P-309 Prediction of peritoneal endometriosis in symptomatic women with negative examination and imaging

Abstract Study question Can we predict peritoneal endometriosis (PE) in women with pain symptoms who have no signs of endometriosis on examination or imaging studies? Summary answer A calculated score including dysmenorrhea, dysuria, dyspareunia, no pain outside of menses and presence of pain for >3 years can predict the presence of PE. What is known already Endometriosis is a common cause of chronic pelvic pain, characterized by dysmenorrhea, dyspareunia, dysuria and dyschezia. However, not all women with these symptoms have endometriosis. While endometriomas, deep infiltrating endometriosis lesions and adenomyosis can be detected with increasing accuracy through exams and with transvaginal ultrasound and/or MRI, it is not possible to accurately diagnose superficial peritoneal lesions using these methods. A clinical predictive model for assessing the risk of endometriosis, sparing the need for diagnostic surgery, would be useful. Study design, size, duration This prospective, multicentre, cross-sectional study was conducted between September 2022 and January 2024 at 18 endometriosis centers in Austria, Germany and Switzerland. The study enrolled 838 symptomatic women with suspected endometriosis scheduled for surgery. Clinical data including severity, timing and duration of specific symptoms, family history of endometriosis, and quality of life and sexual function were collected pre-operatively. All women underwent surgical assessment of endometriosis using the ASRM and #ENZIAN classifications by expert surgeons. Participants/materials, setting, methods Of the 838 participants, 79 were found to have only peritoneal endometriosis while 58 women had no signs of endometriosis or adenomyosis at surgery. We compared potential clinical risk factors that could differentiate these two groups. Factors were evaluated using univariate logistic regression analysis and a model was built to predict the presence of endometriosis using multivariate logistic regression analysis. Main results and the role of chance Women with peritoneal endometriosis and no endometriosis did not differ in age. Women without endometriosis reported the four main symptoms with a prevalence of dysmenorrhoea of 67.2%, dyschezia of 27.6% dysuria of 12.1%, and dyspareunia of 48.3%. Compared to those with peritoneal endometriosis with prevalences of 83.5%, 32.9%, 29.1%, and 67.1%, respectively. Dysmenorrhea (OR 2.47 [1.10-5.55]), dysuria (OR 2.99 [1.18-7.56]), dyspareunia (OR 2.18 [1.09-4.38]), and presence of pain for >3 years (OR 3.10 [1.52-6.32]) were associated with peritoneal endometriosis. Sexual function assessed with FSFI and quality of life assessed with EHP-30 did not differ between the two groups. A score for the presence of peritoneal endometrioses was derived from these parameters, assigning points for each symptom corresponding to the OR. The resulting ROC curve had an AUC of 0.74 (0.65, 0.82) for predicting peritoneal endometriosis. Limitations, reasons for caution The is a preliminary study with a relatively small number of subjects. While bootstrapping was used to internally validate the model. external validation of the model using a second population is in progress Wider implications of the findings The prediction model can be easily applied in clinical practice to estimate the probability of peritoneal endometriosis in women with symptoms but no clinical/ imaging signs of disease. If validated, it could decrease the number of diagnostic surgeries, either leading to targeted treatment or further evaluation for non-endometriosis causes. Trial registration number Yes