Objective: To evaluate the accuracy of endorectal ultrasound examination to ascertain the deepest rectal layer involved in rectal endometriosis.
Design: Retrospective study.
Setting: Department of obstetrics and gynecology at a university hospital in France.
Patient(s): Women presenting with rectal endometriosis who had undergone rectal resection during a 22-month period.
Intervention(s): Endorectal ultrasonography.
Main outcome measure(s): The predicted rectal infiltration depth by using endorectal examination was compared with histological findings. The level of agreement was evaluated by using the coefficients of concordance kappa and weighted kappa.
Result(s): Sixteen women were included in the study. Rectal resection was segmental in 14 cases and was limited in 2 cases. The agreement between 2 examinations was considered good in 9 cases (56%). Endorectal ultrasound overestimated the depth of infiltration in 5 cases and underestimated it in 2 cases. The coefficients of concordance kappa (95% confidence interval) and weighted kappa (95% confidence interval) were, respectively, 0.17 (0-0.34) and 0.22 (0.04-0.4), corresponding to poor concordance between the endorectal ultrasonography and histological examination.
Conclusion(s): Accuracy in the prediction of rectal-layer involvement in endorectal ultrasonography appears to be limited. This information should not be considered sufficient when selecting the type of rectal-resection procedure.