Office gel sonovaginography for the prediction of posterior deep infiltrating endometriosis: a multicenter prospective observational study

Authors Organisations
  • S. Reid(Author)
    University of Sydney
  • Chuan Lu(Author)
  • Nigel Hardy(Author)
  • I. Casikar(Author)
    University of Sydney
  • G. Reid(Author)
    St Luke's Private Hospital
    St George Private Hospital
    Prince of Wales Private Hospital
    Liverpool Public Hospital
  • G. Cario(Author)
    St George Private Hospital
    Hurtsville Private Hospital
  • D. Chou(Author)
    St George Private Hospital
  • D. Almashat(Author)
    University of Sydney
  • George Condous(Author)
    University of Sydney
    OMNI Gynaecological Care Centre for Women's Ultrasound and Early Pregnancy
    Norwest Private Hospital
Type Article
Original languageEnglish
Pages (from-to)710-718
Number of pages9
JournalUltrasound in Obstetrics and Gynecology
Issue number6
Publication statusPublished - 01 Dec 2014
View graph of relations
Citation formats


Objective To use office gel sonovaginography (SVG) to predict posterior deep infiltrating endometriosis (DIE) in women undergoing laparoscopy. Methods This was a multicenter prospective observational study carried out between January 2009 and February 2013. All women were of reproductive age, had a history of chronic pelvic pain and underwent office gel SVG assessment for the prediction of posterior compartment DIE prior to laparoscopic endometriosis surgery. Gel SVG findings were compared with laparoscopic findings to determine the diagnostic accuracy of office gel SVG for the prediction of posterior compartment DIE. Results In total, 189 women underwent preoperative gel SVG and laparoscopy for endometriosis. At laparoscopy, 57 (30%) women had posterior DIE and 43 (23%) had rectosigmoid/anterior rectal DIE. For the prediction of rectosigmoid/anterior rectal (i.e. bowel) DIE, gel SVG had an accuracy of 92%, sensitivity of 88%, specificity of 93%, positive predictive value (PPV) of 79%, negative predictive value (NPV) of 97%, positive likelihood ratio (LR+) of 12.9 and negative likelihood ratio (LR–) of 0.12 (P = 3.98E–25); for posterior vaginal wall and rectovaginal septum (RVS) DIE, respectively, the accuracy was 95% and 95%, sensitivity was 18% and 18%, specificity was 99% and 100%, PPV was 67% and 100%, NPV was 95% and 95%, LR+ was 32.4 and infinity and LR– was 0.82 and 0.82 (P = 0.009 and P = 0.003). Conclusions Office gel SVG appears to be an effective outpatient imaging technique for the prediction of bowel DIE, with a higher accuracy for the prediction of rectosigmoid compared with anterior rectal DIE. Although the sensitivity for vaginal and RVS DIE was limited, gel SVG had a high specificity and NPV for all forms of posterior DIE, indicating that a negative gel SVG examination is highly suggestive of the absence of DIE at laparoscopy. Copyright © 2014 ISUOG. Published by John Wiley & Sons Ltd.


  • DIE, laparoscopy, office gel sonovaginography, posterior deep infiltraing endometriosis, sliding sign, SVG