Chapter 171 Imaging
It is essential to view the entire cervical spine and the junction with the thoracic spine and skull in cases of trauma. Incomplete visualization of the cervical spine through the C7-T1 junction is a well-recognized pitfall in acute trauma. If the shoulders obscure the C7-T1 junction on the lateral radiograph, additional imaging, which may include a swimmer’s view or a CT, must be performed.
Multiple levels of spine injury are relatively common. Adjacent vertebrae are frequently fractured. In addition, noncontiguous injuries are also common following significant trauma. Detection of one fracture, therefore, should not end the search for spine injury. Especially with the widespread availability of spiral CT, evaluation of the entire cervical spine from the craniocervical junction through the cervicothoracic junction is very feasible.
Imaging complements the clinical examination. “Clearing” the spine is not a result of completing a set of radiographs or other imaging tests; it must come from a combination of clinical examination and appropriate imaging.
Integrity (stability) of the spine depends on both bone and soft tissue. Although uncommon, it is possible for a patient with no fractures, or seemingly insignificant fractures, to have an unstable spine because of severe ligamentous injury. Even with clearly demonstrated fractures, the outcome may be markedly affected by the extent of accompanying soft tissue injury. Alteration in alignment, worsening neurologic function, or persistent pain may be indications for MRI.
Multiple imaging tools are available for the evaluation of traumatic spine injury. In order to utilize these tools appropriately, it is important to understand the strengths and limitations of the available techniques. The approach to imaging varies, depending on clinical circumstances, availability of imaging tests, and results of any previous imaging.
Radiographs (“plain films”) of the spine have traditionally been the foundation of imaging acute spine trauma. However, this role is changing considerably with the wide availability of multidetector helical CT. Plain radiography should be readily available at any facility treating acute spine trauma. Radiographs can be obtained relatively quickly, and they are relatively inexpensive compared with CT and MRI. Spatial detail is excellent with good technique, although interpretation can be challenging with multiple overlapping structures. One of the strengths of plain radiography is its ability to quickly assess alignment.
There is no universal agreement on the views that should be obtained for acute cervical spine trauma. A lateral view is necessary, and often this is among the first radiographs (along with an anteroposterior [AP] chest radiograph), to be obtained in a patient with multiple trauma. As is discussed later, it is essential to view the entire cervical spine, including the junctions with the skull and thoracic spine. A swimmer’s view may be useful, if needed, for visualizing the lower cervical spine. The lateral view is often followed by AP and odontoid views. The odontoid view, in particular, may be difficult with an uncooperative or intubated patient. Additional views may include oblique or pillar views. These can be particularly helpful in assessing the dorsal elements. The standard method of obtaining these views involves turning the patient’s neck, which is not acceptable in the acute setting. Modified techniques can be used that involve angling the x-ray tube instead of turning the neck; these result in more distortion than a standard view but can still be useful.
AP and lateral views are usually obtained when acute thoracic or lumbar spine injury is suspected. Oblique views have little role in the evaluation of acute lumbar trauma. They add a considerable radiation dose with limited benefit.
Alignment is well assessed with plain films. Soft tissue injury can be inferred from prevertebral soft tissue swelling in the cervical spine. In the lumbar and thoracic spine, paraspinous swelling on the AP view is a sign of acute injury. Otherwise, however, radiographs are insensitive to the detection of significant soft tissue. The sensitivity of plain films to fracture varies depending on the location of the fracture. Vertebral body fractures are usually well visualized on radiographs, but fractures in the dorsal elements can be difficult. Sensitivity of radiographs to dorsal element fractures in the cervical spine has been found to be as low as 50%.1–3 Fractures of the larger vertebrae of the thoracic and lumbar spine are usually well visualized on plain films.
