Chapter 15: Diagnostic and Interventional Radiology

Imaging of the acute trauma patient is an important adjunct to clinical history and the physical examination, which may be challenging to obtain. Trauma imaging, when appropriately obtained and executed, can provide timely and helpful information about these patients and identify injuries that otherwise may not be immediately apparent. Imaging can be used to assist in patient triage and guide the trauma surgeon to any number of management choices. Despite the advanced imaging technology that is available in the modern trauma center, it is important to recognize that imaging alone cannot be used to make management decisions in isolation; that is, the surgeon must treat the patient based on all of the information obtained as a part of the assessment and not merely treat the images.

Imaging strategies are affected by the proximity of available imaging technology to the resuscitation area, the capabilities of the imaging equipment, the experience and availability of radiology technologists performing emergent imaging procedures, and timely access to expert interpretation and reporting.

The timing of diagnostic imaging should reflect the needs of individual patients and the local system, with imaging initiated in the trauma bay and integrated as a part of the clinical survey. Hemodynamically unstable patients should be resuscitated prior to imaging according to accepted guidelines and recommendations, with some exceptions for image-guided endovascular hemostasis in select scenarios. In order to enhance efficiency, triage priorities for imaging should be based on the acute needs for accurate information that can be used to direct treatment of the patient. Close cooperation and open communication between all major stakeholders, including emergency medicine physicians, traumatologists, consultants, nurses, imaging technologists, and radiologists, are essential to optimize any imaging assessment.

A single chapter alone cannot reasonably teach interpretation of diagnostic images. Instead, a general overview of imaging strategies will be presented, with an emphasis on advanced computed tomography (CT) technology and capabilities, focusing on select high-yield and common clinical scenarios.

Trauma Series

Imaging of the trauma patient is integrated as a part of the secondary clinical survey in the resuscitation suite. Anteroposterior (AP) supine chest and AP supine pelvis radiographs are routinely performed as a part of this initial assessment, if clinical evaluation alone is deemed insufficient (Fig. 15-1). A single view cross table, horizontal beam cervical spine radiograph can be obtained to evaluate for gross malalignment, but should not be used to exclude all fractures of the cervical spine. The goals of these initial imaging studies are to identify life-threatening, but clinically occult, findings that require emergent intervention prior to any further imaging, such as an unstable pelvic fracture, hemopneumothorax, or malpositioned/misplaced support lines and tubes.¹

The trauma resuscitation “ABCD” strategy may be extended to this “trauma series.” Verification of the integrity of the airway (and other tubes and lines) should be specifically made on the chest radiograph (and lateral cervical spine radiograph, if obtained). Radiographic pulmonary opacities associated with hypoxemia include pulmonary contusions, aspiration pneumonitis, and atelectasis (including collapse due to aspirated dental or foreign debris). Tension pneumothorax and hemothorax are typically detected on clinical examination, while clinically occult pneumothoraces or hemothoraces are commonly shown by chest x-rays as a “deep sulcus” sign and generalized hemithoracic opacification, respectively. Other injuries, such as rupture of the hemidiaphragm, flail chest, pneumopericardium, and pneumomediastinum and hemomediastinum, are often diagnosed or at least suggested by initial x-ray findings. Hemodynamic instability may arise from any number of causes, including extraperitoneal hemorrhage from pelvic ring disruption or hemoperitoneum from solid organ injuries. Biomechanically unstable disruptions of the pelvic ring are almost always shown on AP radiographs and may be associated with injuries to the bladder and urethra. In addition, pelvic x-rays may show hip dislocations and fractures of the acetabulum and proximal femur.

A technically adequate (C1–T1) lateral x-ray of the cervical spine is a limited examination but may provide helpful information in the setting of a cervical spine fracture/dislocation or other gross malalignment, and may confirm that spinal shock from a vertebral fracture is the cause of unexplained hypotension.

Focused Abdominal Sonography For Trauma (FAST)

FAST is performed as part of the secondary survey in victims of torso trauma (see Chapter 16) to identify fluid accumulations in the chest and abdomen, with these abnormal fluid collections serving as a proxy for significant internal injury. The ultrasound probe is used to search for fluid in the pericardial sac, both upper abdominal quadrants, and the intraperitoneal recesses in the pelvis adjacent to the urinary bladder. Scanning is optimally performed in two orthogonal planes (eg, longitudinal and transverse) and may be extended to detect a hemothorax and a pneumothorax. In patients who have severe hemodynamic compromise or obvious hemorrhagic shock, FAST can establish the abdomen as a source of hemorrhage within a few seconds. It is important to recognize that FAST is a limited anatomical examination and, as such, the identification of intraparenchymal and retroperitoneal injuries is not a goal, but certainly occurs. FAST is widely available, inexpensive, noninvasive, uses no ionizing radiation, and can be repeated serially. Commercially available, portable handheld real-time imaging devices are technically adequate to perform FAST.

FAST is uniformly accurate for the detection of intraperitoneal fluid with moderately large volumes greater than 400 cm³ (at smaller volumes, accuracy varies with user experience).^2,3,4 Unfortunately, isolated hepatosplenic injuries with minimal or no hemoperitoneum represent as many as one-third of solid organ injuries.^5,6 Fortunately, small isolated intraparenchymal lesions with less than 250 mL of intraperitoneal blood often do not require endovascular or surgical intervention (liver <1%, spleen <5%).⁷ False-positive interpretation of FAST images can result from improper machine settings (gain), sonolucent perinephric fat (which is rarely sonolucent in both axial and coronal scanning planes), fluid-filled loops of bowel, bladder, various types of fluid-filled intra-abdominal cysts, and physiological or nontraumatic free fluid (ascites).

The FAST does, however, require operator training and experience for reliable performance and interpretation, which limits its value. In addition, FAST lacks value in patients with preexisting nontraumatic ascites, bowel and mesenteric injuries, injuries to the retroperitoneum, and hemoperitoneum due to pelvic fractures. While a systematic review of the literature would not support the use of FAST as a replacement for DPL and CT in blunt abdominal trauma, many trained surgeon-sonographers use it on a daily basis with great accuracy.⁸

Hemodynamically stable patients who have suffered blunt trauma with a clinical presentation suspicious for injuries, regardless of FAST result, are candidates for intravenous contrast-enhanced CT of the abdomen and pelvis to identify or exclude internal injuries.

COMPREHENSIVE IMAGING FOR BLUNT POLYTRAUMA

Blunt polytrauma victims are triaged based on the physical examination, initial imaging assessment and hemodynamic status to either the operating room, intensive care unit, or angiography suite for ongoing resuscitation. Some remain in the trauma center for completion of secondary and tertiary clinical surveys and advanced targeted imaging. Typically, this consists of CT for clinically appropriate imaging (eg, head, neck, chest, abdomen, and pelvis) for the most severely injured, but hemodynamically stable patients. Subsequently, conventional x-rays of the extremities may be obtained. Less severely injured individuals may take a slightly different route with conventional x-rays preceding CT and directed at abnormalities found by clinical examination or prior imaging.

There has been recent increased utilization of the “pan scan” or other variations of whole body CT imaging techniques as a method of screening patients for clinically unsuspected injuries or to further define injuries identified on physical examination and initial imaging. This paradigm advocates rapid acquisition of high-resolution thin collimation images to include CT of any or all body parts, including the extremities in some instances. Recent literature suggests that integrating whole body CT as a part of the comprehensive evaluation of the trauma patient decreases morbidity and mortality.^9,10 Employing this strategy provides excellent anatomic coverage with high accuracy for the detection and exclusion of injuries and is frequently utilized in high-volume facilities with considerable imaging resources.

Multidetector Computed Tomography (MDCT)

The widespread availability of MDCT scanners has revolutionized the way that trauma patients are imaged.^11,12,13 Modern MDCT scanners, particularly those with 16 or more detector rows, acquire imaging data at sub-millimeter slice thickness. Because of this isotropic spatial resolution, two-dimensional (multiplanar) images in any arbitrary plane can be reconstructed with the same resolution as the axial images. Three-dimensional images with surface rendering approximate that of anatomic specimens, and some of the newest software packages can reconstruct CT raw data sets to generate “virtual” radiographs (Fig. 15-2).

This, coupled with rapid intravenous contrast injection and increased table speed, allow for even shorter scan times as well as imaging large portions of the body during multiple phases of vascular and organ enhancement. As a result, patients can be scanned from the cranial vertex to feet in only a few seconds.^14,15 Reconstructions of any scanned body part, including the face, thoracolumbar spine, pelvis and extremities, can be generated in multiple planes, thus reducing repeat scanning. The previous practice of detailed specialized imaging of the face and spine can be abandoned. As such, individual repeat thin-section imaging of parts of the body in addition to the CT survey is superfluous, increases ionizing radiation unnecessarily, and increases time away from the critical care area.

Instead of sending all of the sub-millimeter thin section data to a picture archiving and communication system (PACS) for review (which can be quite time-consuming, despite improved image transfer time), the thinner slices are stacked together so that maximum useful information can be extracted without overwhelming the reviewer with high image volume. It is typical for images though the chest, abdomen and pelvis to be reconstructed at 2–5 mm slice thickness in axial, sagittal, and coronal planes, though any slice thickness can be prescribed and tailored to local preferences. It is common practice for the axial images to be reconstructed using thicker slices (such as 3–5 mm), with sagittal and coronal images reconstructed using thinner slices (such as 1.5–3 mm) to provide finer spatial resolution of the pelvis and spine. Open communication with the radiology team is strongly encouraged so that optimal image quality and the imaging needs of the trauma surgeon and patient can be achieved. Consistency and familiarity with the scanning process and the image reconstructions can help improve patient throughput once a standard trauma CT survey is agreed upon, thus decreasing patient time away from the trauma care area.

Computed Tomography of the Head

Axial noncontrast CT scanning remains the reference standard in patients with acute craniocerebral trauma, guiding initial decisions regarding clinical management.^16,17 Indications for CT of the cranium include the following: (1) objective evidence of closed injury to the brain, including decreased level of consciousness; (2) cranial or facial deformity; (3) hemotympanum; or (4) evidence for leakage of cerebrospinal fluid (Figs. 15-3, 15-4, 15-5). Clinical criteria reliably predict significant intracranial injury (ICI) and help determine which patients will require CT scanning of the head. In children, significant ICI is extremely unlikely in any child who does not exhibit at least one of the following high-risk criteria: (1) evidence of significant skull fracture; (2) altered level of alertness; (3) neurological deficit; (4) persisting vomiting; (5) scalp hematoma; (6) abnormal behavior; or (7) coagulopathy.¹⁸ More generically, minor trauma to the head may lead to surgically important injuries to the brain, and liberal utilization of CT is appropriate among individuals who have sustained “high-risk” mechanisms. These clinical criteria are not as reliable in the elderly patient.¹⁹

Head CT images are traditionally acquired using a conventional “step and shoot” axial technique. Most modern MDCT scanners, however, have the capability of scanning the head using helical acquisition, as utilized for other body parts, and can thus reconstruct head CT images in both sagittal and coronal planes. Traditionally, axial images are reconstructed at 5-mm slice thickness using both bone and soft tissue algorithms, and viewed at bone windows and at least two different soft tissue windows (including “brain” and “blood” window settings).

