基于多模态超声对中孕期胎盘功能不全预测模型的建立

Establishing Multi-Modal Ultrasound-Based Prediction Model for Placental Insufficiency of Second Trimester

目的使用常规超声、三维能量多普勒及剪切波弹性成像建立中孕期对胎盘功能不全的预测模型。资料与方法前瞻性分析2021年4月—2022年4月在三峡大学附属仁和医院行超声检查的前壁胎盘孕妇209例,依据临床诊断分为正常组118例和胎盘功能不全组91例,从血流动力学(脐动脉、子宫动脉的血流参数和脑胎盘率)、血流灌注(血管指数、血流指数和血管-血流指数)、机械组织特性(弹性模量平均值)等方面评估胎盘功能。对组间差异有统计学意义的指标进行Logistic回归分析,构建预测模型并评估其效能。结果两组子宫动脉的血流参数、脑胎盘率、血管指数、血管-血流指数和弹性模量平均值等差异均有统计学意义(t=-12.74、-10.28、-11.01、7.02、7.00、-11.97,P均<0.05)。Logistic回归分析发现,子宫动脉阻力指数(≥0.65)、脑胎盘率(≤1.53)及弹性模量平均值(≥4.14 KPa)为独立危险因素,其中弹性模量平均值的诊断效能最大,曲线下面积为0.904。联合3项参数建立预测模型的曲线下面积为0.945,敏感度为87.91%,特异度为94.92%,诊断效能较单一参数明显提高。结论基于多模态超声建立的预测模型对胎盘功能不全有一定的预测价值和高度特异性,可指导临床尽早干预以改善围生期结局。

Purpose Using conventional ultrasound,3D energy Doppler and shear wave elastography to develop one multi-modal ultrasound-based prediction model for placental insufficiency of second trimester.Materials and Methods A prospective analysis was conducted by reviewing 209 pregnant women with anterior wall placenta who underwent multi-modal ultrasound examination in the Affiliated Renhe Hospital of China Three Gorges University from April 2021 to April 2022,and all participants were divided into 118 cases in the normal group and 91 cases in the placental insufficiency group based on clinical diagnosis.Their placental function was evaluated through hemodynamics[blood flow parameters of umbilical artery(UA),uterine artery(Ut A)and cerebral placental rate(CPR)],blood stream perfusion[vascular index(VI),blood flow index(FI)and vaso-blood flow index(VFI)]and mechanical tissue properties[mean of elasticity modulus(Emean)].Logistic regression analysis was performed on the indicators with statistically significant differences between groups,establishing a prediction model of placental insufficiency and evaluating its efficacy.Results There were statistically significant differences in Ut A blood flow parameters,CPR,VI,VFI and Emean between the two groups(t=-12.74,-10.28,-11.01,7.02,7.00,-11.97,all with P<0.05).In regression analysis,Ut A-RI(≥0.65),CPR(≤1.53)and Emean(≥4.14 KPa)were found to be the independent risk factors,of which Emean had the greatest diagnostic efficacy(area under the curve,AUC=0.904).The AUC of the prediction model established by combining the three parameters was 0.945,with a sensitivity of 87.91%and a specificity of 94.92%,which has significantly higher diagnostic efficacy than that of a single parameter.Conclusion The prediction model based on the multi-modal ultrasound parameters has certain predictive value and high specificity for placental insufficiency of second trimester and can guide early clinical intervention to improve perinatal outcomes.