PWE-378 Total pelvic floor ultrasound in pelvic floor defaecatory dysfunction

Introduction Defaecation proctography is gold standard for pelvic floor defaecatory dysfunction assessment but entails radiation. Total pelvic floor ultrasound (USS) allows a safe, dynamic assessment in clinic. The aim was to determine accuracy of USS, compared to proctography, for anatomical findings in women with pelvic floor defaecatory dysfunction. Secondary aims were identification of USS predictors of a trapping rectocoele, significant enterocoele and surgical treatment. Method All consecutive women undergoing investigation for defaecatory dysfunction between May 2011 and November 2014 were retrospectively reviewed by an interpreter blinded to symptoms and results. The following data were collated; The presence/absence of a rectocoele, intussusception and enterocoele, and grade of intussusception for both modalities. The presence of a trapping rectocoele and/or significant enterocoele for proctography. Subsequent treatment (biofeedback/surgery) was recorded. Results 393 underwent USS and proctography. 69 were excluded (images unavailable (24), incomplete (38) or poor quality (5) or tests performed over 2 months apart (2)). Of the 324 included mean length of follow up was 22 months. On proctography 283 had a rectocoele (111 trapping), 142 intussusception and 80 an enterocoele (40 significant). On ultrasound 284 had a rectocoele, 105 intussusception and 55 an enterocoele. 46 underwent surgery. Using proctography as gold standard USS had the following results. Abstract PWE-378 Table 1 USS Sensitivity Specificity Positive Predictive Value Negative Predictive Value --- Rectocoele (Transperineal) 93.84% 52.50% 93.31% 53.85% Rectocoele (Transvaginal) 68.10% 65.91% 92.68% 24.58% Intussusception 58.45% 86.83% 79.05% 71.08% Enterocoele 58.75% 96.64% 85.45% 87.45% Rectocoeles detected on both transperineal and transvaginal USS were more likely to produce barium trapping on proctography (p < 0.05) and require surgery (p < 0.05). An enterocoele detected on USS was more likely a significant enterocoele on proctography (p < 0.05). Agreement between the grading of intussusception by USS and proctography was moderate (intra-class correlation coefficient 0.52). Conclusion Transperineal USS is highly sensitive for detecting rectocoele. A rectocoele visible on both transperineal and transvaginal USS is significantly more likely to produce trapping on proctography and require surgery. Also, an enterocoele detected on USS is significantly more likely to descend onto the l