Although computed tomography (CT) and conventional radiography can be used to show traumatic alterations affecting the spinal column and the perispinal soft tissues, the contents of the spinal canal are not delineated accurately using these techniques. ¹ Magnetic resonance (MR), on the other hand, excellently images the spinal cord and subarachnoid space, and often the individual spinal nerve roots. In many respects, MR is also able to evaluate accurately and sensitively the elements of the spinal column and perispinal soft tissues. Coupled with an IV contrast agent (e.g., gadolinium [Gd]), MR can assess the integrity of the blood-central nervous system (CNS) barrier (i.e., the combined blood-cord, blood-nerve, and relative blood-meningeal barriers). This two-part article series will describe and illustrate the MR imaging findings in symptomatic patients presenting with clinical histories of acute or remote spinal trauma. ^2-6

Practical considerations

The spinal structures often involved in trauma include: the spinal column, related ligaments, and attached muscles; dura, arachnoid and pia; spinal cord, nerve roots/spinal nerves; and regional blood vessels. It is important to note that what is observed on imaging will depend not only upon the structures involved, but also upon the severity of the injury and the timing of the imaging study with relation to the traumatic incident. As is true of all trauma to tissues, there will be a temporal evolution of the pathophysiology, and this will be reflected, in part, in the imaging findings.

In general, the goals of imaging in the patient with trauma include: detection of the specific structures/tissues involved, determination of the extent of the injury, assessment of general patient prognosis, and enhancement of therapeutic planning. These goals are attempts to improve clinical patient outcome. The results of these and other outcomes (e.g., economic) will define how medical care, including imaging studies, is dispensed in the future for patients with trauma; in turn, these decisions will impact everyone involved in providing healthcare services to these patients.

Technical considerations

In most radiology emergency departments, a patient with spinal trauma is evaluated initially by conventional radiography. If spinal column fracture or derangement is known or suspected, conventional radiography followed by CT utilizing thin, contiguous sections and multi-planar reconstructions may be indicated. If spinal instability is suspected, radiographic dynamic flexion-extension studies may be warranted but should be performed under physician (e.g., radiologist and neurosurgeon/ orthopedic surgeon) supervision. This is to ensure that the patient is monitored carefully for evidence of sudden subjective or objective deterioration during the kinetic radiographic study. If the patient on presentation has a neurologic deficit (i.e., more than simply subjective spinal pain and objective local tenderness), however, evaluation of the spinal meninges, spinal cord, and nerve roots is in order. While water soluble contrast myelography, often coupled with post-myelographic CT, may yield pertinent information, nevertheless it is invasive and, with few exceptions, cannot analyze critically the intrinsic condition of the spinal cord or nerve roots.

MR, on the other hand, noninvasively examines the spinal column, cord, nerve roots, and perispinal soft tissues in multiple planes. In addition, coupled with an IV paramagnetic contrast agent (i.e., Gd), MR is capable of challenging the integrity of the blood-CNS barrier. At present, MR can analyze the spine and its contents with greater sensitivity, greater spatial and contrast resolution, and greater spatial coverage than any other diagnostic imaging technique.

Conventional spin echo, fast spin echo, gradient recalled echo, fat suppression techniques, magnetization transfer, and magnetic resonance angiography can be used to yield clinically relevant information in the patient with trauma to the spine. In order to maximize pertinent information attained from the imaging studies, several factors must be decided without delay: selection of the type of study to be performed, which se-quence to select, which planes to choose, if a contrast agent should be used, and whether or not kinetic studies should be employed. These considerations must be decided properly and quickly since they can affect the clinical outcome in the patient with a neurologic deficit emanating from spinal trauma.

In general, T1-weighted images (T1WI) utilizing conventional spin echo acquisitions in the sagittal planes should be obtained initially. On this imaging sequence, spinal column alignment can be judged and products of hemorrhage can be analyzed. Coupled with conventional spin echo T1WI and T2-weighted imaging (T2WI) acquisitions in the sagittal and axial plane the nature of hemorrhage associated with some injuries to the spine can be evaluated. On the other hand, while fast spin echo T2WI depicts anatomy and cord edema very well, it does not reveal paramagnetic forms of acute hemorrhage such as deoxyhemoglobin, and thus hemorrhage may be overlooked. In fact, the best imaging sequences for depicting acute and chronic blood products are T2*-weighted image (T2*WI) acquisitions utilizing gradient recalled echo sequences. These sequences are sensitive to paramagnetic blood products (e.g., deoxyhemoglobin) and blood products that have relatively greater magnetic susceptibility (e.g., hemosiderin).

