VP62.28: Cloudy with no chance of ligation: predicting the need for ureterolysis during laparoscopy for endometriosis

Considered an advanced surgical skill, ureterolysis is necessary to ensure safe surgical treatment of endometriosis. The objective of this study was to develop a decision tree model to predict the need for ureterolysis during laparoscopy for endometriosis. We undertook a retrospective observational study of 94 women who had an advanced transvaginal ultrasound (TVS) and subsequent laparoscopy +/- excision of endometriosis (January 2017–September 2019). Demographic information, clinical symptoms, advanced TVS, intra-operative findings, and histopathology were all extracted for analysis through clinical and operative notes. 133 clinical, TVS and operative end-points were recorded. Univariate/multivariate analyses were performed to determine significant prognostic variables associated with ureterolysis. The decision tree model was developed on the training data, using a J48 algorithm in Weka 3.8, and validated using 5-fold cross-validation. Of 96 patients analysed, 48 (51.1%) had a ureterolysis performed at surgery. The major pre-operative clinical predictors for ureterolysis were dyschezia, non-cyclical or generalised pain, and lower abdominal pain. Ultrasonographic predictors were deep endometriosis (DE) on the right uterosacral ligament (RUSL) and rectum, and a negative sliding sign indicating pouch of Douglas obliteration. Our initial decision tree has a positive predictive value (PPV) of 81.6% and a negative predictive value (NPV) of 69.6% with a satisfactory c-statistic of 0.779 indicating good concordance. On stratified 5-fold cross-validation, the PPV was 64.7%, the NPV was 65.1% and the c-statistic was 0.685. Our study suggests pre-operative clinical and ultrasound features may be used to predict the need for ureterolysis in laparoscopy for endometriosis. In this way, pain symptomology, a negative sliding sign, rectal and/or RUSL deep disease on ultrasound may be beneficial for the gynecology team advanced preoperative surgical planning.