P17.02: The reproducibility of the three dimensional (3D) ultrasonography in the diagnosis of deep endometriosis: effect of a specific training

To investigate the intraobserver and interobserver agreements of three-dimensional ultrasonography (3D US) in the diagnosis of deep endometriosis evaluating also the effect of specific training. Two sonographers (one expert with more than 15 years of experience in gynecologic ultrasonography, operator A and one with 3 years of experience in ultrasonography but one year in gynecologic ultrasonography and use of 3D, operator B) performed a retrospective review of 3D volumes from a random sample of 35 patients with clinical suspicion of deep pelvic endometriosis submitted to surgery. Each examiner was to review the 3D volumes using virtual navigation and multiplanar display twice, 1 week apart for assessing intraobserver agreement. Reviewing was performed with examiners blinded to each other. After six months of specific training in a II levels ultrasonography laboratory the less expert operator re-evaluated the same volumes using the same modalities. Intraobserver and interobserver agreement were assessed by calculating the kappa index (κ) with 95% confidence intervals (CIs). Intraobserver agreement was good for both examiners with different degrees of experience (kappa 0.8754, 95% CI 0.7013-1 for operator A and 0.7087 95% CI 0.5002- 0.9172 , for operator B respectively ) but increases in the less expert operator to 0.8794, 95%CI 0.7423-1. Interobserver agreement was good (kappa 0.7094, 95% CI 0.5007-0.9181) but increases after the training of the less expert operator to 0.8394, 95% CI 0.6808-0.998. The use of 3D in the diagnosis of deep endometriosis is a reproducible technique and, although the difference was not statistically significant, the use of specific training seems further increase the reproducibility. This study was supported in part by the Regione Autonoma della Sardegna (project code CPR-24750).