Therefore, if fast spin echo T2WI is used to evaluate a patient with spinal trauma, it is wise to supplement the imaging protocol with T2*WI utilizing a gradient recalled echo acquisition for comparison. Finally, fast spin echo T2WI is especially useful in combination with fat suppression techniques for the evaluation of posttraumatic tissue edema and ligamentous injury/disruption. Given these considerations, a full protocol for the evaluation of spinal trauma might include: sagittal and axial T1WI, sagittal fast spin echo T2WI with fat suppression, and sagittal and axial T2*WI.

Nonpenetrating/blunt spinal trauma

Much of the spinal trauma evaluated by MR imaging falls into this category. Injuries from motor vehicle accidents and falls are among the most common types of spinal trauma encountered in practice. The types of pathology that may result from nonpenetrating injuries include acute traumatic disc herniation (figure 1), vertebral fracture/dislocation, ligamentous injury, and neural and vascular trauma.

Although cortical bone is not well evaluated on MR imaging, some vertebral fractures can be reasonably assessed by virtue of two phenomena. First, fracture or dislocation can be suspected because the expected normal spatial relationships of the bony structure(s) are altered. This might be seen in a fracture in which the fracture fragments are distracted (figure 2) or in cases of vertebral column dislocation. Second, some relatively nondistracted acute or chronic fractures can be seen on T1WI because of the visibility of the actual hypointense fracture line as it traverses the relatively hyperintense MR signal of the marrow fat. The hypointense fracture on T1WI is often obscured by the surrounding hypointense posttraumatic edema within the marrow fat; however, rendering it indiscernible (figure 3). This is why MR can miss even major relatively nondisplaced vertebral fractures. Therefore, in general, MR should not be used as a primary method of detecting vertebral fractures; this pathology is much better delineated by conventional radiography and CT. ^7-10

The role of MR in the patient with blunt trauma, then, is to find abnormalities that underlie bony spinal column pathology in the patient with neurologic signs and symptoms. One example of posttraumatic sequelae is hematoma formation within the epidural and subdural spinal compartments. These hemorrhages are be-lieved to result from a rupture or tear of epidural or subdural veins and/or arteries. Spinal epidural hematomas tend to be lobulated and somewhat restricted in their longitudinal extent (figure 4B). Subdural spinal hema-tomas, on the other hand, are generally more sheetlike in configuration, tend to extend extensively in a longitudinal direction within the spine, and may have a crescent shape on axial imaging (figure 4A). Practically speaking, however, the two types of spinal hematomas may be indistinguishable from one another, may coexist, and may have similar or identical clinical implications. The signal intensity of the spinal hematoma will vary on T1WI and T2WI, depending upon the age of the hemorrhage and the blood products present. This variance seems to be in accord with that of the observations described in cranial hemorrhage. As noted above, acute spinal hemorrhage (i.e., deoxyhemoglobin) is best evaluated with gradient recalled echo T2*WI, since the acute hematoma will be relatively hypointense relative to other intraspinal tissues. ^11-13

Subacutely, the patient with blunt lumbosacral spinal trauma may reveal a diffuse breakdown in the blood-nerve barrier within nerve roots of the cauda equina. This is seen on MR imaging as extensive traumatic nerve root enhancement following IV Gd administration. This may represent a type of generalized traumatic spinal "shock" because it can be seen in the absence of severe mechanical compression of the cauda equina by retropulsed bony fragments. Of course, frank traumatic compression of the cauda equina may also result in this enhancement pattern.

With or without radiographic evidence of fractures and dislocations of the spinal column, the spinal cord may become contused (figure 5). It is important to note that spinal cord contusion can be hemorrhagic or nonhemorrhagic. This distinction is significant because hemorrhagic contusions carry a worse prognosis. ^14-40 This is part of the reason that these patients may receive aggressive surgical therapy. Although acute parenchymal hemorrhage (e.g., deoxyhemoglobin) can often be seen clearly on conventional spin echo T2WI (figure 6), it can even be better demonstrated on gradient recalled echo T2*WI (figure 7). As stated earlier, however, fast spin echo T2WI can miss such acute hemorrhagic change. To reiterate, as a general rule in the trauma patient, fast spin echo T2WI must be supplemented with a second T2WI set, preferably a gradient recalled echo T2*WI. ^14-40

