摘要
目的探讨MRI征象回归模型在植入型凶险性前置胎盘的应用价值。方法回顾性分析本院350例具有完整临床资料的凶险性前置胎盘患者的MRI资料,包含20例植入型凶险性前置胎盘和330例非植入型凶险性前置胎盘。采用Logistic回归模型评估胎盘MRI征象与植入型凶险性前置胎盘相关性。结果植入型凶险性前置胎盘的单变量分析结果提示各因素预测胎盘植入风险的能力由大到小依次为子宫局限性隆凸、胎盘内T-2WI低信号带、子宫肌层变薄或局限性中断、胎盘信号不均匀、胎盘突出征、剖宫产≥2次、宫颈缩短、胎盘下异常血管、子宫下段及宫颈血管丰富。多变量分析结果提示,子宫局限性隆凸(X1:OR=25.41,P=0.02)、剖宫产≥2次(X2:OR=9.1,P=0.02)、宫颈缩短(X3:OR=7.77,P=0.04),胎盘内T-2低信号带(X4:OR=2.87,P=0.43)4个因素与凶险性前置胎盘植入有关,为凶险性前置胎盘独立危险因素。建立评估植入型凶险性前置胎盘Logistic的回归模型,Logistic(P)=-7.61+25.41X1+9.1X2+7.77X3+2.87X4,该模型预测植入型凶险性前置胎盘的准确率为98.0%。结论 MRI征象回归模型评估植入型凶险性前置胎盘具有一定的指导意义。
Objective To explore the relationship between the MRI signs and the risk of implantable pernicious placenta previa. Methods Retrospective analysis of clinical data and MRI of 350 patients diagnosed with placenta previa,including 20 implantable preclinical placenta and 330 non-implantable preclinical placenta. Logistic regression model is used to analyse the association between characters of images and the risk-classification of implantable pernicious placenta previa.Results Univariate analysis of implantable preclinical placenta suggest the ability of various factors to predict the risk of placental implantation: uterine focal bulging dark intraplacental bands on T-2 WI thinning or limited interruption of the myometrium Uniform MRI signal of the placenta placental protrusion cesarean section greater than or equal to 2 times cervical shortening placental abnormal vascular lower uterine segment and rich vascularity of the cervix. The results of multivariate analysis showed that uterine focal bulging( X1: OR = 25. 41,P = 0. 02),≥2 times cesarean section performed( X2: OR = 9. 1,P = 0. 02),cervix was shortened( X3: OR = 7. 77,P 0. 05),and dark intraplacental bands on T-2 WI( X4: OR = 2. 87,P = 0. 43),were independent risk factors for the preclinical placenta. Logistic model built by these four factors was the Logistic( P =-7. 61 + 25. 41 X1 + 9. 1 X2 + 7. 77 X3 + 2. 87 X4),the accuracy of prediction for the implanted preclinical placenta was 98. 0%. Conclusion MRI can be used for diagnosis of preclinical placenta and predicting the risky classfications.
作者
郭吉敏
曹满瑞
刘小平
刘炳光
张方璟
严芳
GUO Jimin;CAO Manrui;LIU Xiaoping(Department of Radiology,the Affiliated Shenzhen Maternity and Child Healthcare Hospital of Southern medical university,Guangdong 518028,P.R.China)
出处
《临床放射学杂志》
CSCD
北大核心
2018年第8期1325-1328,共4页
Journal of Clinical Radiology