CT scanning is highly sensitive and specific for the detection of extra- and intra-axial hemorrhage, mass effect and midline shift, soft tissue injuries to the scalp and globes, as well as fractures to the calvarium, skull base, bony orbits, and paranasal sinuses. In patients with diffuse axonal injury, however, CT may even be normal and discordant between the severity of the clinical brain injury and radiographic findings. On a cautionary note, skull fractures that are aligned in the plane of scanning may easily elude detection, and review of multiplanar reconstruction (MPR) images or the scout views used to plan the CT study of the head may alert the clinician to such fractures. Except for medicolegal imaging of nonaccidental trauma (child abuse), conventional x-rays of the skull are usually not necessary.²⁰

Computed Tomography of the Maxillofacial Skeleton

Indications for specific facial CT scans include the following: (1) deformity or instability of the maxillofacial structures found by physical examination; (2) deformity, opacification, or fracture of the periorbital or paranasal sinus shown on head CT; and (3) clinical evidence for leakage of cerebrospinal fluid (Fig. 15-6). The mnemonic “LIPS-N” (lip lacerations, intraoral lacerations, periorbital contusions, subconjunctival hemorrhage, or nasal lacerations) provides a helpful tool during clinical examination of trauma patients given the high association between LIPS-N lesions and facial fractures.²¹ Fortunately, physical examination is highly sensitive for selecting which patients need maxillofacial CT imaging.²² Absence of paranasal or periorbital sinus fluid on CT (the “clean sinus” sign) generally excludes surgically important injury to the maxillofacial skeleton, with only very rare exceptions.²³

Images of the face are acquired using helical scanning technique and are reconstructed in the axial, sagittal, and coronals planes at 1–2-mm slice thickness. It is important to recognize that some institutions do not routinely include the mandible as a part of the maxillofacial CT scan protocol, thus open communication with the radiology team is important in order to include the relevant anatomy for evaluation.

While CT is highly accurate at detecting and characterizing surgically important injuries, it does not show the magnitude of osseous fragmentation found at surgery in complex fractures. Orbital and maxillary fractures should heighten suspicion of more complex fractures. Helical acquisition with MDCT technology can be used to generate 3D surface rendered images of the bones, which can be quite helpful in determining the type of facial fracture patterns, particularly for patients with highly comminuted and disorganized high energy midface smash fracture patterns. This technique provides valuable information on spatial relationships and is particularly useful in planning operative treatment.^24,25 On the other hand, axial and 2D reformations best portray defects in soft tissue, interposition of osseous fragments, and herniations of soft tissue (eg, orbital floor blowout fractures) and are best for quantifying size of fractures.^26,27

Radiation reduction techniques for CT continue to evolve; however, maxillofacial CT can still deliver significant radiation to the orbits and to the soft tissues of the neck (eg, thyroid), which is an important consideration when imaging children.²⁸

Imaging for Soft Tissue Injuries of the Neck

Clinical findings for blunt injury to the aerodigestive tract are nonspecific, but include subcutaneous crepitus, hemoptysis, hoarseness, neck pain, and abrasions or hematomas (eg, from the shoulder harness of a three-point restraint; Figs. 15-7 and 15-8). X-ray findings include parapharyngeal or precervical emphysema, soft tissue swelling, or fracture of the larynx or hyoid bone on the lateral image of the cervical spine. Soft tissue injuries are inadequately evaluated using radiography; hence, CT is the accepted imaging standard for imaging this area.

MDCT of the neck with intravenous contrast can be used to evaluate for not only upper aerodigestive tract injuries, but can also be used to simultaneously evaluate the cervical vasculature and spine (see later). Images are routinely reconstructed in three planes using both bone and soft tissue algorithms.

CT is often most helpful in the evaluation of laryngotracheal injuries when physical examination and endoscopy are technically difficult. A careful search for direct signs of a laryngotracheal injury, manifested by focal defects, guides the need for intervention for debridement and mucosal closure. Soft tissue gas may be focal adjacent to the site of injury, but larger defects, particularly if the patient is under positive pressure ventilation, can result in massive soft tissue emphysema extending into the skull base and head superiorly or into the mediastinum and beyond inferiorly.²⁹ CT can assist in the grading of laryngotracheal injuries, but tends to understage compared to endoscopy or open exploration. The most common injuries are those to the thyroid cartilage, which typically occur within 2–3 mm of the anterior crest of the two lateral laminae. Comminuted fractures of the laminae generally result from higher energy and direct impact injuries to the larynx and are more commonly associated with thyrocricoid dislocation. Also, CT can demonstrate subluxations and dislocations of the arytenoid cartilage. Most tracheal disruptions are in the membranous portion of the trachea and will heal with conservative therapy. Soft tissue emphysema immediately adjacent to the trachea suggests an injury in this location, but again, could be massive in volume and anatomic distribution. A search for a fracture of a tracheal ring is often fruitless unless the fracture is displaced. Coronal reformations through the trachea may show separation (vertical diastasis between tracheal rings) of the trachea compatible with more serious grades of injury.

Blunt esophageal injuries are very uncommon, but, when they do occur, are most often seen in the proximal third of the esophagus. On the other hand, a penetrating injury to the esophagus is more common and may involve any portion of the esophagus. Any focal gas adjacent to the esophagus, without an obvious airway source, should raise the possibility of an injury, particularly if a wound tract extends up to or through the region of the esophagus.

If there is a concern for an esophageal injury, contrast esophagography, endoscopy, or both should be performed. Both tests have a sensitivity of 80–90%, but increases to ~95% when both examinations are performed.^30,31 Iso-osmolar water-soluble contrast material is the preferred initial contrast agent of choice for suspected esophageal injuries, as it is less toxic to the lungs if aspirated, and is better tolerated if leaked into the mediastinum or peritoneum. Unfortunately, iso-osmolar contrast material is less dense on fluoroscopy, with the potential to miss subtle injuries. Barium sulfate is easier to see but can be irritating to the soft tissues if an esophageal injury is present. Gastrografin should be avoided because of its hyperosmolarity and can cause life-threatening pulmonary edema if aspirated.

Esophagography performed on patients who are awake and able to cooperate with the examination is performed as any other esophagram, but with the contrast agent modifications as above. Performing esophagography on intubated and/or sedated patients is technically challenging. A nasogastric tube is necessary to perform this examination on uncooperative patients. The patient is placed in reverse Trendelenberg position (10–20^° as tolerated) with the fluoroscopy table floor board in place. The radiologist manipulates the nasogastric tube to the distal esophagus, and, under real time fluoroscopic surveillance, injects the contrast agent in an attempt to achieve full distention of the esophagus. The contrast is then aspirated, and the NG tube is repositioned more proximally. This process is repeated to evaluate all segments of the esophagus. Care should be taken to avoid over distention of the proximal esophagus, which can result in reflux and aspiration, even in intubated patients. Rapid imaging is necessary, with fluoroscopic cineradiographs are obtained at 2–3 frames/sec.

Neck CT angiography with MDCT has a sensitivity of 90–100% for blunt cervical vascular injuries and is, therefore, the imaging test of choice in patients without indications for immediate surgery or catheter angiography. Neck CTA can be performed as its own stand-alone examination (as in the setting of penetrating neck injuries), added on to the standard trauma CT protocol, or can be integrated as the part of a single pass whole body CT (WBCT) trauma protocol.^{32,33,34,35,36} Neck CTA protocols vary widely based on local preferences and mechanism of injury (ie, blunt versus penetrating). Positioning of the patients arms is an important consideration when integrating neck CTA into the imaging evaluation of these patients, as increased image noise and decreased image quality will result when the arms are included in the scanned region.³⁶ Optimal cervical arterial opacification can be obtained following either a timing bolus or using a fixed scan delay. These images are reconstructed at 1–3-mm slice thickness from the aortic arch to the skull base (or to the cranial vertex if needed), with MPR reconstructed in axial, sagittal and coronal planes. Coronal maximum-intensity projections (MIPs) are particularly helpful for visualizing the cervical vasculature, but should not be relied upon for evaluation of the vertebral arteries due to volume averaging effects and the potential for a missed injury. Liberal use of a dedicated 3D volume viewing station for manipulation of the data set is strongly encouraged, as it is often the case that a cervical vascular injury may be best visualized and displayed using nontraditional planes.

Catheter angiography is indicated emergently in patients with an expanding cervical hematoma, active extravasation from the nose, mouth, or ears, or a cervical bruit in individuals younger than 50 years old.³⁷ Angiography is not only a diagnostic tool in symptomatic blunt carotid and vertebral artery injuries, but it is also a therapeutic tool as it allows for rapid utilization of endovascular therapy for bleeding and identifies patients at risk for embolic stroke.^38,39 The sensitivity of duplex Doppler US is inadequately low (38.5%) for directly depicting vascular injuries in the neck and is, therefore, not a primary tool for evaluation of the cervical vasculature.⁴⁰ Existing literature for the use of magnetic resonance angiography (MRA) demonstrates limited sensitivity for cervical vascular injuries compared to CTA and catheter angiography.⁴¹ This, in addition to the additional technical and practical limitations of MRI in the setting of trauma, limits the utility of MRI/MRA in the acute setting.

Blunt carotid and vertebral injury (BCVI) (Fig. 15-9A–C) is now known to be much more common than previously appreciated, occurring in 1–3% of screened asymptomatic patients.³⁴ Vascular imaging of the neck is warranted in injury patterns associated with a high risk of blunt injury to the carotid and vertebral arteries, including the following: fracture of the cervical spine (especially involving C1–C3, involving the transverse foramina or extension into the foramen magnum), neurological deficits not explained by findings at brain imaging, new-onset Horner’s syndrome, high-energy facial fractures (Le Fort II or III), fracture of the skull base involving the foramen lacerum, or soft tissue injury in the neck (neck belt sign and “hangings” sufficient to cause central anoxia).^38,42 It should be noted, however, that approximately 1/5 of patients with blunt cervical vascular injuries do not have these classically described risk factors and thus may elude detection if selective screening is employed.⁴³

Penetrating injuries to the neck are associated with high morbidity and mortality, primarily due to injuries of the cervical vasculature, upper aerodigestive tract, and spine. As such, any patient with “hard” signs of a cervical vascular injury (such as expanding pulsatile hematoma, active bleeding from the wound, hemorrhagic shock, etc.) or compromise of the upper aerodigestive tract injury/airway (such as stridor, hoarseness, etc) requires urgent intervention by individuals with experience managing such injuries. For patients without indications for immediate surgery or intervention, imaging with CT angiography has proven highly useful in identifying and excluding injuries that require intervention and can significantly reduce the rate of nontherapeutic cervical explorations.^44,45,46,47 Indeed, there has been a recent shift away from mandatory evaluation with surgery and other invasive tests based on zonal wound location toward a “no zone” selective approach with CTA evaluation of the neck.⁴⁸ Many factors must be considered when approaching the patient with a penetrating wound to the neck, including the mechanism of injury (high vs low energy), the number and location of surface wounds, hemodynamic and neurological status, and the likelihood of injury. Gunshot wounds can result in significant soft tissue cavitation that may result indirect or indirect vascular injuries. Also, fragments of bone and metal can cause artifacts in the area that make interpreting a CT difficult. Catheter angiography (particularly digital subtraction angiography [DSA]) is extremely useful for the evaluation of the carotid and vertebral arteries when they are partially obscured by beam hardening artifact from metallic bullet fragments. CT angiography has a major role in evaluation of stable patients with clinically occult injuries (Fig. 15-9D-E).

