Traumatic and Nontraumatic Spine Emergencies

Background and Imaging Algorithms

Spine injuries are quite common, with more than 137,000 cases entered into the American College of Surgeons National Trauma Data Bank in 2014. Major causes include motor vehicle collisions or pedestrian strikes, falls, acts of violence, and participation in sports. Of these cases, 35,000 had an abbreviated injury score of serious or higher, and almost 3000 fatalities occurred. According to the National Spinal Cord Injury Statistical Center, approximately 12,500 spinal cord injuries occur each year, resulting in 7500 new cases of quadriplegia. Imaging is liberally applied with a positive yield of 1% to 3% of all examinations, resulting in an estimated annual national cost of approximately $3 billion. The lifetime cost to individuals and society is enormous. Understandably, attempts have been made to develop evidence-based diagnostic algorithms in this area of medicine.

Multivariate analysis of data derived from two major clinical research initiatives, the National Emergency X-Radiography Utilization Study (NEXUS) and the Canadian C-Spine Study, have provided the basis for acute spine trauma imaging pathway development. Decision rules have been created that allow for discrimination of patients in need of imaging and those for whom imaging can be safely avoided, thereby reducing costs when possible. Once the decision to perform imaging has been made, the next step is to select the most appropriate modality. Plain film radiography traditionally has been the initial examination to evaluate for possible fracture or malalignment because it is readily available, relatively inexpensive to perform, and highly sensitive. It continues to be a cost-effective option for patients with a low probability of injury. However, it has been supplanted by computed tomography (CT) in the setting of moderate to high probability of injury, based on cost-effectiveness analysis that takes into account the high medical and legal costs of the rare missed fracture that leads to severe neurologic deficit. Studies of CT as the initial modality have demonstrated higher sensitivity for the detection of fractures; however, the clinical significance of many of the radiographically occult injuries is uncertain because of the lack of studies addressing outcomes. Another clinical prediction rule that has been developed, based on data from the Harborview study, may be used to stratify risk based on the injury mechanism and other clinical parameters. This type of approach is supported by trauma surgery societies and is commonly applied at trauma centers.

Multidetector-row CT (MDCT) has become the standard of care, even for patients with a low probability of injury. A zero-tolerance (for missed injury) approach to diagnosis using the fastest, most accurate examination rather than an outcome-based approach is easily adopted by the emergency department or trauma team. It used to be that for patients who had negative cervical radiographs but persistent pain, tenderness, or limited range of motion, symptomatic treatment with analgesics, cervical collar application, and clinical follow-up were the rule. At the follow-up visit, if symptoms had not resolved, flexion-extension radiographs were obtained to evaluate for the possibility of ligamentous injury or instability. Flexion-extension views have been shown to have little utility in the acute setting, primarily because of limited range of motion as a result of muscle spasm. In low-probability settings, this approach may still be followed, but more commonly, CT is requested to exclude occult fracture. The higher direct cost of CT may be offset by the increase in emergency department throughput, albeit at a higher radiation exposure. On some occasions, after a negative CT examination, magnetic resonance imaging (MRI) may be pursued to screen for signs of potential ligamentous injury before allowing the patient to be discharged.

The cost-effectiveness of MR for the detection of clinically significant ligamentous injuries has not yet been definitively determined. This is another instance in which technology is being applied because of its availability and perhaps for theoretical limitation of liability. Clearly, when a neurologic deficit is present and CT fails to identify a cause, MRI may offer additional sensitivity for the detection of soft tissue injuries, including disk extrusion, hematoma within the spinal canal, cord compression or contusion, and unstable ligamentous injury. Other injuries such as unsuspected bone marrow edema (microfractures) and vascular injuries also may be detected. One study that correlated MRI and intraoperative findings found that MRI had moderate to high sensitivity for detection of injury to specific ligamentous structures but suggested that it may overestimate the extent of disruptive injury. The increased sensitivity of MRI is accompanied by a small false-positive rate that may lead to added costs related to treatment/workup of clinically insignificant or unrelated abnormalities, such as thyroid lesions and lymphadenopathy. Special consideration has been given to the obtunded patient, because some studies have shown a 2% incidence of unstable cervical spine injuries that were not detected by radiography and CT because of the lack of an associated fracture or malalignment. However, many authors have recently suggested that MRI “clearance” of the cervical spine is not necessary in this setting. At our institution, we generally adhere to the practice management guideline from the Eastern Association for the Surgery of Trauma as outlined in Figure 7-1 . This approach, which no longer advocates MRI for obtunded patients with negative results of a high-quality CT scan, is from an evidence-based, meta-analysis approach using a 0.3% upper acceptable limit for missed unstable cervical spine injury.

