Objective: To investigate the diagnostic value of MRI for placenta previa complicated with placenta accreta or not. Methods: A total of 220 placenta previa patients were diagnosed by prenatal ultrasound and MRI in The Second Affiliated Hospital of Wenzhou Medical University from May 2014 to May 2017.The MRI images of 220 placenta previa patients suspicious of placenta previa were interpreted by two radiologists who majored on gynecological radiology. The sensitivity, specificity, accuracy, positive predictive value and negative predictive value of two radiologists in diagnosis of placenta accreta were calculated respectively. Kappa test were used to verify the consistency between two doctors, as well as their MRI diagnosis and pathological results. The diagnostic value of MRI and pathological were assessed by Chi-square test and receiver operating characteristic (ROC)curve. Results: The 220 patients were all confirmed with placenta previa by surgical pathology.Out of 220, 71 cases were diagnosed as placenta accreta, and 149 cases were diagnosed without placenta accreta. The sensitivity, specificity, accuracy, positive predictive value and negative predictive value with physician A and physician B were 90.1%/84.5%, 88.6%/89.9%, 89.1%/88.2%, 79.0%/80.0% and 95.0%/92.4%, respectively.The consistency between MRI diagnosis and pathological results was excellent in physician A (κ=0.759), and good in physician B (κ=0.734). However, the sensitivity, specificity and accuracy of diagnosis between two physicians had no significant difference (all P>0.05). The area under the curve (AUC) of ROC in physician A and B were 0.858 and 0.847 (P=0.980). Conclusion: MRI is feasible for patients with placenta previa, as the sensitivity, specificity and accuracy of MRI are high in assessing whether placenta previa complicated with placenta accreta or not.
目的： 探讨磁共振成像(MRI)对前置胎盘是否合并胎盘植入的诊断价值。 方法： 回顾性分析2014年5月至2017年5月在温州医科大学附属第二医院产前经超声和MRI诊断为前置胎盘的患者共220例。选取两名妇产影像亚专科医师(医师A、B)分析220例前置胎盘是否存在胎盘植入，分别计算两位医师诊断胎盘植入的敏感度、特异度、准确度、阳性预测值、阴性预测值。由kappa检验验证两者诊断结果与手术病理的一致性，两者的诊断价值比较采用χ(2)检验及受试者工作特征(ROC)曲线分析。 结果： 手术病理证实220例均存在前置胎盘，其中胎盘植入71例、无胎盘植入149例。医师A/B诊断胎盘植入的敏感度、特异度、准确度、阳性预测值、阴性预测值分别为90.1%/84.5%、88.6%/89.9%、89.1%/88.2%、79.0%/80.0%、95.0%/92.4%。医师A和病理诊断胎盘植入的一致性好(κ＝0.759)，医师B和病理诊断胎盘植入的一致性较好(κ＝0.734)。医师A、B诊断胎盘植入的敏感度、特异度及准确度差异均无统计学意义(均P>0.05)。ROC曲线分析显示医师A、B的曲线下面积(AUC)分别为0.858、0.847，两者的AUC差异无统计学意义(P＝0.980)。 结论： MRI判断前置胎盘是否合并胎盘植入的敏感度、特异度及准确度较高，前置胎盘患者需要MRI检查。.
Keywords: Magnetic resonance imaging; Placenta diseases; Placenta previa.