EP537 - Accuracy of Surgeon Interpretation in Novel Intraoperative Ultrasound Technique for Deep Infiltrating Endometriosis

Image or Table (Do Not add text here):

Purpose/Background: Endometriosis is a complex, elusive, spectral disease process that affects at least 10% of reproductive aged women, although this is likely underreported. Accurate, expedient diagnosis as well as appropriate management is often challenging due to many factors, including the non-specific presentation, lack of effective, non-invasive diagnostic tools, and absence of consensus regarding indication or timing for surgical intervention.

Appropriate intraoperative technique for deep infiltrating endometriosis (DIE) lesions remains debated. There are three methods for surgical management of DIE: shaving, discoid excision, or segmental resection. Historically, decisions regarding surgical approach are based on extent of lesion. Complete surgical excision of all disease is critical to minimize risk of recurrent symptoms, which highlights the importance of adequate interpretation of extent of lesion, including depth of invasion. Our institution has implemented a novel surgical approach utilizing intraoperative ultrasound (IOUS) to delineate depth of invasion and extent of DIE lesions involving the bowel to assist with surgical decision-making. In this study, we sought to determine the accuracy of surgeon interpretation of this IOUS technique.

Methods/Interventions: A retrospective review of a minimally invasive gynecologist’s database of DIE cases was reviewed. Patients who underwent surgical intervention utilizing our IOUS technique were selected for review. Ultrasound images from their charts were de-identified and placed into a password-protected folder. Patients were randomly assigned to two surgeons for independent, blinded review. Surgeons reviewed ultrasound images for depth of invasion and answers were recorded on an answer sheet. Primary outcome was accuracy of ultrasound interpretation versus pathologic depth of invasion of each individual surgeon. A secondary outcome measure was inter-rater reliability of ultrasound image interpretation. All data collection and analysis were performed by a third independent analyst.

Results/Outcomes: 35 patients were identified meeting inclusion criteria. 12 were excluded from this initial analysis due to missing pathologic information. Mean age was 37.1 years (26-50) and mean BMI was 27.7 (17.6-35.9). 76% had previous abdominopelvic surgery. Pathologic depth of invasion was noted to be to the serosa in 2 (8.7%) patients, muscularis in 14 (60.9%) patients, submucosa in 2 (8.7%) cases, and the mucosa in 5 (21.7%) cases. Surgeon A and B correctly predicted pathologic depth of invasion based on ultrasound images 65.2% and 50% of time, respectively. Of the 17 cases they both independently evaluated, there was concordance between surgeon interpretation 64.7% of time.

Conclusion/Discussion: IOUS provides useful information to assist with improving the accuracy of defining the extent of DIE lesions, especially regarding depth of invasion. The inter-rater reliability of 64.7%, and individual surgeon rates of correct prediction of pathologic depth of invasion noted in our initial experience suggests that we are still in our learning curve of interpreting these images. Additionally, further development of this technique with adjuncts such as intraluminal saline instillation may improve the accuracy of surgeon depth of invasion prediction, which ultimately could result in downstaging of surgical intervention and improved patient outcomes without negatively impacting recurrence.