How repeatable is assessment of external anal sphincter trauma by exoanal 4D ultrasound?

Three-/four-dimensional translabial ultrasound (TLUS) is gaining popularity for the assessment of anal sphincter trauma, although repeatability data are lacking. This study aimed to determine the repeatability of tomographic ultrasound imaging (TUI) on TLUS for the diagnosis of external anal sphincter (EAS) trauma and compare the performance of a novice with that of an experienced investigator.

This was a retrospective study of archived ultrasound datasets of patients who presented with symptoms of pelvic floor dysfunction and were examined twice between 2012 and 2016 at an average interval of 260 (range, 1-1100) days. All volumes were obtained using a GE Medical Systems Voluson 730 Expert or E8 ultrasound system. Post-processing of volumes was performed independently by two investigators, one with over 1 year's experience and another with no prior experience in using TUI, who were blinded to clinical data, each other's results and the results obtained at the first timepoint. Significant trauma on EAS was diagnosed if four of the six TUI slices showed a defect of ≥ 30°. Intra- and interobserver agreement were determined using Cohen's kappa (κ) and intraclass correlation coefficients.

During the study period, 105 women underwent two TLUS assessments of the anal sphincter. Of these, 103 patients with ultrasound volumes available for both timepoints were included in the analysis. The novice investigator demonstrated average repeatability for assessment of significant EAS trauma and single-slice defect (κ, 0.30 and 0.22, respectively) despite relatively high agreement between measurements obtained at the two timepoints (84.5% and 79.3%, respectively). The experienced investigator demonstrated good to very good repeatability for significant EAS trauma and single-slice defect (κ, 0.91 and 0.78, respectively) between the two assessments, which equates to 98.1% and 94.7% agreement, respectively.

The repeatability of TLUS measurements for diagnosis of EAS trauma seems to be very good when imaging is undertaken with state-of-the-art equipment and the analysis is performed by an experienced observer; however, the performance of a novice investigator is much poorer. Copyright © 2018 ISUOG. Published by John Wiley & Sons Ltd.