MRI of the lumbar spine: comparison of 3D isotropic turbo spin-echo SPACE sequence versus conventional 2D sequences at 3.0 T

Background: Three-dimensional (3D) fast spin-echo sequence with variable flip-angle refocusing pulse allows retrospective alignments of magnetic resonance imaging (MRI) in any desired plane. Purpose: To compare isotropic 3D T2-weighted (T2W) turbo spin-echo sequence (TSE-SPACE) with standard two-dimensional (2D) T2W TSE imaging for evaluating lumbar spine pathology at 3.0 T MRI. Material and Methods: Forty-two patients who had spine surgery for disk herniation and had 3.0 T spine MRI were included in this study. In addition to standard 2D T2W TSE imaging, sagittal 3D T2W TSE-SPACE was obtained to produce multiplanar (MPR) images. Each set of MR images from 3D T2W TSE and 2D TSE-SPACE were independently scored for the degree of lumbar neural foraminal stenosis, central spinal stenosis, and nerve compression by two reviewers. These scores were compared with operative findings and the sensitivities were evaluated by McNemar test. Inter-observer agreements and the correlation with symptoms laterality were assessed with kappa statistics. Results: The 3D T2W TSE and 2D TSE-SPACE had similar sensitivity in detecting foraminal stenosis (78.9% versus 78.9% in 32 foramen levels), spinal stenosis (100% versus 100% in 42 spinal levels), and nerve compression (92.9% versus 81.8% in 59 spinal nerves). The inter-observer agreements (k = 0.849 vs. 0.451 for foraminal stenosis, k = 0.809 vs. 0.503 for spinal stenosis, and k = 0.681 vs. 0.429 for nerve compression) and symptoms correlation (k = 0.449 vs. k = 0.242) were better in 3D TSE-SPACE compared to 2D TSE. 3D TSE-SPACE with oblique coronal MPR images demonstrated better inter-observer agreements compared to 3D TSE-SPACE without oblique coronal MPR images (kappa = 0.930 vs. kappa = 0.681). Conclusion: Isotropic 3D T2W TSE-SPACE at 3.0 Twas comparable to 2D T2W TSE for detecting foraminal stenosis, central spinal stenosis, and nerve compression with better inter-observer agreements and symptom correlation.