In the chronic stages of severe spinal cord trauma, the cord itself may cavitate. Known as posttraumatic syringomyelia, this alteration can be seen within months of the traumatic episode. In some instances, the syringomyelic cavity may be static and remain localized to the area of involvement of the actual spinal cord trauma (figure 8). This is mirrored in the stabilization of the clinical syndrome shortly after the injury. However, in other cases, there is an advancement in the signs and symptoms related to the spinal cord after an initial period of clinical stabilization. Clinically, this is seen as a progressively ascending sensory level of involvement and a deterioration of remaining spinal cord function. If followed on MR, these cases show a progressively enlarging syringomyelic cavity. While the cyst may enlarge in both the caudal and cranial directions within the spinal cord, for unknown reasons there is a strong tendency for the enlargement to extend superiorly. This results in an expansion of the syrinx into relatively more normal areas of the spinal cord craniad to the injury and accounts for the temporally progressive signs and symptoms in these cases. On MR, posttraumatic syringomyelia is seen as an intramedullary cavity that is typically hypointense on T1WI and hyperintense on T2WI. Some variation in intensity may be seen if chronic blood products (e.g., hemosiderin) are present in the area of the initial cord trauma. If progressive, some posttraumatic syringomyelic cavities may have an area of parenchymal hyperintensity within the spinal cord adjacent to the leading edge of the cyst (figure 9). This spinal cord hyperintensity on T2WI is usually seen in the cord suprajacent to the craniad extent of the syringomyelic cavity and probably represents escape of cyst fluid into the spinal cord in the direction of syringomyelic cavity growth. In cases where the cyst cavity is followed on MR after surgical stenting/shunting with successful collapse of the cavity, the leading edge of edema may disappear with resolution of the progressive signs and symptoms or even a return of some recently lost spinal cord function. Therefore, hyperintensity within the spinal cord adjacent to a posttraumatic spinal cord cavity on T2WI may be a sign on MR imaging of advancing posttraumatic syringomyelia. ^33,41-48

Penetrating spinal trauma

Penetrating spinal trauma can be caused by a variety of means including metallic projectiles, knives, and other sharp objects (e.g., glass, wood, etc.) introduced during a traumatic incident. Associated hemorrhage, contusion, and disruption of bony and ligamentous structures can usually be well delineated on MR (figure 10). It should be repeated that when looking for hemorrhage in cases of acute trauma, conventional spin echo T2WI, or better yet gradient recalled echo T2*WI, should be performed so that acute hemorrhagic products (e.g., deoxyhemoglobin) will not be missed. When the penetrating object remains in place during imaging, the foreign body may be seen on MR by its relative lack of mobile protons and/or its magnetic susceptibility properties (figure 11). If the object is not ferromagnetic (e.g., lead, wood, glass), it will appear on MR as an area of signal void; if the object is ferromagnetic (e.g., steel), it will exhibit varying degrees of metallic susceptability artifact. Often the severity of the metallic artifact produced by a given foreign body cannot be predicted until the time of the actual MR examination. The degree of metallic artifact will be relatively magnified on gradient recalled echo acquisitions, and somewhat lessened on fast spin echo sequences. ^49-54

Conclusion

Direct nonpenetrating and penetrating trauma are two of the most common forms of trauma encountered in clinical practice. Much of the injury to the spinal column and neural tissue, involved are excellently evaluated in the acute and chronic phases by MR imaging. The one major weak area of MR diagnosis in trauma is in demonstrating nondisplaced, non-deforming fractures of the bony spine accurately; these diagnoses are still well made by conventional radiography and high-resolution CT with multiplanar reconstructions. One other relative problem with MR imaging of the acutely traumatized patient is in the case of the individual on life support devices and/or in spinal traction. It may not be feasible to image these cases by MR, especially when considering high field MR units. Otherwise, deforming fracture/dislocations, neural contusion, intra- or extra-axial hemorrhage, and major vascular trauma are properly imaged noninvasively by MR. Utilizing appropriate techniques, MR may substantially, and quickly, benefit treatment in order to improve outcome for the patient with spinal trauma. AR

Reader Note

In the June 2001 issue of Applied Radiology, the second part of this article will address Distraction Spinal Trauma.