A comparison of two methods to evaluate a narrow spinal canal: routine magnetic resonance imaging versus three-dimensional reconstruction.

Abstract

In routine clinical practice, the presence of lumbar spinal stenosis (LSS) is assessed on axial magnetic resonance images (MRI) typically acquired using a preselected spine sagittal angle. Given the natural lordosis of the lumbar spine, not all axial slices will be parallel to the disc and perpendicular to the spinal canal and, thus, are not optimal for the assessment of dural sac cross-sectional area (DCSA). The objective of this study was to compare DCSA measurements from routinely acquired clinical images with three-dimensional (3D)-reconstructed images. This is a cross-sectional study. The sample consists of 390 patients referred for lumbar imaging with some aspect of anatomical LSS found, with no prior back surgery, 40 years of age or older, and with available volumetric MR images to allow 3D reconstruction of the spine. The outcome of interest in this study was dural sac cross sectional area. Spine images were 3D reconstructed at the level of the disc, perpendicular to the spinal canal. Dural sac cross-sectional area was measured for both 3D-reconstructed and routinely acquired clinical images using the slice orientation captured. Dural sac cross-sectional area for the lower lumbar levels (L4-L5 and L5-S1) was significantly different between routinely acquired clinical images and 3D-reconstructed images, with a standard error of measurement of 12.98 and 19.73 mm(2), respectively. When canal size is of interest, particularly when LSS affecting the lower lumbar levels is of concern, 3D reconstruction of clinical images should be considered.