Stenvers与Pöschl平面CT评估上半规管裂的诊断效能

BACKGROUND AND PURPOSE: CT in the evaluation of superior semicircular canal dehiscence (SSCD) is important for diagnosis and preoperative planning. Various imaging planes of the temporal bone have been used though intraoperative correlation has been limited. This study aims to compare interreader reliability and diagnostic accuracy of the Stenvers and Pöschl views against the surgical criterion standard.

MATERIALS AND METHODS: Patients who underwent preoperative temporal bone CT and surgery for suspected SSCD between May 2007 and July 2021 at a single head and neck tertiary care institution were included. CT data were reformatted into Pöschl and Stenvers views and analyzed by 2 independent reviewers for SSCD. For Stenvers, the number of slices showing dehiscence was used for determination of the defect size. For Pöschl, defect size was directly measured on images showing SSCD by 2 additional reviewers. Imaging defect size was compared with intraoperative surgical defect size. Interrater reliability for defect presence was evaluated by the Cohen κ coefficient; interrater reliability for defect size was measured by the Pearson correlation. Diagnostic accuracy was evaluated against the surgical criterion standard by 2-way ANOVA.

RESULTS: Sixty-one ears with intraoperative evaluations were included. Interreader agreement regarding defect presence was excellent on either Pöschl or Stenvers views (κ = 0.78 [95% CI, 0.32-1.00] and 0.79 [95% CI, 0.38-1.00], respectively). Interreader agreement regarding the measured defect size was moderate on both Pöschl and Stenvers views (r = 0.86 [95% CI, 0.75-0.92] and 0.64 [95% CI, 0.47-0.77], respectively). Stenvers outperformed Pöschl in diagnostic accuracy (97.5% versus 93.4%, P = .03) when compared with the surgical criterion standard. Intraoperative defect size moderately correlated with both Pöschl and Stenvers measured defect sizes (r = 0.60 [95% CI, 0.34-0.77, P < .0001] versus r = 0.50 [95% CI, 0.21-0.71, P = .002], respectively).

CONCLUSIONS: Pöschl and Stenvers views are useful reformats for SSCD assessment, offering excellent interreader reliability. Stenvers slightly outperforms Pöschl in diagnostic accuracy, particularly in the setting of pinpoint or off-meridian dehiscence. The measured defect size on both Pöschl and Stenvers views moderately correlated with the intraoperative defect size.