The role of a new three-dimensional ultrasound technique in the diagnosis of levator ani muscle avulsion.

Aims To evaluate the intermethod agreement between the tomographic ultrasound imaging (TUI), considered as the gold standard, and the OmniView-VCI in the diagnosis of levator ani muscle (LAM) avulsion and in the measurement of levator-urethral gap (LUG). Methods We acquired dynamic 4D transperineal ultrasound volumes from 114 women. Each data set was analyzed on maximal pelvic floor contraction by TUI and OmniView-VCI techniques to check for LAM avulsion. Moreover, we measured LUG using both TUI and OmniView-VCI, twice by an operator and once by another to assess intraobserver and interobserver reproducibility. Reproducibility and intermethod agreement were studied by means of intraclass correlation coefficient (ICC) and Cohen's kappa coefficient. Results In the diagnosis of ani levator avulsion, the two techniques showed a good agreement (Cohen's kappa = 0.691, 95% confidence interval [CI], 0.522-0.860; P < .001); we also reported a good intraobserver and interobserver agreement (Cohen's kappa = 0.738, 95% CI, 0.597-0.879; P < .001, and Cohen's kappa = 0.864, 95% CI, 0.750-0.978; P < .001, respectively). LUG measurements by OmniView-VCI technique showed high intraobserver (ICC 0.895; 95% CI, 0.866-0.918) and interobserver (ICC 0.821; 95% CI, 0.774-0.858) reproducibility. High intermethod agreement was demonstrated between the two methods (ICC 0.813; 95% CI, 0.764-0.853). The area under the receiver-operating characteristic curve of LUG in predicting avulsion was 0.931 (0.868-0.994, 95% CI; P < .001) with 24 mm showing the best sensitivity (82%) and specificity (97%). Conclusions OmniView-VCI is a reliable method for LUG measurement and for levator avulsion diagnosis.