POMEN ENDOSKOPSKEGA ULTRAZVOKA V DIAGNOSTIKI ENDOMETRIOZE cREVESA

Abstract Background: Deep inltrating endometriosis (DIE) aects rectosigmoid in up to one third of patients. In these cases bowel resection with endto-end anastomosis is indicated. Our aim was to determine sensitivity, specicity, positive and negative predictive value and accuracy of rectal endoscopic ultrasound (REUS) for preoperative assessment of bowel in patients with DIE. Methods: In this retrospective study we included 72 patients who underwent surgery for DIE between 2004 and 2010 in University Medical Centre Ljubljana, Slovenia and had REUS preoperatively. REUS ndings were compared with intraoperative ndings and in case of bowel re-section also with patohistological ndings. Results: Bowel inltration was found in 29/72 (40 %) of patients during surgery. Bowel resection was performed in 23 patientsin all cases patohistological examination conrmed endometriosis. REUS correctly conrmed or excluded endometriosis in 60/72 (83 %) of patients, in one patient (1,4 %) it was false positive and in 11 patients (15 %) it was false negative. Sensitivity and specicity of REUS were 62 % and 98 %, respectively. Coefficient Kappa between REUS and surgery was 0,63. Conclusions: REUS should be used during pre-operative workup of patients with DIE as it can accurately diagnose bowel lesions in approximately 80 % of patients. The positive result of REUS is very reliable (specicity 98 %), however, negative result should be interpreted with caution due to limited sensitivity (62 %) of the examination. Indeed, REUS is false negative in approximately one third of patients. Therefore, additional preoperative examinations should be performed in case of negative result of REUS when bowel endometriosis is suspected.