Neck zonal anatomy is an important consideration for patients with “hard” clinical signs of significant internal injury (see Chapter 41); however, it should be noted that clinically significant internal injuries may be present in a zone differing than that of the physical examination, and that penetrating wound tracts may extend beyond the neck into the skull base, cranial vault, mediastinum or upper thorax. Zone III, above the angle of the mandible, is difficult to assess clinically and to explore operatively. There are numerous vessels at risk, including the internal carotid artery, the external carotid artery and branches, the vertebral artery, and the accompanying veins. Thus, imaging plays a vital role in these patients. Neck CTA can be used to evaluate these vessels, but a low threshold for catheter angiography should be maintained for patients with possible injuries in Zone III. Selective internal and external carotid arteriography, vertebral arteriography, and four-vessel intracranial arteriography all play important roles in detecting and evaluating Zone III vascular injuries. Indeed, arteriography is so valuable for the evaluation and treatment of bleeding in this zone that aggressive steps to resuscitate the unstable patient and control bleeding by packing are warranted to allow angiography to proceed. Most vascular injuries in Zone III are best managed by the interventional radiology service.

Zone II, between the inferior margin of the angle of the mandible and the manubrium of the sternum, is more easily and reliably evaluated by physical examination. If the common carotid artery or the internal jugular vein require repair, operative exposure is relatively simple and unobstructed. Thus, little imaging is necessary after penetration when “hard” signs of a vascular injury (see Chapter 41) are present. The detection and treatment of vascular injuries in asymptomatic patients is more controversial.

In Zone I, the area inferior to the manubrium of the sternum, the brachiocephalic vessels, trachea, or esophagus may be injured, and rapid exsanguination may occur. As a result, symptomatic patients may undergo urgent exploration without imaging.

Neck CT angiography is a valuable screening tool for asymptomatic patients because it can be used to determine trajectory and identify direct signs of vascular injury. When patients are stable or asymptomatic, immediate CT angiography has great value in excluding injury, in detecting vascular injuries that can be treated nonoperatively by embolization or insertion of a stent or stent graft, and in facilitating surgical exploration. Penetrating injuries to Zones I and III of the neck are best evaluated with catheter-based imaging in the hemodynamically stable patient, particularly if bullet fragments are present. CT of the neck and chest may allow analysis to determine whether penetration is in proximity to the great vessels or even whether vessels are injured. As intimal injuries to the great vessels may be life-threatening, these artifacts limit the usefulness of CTA in Zone I and, possibly, in Zone III.

Imaging of spinal trauma typically begins with radiography to evaluate for biomechanically unstable injuries, such as burst fractures and dislocations. It should be noted that adequate radiographs of the spine may be technically difficult to obtain, may have limited quality in obtunded or immobilized patients, and may not detect all fractures. A meta-analysis of the sensitivity of radiography for cervical spine fractures by Holmes et al⁴⁹ documented a pooled sensitivity of 52%, compared to 98% for cervical spine CT. Thus, the decision to image with radiography or CT must be made based on the mechanism of injury and the likelihood of a spinal fracture and must be balanced with the possibility of other more life-threatening injuries. If an abnormality is identified on conventional x-rays, CT and/or MRI may be used to further characterize the injury for planning of treatment and provide information on the patient’s prognosis.

Conventional Radiographs of the Spine

Validated clinical prediction rules (ie, National Emergency X-Radiography Utilization Study Group [NEXUS] and the Canadian Cervical Spine Rule) for adults and older children (≥10 years old), can reliably identify trauma victims who need imaging of the cervical spine.^50,51 In essence, oriented asymptomatic individuals without findings on a physical examination following trauma do not require subsequent imaging. Imaging of the thoracic and lumbar spine following blunt trauma is indicated when patients present with one or more of the following: (1) signs or symptoms of local injury (pain, tenderness, interspinous step-off); (2) depressed level of consciousness, including intoxication; (3) acute myelopathy or radiculopathy referable to the thoracolumbar spine; and (4) major distracting injury, including concomitant injuries to the cervical spine.^52,53

The cross-table lateral cervical spine radiograph is often obtained in the setting of acute trauma, though its use is declining due to the easy availability of MDCT and relative lack of sensitivity compared to CT. Single view lateral cervical spine films may not reveal a significant percentage of fractures. In fact, a study by Brohi et al⁵⁴ in 2005 reported a sensitivity of 39.3% for single lateral view radiography and a sensitivity of 51.7% for unstable cervical injuries. This same study reported a sensitivity, specificity, and negative predictive value of CT to be 98.1%, 98.8%, and 99.7%, respectively.⁵⁴ The standard three-view cervical spine series includes an anterior-to-posterior (AP) open mouth odontoid view (an AP image of the craniocervical junction), an AP view of the subdental cervical spine, and a lateral view of the cervical spine that extends down to the C7–T1 interspace. To supplement these three views, a swimmer’s lateral image (one arm elevated above the head and the other arm in caudal traction) is often obtained to better visualize the cervicothoracic junction, and bilateral trauma oblique views obtained to evaluate the facet joints. The addition of these supplemental images adds time and radiation exposure and may still be technically limited, despite maximum effort on the part of the x-ray technologist and the patient.

Given the differences in the incidence of injury to the cervical spine in children and the differences in their distribution (far more common in the upper cervical spine) relative to adults, an examination limited to frontal and lateral views is acceptable.⁵⁵ In infants 0–4 years of age and in children 5–9 years of age, AP, lateral, and open mouth views are satisfactory. All patients aged 10 and over require a minimum of three views to as many as five or six views to adequately survey the cervical spine.

Films of the thoracic and lumbar spine are usually obtained as separate sets of frontal and lateral projections. The upper thoracic spine may require the addition of a swimmer’s lateral view, if one has not been previously obtained as part of a cervical spine series. In general, if pathology is identified, then further evaluation with CT is indicated to further define the fracture pattern and to identify potential adjacent spinal injuries.

Technically inadequate examinations of the cervical or thoracic spine are almost always due to the inability to visualize the cervicothoracic junction. It is necessary to see the top of T1 on the cervical spine and the bottom of C7 on radiographs of the thoracic spine. A careful count of the vertebral bodies on the frontal examination to establish the correct levels is recommended to avoid wrong level surgery and is based on the number of rib-bearing (thoracic) and non-rib-bearing (lumbar) vertebrae. On the lateral images, it is important to look at the corners of the vertebral bodies, especially the anterior superior corner, which is affected in approximately 90% of all vertebral body fractures. On the frontal images, it is most important to evaluate the adjacent end plates for continuity, the lateral margins of the vertebrae, the posterior elements for pathological interspinous and interpediculate widening, and horizontal lucencies that would suggest horizontal soft tissue and/or osseous disruption of a flexion-distraction-type injury.

One of the common errors made in evaluating the cervical spine is mistaking developmental variations for pathology. Common variations at the craniocervical junction include the following: (1) fusion of C1 to the occiput (which may be partial or complete); (2) failure of fusion or development of the posterior elements of C1; (3) pseudospreading of C1 relative to C2, which may mimic Jefferson burst fractures (most common in the 0- to 4-year age range, but may be seen up through puberty); (4) pseudosubluxation of C2 on C3 in pediatric patients, which can be recognized as normal by a normal C1–C3 spinolaminar line; and (5) os odontoideum (an anomalous bone that replaces all or part of the dens axis and is not attached to the atlas).

Flexion and Extension Cervical Spine Radiographs

Flexion and extension radiographs may be used to assess for potential ligamentous instability in patients who are completely alert and who have normal x-rays but exhibit posterior midline tenderness (Fig. 15-10). Some centers have recommended the use of passive (guided by the physician) flexion and extension studies using fluoroscopy and, though this may be appropriate in very limited circumstances in the hands of physicians with considerable experience, the published data are not sufficiently strong to warrant generalization.

A qualified physician should be in attendance if the examination is performed shortly after injury (hours to days), and the patient needs to be completely alert and able to assume an upright posture and precisely follow commands. An initial lateral radiograph is obtained with the patient upright and with the cervical spine in a neutral position. This image is reviewed by the physician overseeing the examination and, if normal, the patient is asked to actively extend the neck. An additional lateral image in maximum extension is obtained. The patient’s cervical spine is then returned to a neutral position. This extension x-ray is evaluated in a manner similar to that taken with the cervical spine in a neutral position. If normal, the examination is repeated with maximum effort at flexion, and an x-ray is obtained. Standards for an adequate examination vary from a range of motion of 30° to 90°. The test is intended to study the capacity of the spine to resist physiological stresses, however, and the mean normal range of motion in adults is approximately 90°. If the flexion–extension x-rays are abnormal in the acute or subacute setting, MRI is pursued. If the examination does not show an adequate range of motion, the patient’s spine is immobilized, and the examination is repeated in 2 weeks if the patient remains symptomatic. It should be noted that approximately one-third of flexion-extension x-rays exhibit suboptimal flexion and/or extension, thus limiting its utility.⁵⁶

Computed Tomography of the Cervical Spine

Blackmore et al⁵⁷ developed a clinical prediction rule to determine preimaging risk of fracture in select patients about to undergo a helical CT to survey the cervical spine for fracture and soft tissue injuries (Figs. 15-11 and 15-12). For the application of this clinical prediction rule, it is assumed that the patient will already be undergoing CT of the cranium. Any one of three mechanisms of injury or any one of three clinical findings puts the patient at a pretest risk of greater than 5% of harboring an injury in the cervical spine. High-risk mechanisms of injury include a high-speed motor vehicle crash greater than 35 mph or greater than 50 km/h combined impact, motor vehicle crash with a death at the scene, or a fall from a height of greater than 10 ft or greater than 3 m. Clinical parameters suggesting an increased risk for injury to the cervical spine that are associated with a high-risk mechanism include a significant closed injury to the brain (or intracranial hemorrhage shown on head CT), acute neurological deficits referable to the cervical spine (acute myelopathy or radiculopathy), or either pelvic fracture or multiple extremity fractures. Hanson et al⁵⁸ validated the clinical prediction rule prospectively and showed its application by separating victims of blunt trauma into a high-risk group (12% prevalence of acute cervical spine injury) and a low-risk group (0.2% prevalence of cervical spine injury). No validated clinical decision rules exist for infants (<1 year old) and younger children (<9 years old). In general, patients with greater severities of injury (ISS >25) have an elevated risk of injury to the cervical spine. For patients aged 9 years or younger, conventional x-rays will depict essentially all clinically important fractures and dislocations. CT is generally not indicated in younger children and infants to screen the cervical spine, nor in search for other occult injuries causing neurological deficits.⁵⁹ CT should be reserved as a staging/treatment planning procedure among patients with a known bony abnormality.

