临床指标联合MRI评估胎盘植入高危患者术中子宫切除风险  

Combining clinical characteristics and specific magnetic resonance imaging features to predict the risk of hysterectomy in gravid patients at high risk for placenta accreta spectrum disorders

作　　者：钟淑媛 丁志广 徐坚民[1] 胡根文 苏放明[2] 成志强[3] ZHONG Shuyuan;DING Zhiguang;XU Jianmin;HU Genwen;SU Fangming;CHENG Zhiqiang(Department of Radiology,Second Clinical Medical College of Jinan University,Shenzhen People's Hospital,Shenzhen 518000,China;Department of Obstetrics,Second Clinical Medical College of Jinan University,Shenzhen People's Hospital,Shenzhen 518000,China;Department of Pathology,Second Clinical Medical College of Jinan University,Shenzhen People's Hospital,Shenzhen 518000,China)

机构地区：[1]暨南大学第二临床医学院(深圳市人民医院)放射科,深圳518000 [2]暨南大学第二临床医学院(深圳市人民医院)产科,深圳518000 [3]暨南大学第二临床医学院(深圳市人民医院)病理科,深圳518000

出　　处：《磁共振成像》2021年第5期35-39,共5页Chinese Journal of Magnetic Resonance Imaging

基　　金：深圳市科技计划基础研究项目(编号:JCYJ20180228164641207)。

摘　　要：目的探索临床指标联合MRI征象评估胎盘植入谱系疾病(placenta accreta spectrum,PAS)高危患者术中子宫切除风险的价值。材料与方法回顾性分析251例妊娠晚期(32周以上)PAS高危孕妇的MR图像及临床资料,包括64例子宫切除患者及187例子宫保留患者,对相关临床指标及MRI征象进行单因素及多因素分析,构建预测PAS高危患者子宫切除风险Logistic回归模型。结果单因素分析显示,剖宫产次数、前置胎盘、T2WI胎盘内暗带、胎盘膨出、子宫肌层变薄或消失、膀胱壁T2WI低信号中断、局部外突性肿块和子宫浆膜面异常血管在两组之间差异具有统计学意义(P<0.01)。Logistic回归分析显示,剖宫产次数(X1)、T2WI胎盘内暗带(X2)、胎盘膨出(X3)、子宫浆膜面异常血管(X4)是PAS高危患者子宫切除的独立危险因素,联合预测模型为Logistic(P)=-4.713+0.960X1+1.477X2+1.569X3+1.901X4,模型AUC值0.915(95%CI:0.873~0.946),敏感度、特异度分别为87.50%、81.82%。校准曲线显示模型校准能力好。结论基于临床指标及MRI征象,有望产前为PAS高危患者子宫切除风险的评估提供依据,改善患者预后。Objective:To explore the value of clinical characteristics combined with MRI features for predicting the risk of intraoperative hysterectomy in gravid patients at high risk for placenta accreta spectrum(PAS).Materials and Methods:Retrospectively analyzed the MRI images and clinical data of 251 patients(including 64 patients who required hysterectomy and 187 patients did not)who underwent MRI during the third trimester from January 2010 to March 2020 with high risk for PAS disorders according to FIGO guidelines.Clinical characteristics included age,gestational age at delivery,number of cesarean deliveries,gravidity,prior other uterine surgeries and placenta previa.Two radiologists independently evaluated the following MRI features according to the consensus from Society of Abdominal Radiology and European Society of Urogenital Radiology:T2-dark intraplacental bands,placental bulge,loss of retroplacental T2-hypointense line,myometrial thinning,uterine serosa hypervascularity,focal exophytic mass and disruption of low-T2 bladder wall.Univariate analyses of clinical characteristics and MRI features were performed between patients with hysterectomy and those who without.Absence or uncertainty of MRI features was recorded as negative while presence as positive.Logistic regression was used to identify any clinical or MRI features in predicting hysterectomy.ROC analysis and calibration curve were performed to test the predictive power.Results:Significant differences were found in number of cesarean deliveries,placenta previa and all seven MRI features except for loss of retroplacental T2-hypointense line between patients with hysterectomy and those who without(P<0.01).The number of cesarean deliveries(X1:OR=2.611,P=0.017),T2-dark intraplacental bands(X2:OR=4.379,P=0.001),placental bulge(X3:OR=4.804,P=0.000)and uterine serosa hypervascularity(X4:OR=6.691,P=0.000)were independent risk factors for intraoperative hysterectomy.The Logistic regression model combining the four independent risk factors to forecast intraoperative

关 键 词：胎盘植入谱系疾病 磁共振成像 产前诊断 高危妊娠患者 子宫切除 

分 类 号：R445.2[医药卫生—影像医学与核医学] R657.51[医药卫生—诊断学]