Comparison Between MRI and Myelography in Lumbar Spinal Canal Stenosis for the Decision of Levels of Decompression Surgery

Study Design: A retrospective case-control study. Objective: To investigate differences between number of levels of decompression surgery for lumbar spinal canal stenosis (LSCS) as decided by magnetic resonance imaging (MRI) and myelography with computed tomographic myelography (CTM), and to investigate intraobserver reliability and interobserver reproducibility of MRI and myelography with CTM. Summary of Background Data: Although the decision to carry out decompression surgery for LSCS largely depends on the severity of stenosis on imaging along with clinical findings, no universally accepted imaging criteria have been defined for determining severity of LSCS. The decision to carry out decompression surgery for LSCS thus seems relatively subjective, and would differ according to both observer and the findings of preoperative imaging. Methods: We retrospectively selected 50 patients with LSCS who underwent decompression surgery. Mean patient age was 69 years, and mean number of levels carried out decompression surgery was 2. Each of 4 spine surgeons retrospectively decided levels of decompression surgery 4 times, according to the findings of MRI or myelography with CTM (MRI-1, MRI-2, CTM-1, CTM-2). We investigated differences between number of levels of decompression surgery decided by MRI and myelography with CTM, and also investigated intraobserver reliability and interobserver reproducibility of MRI and myelography with CTM. Results: The number of levels of decompression surgery as decided by myelography with CTM was significantly greater than that by MRI. kappa coefficients for intraobserver reliability and interobserver reproducibility revealed myelography with CTM as more reliable and reproducible than MRI. Conclusions: The number of levels of decompression surgery for LSCS as decided by MRI would be less than that by myelography with CTM. Myelography with CTM is more reliable and reproducible than MRI for preoperative evaluation of LSCS.