OC02.04: Three dimensional (3D) ultrasonography improves the accuracy of two dimensional (2D) ultrasonography in the diagnosis of deep endometriosis

To compare the diagnostic accuracy of 2D and 3D “tenderness-guided” transvaginal ultrasonography in the identification of deep endometriosis. Two-hundred and two patients (mean age, 34 years) with clinical suspicion of deep pelvic endometriosis were prospectively enrolled. All women underwent modified transvaginal ultrasonography. One single examiner performed all ultrasound examinations. Two locations of deep endometriosis were considered: rectosigmoid involvement and other posterior lesions (including retrocervical location, rectovaginal septum, and uterosacral ligaments). Once 2 D ultrasonography was done a 3D volume was acquired and stored. After six months from the last examination to avoid the risk of recall bias the same operator evaluated the 3D volumes using virtual navigation to provide a presumptive diagnosis of presence and localization of deep endometriosis. Sensitivity, specificity, and likelihood ratios were calculated with 95% confidence intervals (CIs). The area under the receiver operating characteristic curve (AUC) of 2D and 3D for the two locations was calculated by using five levels of diagnostic confidence. Surgery revealed deep endometriosis in different pelvic locations in 129 patients. For the rectosigmoid involvement, the specificity, sensitivity, and LR+ and LR− were 93%, 95%, 13 and 0.06, respectively, for 2D and 97%, 91%, 25 and 0.09, respectively, for 3D. For other posterior locations, the specificity, sensitivity, and LR+ and LR− were 88%, 71%, 6.10 and 0.32, respectively, for 2D and 94%, 87%, 14 and 0.14, respectively, for 3D. The AUCs for rectosigmoid location were similar for both techniques. The AUCs for other posterior locations were significantly different (0.789 for 2D and 0.891, respectively, P = 0.01). 3D ultrasonography is useful only to increase the diagnostic accuracy of 2D in the diagnosis of other posterior localizations of deep endometriosis. This study was supported in part by the Regione Autonoma della Sardegna (project code CPR-24750).