Comparison of endoscopic ultrasound and magnetic resonance imaging in severe pelvic endometriosis.

UNLABELLED

Deep pelvic endometriosis may lead to severe pain, the treatment of which may require complete surgical resection of lesions. Digestive infiltration is a difficult therapeutic problem. Preoperative diagnosis is difficult and digestive infiltration may remain unknown with incomplete resection and sometimes repeated surgery. Both magnetic resonance imaging (MRI) and endoscopic ultrasonography are able to detect rectosigmoid infiltration but their usefulness in the preoperative staging is still to be evaluated. The aim of this work was to evaluate and compare both techniques in the preoperative detection of deep pelvic endometriosis, particularly digestive infiltration.

PATIENTS AND METHODS

From 1996 to 1998, 48 women with painful deep pelvic endometriosis had preoperative imaging exploration with endoscopic ultrasonography and MRI, and were operated on in order to attempt complete endometriosis resection. Patients were proposed for laparoscopic resection if endoscopic ultrasonography and/or MRI did not reveal digestive infiltration or for open resection if endoscopic ultrasonography and/or MRI were positive for digestive infiltration.

RESULTS

Endoscopic ultrasonography and/or MRI led to suspicion of digestive endometriosis in 16 patients. Surgical resection was performed in 12 and digestive wall invasion was histologically demonstrated. At final follow-up, all patients had a dramatic decrease of their symptoms. The remaining 4 patients refused digestive resection and had only laparoscopic gynecologic resection. Infiltration although not histologically proven was very likely both on operative findings and clinical evolution. Digestive infiltration was preoperatively excluded in the 32 other patients. All had a laparoscopic treatment without digestive resection and pain diminished in all patients. In the 12 patients group who had digestive resection, digestive infiltration was correctly diagnosed by endoscopic ultrasonography in all cases (no false negative) whereas MRI, even with the use of endocoil antenna, led to correct diagnosis in 8 out of 12 cases. When endoscopic ultrasonography was negative for digestive infiltration, laparoscopic resection of lesions at surgery appeared complete in all cases. For the 16 patients with presumed digestive infiltration, sensitivity of endoscopic ultrasonography and MRI was 100 and 75% respectively, with a 100% specificity in both cases. MRI appeared very accurate for the detection of ovarian endometriotic locations. MRI was more sensitive but less specific than endoscopic ultrasonography for the diagnosis of isolated endometriotic recto-vaginal septum and utero-sacral ligaments lesions.

CONCLUSION

Endoscopic ultrasonography was the best technique for the diagnosis of digestive endometriotic infiltration, which complicates the therapeutic strategy. MRI, however, allows more complete staging of other pelvic endometriotic lesions.