Accuracy of transvaginal ultrasound for diagnosis of deep endometriosis in uterosacral ligaments, rectovaginal septum, vagina and bladder: systematic review and meta-analysis.

OBJECTIVE: To review the diagnostic accuracy of transvaginal ultrasound (TVS) in the preoperative detection of endometriosis in the uterosacral ligaments (USL), rectovaginal septum (RVS), vagina and bladder in patients with clinical suspicion of deep infiltrating endometriosis (DIE).

METHODS: An extensive search was performed in MEDLINE (PubMed) and EMBASE for studies published between January 1989 and December 2014. Studies were considered eligible if they reported on the use of TVS for the preoperative detection of endometriosis in the USL, RVS, vagina and bladder in women with clinical suspicion of DIE using the surgical data as a reference standard. Study quality was assessed using the PRISMA guidelines and QUADAS-2 tool.

RESULTS: Of the 801 citations identified, 11 studies (n = 1583) were considered eligible and were included in the meta-analysis. For detection of endometriosis in the USL, the overall pooled sensitivity and specificity of TVS were 53% (95%CI, 35-70%) and 93% (95%CI, 83-97%), respectively. The pretest probability of USL endometriosis was 54%, which increased to 90% when suspicion of endometriosis was present after TVS examination. For detection of endometriosis in the RVS, the overall pooled sensitivity and specificity were 49% (95%CI, 36-62%) and 98% (95%CI, 95-99%), respectively. The pretest probability of RVS endometriosis was 24%, which increased to 89% when suspicion of endometriosis was present after TVS examination. For detection of vaginal endometriosis, the overall pooled sensitivity and specificity were 58% (95%CI, 40-74%) and 96% (95%CI, 87-99%), respectively. The pretest probability of vaginal endometriosis was 17%, which increased to 76% when suspicion of endometriosis was present after TVS assessment. Substantial heterogeneity was found for sensitivity and specificity for all these locations. For detection of bladder endometriosis, the overall pooled sensitivity and specificity were 62% (95%CI, 40-80%) and 100% (95%CI, 97-100%), respectively. Moderate heterogeneity was found for sensitivity and specificity for bladder endometriosis. The pretest probability of bladder endometriosis was 5%, which increased to 92% when suspicion of endometriosis was present after TVS assessment.

CONCLUSION: Overall diagnostic performance of TVS for detecting DIE in uterosacral ligaments, rectovaginal septum, vagina and bladder is fair with high specificity.