Using Reformatted Axial Computed Tomography Images in Isolation Will Miss Narrow S1 Transsacral Screw Corridors

Abstract

Objective:To evaluate S1 transsacral (TS) corridors on reformatted and nonreformatted computed tomography (CT) images to determine which most reliably identifies narrow corridors.Design:Retrospective cohort.Setting:Level 1 trauma center.Patients:Two hundred forty-five patients with operative pelvic ring injuries.Intervention:CT scan review.Main Outcome Measurements:Preoperative CT scans were evaluated to determine the width of the S1 TS corridor on standard axial nonreformatted (ANR), axial reformatted (AR), and coronal reformatted images. Sensitivity and specificity of each format to detect a narrow corridor (<10 mm on AR or coronal reformatted) were calculated. Patients with S1 TS screws were evaluated to determine the rate of screw breach with narrow corridors.Results:The axial width of the S1 TS corridor was consistently smaller on ANR versus AR images (mean difference 1.4mm, 95% confidence interval 1.1–1.5). The corridor width on ANR images was on average 86% of the AR measurement. ANR images had the highest sensitivity and specificity (100% and 98%) for detecting S1 TS corridors <10 mm. Fifty-three S1 TS screws were placed in corridors ranging 10–23 mm on AR images and 7–19 mm on ANR images. Four (57%) of the 7 screws placed in corridors less than 10 mm in width on ANR images breached the sacral cortex.Conclusion:Using ANR images to measure the S1 TS corridor consistently measured smaller widths than AR images and identified all narrow corridors. A high rate of screw breach was noted with screw placement in narrow corridors.Level of Evidence:Diagnostic Level III. See Instructions for Authors for a complete description of levels of evidence.