Comparison of Diagnostic Accuracy of Transvaginal Ultrasound for Deep Endometriosis and Pouch of Douglas Obliteration in The Presence and Absence of Ovarian Endometriomas

Study Objective We aim to evaluate and compare the diagnostic accuracy of deep endometriosis transvaginal ultrasound (DE TVS) in the presence and absence of ovarian endometriomas (OEs). Design This is a multicenter observational study of prospectively collected data. Setting Patients attended one of two gynecology-focused ultrasound practices and underwent laparoscopy at one of nine hospitals in the Sydney metropolitan area between 2009 and 2017. Patients or Participants Patients referred with suspected endometriosis. Interventions A DE TVS was performed in all patients and reported in a systematic fashion. Measurements and Main Results Surgical findings were the defined reference standard. Data were analyzed to determine the prevalence of surgically-diagnosed DE and pouch of Douglas (POD) obliteration in the presence or absence of OEs. DE TVS was assessed in terms of accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and positive (LR+) and negative (LR-) likelihood ratios. Comparison of DE TVS prediction of DE and POD obliteration in patients with versus those without OEs was performed. 410 patients were included. An OE was noted in 91, 22.2% patients (left: 66, 16.3%; right: 58, 14.2%; bilateral: 33, 8.2%). When comparing those with and without an OE, DE prevalence was 62/91, 68.1% and 81/313, 25.9%, respectively, and POD obliteration prevalence was 55/91, 60.4% and 31/313, 9.9%, respectively (p<0.001). There was better sensitivity for DE (85.5 vs 59.3%) and POD obliteration (81.8 vs 64.5%) with an OE present versus absent, respectively (p<0.001). Conversely, specificity was higher for DE (92.7 vs 79.3%, p<0.001) and POD obliteration (98.6 vs 88.6%, p=0.13) with an OE absent vs present, respectively. Conclusion The prevalence of DE and POD obliteration in patients without OEs is lower, but not negligible. Diagnostic accuracy of DE TVS is poorer in patients without OEs, which is not well understood and concerning as many patients may be falsely reassured