CT has assumed a major role in imaging spine trauma because of the technological advances of the multidetector helical technique. It is possible to screen very rapidly for cervical spine trauma with the initial imaging evaluation of a patient with major trauma using a high-resolution technique that permits multiplanar reconstruction. Although not completely eliminated, the limitations of earlier CT technique, such as patient motion, lower resolution in reconstructed images, and relative insensitivity of CT for axially oriented fractures, are greatly diminished. Reformatted CT views in sagittal, coronal, or other planes are essential in such cases, as well as for viewing alignment. As CT increases in speed and technical capability, it has become feasible to incorporate it in a routine manner in the evaluation of major trauma, especially in cervical spine evaluation.4–7
In addition to planar (2D) reconstruction views, 3D projection views can be useful for evaluating the position of fractures. Considerable variation is possible in reconstruction techniques. The images can be created using surface reconstruction projection or varying degrees of apparent transparency of the reconstructed image. Each has advantages in specific situations, and the user of CT machines should become familiar with the options available.
The value of MRI for evaluating acute spine trauma may not be obvious. MRI is relatively insensitive to detection of fractures, since cortical bone provides very little signal and appears black on MRI. MRI has significant limitations in the acutely injured patient, including challenges in patient monitoring, longer imaging times than with radiography or CT, and difficulties in using standard coils in a patient with spine immobilization. For these practical reasons, MRI is often more suitable in the first few days than the first few hours after trauma. Nevertheless, MRI has unique advantages in assessing acute spine trauma.
MRI is highly sensitive to soft tissues, especially for edema.8–10 MRI is the best method for visualizing the spinal cord. Compression or deformity of the spinal cord, edema, and hemorrhage are visualized well with MRI. It is also excellent for evaluating the intervertebral discs (Fig. 171-1). For example, detection of an acute disc disruption may alter plans for the surgical approach.11 Cases have been reported in which the reduction of a dislocation worsened symptoms because of further herniation or displacement of disc material. MRI permits detection of such herniation before surgery. Ligament disruption can occasionally be directly visualized with MRI, especially the anterior and posterior longitudinal ligaments. Some evidence suggests that the extent of ligament disruption may correlate with the risk of instability in cervical spine dorsal element fractures.
FIGURE 171-1 Traumatic disc and ligament disruption. Sagittal fast inversion-recovery MRI shows disruption of both anterior and posterior longitudinal ligaments at C5-6 and disc disruption with traumatic disc herniation and ventral and dorsal soft tissue edema. The findings were confirmed at surgery.
The soft tissues around the vertebra are also visualized with MRI. Extensive edema can serve as a marker for acute injury and the need for further evaluation. Deep, interspinous edema in the setting of acute trauma may be an indication of high risk of instability due to flexion injury. Tears or stretching of the anterior and posterior longitudinal ligaments are especially concerning for instability. MRI can be helpful in a variety of clinical situations in which the combination of clinical and initial imaging findings is ambiguous or nondefinitive. For example, degenerative changes are very common in the cervical spine and can make detection of acute fractures difficult. Although CT can help to identify fractures, subluxation or chronic deformity can still be a challenge. A negative MRI in such situations, showing no evidence of any significant nearby soft tissue edema, makes acute injury unlikely as a cause of subluxation. MRI can be especially helpful when clinical assessment is limited, such as in the obtunded or intubated patient.12
Unexpected worsening of neurologic status after spine trauma can be due to a variety of factors. MRI may reveal such causes as epidural hematoma, disc herniation, or spinal cord edema from infarction.
The limitations of MRI mentioned earlier can be overcome in most cases. Monitoring is essential in the acutely injured patient. MRI-compatible monitoring equipment is available. The patient must be screened for the presence of metallic devices or metal within the body that would preclude MRI. Although standard spine coils may not be usable with a cervical collar, other coils can still permit diagnostic images. Faster imaging techniques continue to be developed for MRI, and it is often not necessary or appropriate to use the same sequences for acute trauma patients that would be used for evaluation of degenerative disc disease, for example. The specific sequences to be used can be tailored to the clinical situation.
