Study design: Meta-analysis.
Objective: Meta-analysis of the data to determine the effect of magnetic resonance imaging (MRI) signal change on preoperative and postoperative Japanese Orthopedic Association (JOA) scores and on recovery rate after surgery.
Summary of background data: MRI signal changes are commonly found in myelopathy. There is often an increased T2 signal with or without a decreased T1 signal. The clinical significance of these signal changes remains debated.
Methods: A comprehensive review of the literature was performed to identify all published studies with data on the presence of MRI signal change and JOA scores in myelopathic patients. T tests were performed to determine if there were significant differences between preoperative and postoperative JOA scores in patients with or without MRI signal change. The recovery rate was calculated for all patients undergoing surgery. T tests were performed to determine whether significant differences occurred in recovery rate in patients with or without MRI signal change.
Results: A total of 16 studies were used for the meta-analysis. The total population included 886 patients: 659 with MRI signal change, and 227 without MRI signal change. Preoperative and postoperative JOA scores and the recovery rates were significantly better in patients without MRI signal changes (P < 0.05). The mean preoperative JOA scores were 10.63 and 11.37 for patients with and without MRI signal changes, respectively. The mean postoperative JOA scores were 13.37 and 14.19 for patients with and without MRI signal changes, respectively. The mean recovery rates were 43.87% and 49.31% for patients with and without MRI signal changes, respectively.
Conclusion: A meta-analysis of the literature revealed statistically better preoperative and postoperative JOA scores and recovery rates following surgery in myelopathic patients without MRI signal change. Although the data were statistically significant, the clinical significance of the differences might be less due to the relatively small differences in actual values.