OBJECT: The goal of this study was to determine the long-term clinical significance of and the risk factors for intramedullary signal intensity change on MR images in patients with cervical compression myelopathy (CCM), an entity most commonly seen with cervical spondylotic myelopathy and ossification of the posterior longitudinal ligament (OPLL). METHODS: One hundred seventy-four patients with CCM but without cervical disc herniation, severe OPLL (in which the cervical canal is < 10 mm due to OPLL), or severe kyphotic deformity (> 15 degrees of cervical kyphosis) who underwent surgery were initially selected. One hundred eight of these patients were followed for > 36 months, and the 71 patients who agreed to MR imaging examinations both pre- and postsurgery were enrolled in the study (the mean follow-up duration was 60.6 months). All patients underwent cervical laminoplasty. The authors used the Japanese Orthopaedic Association (JOA) score and recovery ratio for evaluation of pre- and postoperative outcomes. The multifactorial effects of variables such as age, sex, a history of smoking, diabetes mellitus, duration of symptoms, postoperative expansion of the high signal intensity area of the spinal cord on MR imaging, sagittal arrangement of the cervical spine, presence of ventral spinal cord compression, and presence of an unstable cervical spine were studied. RESULTS: Change in intramedullary signal intensity was observed in 50 of the 71 patients preoperatively. The pre- and postoperative JOA scores and the recovery ratio were significantly lower in the patients with signal intensity change. The mean JOA score of the upper extremities was also significantly lower in these patients. Twenty-one patients showed hypointensity in their T1-weighted images, and a nonsignificant correlation was observed between intensity in the T1-weighted image and the mean JOA score and recovery ratio. The risk factors for signal intensity change were instability of the cervical spine (OR 8.255, p = 0.037) and ventral spinal cord compression (OR 5.502, p < 0.01). Among these patients, 16 had postoperative expansion of the high signal intensity area of the spinal cord. The mean JOA score and the recovery ratio at the final follow-up were significantly lower in these patients. The risk factor for postoperative expansion of the high signal intensity area was instability of the cervical spine (OR 5.509, p = 0.022). No significant correlation was observed between signal intensity on T1-weighted MR images and postoperative expansion of the intramedullary high signal intensity area on T2-weighted MR images. CONCLUSIONS: Long-term clinical outcome was significantly worse in patients with intramedullary signal intensity changes on MR images. The risk factors were instability of the cervical spine and severe ventral spinal compression. The long-term clinical outcome was also significantly worse in patients with postoperative expansion of the high signal intensity area. The fact that cervical instability was a risk factor for the postoperative expansion of the high signal intensity indicates that this high signal intensity area occurred, not only from necrosis secondary to ischemia of the anterior spinal artery, but also from the repeated minor traumas inflicted on the spinal cord from an unstable cervical spine. The long-term neurological outcome found in the preliminary study of patients with CCM who had cervical instability and intramedullary signal intensity changes on MR images suggests that surgical treatment should include posterior fixation along with cervical laminoplasty or anterior spinal fusion.