Many ligaments are seen normally at the craniocervical junction (Fig. 1). However, only three are considered the major stabilizers. These are the tectorial membrane (Fig. 2), the transverse ligament, and the alar ligaments (Fig. 3). The normal tectorial membrane and transverse ligament are routinely seen on MR imaging, whereas the normal alar ligaments can be more difficult to visualize because of lack of contrast from adjacent tissues (Fig. 3). In most individuals, each alar ligament arises from the lateral margin of the dens, then courses laterally in a near-vertical plane, attaching to both the ipsilateral occipital condyle and the subjacent superior margin of the lateral mass of the atlas (C1). However, in about a third of individuals, these ligaments insert solely onto the occiput. The alar ligaments limit axial rotation at the occipitoatlantoaxial complex. Blood or edema adjacent to an acute alar ligament tear (Figs. 4 and 5A,5B) improves visualization of these ligaments. Secondary evidence of ligamentous injury to one of the alar ligaments is displacement of the dens to the contralateral side. Isolated posttraumatic alar ligament tears have been classified. These are clinically significant because hypermobility at the atlantoaxial joint can reduce blood flow in the contralateral vertebral artery. Hulse [9] describes “cervical nystagmus as a manifestation of vertebral artery insufficiency due to rotatory hypermobility at the occipitoatlanto-axial complex.” Figure 6A,6B shows displaced ligament injuries at the craniocervical junction associated with a type II dens fracture.

Figures 7 and 8 show ligamentous injury associated with a burst fracture of the cervical vertebrae. Figures 9A,9B,9C,9D,10,11 picture injuries caused by cervical hyperextension. Injuries associated with interfacetal dislocations and teardrop fractures are also shown in Figures 12,13,14.

The concept of three columns of support in the thoracic and lumbar spine is well accepted. The same principles have been applied to the C3-C7 vertebral levels in the cervical spine. Stability is provided by intact osseous and ligamentous structures. The anterior column consists of the anterior vertebral body, the anterior longitudinal ligament, and the anterior annulus fibrosus. The middle column comprises the posterior vertebral body, the posterior longitudinal ligament, and the posterior annulus fibrosus. Hyperextension can result in injury to the anterior column (Fig. 10) or to both the anterior and middle columns (Figs. 11 and 15). The posterior column consists of the posterior elements of the spine, ligamentum flavum, interspinous ligaments, supraspinous ligaments, and facet joint capsules. Hyperflexion may result in injury to the middle and posterior columns (Figs. 9A,9B,9C,9D and 16A,16B). Injury to any two adjacent columns will result in instability. Disruption of all three osseous or ligamentous supporting columns is shown in association with burst fractures in Figures 7 and 8, bilateral interfacetal dislocation is shown in Figures 12 and 13, and teardrop fractures of C7 are shown in Figure 14.

Successful MR imaging of spinal trauma depends on several factors. One of these is the timing of the study. Although no research has yet, to our knowledge, defined the optimal time interval between injury and MR imaging, it should probably be less than 72 hr [8]. Beyond this time, resorption of the edema or hemorrhage reduces sensitivity of MR imaging to reveal injuries. Specifically, the T2 signal hyperintensity produced by edema or extravasation of blood into injured extradural tissues provides an excellent contrast medium, improving the conspicuity of the ligaments that are usually of low signal intensity on all imaging sequences.

The use of appropriate sequence parameters for MR imaging is also important. These parameters vary widely according to the field strength, coil design, gradient strength, and software capabilities of the MR imaging system used. Thus, each system requires an individualized approach, fine-tuned by trial and error. In general, field of view, slice thickness, matrix, and signal averages must be chosen to balance the effects on signal-to-noise ratio, spatial resolution, and imaging times. For example, longer imaging times may improve scan quality but provide more opportunity for patient motion. A typical MR imaging protocol for spinal trauma should include the following sequences in the sagittal plane: T1-weighted, fast spin-echo T2-weighted, gradient-echo, and fast spin-echo inversion-recovery images. In the axial plane, protocol should include gradient-echo or T2-weighted images. Optional coronal T1-weighted or gradient-echo sequences can aid in evaluation of the cranioatlantoaxial segment, especially with regard to alignment and dens fracture. Figure 9A,9B,9C,9D compares the relative merits of the four sagittal sequences described previously. T1-weighted images provide the best anatomic detail, accurately depict alignment, and are invaluable for detection of fracture. Ligaments are usually best seen on gradient-echo and T2-weighted sequences. At high field strength, the heterogeneity effect produced by gradient-echo imaging techniques results in greater sensitivity for detection of blood products within the spinal cord but also reduces signal intensity within bone and makes fracture detection more difficult. At low field strength, the gradient-echo heterogeneity effect is weaker so that blood products are less easily detected but fracture detection is somewhat improved. T2-weighted and fast spinecho inversion-recovery sequences are most sensitive for bone marrow edema (caused by fracture or trabecular contusion), spinal cord injury, and soft-tissue edema. Cerebrospinal fluid pulsatility artifacts and truncation artifacts can sometimes interfere with spinal cord evaluation on T2-weighted and fast spin-echo inversion-recovery sequences.

MR imaging of the posttraumatic spine is a rapidly evolving technique with the potential to revolutionize the evaluation and treatment of ligamentous injuries. In our clinical experience, it has been an invaluable adjunctive technique, particularly in patients with relevant neurologic deficits and those requiring closed reduction of a posttraumatic spinal subluxation. It has also been helpful in evaluating spinal trauma complicated by altered sensorium, extreme obesity, or even malingering.

Presented at the annual meeting of the American Roentgen Ray Society, New Orleans, May 1999.

Address correspondence to L. A. Hayman.