Modern helical MDCT scanners allow for high resolution imaging of the cervical spine, with the quality of sagittal and coronal images equating that of the axial images. Axial slices of the cervical spine are obtained at 1.0–3.0-mm slice thickness from the skull base through the upper thoracic spine (T2–T4 vertebral body level). Coronal and sagittal reformations are typically 1.0–3.0-mm thick images and can be reconstructed using either a bone or soft tissue algorithm. Generally, the information gathered from these reformations is sufficient and makes plain x-rays unnecessary. PACS workstations can facilitate the rapid review of these large image sets, and the use of cross-referencing tools assists with identification of specific vertebral levels.

Careful review of the entire cervical spine in all three planes is necessary for comprehensive evaluation. Some of the more common fracture patterns (such as type II fractures of the base of the odontoid process or horizontally oriented fractures of the spinous process) have “in plane” axial fracture lines and thus may be difficult to identify when reviewing the axial images alone. The coronal reformations may be viewed using the same approach as standard AP x-rays, while viewing of the sagittal reformations can be performed with guidelines used for the lateral cervical spine. Nonetheless, careful attention to axial images is necessary to detect fractures involving the craniocervical junction, transverse processes (which increase the likelihood of a vertebral artery injury), margins of vertebral bodies, pedicles, lateral mass, or lamina and spinous processes.

The sensitivity and specificity of cervical spine CT for acute bony injuries is routinely in excess of 95%. Although CT does not directly show soft tissue injuries to the spine, focal kyphosis, focal lordosis, and widening of the disk space can be used as with conventional x-rays to suggest associated injuries to soft tissue. Occasionally, acute disc herniations can be identified on axial or sagittal soft tissue images. Some authors feel that clinically important injuries to soft tissue causing biomechanical instability are almost always evident on technically adequate CTs of the cervical spine (especially on the sagittal and parasagittal reformations).⁶⁰

Computed Tomography of the Thoracolumbar Spine

A low threshold should be maintained for the use of CT to further evaluate vertebral body deformities identified on conventional x-rays. In addition, patients with high-risk mechanisms and impressive signs or symptoms without abnormal x-rays should undergo a thoracolumbar CT to detect occult minimal burst fractures or Chance or flexion-distraction-type injuries. Dedicated radiographs of the thoracolumbar spine can be omitted if a CT of the chest and/or abdomen is obtained. The CT raw data of the chest and/or abdomen can be used to generate dedicated coned down images of the thoracolumbar spine, reconstructed in axial, sagittal, and coronal planes using a bone algorithm, without the need to re-image the patient.

Limited imaging of specific vertebral body levels based on radiographic findings is being replaced by more comprehensive spinal imaging. Typically, 2.0–3.0-mm thick axial slices are acquired and are reconstructed using both bone and soft tissue algorithms. Imaging of the entire thoracic or lumbar spine is advised as it allows more accurate determination of the location of injury. Sagittal reformations are made in both algorithms and viewed at bone and soft tissue windows, respectively. Careful attention to the paraspinal structures, such as the ribs, mediastinum, retroperitoneum, is strongly recommended if a flexion-distraction or chance fracture is detected on dedicated thoracolumbar CT imaging. This diagnostic pitfall is becoming less of a factor due to the increased use of contrast-enhanced whole body CT imaging in trauma.

Magnetic Resonance Imaging of the Spine

The principle indications for MRI are to characterize soft tissue injuries and potential spinal cord injuries in patients with associated fractures and/or luxations of vertebrae (Fig. 15-13). An MRI is indicated in individuals who have no conventional x-rays or CT abnormality, but who have an acute myelopathy or radiculopathy (spinal cord injury without radiographic abnormality [SCIWORA]). The use of MRI is frequently controversial in the setting of dislocation of bilateral or unilateral facets (see Chapter 23). Some advocate initial reduction followed by MRI, but this is a practice that varies from institution to institution. In general, urgent MRI is appropriate when there is an evolving neurological deficit or neurological deficits without explanation.

Edema of the spinal cord has a much better prognosis than hemorrhage into the cord. MRI is helpful in making this distinction, as well as detecting epidural hematomas that may require decompression. Evaluations of the disk spaces, ligaments of the spinal column, facet joints, and the cranio-cervical junction are best made with MRI.

Spine MRI imaging protocols in trauma vary among institutions, but sagittal and axial T1-weighted and fluid-sensitive sequences (eg, T2-weighted STIR) are standard. Gradient echo (GRE) sequences are useful in detecting magnetic susceptibility artifact due to blood products. When assessing the transverse atlantal ligament, images should be obtained in the axial plane parallel to Ranawat’s line (a line from the anterior most portion of the anterior tubercle of C1 to the most posterior aspect of the posterior arch of C1).

MRIs have been used to “clear” the cervical spine in obtunded or unexaminable patients with otherwise normal imaging (CT or high-quality conventional radiographs).^61,62 The absence of abnormal high signal intensity in ligaments and discs effectively excludes biomechanically significant injuries; however, an abnormal signal does not imply instability. At this time, MRI has not been shown to accurately grade injuries to the posterior longitudinal ligament (PLL) or anterior longitudinal ligament (ALL), though injuries may be identified.

Computed Tomography of the Chest

Chest CT has high sensitivity, specificity, and negative predictive value in the setting of trauma and can provide significant information about the lungs, pleural cavities, and chest wall.⁶³ Chest CT is generally performed to evaluate adult victims of high-energy blunt trauma (especially those with chest pain, deformity, or hypoxia), with particular attention to the mediastinal contents (Figs. 15-14, 15-15, 15-16, 15-17, 15-18). Children presenting with hypotension, elevated respiratory rate, abnormal physical examination, depressed consciousness, and femur fractures after blunt trauma are at a substantially increased risk for an intrathoracic injury.⁶⁴

Chest CT is performed with intravenous contrast, preferably with the patient’s arms abducted above the head, from the thoracic inlet to the mid abdomen to include the entire thoracic cavity. Continuing the scan through the abdomen and pelvis can be easily performed. Images are acquired during the systemic arterial phase to maximize aortic enhancement and are reconstructed in the axial, sagittal and coronal planes at 2.0–4.0-mm slice thickness. Multiplanar reconstructions (MPR) are used to evaluate not only the thoracic aorta and mediastinum, but also the thoracic spine (see earlier discussion). Tailored MPRs, such as sagittal oblique (“candy cane” view of the aorta) may be helpful in visualization of an aortic injury. Noncontrast chest imaging has very little utility in the setting of trauma, unlike CT protocols for nontraumatic acute aortic pathology. In patients unable to receive contrast due to a known severe allergy, noncontrast CT can be effective in detecting a mediastinal hematoma and guiding the patient to evaluations such as transesophageal echocardiography, magnetic resonance angiography, etc. Contrast enhanced chest CT is a relative contraindication in patients with renal insufficiency; however, the risk of contrast-induced nephropathy or worsening renal function should be weighed against the risk of missing a major thoracic injury.

Chest CT is believed to be most cost-effective when patients are already undergoing another CT examination (eg, CT of the head, abdomen, and pelvis); are at risk for injury to the thoracic aorta because of high-energy mechanism, associated injuries, or age (>50 years old); or have a previously abnormal chest radiograph. A decision rule proposed by Blackmore et al⁶⁵ in which individuals with two or more of the following are at high risk for aortic injury: age greater than 50, unrestrained occupant in motor vehicle crash, hypotension, thoracic injury (rib fracture, pneumothorax, pulmonary contusion, or laceration), abdominopelvic injury (fracture of lumbar spine or pelvic ring, injury requiring laparotomy), fractures of appendicular skeleton, or injury to the brain.

The role of MDCT in the diagnosis of acute traumatic aortic injury has undergone significant evolution over the past 25 years, with sensitivity and negative predictive value of contrast-enhanced chest CT routinely in excess of 98% for modern scanners. In fact, patients who have direct signs of a thoracic aortic injury on MDCT no longer require catheter angiography confirmation, as was often the case in the past, and a normal thoracic aorta on chest CTA reliably excludes a traumatic aortic injury.⁶⁶ CT imaging features of a traumatic aortic injury can be divided into direct and indirect findings. Direct findings of traumatic aortic injuries include pseudoaneurysms, intimal flaps, pseudocoarctation (due to subadventitial dissection), and active bleeding. A mediastinal hematoma, however, is an indirect finding and can be present in the absence of an aortic injury. To suggest an aortic injury, a mediastinal hematoma should be contiguous with the aortic wall and should not be separated from the aorta by a rim of fat (Fig. 15-17D). Determination of whether or not a mediastinal hematoma has obliterated juxta-aortic fat can be difficult in thin patients or in patients with extensive soft tissue edema. Complex atheromatous disease can make interpretation of the examination difficult, particularly for more subtle injuries. An emerging form of an acute aortic injury, called the “minimal” or “minor” aortic injury, is an injury pattern thought to be secondary to advancing imaging technology (Fig. 15-19).^67,68

There are no universally accepted definitions for a “minimal” aortic injury, but those that exist in the literature include an intimal injury with thrombus, intimal flap less than 1 cm in length, and pseudoaneurysm less than 1 cm in diameter. There is often very little or no mediastinal hemorrhage with this type of injury.^67,68 Finally, aortic pulsation artifact (particularly involving the aortic root and ascending aorta) and beam hardening artifact due to dense contrast in adjacent venous structures may make interpretation difficult. Further evaluation with catheter angiography can be performed in equivocal or suboptimal examinations, but is of low diagnostic value in the setting of a high quality diagnostic chest CT angiogram.

Multiplanar and 3D reformations of the injured thoracic aorta are particularly helpful for planning treatment. It is important to delineate the anatomy of interest to the trauma surgeon, such as the distance from the most proximal point of injury to the takeoff of the left subclavian artery or any anomalous branches. These capabilities have significantly altered the role of catheter angiography from its traditional role of diagnosis, staging, and pretreatment planning, to one more often used for resolving diagnostic conundrums raised by CT or transesophageal echocardiography or as part of treatment using an endovascular stent graft.

CT is the most sensitive diagnostic method for detection of acute blood in the pericardium.⁶⁹ It is also among the most sensitive methods for detection of injuries to the chest wall, pleural cavities, or lungs.⁶³ Older CT technology exhibited relatively low sensitivity for the detection of injuries to the hemidiaphragm (sensitivity is 65–70%) but have vastly improved with helical MDCT technology, now with sensitivities of 71–90% and specificities of 98–100%.⁷⁰ For suspected diaphragmatic injuries, especially herniations through a diaphragmatic tear, coronal and sagittal multiplanar reformations are useful, as they better display characteristic findings (Fig. 15-20).

CT is helpful for diagnosing injuries to the tracheobronchial tree and has a sensitivity of 85% for tracheal rupture.²⁹ Nearly all patients with tracheobronchial injuries exhibit a massive pneumomediastinum or soft tissue emphysema.

