Identifying spinal injuries in trauma patients with altered mental status can be difficult. CT scanning and clinical examination are the basis of our spinal clearance, but screening "trauma protocol" spinal MRI is used to exclude occult injuries. We sought to evaluate the sensitivity of CT scanning for spinal injuries compared with our MRI protocol. Ninety-seven patients underwent MRI cervical spine trauma protocol during 2004. Twenty-nine patients were obtunded, 29 had neurologic symptoms, and 39 had spine pain. MRI confirmed the initial CT findings without new injuries in 83 cases. MRI reclassified fractures as degenerative changes in 12 cases. In 2 cases, the MRI identified new injuries: one a stable partial ligament tear, the second a T7 Chance fracture with ligamental disruption requiring operative fixation. There was no morbidity or mortality documented in obtaining the MRI studies. Overall negative predictive value of CT scanning of the spine was 98 per cent, the positive predictive value was 78 per cent, and the sensitivity and specificity was 94 per cent and 91 per cent, respectively. CT scanning of the cervical and axial spine is sensitive for spinal trauma but not specific. MRI trauma protocol should be reserved for cases when initial CT scanning is suggestive of traumatic injury.