Lumbar spinal canal dimensions measured intraoperatively after decompression are not properly rendered on early postoperative MRI

Abstract

Background: In cases of lumbar spinal stenosis (LSS) treated with surgical decompression, a postoperative magnetic resonance imaging (MRI) is sometimes required. In the experience of the investigators of this study, the obtained decompression observed on early postoperative MRI tends to be disappointing compared to the decompression achieved intraoperatively. This raises the question of whether the early postoperative MRI, performed after lumbar decompression, is a fair representation of the ‘real’ decompression. This study investigated the correlation between intraoperative and postoperative measurements of the lumbar spinal canal. Method: Twenty patients with LSS underwent surgical decompression on a single level. The orthopaedic surgeon performed direct intraoperative measurements of width, length and height of the spinal canal. Preoperative supine MR images and postoperative prone and supine MR images were acquired. Two radiologists (R.B. and M.G.) measured width, length and height of the spinal canal on the preoperative and postoperative MRIs. Intraoperative measurements were compared to measurements on postoperative MRI in prone position (thus reproducing the intraoperative situation) to avoid positioning bias. Preoperative and postoperative measurements on MR images were also compared. In addition to this, postoperative measurements on supine and prone MR images were also compared. Results: Interobserver reliability for MRI measurements by both radiologists was generally excellent (intraclass correlation coefficients ≥0.71). The postoperative spinal canal dimensions improved on both prone and supine MRI compared to the preoperative imaging (P < 0.05). Intraoperatively measured dimensions demonstrated a significantly greater height (difference 2.8 ± 3.3 [R.B.] and 1.9 ± 3.7 [M.G.]) and greater width (difference 2.1 ± 3.2 [R.B.] and 2.5 ± 2.7 [M.G.]) compared to postoperative MRI in the prone position (P < 0.05). Postoperative dural sac height was greater on the supine MRI compared to the prone MRI (P < 0.05). Conclusions: Surgical decompression of the spinal canal effectively decreases the compression of the dural sac. However, early postoperative MRI after lumbar decompression does not adequately represent the decompression achieved intraoperatively.