Computed Tomography of Abdomen and Pelvis

Abdominopelvic CT is one of many adjunctive tests to assist the trauma surgeon in the evaluation of hemodynamically stable patients with suspected occult intra-abdominal injuries, or to aid in more definitive characterization of injuries previously detected by other diagnostic tests (eg, DPL or FAST; Figs. 15-20, 15-21, 15-22, 15-23, 15-24, 15-25). Usual indications include abdominal signs (eg, lap belt sign) or symptoms (eg, pain and tenderness) following high-energy blunt trauma. The combination of pleuritic chest pain at the left costal margin and left lower rib fractures is an independent predictor of splenic injury and warrants diagnostic evaluation.⁷¹ Abdominopelvic CT is extremely helpful in helping to guide management and has been shown to significantly decrease the incidence of a negative or nontherapeutic laparotomy. There has been a recent increase in the utilization of “triple contrast” abdominopelvic CT in the setting of penetrating trauma, in the absence of indications for immediate laparotomy (eg, peritonitis or hemodynamic instability).⁷²

It is estimated that clinical signs and symptoms of intra-abominal injuries may be misleading in 20–50% of patients, that physical examination has a 55–65% sensitivity for the detection of internal injuries, and that up to 19% of patients with blunt abdominal trauma have unsuspected injuries.^9,73,74 This, coupled with a negative predictive value of greater than 97% for abdominopelvic CT, has resulted in overall increased utilization of CT in the setting of blunt abdominal trauma. In addition, abdominopelvic CT is the principle means of both detection and characterization of renal injuries among adults with gross hematuria, children with microscopic hematuria (>50 red blood cell count per high-power field), or microscopic hematuria among adults who have had one or more episodes of systolic hypotension.

CT of the abdomen and pelvis acquired during the portal venous phase (60–75 second scan delay) has been shown to have a high sensitivity, specificity, and negative predictive value for the diagnosis, grading, and exclusion of solid organ injuries. The addition of an arterial phase aids not only in the detection of arterial injuries and active bleeding, but also for the detection of pseudoaneurysms in solid organs which may not be visible on portal venous phase images (Fig. 15-26).^75,76

Portal venous or delayed phase imaging can also assist in the detection of slower bleeding from the organs or pelvis. CT is accurate in the detection of urinary leaks from the upper genitourinary tract, and delayed phase (>5 minutes) imaging is recommended in the setting of renal lacerations to evaluate for urinary leaks. Isolated fluid around the ureters on initial scanning should raise the possibility of an injury, and 5 minute delayed scanning is indicated (Fig. 15-27).

Portal venous or delayed phase scanning through the pelvis is helpful for the detection of pelvic active bleeding. Images of all acquired phases should be reconstructed at 2–5-mm slice thickness in the axial, sagittal, and coronal planes.

The arterial-weighted phase images are particularly useful for the detection of brisk arterial bleeding and may detect an intraparenchymal pseudoaneurysm in a solid organ or an arteriovenous fistula, which may otherwise have gone undetected if portal venous phase imaging alone is acquired (Fig. 15-26). Recent literature suggests that a comprehensive evaluation to assess for the presence of a splenic injury requires both arterial and portal venous phases. The arterial phase performs better for the detection of splenic pseudoaneurysms, and portal venous phase has improved detection for parenchymal lacerations and active bleeding over arterial phase alone.⁷⁵ Extravasation of venous contrast is seen in 5–10% of victims of high-energy blunt trauma and may be difficult to differentiate from slow arterial bleeding on CT. The spleen is the most common region of isolated active extravasation of contrast; however, fractures of the pelvic ring are most commonly associated with multiple sites of extravasation.⁷⁷ The amount of hematoma associated with disruptions of the pelvic ring directly correlates with the likelihood of an angiographically demonstrable arterial injury (200 cm³, 5% arterial injury; more than 500 cm³, approximately 50% arterial injury).⁷⁸ Nonetheless, otherwise unexplained continued hemodynamic instability in patients with blunt pelvic fractures warrants angiographic evaluation, even in the absence of a pelvic hematoma or active pelvic bleeding on the initial CT (Fig. 15-28).⁷⁹

Active bleeding identified on a CT is an independent predictor for the failure of nonoperative management, particularly when bleeding is arising from the spleen, liver, mesentery, or pelvis. Active bleeding from the spleen or liver is a stronger indicator for the need for intervention than the organ injury grade alone (Fig. 15-21). The detection of lacerations that extend to the hepatic veins is of particular importance in the liver, as these have a strong predictive value for failure of nonoperative management when associated with large (>10 cm) hypoperfused regions.⁸⁰ Adrenal hemorrhage is relatively common, particularly on the right, and is not of clinical importance unless bilateral. Even then, posttraumatic hypoadrenalism is rare. It should be noted that CT grading scales for the solid organs do exist, and though there is significant overlap with the American Association for the Surgery of Trauma (AAST) solid organ grading scales, they are not identical. Thus, open communication with the interpreting radiologist is recommended to avoid any miscommunication.

CT cystography typically is performed after the administration of intravenous contrast, as a large amount of leaked extraperitoneal contrast may obscure active arterial extravasation in the pelvis. CT cystography can be easily included in the initial evaluation of the trauma patient after the traditional trauma CT survey and is highly accurate for the diagnosis, exclusion and comprehensive characterization of bladder injuries (Figs. 15-29 and 15-30).⁸¹ Passive physiologic filling of the bladder (such as a 10 minute delay with a clamped Foley catheter) may detect injuries to the bladder, but has been shown to be insufficient for the reliable exclusion of all injuries.^81,82

To perform CT cystography, the bladder is first emptied of unopacified urine to diminish the effects of contrast dilution. Scanning through the pelvis is then performed. Administration of dilute water soluble contrast via low gravity drip can proceed only after the positioning of the Foley catheter within the bladder is confirmed. It is not uncommon for a Foley catheter to be inserted through a laceration of the bladder dome in a patient with an intraperitoneal bladder rupture. The bladder should then be filled with at least 250–300 mL of contrast, or until the patient can no longer tolerate the degree of bladder distention, or until the contrast stops passively flowing into the bladder. It should be noted that contrast may continue to freely flow in patients with intraperitoneal bladder injuries and infusing over 350–400 mL without scanning is not advised. Scanning through the pelvis with maximum bladder distention is then performed. All scan series are reconstructed in axial, sagittal, and coronal planes at 1–3-mm slice thickness. Review of the images using “bone” windows may make the identification of the source of urinary leak more apparent, depending on the density of the contrast material in the bladder. Post-void scanning may identify small bladder leaks, but this is rare. Indications for CT cystography include hematuria and fracture of either the pelvic ring or acetabulum or hematuria and free intraperitoneal fluid in the absence of a clearly identifiable source. Cystography is low yield in the absence of hematuria and pelvic fractures.

Injuries to the bowel and mesentery are thought to occur in 1–5% of patients with blunt trauma, but, when present, are associated with increased morbidity and mortality primarily due to a delayed diagnosis.^83,84 The physical examination alone lacks sensitivity and specificity, particularly in patients with an unreliable physical examination.⁸⁴ The sensitivity and specificity of CT for bowel injuries are modest, ranging from 70–95% to 91–100%, respectively.⁸⁴ Specific CT findings of bowel injuries may include thickened bowel wall, asymmetric mural enhancement, pneumoperitoneum, leak of oral contrast, focal defect in the bowel wall, and free fluid not explained by other injuries (Fig. 15-31). Unfortunately, the specific signs of a bowel perforation are relatively uncommon and are not sensitive.^84,85 Active bleeding from the bowel mesentery, as with active bleeding elsewhere in the body, is an indication for intervention and is associated with a high likelihood of injury to the bowel or delayed necrosis of the bowel.⁸⁶ Fluid present in the mesentery or between loops of bowel is very suspicious for a transmural injury to the bowel, even in the presence of injury to a solid organ and even immediately following DPL (Fig. 15-31). One note of caution regarding free intraperitoneal fluid is that women of childbearing age have small amounts of fluid (±50 mL) in their pelvis. Recent observations also suggest that a small volume of isolated free fluid in the pelvis in male patients with blunt abdominal trauma, which can be seen in up to 4.9%, is unlikely to be due to a bowel injury.⁸⁷ Patients who have been vigorously resuscitated (especially if they are >24 hours from their injury) may have ascites and interloop fluid present due to a capillary leak syndrome. Acute and subacute hemorrhage typically measures 40–70 Hounsfield Units (HU), while urine, bowel contents, and ascites measure closer to water (eg, 0 +/–20 HU). If a patient with blunt trauma has free fluid in the peritoneal cavity of uncertain source and lacks signs of peritonitis on a physical examination, a repeat abdominopelvic CT in 4–6 hours may be warranted.⁸⁸

CT performs well in the diagnosis of peritoneal violation and intra-abdominal injuries in the setting of penetrating abdominal trauma.⁷² The diagnosis of injuries to the bowel following penetrating trauma is aided by the use of the “triple contrast” protocol utilizing intravenous, oral, and rectal contrast material. Leakage of enteric contrast is highly specific for bowel injury (Fig. 15-21F). A wound tract that extends to the bowel surface is the most sensitive sign of an associated bowel injury, in the absence of other definitive findings.

CT performs slightly better at detection of diaphragmatic ruptures than conventional x-rays, though the initial sensitivity of conventional CT was only up to 61% and a specificity of 76–99%. Overall performance improves with the use of helical CT, which has a sensitivity of 56–87% and a specificity of 75–100%. The addition of high quality multiplanar reformatted images with MDCT improves performance over helical CT alone, with a sensitivity of 71–90% and a specificity of 98–100%.⁷⁰ Direct CT signs of diaphragmatic rupture include focal diaphragmatic defect, the “dangling” diaphragm sign (from a free floating diaphragm flap), and herniation of abdominal contents through a diaphragmatic defect. A normal contour of the diaphragm and no pleural collections or adjacent airspace disease effectively excludes diaphragmatic injury.

Computed Tomography of the Pelvis and Acetabulum

CT is extremely valuable in the evaluation of the pelvic girdle, including the pelvic bones, acetabulum and sacrum, following both blunt and penetrating trauma. There are several indications for pelvic CT, including the following: (1) evaluation of unstable fractures of the pelvic ring as determined by physical examination or appearances on conventional x-rays; (2) to evaluate for radiographically occult fractures in the setting of high clinical suspicion; (3) to detect entrapped intra-articular debris/bone fragments following hip dislocation; and (4) for preoperative surgical planning (Figs. 15-33, 15-34, 15-35, 15-36, 15-37). Pelvic CT can be performed on its own, or be reconstructed from the raw CT data of the abdomen and pelvis performed as a part of a trauma CT “pan scan.” Three-dimensional volume rendered images can be generated from the CT data set, and dedicated evaluation of acetabular fractures can be performed with femoral head subtraction. In fact, some of the most recent CT scanner software packages can generate “virtual pelvis” images from the CT raw data that approximate that of conventional pelvic radiographs (Fig. 15-2D).

The primary goal of CT evaluation of disruption of the pelvic ring and acetabular fractures is to aid in surgical planning. CT scans of acetabular fractures are performed for the following: (1) assessment of fracture types;⁸⁹ (2) secondary congruence of the hip (eg, are the fracture fragments symmetrically oriented about the intact femoral head); (3) evidence for marginal impaction (eg, subarticular bone depressed or impacted, and not showing secondary congruence); (4) detection of a fracture of the femoral head; and (5) detecting intra-articular debris. There is approximately a 15% concurrent rate for fractures of the pelvic ring and acetabulum, so any fracture of the pelvis or acetabulum should initiate a search for other fractures in the pelvic girdle.

