P28.11: How reliable is office gel sonovaginography in predicting depth of infiltration of bowel DIE?

To determine the accuracy of office gel sonovaginography in predicting the depth of infiltration of bowel DIE. A retrospective review of 20 patients who had segmental resections for bowel endometriosis over a 4 year period. Each patient had an office gel sonovaginography as part of a mandatory preoperative work up protocol in our unit. During office gel sonovaginography, the depth of invasion of the bowel DIE was ascertained and recorded as involving the serosa/muscularis propria alone or extending to the submucosa or mucosa. The reports of histopathology examination of each resected bowel specimen were then reviewed and the depth of bowel infiltration determined from the pathologist reports. These were recorded and compared with the ultrasonographic prediction of depth of infiltration of bowel DIE. Data was complete for 18/20 patients. The sensitivity, specificity, PPV, NPV, test accuracy were calculated for the ability of gel sonovaginography to predict depth of infiltration of bowel DIE. The PPV of gel SVG for the prediction of presence of DIE affecting the muscularis propria is 100%. The sensitivity, specificity, PPV, NPV, and test accuracy for submucosal/mucosal layer involvement were 71.4%, 64%, 55.5%, 78%, 67% respectively. In our unit, gel SVG is a valuable tool in the preoperative management of bowel endometriosis. Whereas it can accurately predict infiltration of the bowel muscularis propria by endometriosis, it is less valuable in predicting submucosal/mucosal involvement.