This study was designed to clarify the sonographic anatomy of the normal anal canal by comparison with endoanal magnetic resonance imaging, to determine agreement between these imaging modalities and interobserver error in measuring layer thickness. Three-dimensional endosonographic and endocoil magnetic resonance images of the anal canal were obtained in four males and five nulliparous females aged 22 to 34 years. Images were analyzed at similar levels throughout the canal using a graphics-overlay technique to compare sonographic with magnetic resonance images. Measurements were taken at one level for agreement analysis between modalities and for interobserver variability in the measurement of the thickness of the main anal canal layers. The muscularis submucosae ani, muscle bundles in the longitudinal muscle layer, and puboanalis were identified on sonography. The outer border of the external sphincter was demarcated by an interface reflection with ischioanal fat. Clarification of the external sphincter anatomy allowed excellent correlation (Ri = 0.96) for the assessment of thickness. There was excellent correlation for the interobserver measurement of the external and internal sphincters and the submucosal width on endosonography, but there was poor correlation for the longitudinal muscle (0.12). The overlay technique has improved endosonographic interpretation, and measurement of external sphincter thickness has been validated both by comparison with magnetic resonance and on interobserver agreement.
Read full abstract