CT of the pelvis and acetabulum typically uses a slice thickness of 1–3 mm in the axial plane, and images are reconstructed using a bone algorithm and reviewed at bone and soft tissue windows. Sagittal and coronal reformatted images can be helpful in fully characterizing a fracture of the pelvic ring or acetabulum.

A systematic approach should be employed when reviewing pelvic CT images, with special attention made to the fractures that one would expect to see in the setting of AP compression, lateral compression, and vertical shear force. Correlation with at least an AP pelvic radiograph is recommended to provide an overview of the type of pelvic ring disruption. Fractures of the iliopubic and ischiopubic rami, as well as the sacral wings, are typical of a lateral compression type injury. The anterior surface of the sacrum is carefully evaluated for “buckle” fractures due to internal rotation of the hemipelvis due to a lateral impact. Fractures of the iliopubic and ischiopubic rami are assessed for their orientation (lateral compression fractures typically show orientation in the axial plane or coronal plane, while AP compression and vertical shear fractures will show orientation in the sagittal plane). Diastasis of the pubic symphysis and sacroiliac joints (eg, “open book” pelvis) are typical of an AP type injury. The normal pubic symphysis is never more than 1 cm in width in normal subjects, regardless of age. When the pubic symphysis is traumatically wider than 2.5 cm, disruptions of both sacrospinous and sacrotuberous ligaments are assumed. In the posterior ilium, it is important to look for avulsion fractures of the posterior superior iliac spine, so-called crescent fractures, as these are strongly associated with biomechanically unstable fractures in the presence of disruption of the anterior pelvic ring. The axial images give good evaluation of the amount of internal or external rotation, but underestimate the amount of flexion or extension of a hemipelvis relative to the intact pelvis. Furthermore, evaluation of the amount of pelvic hematoma may be helpful in determining the need for angiography for embolization. Localization of these hematomas is a good predictor of associated pelvic vascular injuries. Indeed, as above, a search for active bleeding in the pelvis should be sought if the patient undergoes the contrast-enhanced trauma CT series.

It should be noted that CT is not a dynamic examination and provides information for a single snapshot in time. As such, the assessment of potential biomechanical instability of the pelvis may be difficult. Certainly, the combination of a crescent fracture from the posterior superior iliac spine and displaced, anterior pelvic ring fractures will be unstable under anesthesia. The stability of other patterns, however, is not so predictable, even though CT images provide a great deal of information about the injury pattern. Therefore, conventional x-rays are necessary as guides to intraoperative reduction.

If an acetabular fracture is present, the goal is to determine what remains attached to the intact hemipelvis and to describe and characterize the major fracture fragments and their relation to each other and the femoral head. Assessment of the posterior through anterior walls, the ischium for the posterior column, the pubis and symphysis for evaluation of the anterior column, the iliac wing for superior extension, and the secondary congruence between the femoral head and tectum allow for a more complete recognition of fracture fragments. There are 10 types of acetabular fractures based on the Judet and Letournel classification scheme. Fortunately, approximately 90% fall into one of five fracture patterns, divided into two groups based on the presence (both-column and T-shaped) or absence of (transverse, transverse with posterior wall, and isolated posterior wall) involvement of the obturator ring.⁹⁰

Conventional Radiography

Conventional radiography remains the imaging standard for the evaluation of long bones showing obvious deformity, instability, palpable crepitus, pain, and swelling. For periarticular regions, conventional x-rays are indicated for deformity, instability, decreased range of motion, pain, and swelling. The Ottawa ankle and knee clinical prediction rules add considerable precision to specificity.^91,92,93

For long bones (Fig. 15-38), two orthogonal views are obtained, including an AP view and lateral projection centered at the midshaft. Projections should include both the proximal and distal joints. Joints should be imaged with two orthogonal views at a minimum, with additional optional one or two oblique views, centered at the midportion of the articulation (Figs. 15-39, 15-40, 15-41, 15-42, 15-43).

Analysis of the long bone should allow assessment of the direction of the force that created the fracture pattern (eg, twisting injuries result in spiral fractures; bending injuries result in wedge fractures). In general, higher-energy injuries tend to be more comminuted and displaced. If there is a mismatch between the apparent amount of comminution and the reported energy of the injury, osteoporosis or otherwise pathological bone should be suspected. The degree of displacement and angulation of the predominant fracture fragments, as well as fracture line involvement of an articular surface, should be noted.

Periarticular regions should be evaluated for fracture involvement as well as for partial or complete loss of congruity of the joint. The appearance of a superimposed white line due to overlapping of bones may indicate a dislocation and disruption of expected alignment of adjacent articulating structures.

Careful attention to soft tissues (eg, focal swelling, obliteration of normal fat pads, joint effusions) is helpful for subtle or otherwise occult fractures (eg, elbow, knee, and wrist). A search for foreign bodies in soft tissue in the setting of an open fracture or soft tissue injury should be performed.

Computed Tomography of Appendicular Joints

CT of appendicular joints, particularly the shoulder, supracondylar femur, tibial plateau, pilon, midfoot and calcaneus, is indicated for “displaced” intra-articular fractures (eg, 1–2 mm at the wrist or scapula, and glenoid, 5–10 mm at the tibial plateau) or unstable fracture patterns (Figs. 15-37, 15-38, 15-39, 15-40, 15-41). CT is a valuable surgical planning tool and is very helpful in the detection of otherwise occult fractures, particularly of the midfoot.

In most patients, CT should be performed after provisional placement of traction or reduction, if feasible. Because acquisition of isotropic image data of modern MDCT scanners allows for reconstruction of images in any conceivable plane, the joint in question no longer needs to be precisely positioned in the CT scanner gantry. Thus, “axial” images of a joint are not necessarily oriented in the axial plane of the body, but instead are oriented in the “axial” plane of the joint. Orthogonal images in both the sagittal and coronal planes are reconstructed from the joint specific axial plane at 1–3-mm slice thickness.

Use of traction prior to imaging allows ligamentotaxis to indirectly reduce fracture fragments and support indirect assessment of the integrity of soft tissue attachments to major bony fragments. Specifically, bone fragments that do not move or reduce on stretch are presumed to be no longer attached to soft tissue and may require debridement or direct repositioning. In addition, CT facilitates the assessment of intact bone and the integrity of subchondral bone (eg, need for bone grafting).

Peripheral Vascular Injuries

Vascular injuries in the extremities comprise between 40–75% of vascular injuries seen in civilian trauma centers, with a majority of these (~2/3) involving the femoral or popliteal arteries. Approximately three-fourth of peripheral vascular injuries are the result of penetrating trauma, with 70–80% secondary to projectiles such as gunshot wounds and shrapnel. Fatal exsanguination, multiorgan failure secondary to hemorrhagic shock, and limb loss are catastrophic consequences of peripheral vascular injuries, thus rapid identification and appropriate triage are of paramount importance. Vigorous or pulsatile external active hemorrhage, a rapidly expanding hematoma, or loss of distal pulses mandate emergent operative exploration.⁹⁴

The imaging strategy for patients with suspected peripheral vascular injuries depends on a variety of factors that are primarily related to clinical presentation, hospital course and to associated injuries, mechanism of injury, and signs of circulatory shock. Imaging modalities available include catheter angiography, CT angiography, Doppler ultrasound, and MR angiography (MRA).

Doppler ultrasound may be a useful adjunct but is not widely used, primarily because of its decreased sensitivity and accuracy for peripheral vascular injuries when compared to catheter angiography. This, coupled with the operator dependent nature of the examination and limited visualization of the entire vascular system for a variety of reasons (such as external fixator devices, overlying split or cast material, and associated soft tissue injuries and gas), make the utility of Doppler ultrasound for peripheral vascular injuries quite limited.

MR angiography is useful for the assessment of chronic peripheral vascular disease, but is not widely used in the acute setting for the following reasons: (1) may not be available at all centers at all times of the day; (2) it is technically challenging to effectively manage severely traumatized patients while they are in the MR magnet; (3) the presence of metallic foreign bodies may limit patient and operator safety; and (4) even if not ferromagnetic, foreign bodies may still cause extensive artifact obscuring the adjacent soft tissues. Thus, CT or catheter angiography are the primary imaging choices for patients with extremity injuries and suspected peripheral vascular injuries.

CT Angiography for Suspected Peripheral Vascular Injuries

CT angiography is a useful adjunct for suspected peripheral vascular injuries in patients without indications for immediate surgical exploration. As stated previously, CT is readily available, quick, convenient, noninvasive and reliable. In addition, CTA can be easily integrated into the imaging work-up of a traumatized patient, performed either as a stand-alone examination or integrated as part of a whole body CT for polytrauma. CTA has high sensitivity, specificity and accuracy when compared to catheter angiography (which remains the reference “gold” standard). In addition, CT capabilities allow for global evaluation of the adjacent structures and can thus be used to detect and characterize associated nonvascular injuries. The inability to immediately perform interventions on an injured vessel is a major limitation of CTA compared to catheter angiography.

The overall performance of CTA for peripheral vascular injuries compared to catheter angiography is certainly a concern for the trauma surgeon, as there is very little tolerance for missed or mischaracterized vascular injuries. A meta-analysis by Jens et al⁹⁵ pooling data from 11 studies from both the radiology and surgical literature and using a variety of scanner types, documented that peripheral CTA for extremity vascular injuries had a sensitivity of 96.2% and specificity of 99.2% compared to catheter angiography. The rate of nondiagnostic CTA examinations was 4.2% in this meta-analysis. Thus, in patients without indications for immediate surgical exploration, CTA is useful to exclude peripheral vascular injuries and to characterize injuries that are present.

The CTA signs of a vascular injury include pseudoaneurysm, active bleeding, occlusion, intimal injury, dissection, and arteriovenous fistula. Imaging findings that raise the likelihood of an injury include a perivascular hematoma, a vessel within a projectile wound tract, and projectile fragments within 5 mm of a vessel. The latter two findings should prompt further investigation with catheter angiography.

Extremity CTA should be performed using MDCT with thin section acquisition, peripheral IV access or a power injectable central line, a high contrast injection rate (4–5 mL/sec) and be acquired during the systemic arterial phase. Images are reconstructed in the true axial plane to the long axis of the extremity at 1–3-mm slice thickness, with additional sagittal and coronal images reconstructed with respect to the axial plane. Coronal maximum intensity projections (MIP, or thick slab images) at 5–10-mm slice thickness are extremely valuable in displaying vessels along their long axis using a relatively small number of slices. Thin images should be reviewed, as well, as small intimal vascular injuries may not be apparent on MIP images due to volume averaging. Liberal use of a 3D volume viewing station is encouraged, as some vascular injuries may be best displayed on nontraditional planes.

