Diagnostic Accuracy of Ultrasound and MRI in the Mapping of Deep Pelvic Endometriosis Using the International Deep Endometriosis Analysis (IDEA) Consensus

Abstract

Objectives. The primary aim was to investigate the diagnostic accuracy of transvaginal ultrasound (TVS) and magnetic resonance imaging (MRI) in the mapping of deep pelvic endometriosis (DE) in a diseased population. The secondary aim was to offer first insights into the clinical applicability of the new International Deep Endometriosis Analysis group (IDEA) consensus for sonographic evaluation, which was also adapted for MRI and surgical reporting in this study. Methods. The study was a prospective observational cohort study. In this study, consecutive women planned for surgical treatment for DE underwent preoperative mapping of pelvic disease using TVS and MRI (index tests). The results were compared against the intraoperative findings with histopathological confirmation (reference standard). In case of disagreement between intraoperative and pathology findings, the latter was prioritised. Index tests and surgical findings were reported using a standardised protocol based on the IDEA consensus. Results. The study ran from 07/2016 to 02/2018. One-hundred and eleven women were approached, but 60 declined participation. Out of the 51 initially recruited women, two were excluded due to the missing reference standard. Both methods (TVS and MRI) had the same sensitivity and specificity in the detection of DE in the upper rectum (UpR) and rectosigmoid (RS) (UpR TVS and MRI sensitivity and specificity 100%; RS TVS and MRI sensitivity 94%; TVS and MRI specificity 84%). In the assessment of DE in the bladder (Bl), uterosacral ligaments (USL), vagina (V), rectovaginal septum (RVS), and overall pelvis (P), TVS had marginally higher specificity but lower sensitivity than MRI (Bl TVS sensitivity 89%, specificity 100%, MRI sensitivity 100%, specificity 95%; USL TVS sensitivity 74%, specificity 67%, MRI sensitivity 94%, specificity 60%; V TVS sensitivity 55%, specificity 100%, MRI sensitivity 73%, specificity 95%; RVS TVS sensitivity 67%, specificity 100%, MRI sensitivity 83%, specificity 93%; P TVS sensitivity 78%, specificity 97%, MRI sensitivity 91%, specificity 91%). No significant differences in diagnostic accuracy between TVS and MRI were observed except USL assessment (p=0.04) where MRI was significantly better and pouch of Douglas obliteration (p=0.04) where TVS was significantly better. Overall, there was a good agreement between reference standards and both index tests (TVS kappa value (κ) = 0.727 [p≤0.001), MRI κ = 0.755 [p=p≤0.001]). Conclusion. We found that both imaging techniques had overall good agreement with the reference standard in the detection of deep pelvic endometriosis. This is the first study to date involving the IDEA consensus for ultrasound, its modified version for MRI, and intraoperative reporting of deep pelvic endometriosis in clinical practice.