The Diagnostic Value of Multivariate Logistic Model with Transvaginal Power Doppler Ultrasonography for the Prediction of Ectopic Pregnancy

  • Kun Liang, From the Department of Obstetrics and Gynecology, The Third Affiliated Hospital of Guangzhou Medical College, Institute of Obst, China
  • Dr Hua Liu, From the Department of Function Imaging and Ultrasonography, Guangzhou First Municipal People's Hospital, Guangzhou Medical Col, China
  • Dr Yi Chen, Department of Medical Ultrasound, The Third Affiliated Hospital of Guangzhou Medical College, Guangzhou, China
  • Objective: This study established a regression model with the relative factors, evaluated its clinical applied value, and explored the diagnostic criterias of transvaginal power Doppler ultrasonography (TV-PDU) in predicting ectopic pregnancy (EP) of the women with pregnancy of unknown location (PUL).
    Study Design: 129 suspected early EP were studied retrospectively. The thickness, symmetry, resonance, pattern, blood flow display of endometrial and resistance index value (RI) were analyzed and observed by TV-PDU. The significant diadynamic criterias with multivariate Logistic regression analysis were selected in order to establish the multivariate predictive model for diagnosis of early EP.
    Results: The multivariate analysis indicated that the final predictive model included three factors as follow: whether the endometrial thickness was more than 9mm, whether the endometrial pattern was type A, and whether the endometrial spiral arteriae were detected by TV-PDU. The area under the ROC curve of the TV-PDU multivariate predictive model was 0.980. The accuracy of diagnosis was high, when the RI value was 0.65 and predictive regression diagnosis ratio was more than 0.50. The coincidence rate was 94.5%, when this model was used to diagnose early EP.
    Conclusion: We conclude that the multivariate regressive predictive model established in this study has significant clinical value in the differential diagnosis of early EP of PUL.