Regarding the thoracolumbar spine, clinical prediction rules have been evaluated but provide only a very small decrease in the number of examinations performed. In patients with blunt trauma who undergo CT of the chest, abdomen, and pelvis with thin-section imaging (2.5 mm or less), sagittal and coronal reformats have been shown to be more sensitive and specific for detection of fractures, and therefore radiography can be avoided. When the viscera are not in need of examination, the role of CT for screening the spine is not as clear. The mechanism of injury is an important determinant for further workup in this category of patients. Similar to the logic applied to the cervical spine, screening is warranted if a high-energy mechanism of injury is known or suspected, including falls from significant height (greater than 10 feet), a motor vehicle or bicycle crash, pedestrians who have been struck, assault, a sports-related or crush accident, and a concomitant cervical spine fracture. Other valid indications are altered mental status, evidence of intoxication with ethanol or drugs, painful distracting injuries, neurologic deficits, and spine pain or tenderness to palpation.

For patients with neurologic deficits referable to a thoracolumbar spine injury, current Eastern Association for the Surgery of Trauma guidelines recommend obtaining an MRI examination as soon as possible after admission to the emergency department. Early decompression of a mass lesion, such as a traumatic herniated disk or epidural hematoma, is more likely to improve the neurologic outcome.

It is somewhat counterintuitive that the absence of symptoms does not exclude injury to the thoracolumbar spine. In one study, only 60% of trauma patients with a confirmed fracture were symptomatic. In a review from Maryland’s Shock Trauma Center of 183 fractures in 110 patients who were neurologically intact with a Glasgow Coma Scale score between 13 and 15 and who were considered amenable to clinical examination, 31% were recorded as having no pain or tenderness. The evidence would suggest that many of these fractures were not truly asymptomatic but rather occult as a result of intoxication or an unreliable physical examination. It is clear from the literature that no imaging modality is accurate 100% of the time. Most studies have found that radiographs of the thoracolumbar spine are commonly inadequate, especially in obese patients, and provide a sensitivity and specificity of only 60% to 70%.

Studies to develop specific guidelines for imaging of pediatric trauma patients are ongoing. The increasing use of MDCT and the long-term effects of radiation exposure are topics of special concern in this population.

Patterns of Cervical Spine Injury and Imaging Findings

In this section we include a brief review of the many different types of spine injuries that one must be familiar with when evaluating victims of trauma. Many texts are devoted solely to the imaging of spine trauma, with a few that truly reward the reader with insight into the anatomy, physiology, biomechanics, and pathology of this extensive topic. This section should serve as a valuable aid to the radiologist on call and can be used as a starting point for further study. Rather than taking a how-to approach to the interpretation of spinal imaging, in this section we rely on a working knowledge of the normal anatomy and basic principles of plain film, CT, and MR analysis. The general classifications of injuries are covered through a review of classic examples, using primarily CT with important plain film and MR correlations where appropriate.

Imaging of the spine can be thought of as a continuum, with radiography providing an overview of alignment and soft tissues, CT adding greater detail regarding fractures, and MR yielding finer detail with respect to soft tissues, including the spinal cord. Attention must be directed to the technical factors necessary to achieve a satisfactory (and safe) examination, including patient immobilization and positioning, image acquisition parameters, and multiplanar analysis.

Lateral, anteroposterior, and open-mouth odontoid views are the minimum requirement for plain films. A “swimmer’s” view may be necessary to adequately demonstrate the cervicothoracic junction. Thin-section CT (with a section thickness of 2 mm or less) with sagittal and coronal reformats of similar thickness is generally sufficient. However, anecdotal cases have arisen in which hairline fractures that were not detected prospectively with the standard technique were detected on scans performed with submillimeter thickness. Clearly a trade-off exists between the level of anatomic detail and the number of images that must be reviewed. With isotropic voxel size now possible with modern scanners, some radiologists have proposed a primary review of sagittal and coronal reformats to increase patient throughput. Thankfully, many of the missed fractures will be clinically insignificant because of their small size and inherent stability. In addition to the standard T1- and T2-weighted sequences used to evaluate the cervical spine, fat-suppressed T2-weighted sequences, with chemical selective or short tau inversion recovery (STIR) technique and gradient-echo sagittal sequence, are useful in the trauma setting. MR angiographic sequences may be indicated in certain circumstances.