It should be noted that vascular injuries near highly mobile joints, such as the knee, are poorly treated with endovascular techniques. Thus, the decision to image with CT angiography over catheter angiography must be weighed against patient stability and the need for urgent surgical vascular repair. Despite this, the wide availability of stent grafts has also increased the utilization of angiography as a prelude to nonoperative management of some clinically significant vascular injuries.

Catheter Angiography

Catheter angiography remains the reference “Gold Standard” for the definitive evaluation of arterial blood vessels for injury, as well as identifying active arterial hemorrhage, false aneurysms, and arteriovenous fistulas. While CT angiography is frequently favored for the initial imaging of many traumatic vascular lesions in stable patients, it is limited by the inability to immediately intervene in the event of a positive study. CT angiography performance is very high, but not yet as accurate as catheter angiography. Catheter angiography is utilized as a problem solving tool for equivocal CT examinations and for definitive therapy of some acute vascular conditions.

There are many advantages of catheter angiography. It allows simultaneous detection and treatment of a wide variety of traumatic vascular injuries. It is a very specific method of identifying bleeding at the sub-millimeter diameter of vessel. It can evaluate many sites of bleeding simultaneously. It has an excellent safety record, especially when using iso-osmolar nonionic contrast agents, coaxial micropuncture access, digital subtraction techniques, coaxial microcatheters, and steerable guidewires.

The disadvantages of catheter angiography are cost, the delay necessary to assemble the interventional radiology team (comprised of radiologists, technologists, and nurses), the lack of suitability as a screening test for most traumatic conditions, and the risks of radiation exposure. Technical expertise is limited to predominantly subspecialty-trained interventional radiologists, although endovascular surgeons and cardiologists may develop this expertise on an individual basis. These disadvantages are magnified when the likelihood of injury is low. Thus, noninvasive vascular techniques such as CT angiography should be explored under controlled studies to further assess their accuracy and appropriateness in such situations.

Transcatheter Endovascular Therapies

Endovascular techniques have become a broadly accepted way of controlling traumatic hemorrhage for a variety of reasons. Catheter-based hemostasis allows precise control from a remote site that avoids exacerbation of venous hemorrhage, introduction of pathogens, and hypothermia that may result from open exposure. It is especially valuable for hemorrhage that is remote or hidden from view and requires laborious time-consuming exposures or that is the result of multiple small bleeding sites that are not easily detected or controlled during operative exploration.

Endovascular techniques include embolization, stenting, stent grafting, and temporary balloon occlusion and may be definitive in nature or an adjunct to operative exposure. The methods of embolization include particulate or microcoil embolization of small vessels, proximal and distal control of a bleeding vessel, and coil occlusion to cause selective temporary hypotension of the bleeding zone.

Stenting, which facilitates blood flow beyond an injury, has largely been replaced by covered stent grafts that exclude lacerations, transections, and arteriovenous fistulae while maintaining flow through the conduit. Endografts are made of a variety of porous materials such as expanded polytetrafluoroethylene and are reinforced by a metallic skeleton that apposes the stent graft to the native artery. Reports of midterm patency, while limited at this time, are beginning to show that these are durable options to vascular repairs.

Contraindications to endovascular techniques are highly dependent on skills, teamwork, and hemodynamics; however, there are some injuries that are difficult for rapid surgical control and endovascular techniques have a role, even in the unstable patient.

Arch Angiography for Acute Blunt-Force Traumatic Aortic Injury

Screening and diagnostic arch angiography for blunt traumatic aortic injuries has largely been replaced by chest CT angiography, particularly with the widespread availability of MDCT technology.⁶⁶ Arch angiography, however, does have a role for the evaluation of equivocal findings on chest CTA and for potential endovascular stent graft therapy. This is particularly true if there is a suspected injury to the ascending aorta or aortic root seen on CT. If patients are going directly to angiography for evaluation of disruptions of the pelvic ring and the mediastinum is not normal on a chest x-ray, catheter arch angiography is the preferred “screening” modality; otherwise, CT is the preferred modality. Modern CT angiographic techniques are quite exquisite in demonstrating aortic injuries as well as providing coronal and sagittal reformations that can illustrate the important relationships and variants necessary for surgeons to create a treatment plan. Among selected patients sustaining aortic injury who are not operative candidates, endovascular stent grafts have been advocated as either temporizing or definitive therapy.

Typically, a 5 French pigtail catheter is guided to the ascending aorta via a femoral arterial approach. Patients are positioned and imaged in both 35° right anterior oblique (RAO) and left anterior oblique (LAO) projections, using injection rates of approximately 25–30 mL/s for a 40–60 mL volume (depending on hemodynamic status) and positioning to include the great vessels and diaphragm.

The arteriographic appearance is classical. Linear filling defects indicate torn and ruffled intimal lining, with expansion of the lumen (typically at the ligamentum arteriosum) indicating the presence of a pseudoaneurysm (Fig. 15-44). The tear of the aortic wall may be segmental or circumferential, and is sometimes associated with distal narrowing of the contrast column (pseudocoarctation). Minimal aortic injuries, as discussed earlier in this chapter, may be very difficult to visualize angiographically, and a “negative” aortogram following a positive chest CTA for a minimal aortic injury will not eliminate the diagnosis.

Associated mediastinal vascular injuries should also be identified. Injuries of the arch arteries may occur instead of, or in association with, an aortic injury. Bleeding from the internal mammary or the intercostal arteries may be easily overlooked without diligence.

Hepatic Angiography for Blunt-Force Lacerations

Visceral catheter angiography is appropriate to evaluate hepatic lacerations (Fig. 15-45), particularly in patients with a labile hemodynamic status, or those with active extravasation or vascular abnormalities as seen on a contrast-enhanced CT. Visceral catheter angiography may have a role after a “damage control” operation. Gross hemodynamic instability and profound shock, however, usually mandates urgent celiotomy.

On CT, hepatic fracture lines that traverse the hepatic triad result in bleeding more often than those that run parallel to the triad. Contrast extravasation on CT tends to be associated with a positive arteriogram, but the decision to use angiography should primarily be based on clinical status rather than the CT appearance. Lack of enhancement of liver segments on CT is a very important finding, as these regions represent an intraparenchymal hematoma, occlusion of the portal triad, or injury of the hepatic outflow from that segment. It is vital to distinguish nonenhancement from a hematoma, which can be challenging. A large hepatic hematoma pushes the hepatic fragments away from each other and lacks hepatic vessels. If the area of nonenhancement has vessels running through it, it suggests an occlusion of the portal vein and hepatic artery or injury to a hepatic vein. Therefore, CT nonenhancement of the liver is an indication for angiography to confirm such injuries and to control arterial hemorrhage. As surgical exploration of damaged hepatic veins may be quite difficult, hepatic embolization and observation of a nonbleeding hepatic venous injury can be lifesaving. And, as noted earlier, hepatic angiography has an important role in the management of penetrating liver injuries that are isolated to the liver.

Selective catheterization of both the celiac trunk and the superior mesenteric artery (SMA) is essential due to the high rate of hepatic vascular variants, particularly the aberrant replaced right hepatic artery from the SMA. Imaging should be continued through the late portal venous phase to evaluate for slowly bleeding vessels. There should be careful inspection for abnormal parenchymal enhancement and intrahepatic portal venous filling, as these are features of hepatic arterial-portal venous fistulae, which can easily be overlooked.⁹⁶

Critical findings include arterial extravasation, spasm and occlusion, or shunting and fistula to portal or hepatic venous structures. Embolization of discretely abnormal vessels can be performed using a number of methods. A diffusely abnormal parenchymal injury with arterial bleeding may be safely embolized with Gelfoam due to the dual blood supply of the liver (hepatic arterial and portal venous). Embolization of hepatic arteries in the absence of portal flow increases the risk of developing an infarction or abscess. Depending on the location of bleeding and on the difficulty with catheterization, particulate embolization is the fastest technique; however, single microcoil embolization is preferred if time and circumstances allow. While formation of a postprocedure abscess is a complication of embolization, outcomes are favorable by integrating percutaneous image-guided drainage into the scheme.

Splenic Arteriography for Blunt-Force Lacerations

A patient who is hemodynamically unstable is not a candidate for angiography and embolization. Patients with splenic injuries diagnosed on CT are candidates for nonoperative management with overall good salvage rates, particularly when combined with splenic angiography and embolization.^76,97 When CT demonstrates active arterial extravasation or a parenchymal vascular abnormality (such as an intra-parenchymal pseudoaneurysm), one should consider angiography, as these are independent predictors of failure of nonoperative management. It is thought that there was poor correlation between the CT grading system and outcome of treatment, as many Grade IV injuries can be observed, and some Grade I injuries become worse, rebleed, and require definitive procedural therapy; however, there is emerging evidence that a CT-based grading schemes, in combination with clinical parameters (such as AIS), may be helpful for patient triage.^98,99 The treatment algorithms for patients with splenic injuries in the absence of active bleeding or other vascular lesions varies by institution. Some centers observe (ie, bed rest) most Grade I injuries, some advocate liberal use of angiography for triage of most other CT-diagnosed splenic injuries (such as those Grade III or higher), some centers advocate angiography in patients with a significant hemoperitoneum or transient hypotension. But in general, patients with high-grade injuries on CT should be imaged by angiography early to avoid transfusion or delayed rupture. The absence of arteriographic extravasation is a highly reliable predictor of successful nonoperative therapy regardless of injury grade. Identification of active arterial extravasation is the standard indication for endovascular treatment.

Diagnostic angiography of the celiac trunk is followed by selective splenic artery catheterization with a 5 French catheter. If splenic artery anatomy permits, and a solitary pseudoaneurysm or focus of extravasation is seen, distal coil embolization at the site of injury can be attempted. This is especially true in a patient in whom the extravasation extends beyond the splenic capsule into the peritoneal cavity. One should note that distal superselective embolization is associated with the development of more postprocedure splenic infarctions and abscess, though these complications are uncommon. Finally, most patients have tortuous splenic arteries and most extravasations are multiple.

Diffuse intrasplenic extravasation is far more common, and superselective occlusion of these multiple sites would be very time-consuming and less effective. Also, the splenic tortuosity that results from medial displacement of the spleen by the perisplenic hematoma often prevents rapid catheterization (Fig. 15-46). In such cases, embolization of the proximal splenic artery by coils or other occlusion devices placed distal to the dorsal pancreatic branch and proximal to the pancreatic magna branches is advocated to reduce the arterial pressure head at the injury site while allowing perfusion through collateral vessels. Such collaterals prevent splenic infarction by maintaining splenic perfusion through connections between the left gastric and the short gastric arteries, between the dorsal pancreatic artery and the pancreatica magna branches, between the right and left gastroepiploic vessels, and others.

Complications are uncommon when proximal splenic artery embolization is performed. A poorly selected coil size may result in hilar occlusion if the selected coil is too small and migrates distally. Too large coils may migrate proximally to occlude the celiac axis or embolize into the aorta. As noted earlier, distal microembolization bypasses the collateral circulation and results in more loss of immune function. Occlusion of the pancreatic branches may result in pancreatic necrosis and pancreatitis, though this complication is rare and can be avoided by careful review of the angiographic images if main splenic artery embolization if considered.

