Comparison of Intraoperative Stitched Fluoroscopic Images to Postoperative Standing Full Spine Radiographs

Abstract

Introduction Spinal deformity surgery requires assessment of correction achieved intraoperatively. Long cassette X-rays have been used in the past. Recent advances include availability of 3D scanners (e.g., O-arm). These are also able to take 2D images which can be stitched together to simulate full length X-rays. The process of “stitching” can be performed quickly and may potentially save time compared with conventional X-ray film developing. This is the first study to look at correlation between stitched O-arm fluoroscopic images and postoperative standing X-rays. Methods Patients who underwent fusion of ≥ 3 levels for coronal/sagittal deformity and who had intraoperative stitched O-arm images were identified. The intraoperative images and postoperative full-spine X-rays were reviewed. Coronal and sagittal curves and balance were measured by and averaged between two independent observers. Correlation between intra- and postoperative parameters was tested by computation of Pearson coefficient. Results Between March 2011 and June 2013, 80 patients met the inclusion criteria. Mean age was 27 years (range, 6–85 years). Correlation between intraoperative stitched images and postoperative full-length films was very strong for focal curve measurements at the operative levels (major curve Cobb r = 0.90, lordosis/kyphosis across instrumented levels r = 0.74, T5–T12 kyphosis r = 0.88). However, global balance parameters and regional lordosis were found to either have poor correlation (shoulder–pelvis balance 0.41, L1–S1 lordosis 0.62) or were not measurable on the stitched images (T1–pelvic angle). Conclusion Intraoperative AP and lateral stitched O-arm images correlated strongly with postoperative full-spine standing X-rays for focal angular measurements, such as major curve coronal Cobb angles, instrumented level lordosis, and mid-thoracic (T5–T12) kyphosis. However, global balance parameters, such as shoulder–pelvis balance and T1–pelvic angle, cannot be reliably measured on stitched images. This may be either because of differences between prone and upright positioning, image quality, technique of stitching, or the inability to include the entire spine and pelvis in the images.