EAS DEFECT SIZE: DOES IT MATTER?
N. SUBRAMANIAM, H. DIETZ;
Sydney Med. Sch. Nepean, Sydney, Australia.
Introduction: Transperinealultrasound (TPUS) is increasingly used for sphincter imaging and hasshown moderate to good correlation with endoanal ultrasound (EAU)1.Equivalent to definitions used on EAU2, a ‘significant’defect on TPUS is diagnosed if 4/6 slices on tomographic ultrasound(TUI) show a defect of ≥30◦ of the circumference of the externalanal sphincter (EAS).
Objective: To determine the optimalcut- off for the definition of 'significant defect of the EAS' asimaged by TPUS in single slices.
Methods: Thisretrospective study involved patients seen consecutively in atertiary urogynaecological unit between 2014 and 2015. All underwenta structured interview including St Mark’s incontinence score(SMIS), clinical examination and 4D TPUS. Volume data was obtained onpelvic floor muscle contraction to evaluate EAS trauma3.Tomographic offline analysis of archived sphincter images wasperformed at a later date, using proprietary software by the firstauthor, blinded against all clinical data. Individual slices wererated positive if there was a defect over at least 30o ofthe EAS circumference in 4/6 slices (Figure 1).
The correlationsbetween mean defect angle and all measures of anal incontinence (AI)were examined using non-parametric analysis. ROC statistics were usedto obtain a possible criterion value for mean defect angle of the EASas predictor of AI.
Figure 1: 3D TPUS of the anal canalshowing EAS defects in 4/6 slices
Results: Of the 890 womenseen during the study period, 842 were analysed after excluding 43for missing TPUS volumes and 5 for insufficient quality of volumedata. Mean age was 54 (16-84) years with a mean BMI of 29 (16-64)kg/m2. Median parity was 2 (0-8), 89 % (n=751) werevaginally parous and 26% (n=219) had a history of forceps. AI wasreported in 123 (15%) patients, with a median St Mark’s score of 12(IQR, 2-23).
Table1: Association between mean defect angle of anal sphincter andmeasures of anal incontinence
MeanDefect Angle of EAS
The associations between mean EASdefect angle and all measures of AI were highly significant (seeTable). However, ROC statistics (Figure 2) were of limited utility inthe identification of an optimal cut- off angle for 'significantdefect'. The maximum sensitivity of any defect angle measurement (>0degrees) was only 0.49. Given this very substantial limitation, theoptimum threshold for significant mean defect angle of the EAS was15o (sensitivity 0.40, specificity 0.79).
Figure 2: ROC for the associationbetween AI symptoms and mean EAS defect angle. AUC 0.60 (CI0.54-0.65, p=0.001); n= 842. The interrupted lines indicates absenceof ties beyond the 0.49 sensitivity level.
Conclusions: Wepropose a mean defect angle of 15o on TUI for analsphincter trauma as optimally predictive of AI. However,the modest characteristics of EAS angle as a test for AI limit theclinical utility of this finding.
1.Ultrasound Obstet Gynecol 2015; 46: 363-366
2. UltrasoundObstet Gynecol 2013; 42: 461-466