Background: The aim was to evaluate the diagnostic accuracy of transvaginal tenderness-guided ultrasonography in the identification of location of deep endometriosis.
Methods: Consecutive women scheduled for surgery in our Department for clinically suspected endometriosis were included in this prospective study. All women underwent modified transvaginal ultrasonography using a stand-off in the week before surgery, which also evaluated the painful sites evocated by a gentle pressure of the probe. Five locations of deep endometriosis were considered: vaginal walls, rectovaginal septum, rectosigmoid involvement, uterosacral ligaments and anterior compartment (anterior pouch and/or bladder). Sensitivity, specificity and likelihood ratios (LR+/-) were calculated with 95% confidence intervals (CIs).
Results: We included 88 women; surgery associated with histopathological evaluation revealed deep endometriosis in different pelvic locations in 72 patients. With respect to the vaginal walls, transvaginal ultrasonography had a sensitivity of 91% (95% CI, 79-97%), specificity of 89% (95% CI, 81-93%), an LR+ of 8.2 and an LR- of 0.09. For endometriosis of rectovaginal septum, transvaginal ultrasonography had a sensitivity of 74% (95% CI, 64-80%), specificity of 88% (95% CI, 4-8%), an LR+ of 6.2 and an LR- of 0.3. For other locations, the sensitivity was lower (ranging from 67% to 33%) with a comparable specificity.
Conclusions: This technique shows a high specificity and sensitivity in the detection of vaginal and rectovaginal endometriosis. Good specificity associated with a lower sensitivity was obtained in the diagnosis of deep endometriosis of uterosacral ligaments, rectosigmoid involvement or anterior deep endometriosis.