Although it is impossible to specify MRI parameters that should be used because of the great variety of manufacturers, machines, and software available, some broad principles apply. A T2-weighted sequence is important for detecting edema. Fast-spin echo imaging, in which multiple echoes are acquired during each pulse sequence, is nearly always used in standard spine imaging. Such sequences are much faster than spin echo sequences and produce excellent signal-to-noise and high-quality images. Fast-spin echo sequences can be excellent for visualizing the spinal cord, for example. However, it is important to note that fat remains very bright on such sequences, even with T2-weighting. Therefore, if adjacent soft tissue edema is to be demonstrated, different sequences must be used that suppress the signal from fat (Fig. 171-2). A fat saturation pulse can be added to fast-spin echo imaging sequences. Alternatively, inversion recovery sequences (short-tau inversion recovery [STIR]) accomplish the same effect of heavy T2 weighting and fat suppression.
FIGURE 171-2 Fat suppression MRI for soft tissue evaluation. MRI was performed on a young man with myelopathic symptoms after a motor vehicle accident. A, Sagittal fast-spin echo T2-weighted image shows slight subluxation at C6-7, but edema from soft tissue injury is difficult to identify because fat also remains bright. B, On a sagittal fast-spin echo inversion-recovery image the fat is now dark, but edema in the dorsal soft tissues, the marrow space of upper thoracic vertebral bodies, and prevertebral space is very conspicuous. Anterior and posterior longitudinal ligaments at C6-7 are stretched (arrows).
Edema within the bone marrow is also well demonstrated with MRI. This appears as low signal (dark) on T1-weighted images, replacing the normally bright signal from fat in the marrow and high signal on fat-suppressed T2-weighted images. Acute fractures, especially those that involve the vertebral body, cause marrow changes, but bone contusions that do not result in cortical bone disruption can also result in marrow edema. Over time, marrow signal intensity usually returns to an appearance close to that of normal vertebral body marrow. Thus, signal intensity in marrow of a compressed vertebral body that matches that of adjacent, normal vertebrae provides evidence for a chronic rather than an acute injury.
The sensitivity of MRI for acute soft tissue injury depends on several factors. Edema resolves over several days; the precise time course has not been defined but clearly depends on the severity of the original injury. In our experience, soft tissue edema associated with an acute cervical spine injury is likely to be less extensive on MRI in the setting of axial load injuries. Presumably, this is due to less stretching and tearing of the soft tissues. For example, a minimally displaced Jefferson burst fracture may result in very little soft tissue edema. When these limitations are understood, however, MRI can play a very useful adjunct role in assessing major acute cervical spine trauma.
Radiographs of the spine in different positions (i.e., flexion and extension views in a lateral position) are excellent for evaluating the stability of the spine in a delayed or chronic setting (Fig. 171-3). Such studies have significant limitations in the acute setting, however. Most importantly, they pose major risks if the spine is in fact unstable. Complications of flexion-extension radiographs are rare but well known. If motion studies are to be undertaken, it is highly desirable that the patient be fully alert, cooperative, and capable of controlling or stopping the motion. If flexion and extension are performed on an obtunded or comatose patient, performing the study under fluoroscopy should improve the safety. The motion can be immediately stopped as soon as subluxation or abnormal movement is visualized. However, data on the safety and accuracy of performing motion radiography on an obtunded patient in the setting of acute injury are very limited. Such studies are often time-consuming, and visualization of the cervicothoracic junction is frequently difficult.
FIGURE 171-3 Instability on flexion film. Flexion lateral radiograph obtained on a delayed basis after cervical spine trauma shows focal kyphosis and dorsal widening. Alignment in neutral position was normal.
A second limitation in the acute setting is a high incidence of muscle spasm or guarding. From one quarter to one third of patients with acute cervical spine injury may have nondiagnostic results because there is inadequate movement of the neck to assess stability.13 Delayed studies, several weeks after trauma and after muscle spasm has subsided, with the patient cooperative and in control of neck motion, remain the gold standard for evaluating the stability of the cervical spine.