Interventions for Renal Trauma

Many renal injuries are usually well-tolerated and do not require angiography, especially when caused by blunt trauma. Initial nonoperative management of blunt renal injuries with an intact pedicle is the current accepted management standard. High-grade injuries that result in massive hemorrhage are usually managed by nephrectomy. Angiography with intent to embolize bleeding is appropriate for hemodynamically stable patients. Angiography is recommended for patients with CT evidence of a major renal injury and ongoing blood loss or persistent gross hematuria. Peripheral wedge-shaped regions of nonenhancement on CT suggest a segmental or distal renal artery injury, often due to avulsion injury or intimal stretch resulting in distal platelet embolization. Penetrating renal injuries are more aggressively approached by angiography if nonoperative management is undertaken. Large perinephric hematomas, areas of nonenhancement, and active bleeding on CT warrant angiography following penetrating trauma.

Aortography is helpful to assess injury of the origin of the renal artery, to exclude renal parenchymal injury perfused by accessory renal arteries, and to look for associated bleeding sites. A selective renal arteriogram using a 5 French catheter is then performed. Most injuries will require use of a coaxial microcatheter and embolization of small branches. Coils are preferred as they can be carefully placed to prevent infarction of adjacent noninjured renal tissue, but surgical gelatin pledgets can be used, as well. Because renal branches are end vessels with little collateralization, infarction is likely, and the goal is to reduce these infarctions to a minimum.

The treatment of vascular injury in the renal pedicle continues to be a vexing problem, especially since delays in revascularization usually result in a renal infarction or renovascular hypertension. Partial wall injuries that result in a pseudoaneurysm or segmental infarction often went unrecognized prior to the widespread use of CT. Such injuries are routinely detected before complete arterial thrombosis and renal infarction occurs. Therefore, arteriography is indicated when an injury in the renal artery is suspected. When such injuries are detected, treatment options are many, including operative revascularization, antiplatelet therapy and observation, and the application of covered stent grafts. Stent grafts can effectively seal full thickness injuries and cover exposed media that results in embolic infarctions. While long-term follow-up of series of these patients is lacking, the midterm (1–5 years) patency of stent grafts throughout the body remains high (Fig. 15-47).

Pelvic Hemorrhage

Pelvic fractures are potentially life-threatening injuries that are caused by high-energy impact trauma and account for about 3% of skeletal injuries. They are the third most common lethal injury following motor vehicle crashes. The majority of patients with pelvic fractures do not require massive transfusion as bleeding in most cases is likely to be venous or osseous in nature and often self-limited. Radiological intervention is not commonly needed in patients with routine pelvic fractures. Severe hemorrhage, however, occurs in 3–10% of patients, and mortality rates may be as high as 50% in patients with unstable pelvic fractures. Thus, the use of angiography in patients with pelvic fractures is highly dependent on the hemodynamic status, the type of pelvic fracture pattern, the transfusion requirements, and the presence or absence of hemoperitoneum.

Blunt Pelvic Fractures

Blunt pelvic fractures with crushing or shearing tear the small branches of the internal iliac artery that accompany ligaments, muscles, and tendons. Injuries tend to be multiple and bilateral, and from several branches. In addition, bony fragments can penetrate or perforate vascular structures. Examples include a fracture of the superior pubic ramus injuring the internal pudendal or obturator artery, a fracture of the iliac wing through the sciatic notch injuring the superior gluteal artery, and disruption of the sacroiliac joints injuring the lateral sacral arteries.

Most of the indications for angiography in blunt pelvic trauma have remained the same for more than 30 years, and include the following:

1. Hemodynamic instability in a patient with a pelvic fracture with no or little hemoperitoneum detected by FAST or diagnostic peritoneal lavage

2. Pelvic fracture and transfusion requirement of greater than 4 U in 24 hours or 6 U in 48 hours

3. Pelvic fracture and a large or expanding hematoma identified during celiotomy

4. CT evidence of large retroperitoneal hematoma with extravasation of contrast

5. Need for detection and treatment of other injuries during angiography

The presence of contrast extravasation on MDCT has been used as an indication for follow-up pelvic angiography. Although it should not delay angiography that is already indicated for pelvic hemorrhage, CT is helpful in localizing the vessels likely to be bleeding not only in the pelvis, but also from the solid organs and thoracic cavity. Correlations of location of the hematoma and site of vascular injury include obturator space and obturator artery, presacral space and lateral sacral artery, space of Retzius and internal pudendal artery, and buttock and gluteal artery.

Femoral access is the preferred approach; however, catheterization may be difficult because of hypotension, tachycardia, and difficulty in palpating the vessels as the pelvic hematoma expands. Ultrasound or fluoroscopic guidance is very helpful in these situations. A 5 French aortic flush catheter is used for flush abdominopelvic aortography. This is valuable to screen the abdominal viscera and mesentery, to exclude aortoiliac and other retroperitoneal bleeding sources, and as a road map of the pelvic vessels. Selective bilateral internal iliac arteriography is mandatory to exclude bleeding sites since aortography may not identify all bleeding. From one access, both internal iliac arteries are sequentially catheterized and opacified. Then, external iliac arteriography is used to evaluate the external pudendal and external obturator vessels.

It is common to identify multiple areas of extravasation during pelvic angiography. These may be bilateral and may involve multiple vascular beds. Extravasation is often punctate, but can be large and coarse, and the size of such extravasations may not correlate with the degree of blood loss. Vascular occlusions due to transection or dissection with subsequent occlusion can be present, as well. It can be difficult to differentiate between arterial injury due to thrombosis and vasospasm. Failure to treat these occlusions may result in recurrent hemorrhage when vasospasm resolves. Arteriovenous fistulas can occur, but are more common in penetrating trauma.

Because bleeding is usually multifocal and originates from multiple small blood vessels, embolization requires small particulate embolization. Large coil occlusion is as ineffective as surgical ligation of the internal iliac artery because bleeding soon resumes through numerous collateral circuits. Surgical gelatin pledgets are ideal because they are inexpensive, readily available, and often temporary lasting only a few weeks and allowing reestablishment of normal blood flow after the tissue has healed (Fig. 15-48A and B). Permanent particulate emboli, however, are often used because of their ease of use through a microcatheter (Fig. 15-49A and B). Embolization is technically successful in more than 90% of patients, and hemorrhage control is highly effective. Survival depends on many other factors including associated injuries, the presence of an open fracture, transfusion requirements, and delays to embolization.

Penetrating Pelvic Trauma

Penetrating trauma is an uncommon indication for pelvic angiography, as many patients are hemodynamically unstable or have clear indications for immediate exploratory celiotomy, often due to injury to a large vessel. Because the extraperitoneal space has been exposed by a penetrating wound, intraperitoneal bleeding is likely and direct exploration is warranted. Occasionally, angiography is valuable when operative control cannot be initially accomplished and damage control has been performed. Angiography and embolization prior to unpacking can aid in decreasing blood loss at a reoperation.

Injuries to large vessels require a very different endovascular strategy than small vessels or end organ vessels. When an injury to a noncritical internal iliac artery or branch has been missed at operation, but detected on postoperative angiography, coil occlusion of both the proximal and distal end of the vessel (whenever possible) is the standard treatment.

Catheter Angiography for Peripheral Vascular Injuries

Almost all peripheral vascular injuries can be reached using a 5 French catheter from femoral access provided a long enough catheter is available. Angiography should be done in multiple projections with opaque marking of the entry and exit wounds demonstrating that the entire course of the wounding agent is within the field of view. Iso-osmolar nonionic contrast medium is the optimal agent for visualization. Multiple images in the arterial, capillary, and venous phases are necessary.

Patients with proximity wounds and stable hematomas, diminished pulses, or signs of an arteriovenous fistula may benefit from either CT angiography or catheter angiography. While debatable to some, “proximity” angiography has value in asymptomatic patients with penetration that has passed close to the estimated path of major vessels. Vascular injuries occur in 3–8% of asymptomatic patients. Failure to diagnose arterial injuries may result in delayed hemorrhage or chronic arteriovenous fistulas with claudication, venous insufficiency, and congestive heart failure. Exclusion angiography avoids the time and effort needed to keep track of patients who may be lost to follow-up.

Vascular injuries resulting from fractures and dislocations are uncommon. Clinical evaluation is often difficult as the hematoma from a fracture may be quite large and indistinguishable from one associated with a vascular injury. Crush wounds, angulation deformities, and fracture hematomas may cause a pulse deficit by kinking, entrapping the vessel, or inducing spasm without an intrinsic vascular injury. A laceration into muscle may result in external blood loss without major vascular injury. Finally, a compartment syndrome may result in tissue ischemia without loss of pulses.

The natural history of many injuries cannot be predicted by the angiographic appearance alone. Therefore, observation of some injuries is warranted. Equivocal findings such as luminal narrowing can be assessed by repeating angiography after infusion of an intra-arterial vasodilator, on a subsequent day. Small irregularities and intimal tears that do not restrict flow may be treated by antiplatelet therapy and will heal (Fig. 15-50A–D).

Treatment of angiographically diagnosed vascular injuries is based on the criticality of the bleeding vessel, its size, location, and accessibility, the hemodynamic condition of the patient, and the type of lesion. Small vessels that are not essential for tissue perfusion can be treated by small particle embolization, using surgical gelatin pledgets or more permanent smaller agents. Permanent agents have no advantage, but in some instances are more easily administered through microcatheters than surgical gelatin. These agents are delivered by flow direction toward the path of least resistance, which is usually toward the bleeding site. Microcoils can be utilized for injury to a small vessel provided they can be delivered near enough to the injury site to avoid collateral recruitment that permits continued bleeding. Examples of vessels that can be treated by embolization of small particles include hemorrhage from a pelvic fracture, multifocal hepatic arterial hemorrhage, and injuries to muscular branches such as those of the profunda femoris artery in the lower extremity.

Injury to larger vessels such as those greater than 3 mm in diameter requires two techniques, one for essential vessels and one for expendable vessels. The treatment of essential vessels requires repair of the bleeding site while allowing continued blood flow. Thus, stent grafts can be deployed to cover the injured segment while allowing antegrade flow (Fig. 15-51).

Nonessential conduits, such as branches of the profunda femoris artery or the brachial artery, or one of the arteries in the shank, can be safely embolized. Particulate embolization will flow past the injury and penetrate deep into the vascular bed. When conduits are injured, this insult to the vascular bed is unnecessary. Therefore, large vessel agents are used to occlude the damaged segment of the conduit while the vascular bed is perfused through collaterals (Fig. 15-52).

Coils in various sizes, some containing threads or fibers to accelerate thrombosis, are the most common devices used to occlude a large vessel. A coil is sized to have a diameter large enough to prevent distal migration, but not too large to end up recoiling into a parent, nontarget vessel.

The technique of arterial isolation attempts to occlude both the proximal and distal vessels around the area of injury by coiling (Fig. 15-52). The goal is to exclude the vascular defect and prevent rebleeding through collateral vessels. This is highly desirable in most circumstances, but mandatory when treating arteriovenous fistulas. The guidewire is carefully maneuvered distal to the injured segment, but proximal to any branches, and coils are delivered. The catheter is then withdrawn, and coils are placed in the proximal segment.