变量分析确定血管内皮生长因子受体1(Flt1)是检出率最高的标志物，胎盘特异性1的检出率最低。对照组的子痫前期平均估计得分为9.4分，先兆子痫患者的平均估计得分为72.5分。以先兆子痫的估计分数作为测试变量得到的ROC曲线，在假阳性率为5%的情况下，DR为84%(95%CI71.891.5%)，曲线下面积为0.927(p&lt；0.001)。同样，每个患者的DR和评分与病情严重程度相关。其次，在细胞RNA分析中，单变量ROC分析确定sFlt-1和endoglin是最敏感的标记物。将Flt-1、endoglin、胎盘生长因子(PlGF)和胎次相结合，得到最佳的多变量模型。相对ROC曲线的敏感度为66%，特异度为10%，曲线下面积为0.884(p&lt；0.001)。总而言之，一组无细胞或有细胞的信使核糖核酸能够检测出患先兆子痫的受试者。

To develop the strategy predicting the invasive placenta, we have been collecting over 660 blood samples. However, since nobody have complicated with invasive placenta in the cases, we could not assess the predictive markers for invasive placenta. Therefore, we employed to predict the preeclampsia. The objective of this study is to predict occurrence of preeclampsia in a series of patients at gestational week 15-20 using a panel of cell-free or cellular mRNA markers. Data from 62 preeclamptic patients asymptomatic at the time of blood testing and 310 controls were analyzed. Multivariable analysis was performed using discriminant analysis. In the analysis of cell-free RNA, univariable analysis identified vascular endothelial growth factor receptor 1 (FLT1) as the marker with the highest detection rate, and placenta-specific-1 with the lowest. Mean estimated score for preeclampsia was 9.4 for controls and 72.5 for subjects who developed preeclampsia. An ROC curve obtained using the estimated score for preeclampsia as test variable yielded a DR of 84% (95%CI 71.8-91.5) at a 5% false-positive rate with an area under the curve of 0.927 (p<0.001). Again, DR and score for each patient for classification as preeclamptic correlated with severity. Next, in the cellular RNA analysis, the univariate ROC analysis identified soluble Flt-1 and endoglin as the markers with the highest sensitivity. The best multivariable model was obtained by combining Flt-1, Endoglin, placenta growth factor (PlGF) and parity. The relative ROC curve yielded a sensitivity of 66% at a 10% 1-specificity rate with an area under the curve of 0.884 (p<0.001). In conclusion, a panel of cell-free or cellular mRNA is able to detect subjects who will develop preeclampsia.