Unilateral coronal diameters of the levator hiatus: baseline data for the automated detection of avulsion of the levator ani muscle

Abstract

Levator ani avulsion injury is associated with pelvic organ prolapse. Although it can be detected by magnetic resonance imaging or translabial ultrasonography, both require substantial training. This retrospective study provides baseline measurements of the unilateral coronal or transverse diameter of the levator hiatus and assesses its use in the prediction of levator ani avulsion injury, as a step towards automated detection. Datasets of 787 patients seen in a tertiary urogynecological unit were assessed, including history, examination and three/four-dimensional pelvic floor ultrasound. Unilateral coronal diameters were measured off-line in stored volume datasets, with the operator blinded to clinical data. Volumes obtained during maximum pelvic floor muscle contraction were used, resorting to volumes obtained at rest for patients unable to contract their pelvic floor muscles. Ultrasound images were available for 761 patients, including 16 (2.1%) who could not contract their pelvic floor muscles, generating 1522 unilateral measurements (both sides for 761 patients). Levator ani avulsion injury was associated with significant increase in the unilateral coronal diameter (2.57 +/- 0.5 vs. 2.1 +/- 0.3 cm, P < 0.0001). Receiver-operating characteristics curve analysis showed an area under the curve of 0.801 (95% CI, 0.765-0.837). A cut-off of 2.3 cm had a sensitivity of 71% and a specificity of 79% for the diagnosis of levator ani avulsion injury. Levator ani avulsion injury is associated with a significant increase in unilateral coronal diameter, a cut-off point of 2.3 cm being most appropriate for alerting clinicians to a high likelihood of this diagnosis. It is hoped that this information will help to automate the detection of levator ani avulsion injury.