ABSTRACT
BACKGROUND: Rectovaginal endometriosis is a severe form of pelvic endometriosis in which pharmacological treatment is relatively ineffective (Vercellini et al., Fertil Steril. 2005;84:1375-87). Laparoscopic surgical treatment is effective, but has the potential risks of bowel perforation and colostomy formation (Darai et al., Am J Obstet Gynecol. 2005;192:394-400). Transrectal ultrasound scanning can be applied as a preoperative tool to predict the presence of rectovaginal endometriosis and bowel wall involvement (Abrao et al., J Am Assoc Gynecol Laparosc. 2004;11:50-4). METHODS: Thirty-two women underwent transrectal ultrasound followed by therapeutic laparoscopy. Likelihood ratios and post-test prevalences were calculated with Fagan's normogram. This was then extrapolated with the aid of a mathematical model to a low-risk population. RESULTS: A positive likelihood ratio was found to be 10.89 (95% confidence ratio (CI): 1.62-73.15) and a negative likelihood ratio was found to be 0.24 (95% CI: 0.1-0.57). The pre-test prevalence of rectovaginal endometriosis was 56%. The positive post-test prevalence probability was 93%, and the negative post-test prevalence probability was 23%. CONCLUSION: Preoperative transrectal ultrasound scanning for rectovaginal endometriosis is an extremely accurate predictive test, and strongly predicts the need for extensive laparoscopic dissection and potential bowel resection.
