Comparing two-dimensional ultrasonography with three-dimensional ultrasonography and MRI for the levator ani defects grading.

Levator ani defect (LAD) closely correlates with pelvic floor disorders (PFD). In general, LAD was graded by three-dimensional ultrasonography (3D-US) and MRI, which could be used hardly in some developing area. Our objective was to determine whether two-dimensional ultrasonography (2D-US), a method that is almost universally accessible, could be used to diagnose the LAD. 129 Chinese women with PFD were recruited for the LAD grading by 2D-US and 3D-US and MRI. LAD was classified into intact, partial and complete avulsions. The puborectalis attachment width (PAW) was measured by 2D-US and with the software on the three-dimensional MRI-based LAD models. The results were compared and analyzed using the weighted kappa and the Pearson's coefficient. Of the 119 patients, 64 were diagnosed with LAD by 2D-US, 70 were identified by 3D-US while 68 were confirmed by MRI. The LAD grading of 2D-US showed good agreement with MRI (kappa = 0.78, 95% CI 0.71-0.86) and 3D-US (kappa = 0.77, 95% CI 0.70-0.84). In regard to the consensus of partial or complete avulsions, 2D-US showed excellent good agreement with MRI (kappa = 0.86, 95% CI 0.73-0.97), superior than 3D-US with MRI (kappa = 0.55, 95% CI 0.36-0.71). Additionally, iliococcygeus avulsions detected by MRI (n = 7) were accompanied by complete puborectalis avulsions. The averaged PAW was 10.42 ± 5.57 mm measured by 2D-US, which correlated well with the results measured by MRI (Pearson's coefficient = 0.90). 2D-US, 3D-US and MRI showed the good agreement on LAD diagnosis. Compared with MRI and 3D-US, 2D-US was comparable in grading LAD, especially complete avulsions.