Careful analysis of the structures (vertebrae, intervertebral disks, spinal cord, and other soft tissues) and their normal and abnormal attributes (size, shape, alignment, density, and signal intensity) requires an understanding of mechanisms of injury, including magnitude and acuity, and underlying diseases. The mechanisms generally can be grouped into hyperflexion, hyperextension, rotation, axial loading, lateral flexion, and others. Box 7-1 attempts to categorize the injuries of the cervical spine based on these mechanisms. Combined mechanisms, such as flexion and rotation, are common and may lead to multiple injuries at different sites and vertebral levels within the same patient. Rather than relaxing one’s vigilance after detecting an injury, the examiner should intensify the search for other lesions.

The determination of instability, which may be associated with or have the potential to progress to neurologic injury, major deformity, or incapacitating pain, is an important part of this process. General principles may apply based on specific imaging findings, but the final determination is probably best made by an expert in the treatment of spine injuries. One should note that the classifications of these injuries are constantly being revised based on new treatment techniques and clinical outcomes.

The following sections provide a top-down review of the major types of injuries and the mechanisms that cause them. It is not possible to describe all of the features of each injury in this abbreviated format; however, the general principles and commonly used classifications are described.

Injuries of the Cervicocranium

The cervical spine can be subdivided into the cervicocranium (including the basiocciput, craniocervical junction, atlas [C1], and axis [C2]) and the subaxial spine (C3 through C7), with each region including associated supporting ligaments.

Injuries of the cervicocranium are usually the result of blunt trauma and often are associated with high morbidity and mortality. Accurate rapid diagnosis is essential for appropriate management and satisfactory outcomes but is challenging because of the subtlety of imaging findings, the low incidence of disease, satisfaction of search (stopping after finding the first injury) in cases of multitrauma, and lack of familiarity with this anatomic region. Patients are often unresponsive or have altered mental status at the time of presentation, thus compounding the difficulty of injury assessment.

Although the relationship of the skull to the cervical spine has always been part of the usual evaluation, findings on radiographs are often subtle or obscured by overlapping structures. Reassessment of formerly relied upon radiographic measures to diagnose craniocervical injuries (including the Power ratio, X-line, and basion-posterior axial line) has shown limited sensitivity and specificity of these methods. Many authors have studied the relationships of the craniocervical junction in healthy subjects and in patients with clinically evident abnormalities and have derived myriad normal CT measurements, some of which differ from traditional radiographic measurements. As in radiography, abnormalities of the prevertebral soft tissues also provide important clues to diagnosis. Familiarity with and working knowledge of the CT measurements and prevertebral soft tissue findings will prepare the radiologist to identify clinically important injuries of the cervicocranium.

Considerable differences exist among authors because of various study methods and the expected variance within a population. Normal variation of the anterior atlantodental interval with age has been accepted for many years. Small variations in the other atlantoaxial relationships have also been reported based on MDCT studies. Therefore, it is not possible to define an absolute value for the upper limit of normal for these measurements. Use of the most conservative measurements may be the safest approach. When in doubt, based on CT, MRI should be considered for further evaluation of potential injury to ligaments and other soft tissues. Table 7-1 provides reasonable limits (based mainly on the work of Chang et al., Radcliff et al., and Rojas et al.—see the Suggested Readings). Based on logistic regression analysis of a series of proven injuries, Chang et al. suggested that an abnormal sum of occipital condylar displacements or C1-C2 spinolaminar widening resulted in a 95% detection rate of craniocervical distractions. Overall, the measurements should be considered with appropriate judgment to avoid missing clinically significant injuries. Occasionally the change from a patient’s previous examination will provide a clue to the diagnosis.

Assessment of the Prevertebral Soft Tissues

Evaluation of the prevertebral soft tissues (PVSTs) is an important component of the radiologic assessment of the cervical spine because subtle alterations in contour and thickness may be a harbinger for adjacent osseous and/or ligamentous injury.

In 2001, Harris described alteration of the normal PVST contour as an important secondary sign of injury. Such analysis increased detection of clinically important abnormalities with high sensitivity (less than 1% false-negative results) but was not very specific (84% false-positive results).

In 2009, Rojas et al. evaluated the PVST utilizing MDCT and established a range of normal values at the various cervical levels (upper limits): 8.5 mm at C1, 6 mm at C2, 7 mm at C3, and 18 mm at C6 and C7. Values were not determined at the C4 and C5 levels because of the varying position of the esophagus and larynx. In the setting of trauma, thickened prevertebral soft tissues may be due to edema and hemorrhage. Anterior displacement of the usually thin stripe of fat between the soft tissue densities of the pharyngeal mucosa/constrictor muscles and the prevertebral ligaments is another clue to injury. As with other measurements, none is perfect or absolute and should be considered with other findings and the clinical setting. Without a history or other signs of trauma, alternate differential diagnoses should be considered.

Specific Injuries by Site