The Jefferson burst fracture of C1 is a relatively common injury. CT can demonstrate the multiple fractures of the ring of C1 and the extent of displacement (Fig. 171-4). Plain films usually show prevertebral soft tissue swelling. Some components of the fractures can be visible on plain radiographs, more so if the fractures are displaced. The lateral masses of C1 are likely to be displaced laterally if the transverse ligament is disrupted. Total displacement of greater than 7 mm as seen on an odontoid view has been suggested as a guideline to the presence of transverse ligament rupture.14 The transverse ligament can be visualized directly on MRI, and fluid signal in place of the expected low signal intensity of the ligament is evidence of rupture.15
The imaging findings just described apply to adults. In at least the first 4 years of life, the lateral masses of C1 often project lateral to the lateral margins of C2 on an AP view.16 Because Jefferson fractures are uncommon in children, CT may be necessary to demonstrate such fractures in children.
Isolated fractures of the dorsal arch of C1 can occur with hyperextension injuries (Fig. 171-5). In such cases, prevertebral soft tissue swelling would likely be absent, and the dorsal arch fracture can be seen on a lateral radiograph. Another type of hyperextension injury at C1 is an avulsion at the ventral caudal portion of the ventral arch, at the attachment of the atlantoaxial ligament. In this case, the fracture is visible on the lateral view, and focal prevertebral soft tissue swelling is usually present.
Transverse ligament rupture can be seen in association with a variety of upper cervical spine fractures, and this possibility should be considered in any such injury. In addition, transverse ligament injury uncommonly may occur without other fractures. Loss of integrity of the transverse ligament can result either from rupture in the midportion of the ligament or from avulsion of the ligament at one of the attachments to the lateral mass of C1. In the latter case, a small fracture is often visible at the tubercle where the ligament attaches (see Fig. 171-26). If the ventral atlantodental space is widened, transverse ligament rupture should be suspected. The ligament itself can be seen directly using MRI. In the case of rupture, fluid signal intensity (bright on T2-weighted images) can be seen in the expected location of the ligament, and fluid is also likely to be present between the dens and ventral arch of C1.
Odontoid fractures can occur from a variety of mechanisms. Anderson and D’Alonzo17 described three types of odontoid fractures: type I, an oblique fracture near the apex of the dens; type II, a transverse fracture through the lower third of the dens but above the body of C2 (Fig. 171-6); and type III, which is a fracture below the base of the dens and through the body of C2 (Fig. 171-7). As with transversely oriented fractures elsewhere, odontoid fractures can be difficult to identify by CT. Axial images may show only a region of lucency or subtle gaps in the cortical margin. Reconstructed images from thin-slice axial images, especially from rapid spiral acquisitions, can be very helpful for identifying odontoid fractures on CT. Plain radiographs should be carefully inspected for signs of odontoid fractures, including prevertebral soft tissue swelling, abnormal angulation of the odontoid process, offset of the dens with respect to the body of C2, and disruption of the cortical margin. The type III fracture (or type 3 C2 body fracture) is a horizontally oriented, rostral fracture at the base of the dens. In such fractures, the lateral radiograph shows a break in the apparent ring that results from the superimposition of densities from the junction of the pedicle and body, dens and body, and dorsal cortex of the C2 body.18 The AP or odontoid view usually shows a fracture with inferior convexity.
FIGURE 171-6 Type II dens fracture. Motor vehicle accident victim with neck pain: A, Anteroposterior radiograph shows fracture across the base of the dens. B, Lateral radiograph shows marked prevertebral swelling, fracture, and dorsal displacement of the dens relative to the body of C2. C, Axial CT discloses a fracture line through the dens. D, Sagittal CT reconstruction view shows the fracture and displacement. E, Sagittal short-tau inversion-recovery MRI shows the fracture, extensive prevertebral and lesser amount of dorsal edema, and relationship of the dens to the spinal cord.