| Objective This article is going to investigate the value of the2D ultrasonography (2D-US) and3D ultrasonography (3D-US) in diagnosing the congenital uterine malformation and recognize the better one, compare the Pros and cons of these two kinds of3D ultrasonic imaging mode (Render mode and Omni view mode), and these two kinds methods for measuring angle, including the three-point method and the two-line method. It is also going to optimize these current ultrasonic parameters for diagnosing uterus malformation.Methods We have collected359cases patients with suspected uterine malformation who came to ultrasound department of Obstetrics and Gynecology Hospital of Fudan University for diagnosis from September2011to January2013.159cases of these359cases had been examined and confirmed by one of the other inspections, including office hysteroscopy, combined hysteroscopy and laparoscopy evaluation, MRI and HSG. All patients had been taken2D-US examination first to draw inspection results, and then adjusted the image to the median sagittal plane through the whole length of the uterus from the fundus to the cervix. At the end of the examination, satisfied three dimensional ultrasound volumes were recorded and stored in the machine computer memory. By using3D-US imaging software, uterine coronal plane was reconstructed and measured respectively in two modes (Render mode and Omni view mode). Each variable had been measured three times to gain average value.3D-US diagnoses were finally obtained. And all data was analyzed by using SPSS17.0.Resultsl.359patients have been analyzed by these two imaging modes (Render mode and Omni view mode). Firstly, the successful detection rate of uterus deformity in3D-US (295/359) is higher than that of2D-US (262/359)(p=0.003), the consistency between2D-US and3D-US is low and they have different composition of uterine anomalies. Secondly,97.77%of359cases had been identified by both Render mode and Omni view mode, and the consistency between these two modes is high (Kappa value is0.971).2. The matching rate of test results between2D-US and golden standard is52.83%in159cases, which is lower than that of78.62%in Render mode and81.76%in Omni view mode (p<0.001). The general diagnostic ability of3D-US and the ability to identify arcuate uterus, subseptate uterus and complete septate uterus is much better than that of2D-US (p<0.05).3. The diagnostic efficiency of2D-US in different types of uterine malformation is different.1) The specificity of arcuate uterus (96.80%) and subseptate uterus (94.93%) is high and so is positive likelihood rate (LR+).2) The accuracy (99.37%) and specificity (100%) of unicornuate uterus combined rudimentary horn are high and so is LR+. All of the accuracy, sensitivity and specificity of didelphys uterus turned out to be100%.By the way, the negative likelihood rate (LR-) and LR+of didelphys uterus are ideal.4.1) In addition, the diagnostic efficiency in different types of uterine malformation is different under both two kinds of3D ultrasonic imaging modes. Firstly, the sensitivity (It’s97.06%under both two modes) and negative predict value (PV-)(It’s98.99%in Render mode and99.04%in Omni view mode) of the arcuate uterus are high, LR-(It’s0.038in Render mode and0.036in Omni view mode) of which is low as well. Secondly, the specificity (It’s96.70%under both two modes), PV+(It’s93.88%and94.44%respectively under two modes) and LR+of the subseptate uterus are high. Thirdly, the sensitivity, specificity and accuracy of complete septate uterus and didelphys uterus are high, LR+and LR-of which are ideal as well, and so is unicornuate uterus combined rudimentary horn. In addition, the area under ROC curve of septate uterus, didelphys uterus and unicornuate uterus combined rudimentary horn are close to1(They are0.989,1.000and1.000respectively).2) The diagnostic efficiency in the same type of uterine malformation is different under different kind of3D ultrasonic imaging mode. For example, the diagnostic efficiency of Omni view mode in arcuate uterus and subseptate uterus is superior to that of Render mode (p<0.05).5. After comparing ultrasonic parameters of the median sagittal plane of uterus in212cases by two different imaging modes, the following conclusion can be drawn.1) The data excluding variable E (E denotes depth of external contour indentation) measured in Render mode is different from Omni view mode. Firstly, the variable F (F denotes the depth of fundal distortion), M (M denotes muscular thickness of the fundal) and D (D denotes the depth of the uterine cavity) measured in Render mode are less than those in Omni view mode (p<0.05).Secondly, the variable A (A denotes angle of fundal distortion) measured in Render mode are greater than that in Omni view mode (p<0.05).2) The two methods for measuring angle are different in Omni view mode:the variable Al measured in three-point method is greater than variable A2measured in two-line method (p<0.05).6. Take the ultrasonic parameters F/(F+M)(F+M denotes the distance between fundal serosal layer and free margin of septum) and F/D in Omni view mode as the reference value for drawing ROC curve, we found that When the cutoff value of variable F/(F+M) is 47.45%and cutoff value of variable F/D is33.5%, the sensitivity and specificity for identifying arcuate uterus and septate uterus are highest (p<0.05).Conclusions The2D-US can exactly detecting didelphys uterus and unicornuate uterus combined rudimentary horn, but it needs to depend on3D-US for diagnosing subseptate uterus, complete septate uterus and bicornuate uterus. With the purpose of identifying arcuate uterus from subseptate uterus, the Omni view mode is much better than the Render mode, and the two-line method is better than three-point method. In order to identify arcuate uterus from subseptate uterus by using3D-US, we propose some new quantitative ultrasonic parameters of uterine anomaly (such as variable F/(F+M) and F/D) on the base of the Modified American Fertility Society Classification (MAFSC) will also improve the diagnosis success rate of3D-US.
|
PDF Full Text Request