Purpose: To determine whether a commercially available automated MR myelogram sequence improves the interpretation and diagnostic yield at MRI of the lumbar spine.
Material and methods: A total of 207 consecutive patients referred for MR examination of the lumbar spine for evaluation of low back pain or spinal radicular symptoms were included for study. All patients had initial imaging with sagittal T1-W and T2-W scans, followed by axial T2-W images. Subsequently an MR myelogram was acquired in each case in coronal, sagittal and oblique planes. MR myelogram images were evaluated initially and a diagnosis was established and recorded. Subsequently, a diagnosis was established by review of conventional images of the lumbar spine in sagittal and axial planes, in conjunction with the MR myelogram. The utility of the MR myelogram in establishing the diagnosis was graded on a 4-point scale, where grade 0 indicated that it contributed no additional information and grade 3 indicated that it was essential for diagnosis. Analysis of the additional value of myelography in patients with multilevel disease was made.
Results: Primary diagnoses were disc herniation in 69 cases (33%), degenerative disc disease in 51 cases (26%), spinal stenosis in 19 cases (9%), isolated lateral recess stenosis in 5 cases (2%), or other diagnoses, including facet degeneration in 48 cases (23%). Scans were normal in 15 cases (7%). MR myelography was graded as grade 0 in 151 cases (73%), grade 1 in 44 cases (21%) and grade 2 in 12 cases (6%). In no case was MR myelography essential for diagnosis (grade 3). In patients with multilevel disease (n=27), in 17 cases MR myelography was graded as grade 1 (63%) and in 8 cases grade 2 (30%).
Conclusion: MR myelography when employed in routine practice was of limited value, assisting in establishing a diagnosis in a minority of cases (6%). The technique was of limited additional value in patients with multilevel pathology and to a lesser extent in patients with scoliosis, where it helped to establish the level most likely to account for pathology.