Correlation of Sonographic and Intraoperative Findings of Deep-Infiltrating Endometriosis

INTRODUCTION: Endometriosis is a chronic inflammatory gynecologic condition affecting approximately 5–10% of reproductive-aged women. Preoperative evaluation typically consists of pelvic imaging such as ultrasound (US) or MRI to assess for deep infiltrating endometriosis (DIE), ovarian endometriosis, and/or adenomyosis. Due to its ease of access and cost-effectiveness, sonography should be the first-line imaging modality; a growing body of literature has demonstrated its high sensitivity and specificity for DIE. OBJECTIVE: While sonography has improved the preoperative diagnosis of endometriosis, the purpose of this study is to determine its accuracy in detecting intraoperative DIE lesions and predict the need for advanced laparoscopic procedures for patients desiring surgical evaluation and treatment. METHODS: An IRB-exempt retrospective study was conducted over a 12-month period at a tertiary-level academic hospital. Inclusion criteria included women >18 years old who underwent a laparoscopic resection of endometriosis with or without hysterectomy for pelvic pain, suspected/confirmed history of endometriosis, and/or infertility by the minimally invasive gynecologic surgery (MIGS) team. All patients had a preoperative office US performed by the same experienced sonographer within 6 months of surgery. Any patient who underwent surgery and did not have a preoperative US by this sonographer was excluded. Based on the 2016 IDEA Consensus Statement, a negative sliding sign, ovarian immobility, presence of endometrioma, and lesions or nodules in the anterior or posterior compartment were considered sonographic markers of DIE. Intraoperative DIE was defined by the presence of lesions >5 mm in depth, peritoneal pockets, ovarian endometrioma, uterosacral nodules, bowel nodules, retrocervical/rectovaginal space abnormalities, and obliterated posterior cul-de-sac. Advanced laparoscopic procedures such as significant adhesiolysis/enterolysis >30 minutes, ureterolysis, rectal shaving, bowel resection, and appendectomy were noted if performed. Demographic data, symptomatology, endometriosis history, fertility status, and pathology reports were collected. Appropriate statistical tests were applied for continuous and categorical variables. RESULTS: Of 117 patient charts, 89 patients met inclusion criteria. The most common presenting symptoms were dysmenorrhea and dyspareunia, and approximately half of the patients desired future fertility. Pathological confirmation of endometriosis was found in 90% of patients. The overall sensitivity and specificity for detecting intraoperative DIE was 61% (49–72%) and 94% (71–100%), respectively. The overall PPV, NPV, and odds ratio (OR) were 98% (88–100%), 36% (22–52%), and 24.4 (3.4–1071) (p<0.001), respectively. Specific ultrasound markers associated with intraoperative DIE were negative sliding sign (OR: 7.12 [1.47 – 69.3], p-value 0.006), ovarian immobility (OR: 3.30 [0.94 – 12.2], p-value 0.040), and rectovaginal space abnormality (OR: 19.9 [1.15 – 345.3], p-value 0.002). The OR for significant enterolysis/adhesiolysis >30 minutes and ureterolysis with the presence of sonographic markers of DIE was 11.3 (3.55–44.0, p<0.001) and 3.29 (1.19–9.70, p-value 0.013), respectively. Appendiceal endometriosis was found in 50% of patients who had US and intraoperative findings of DIE (OR 25.0 [1.24–503.4], p-value 0.006) (Figures 1 and 2). CONCLUSIONS: Sonographic markers of DIE can predict intraoperative DIE lesions and the need for advanced laparoscopic procedures during endometriosis resection. The results of this study support the utilization of ultrasound for the preoperative evaluation and diagnosis of endometriosis to improve surgical planning and patient outcomes (Table 1).