Intra- and interobserver reproducibility of transvaginal ultrasound for the detection and measurement of endometriotic lesions of the bowel.

Abstract

The number and invasion depth of endometriotic bowel lesions, total length of bowel affected by endometriosis, lesion-to-anal verge distance, and extent of pouch of Douglas obliteration are important factors in preoperatively determining risk and complexity of endometriosis surgery. The intra- and interobserver reproducibility of transvaginal ultrasound in the evaluation of many of these parameters has not yet been investigated. Our study aimed to assess the intra- and interobserver reproducibility of transvaginal ultrasound between an experienced and less experienced examiner for all of these parameters. This prospective observational cross-sectional study was conducted between July 2019 and November 2020. Fifty consecutive premenopausal women who underwent transvaginal ultrasound examination in our clinic for the first time, were examined by the same two operators during the same attendance. Outcomes of interest were the inter-rater reproducibility of transvaginal ultrasound for detecting the presence, number, depth and size of bowel endometriotic nodules, lesion-to-anal-verge distance, total length of bowel affected, and pouch of Douglas obliteration. The intraobserver reproducibility was assessed for the continuous parameters. Cohen's kappa (κ) statistic, Cohen's weighted kappa (κ), proportions of agreement, intraclass correlation coefficient (ICC) and Bland-Altman limits of agreement were used to assess the reproducibility of the parameters. The inter-rater agreement and reliability were very good for identifying bowel endometriosis, the number and invasion depth of bowel nodules, determining whether the maximum nodule length was <3 cm, and lesion-to-anal-verge distance <8 cm (proportion of agreement 0.92, 0.94, 0.97, 0.94, 0.96; κ 0.92, 0.91, 0.92, 0.82, 0.89). The inter-rater agreement and reliability were good for assessing pouch of Douglas obliteration (proportion of agreement 0.86, κ 0.80). The intra-rater reliability for the mean nodule diameter (ICC 0.93 and 0.97) and total length of bowel affected (ICC 0.94 and 0.91) were excellent for operators A and B, respectively. The inter-rater reliability for the mean nodule diameter was good (ICC 0.80), and moderate for the total length of bowel affected (ICC 0.70). The Bland-Altman limits of agreement demonstrated clinically acceptable ranges for these two parameters. This study demonstrated a high intra- and inter-rater reproducibility of transvaginal ultrasound in the diagnosis of bowel endometriosis and measurement of its various components.