Imaging of blunt cerebrovascular injuries

Transcription

1 European Journal of Radiology 64 (2007) 3 14 Imaging of blunt cerebrovascular injuries Clint W. Sliker, Stuart E. Mirvis 1 Department of Diagnostic Radiology, University of Maryland Medical Center, 22 South Greene Street, Baltimore, MD 21043, USA Received 5 February 2007; accepted 14 February 2007 Abstract Blunt cerebrovascular injuries are uncommon but potentially devastating injuries that can lead to stroke and death. While uncommon, appreciation of the seriousness of these injuries, a high index of suspicion in high risk patients, and aggressive screening of multitrauma patients leads to early diagnosis of asymptomatic lesions that may be amenable to treatment prior to the onset of ischemia. The radiologist can play a vital role in the early diagnosis, follow-up, and, in some cases, treatment of these challenging injuries Elsevier Ireland Ltd. All rights reserved. Keywords: Blunt trauma; Carotid artery injury; Vertebral artery injury; Diagnosis 1. Introduction Historically, blunt cerebrovascular injuries (BCVI) were considered uncommon in the general blunt trauma population. Although many of these lesions lead to stroke and death, diagnosis prior to onset of symptoms can be challenging. With increased appreciation of these injuries, increasing therapeutic options, and increasing access to ever advancing diagnostic imaging modalities, this group of injuries has received greater attention in the trauma literature and from those caring for acutely traumatized patients. The purpose of this article is to provide the radiologist with a review of blunt cerebrovascular injuries with special attention to diagnosis and those clinical factors that impact diagnosis and the radiologist s role in patient management. 2. Incidence Blunt cerebrovascular injuries (BCVI) are generally considered uncommon injuries. Among all hospital admissions for blunt trauma, blunt carotid artery injuries (BCI) are encountered in % of patients [1 3] and blunt vertebral artery injuries (BVI) are encountered in % [4,5]. Although the overall incidence of BCVI ranges from 1.1 to 1.4% [6,7] of all Corresponding author. Tel.: ; fax: addresses: csliker@umm.edu (C.W. Sliker), smirvis@umm.edu (S.E. Mirvis). 1 Tel.: ; fax: blunt trauma patients, it is more frequently diagnosed in more severely injured patients. For example, Mutze et al. [8] reported an incidence as high as 2.7% in blunt patients with an injury severity score equal to or greater than Outcomes BCVIs result in high mortality and morbidity. For BCI, the overall mortality and morbidity rates are and 32 67% [1 3,5,7,9], respectively. Overall mortality and morbidity rates for BVI are generally lower, although morbidity still ranges from 14 to 24% [5,10] and mortality ranges from 8 to 18% [4,9]. The high morbidity and mortality are, in part, attributable to the severity of concurrent injuries frequently associated with BCVI, although BCVI-specific mortality itself is high. In one retrospective review of 22 patients with BCVI, McKevitt et al. [11] reported a stroke rate of 60% and BCVI-specific mortality of 25%. In a prospective study conducted by Cothren et al. [9], BCI-specific mortality was 21%, while BVI-related mortality was 18%. Moreover, reflecting a stroke s globally deleterious impact on a patient, Cothren reported higher overall mortality in those with BCVI-related stroke (32% for BCI and 18% for BVI) compared to those without stroke (7% for both BCI and BVI). 4. Mechanism of injury Hyperextension and contralateral rotation of the head and neck, with the cervical internal carotid stretched against the transverse processes of the upper three cervical vertebrae, is X/$ see front matter 2007 Elsevier Ireland Ltd. All rights reserved. doi: /j.ejrad

2 4 C.W. Sliker, S.E. Mirvis / European Journal of Radiology 64 (2007) 3 14 Fig. 1. A 82-year old woman with atlanto-occipital dissociation and bilateral vertebral artery injuries following motor vehicle collision. 3D volume rendered (a) and axial images (b) obtained with whole-body multidetector computed tomography (WB-MDCT) protocol. Magnetic resonance angiogram (MRA) (c) and T1 weighted spin echo magnetic resonance image (MRI) (d). WB-MDCT demonstrates right vertebral artery segmental occlusion from C1 to foramen magnum (curved arrow), focal left vertebral artery central narrowing from intramural hematoma at level of foramen magnum, and bare occipital condyles (asterisks). MRA demonstrates absent flow in distal right vertebral artery with focal narrowing of injured left vertebral artery (arrow). Left vertebral artery high T1 signal (arrowhead) indicates intramural hematoma from dissection.

3 C.W. Sliker, S.E. Mirvis / European Journal of Radiology 64 (2007) considered the most common cause of BCI [12]. Also, extension of basilar skull fractures through the carotid canal may cause petrous carotid injury [12]. Less common causes of BCI include a direct blow to the neck (e.g. assault, near hanging, etc.) [12] and intraoral trauma from either a hard foreign object (e.g. falling against a toothbrush) [12] or posteriorly displaced mandible fracture fragments [13]. Since most of the vertebral arteries are fixated within the transverse foramina, they are vulnerable to injuries induced by displaced bone fragments or stretching related to rotation and subluxation [14,15]. In cases of craniocervical junction distraction or dislocation, the vertebral arteries may be crushed against C1 [14] or stretched over the dura [16] (Fig. 1). 5. Clinical presentation and associated injuries The diagnosis of BCVI is typically suspected in patients presenting with overt signs of fixed neurologic deficits that are unexplained by cerebral computed tomography (CT). Other signs, such as transient ischemic attacks, neck hematomas, Horner s syndrome, or cervical carotid bruit, also suggest BCVI, although they can be difficult to identify in the acute multitrauma patient [17], especially those with concomitant traumatic brain injury. Several studies suggest prompt treatment reduces BCVIrelated stroke and mortality rates [4,5,7,18]. To this end, several retrospective studies have described numerous clinical and imaging findings that may serve as triggers for screening of blunt trauma patients for BCVI [5,11,12,17,19 22]. Table 1 lists the authors institutional guidelines for predicting BCVI. Despite a lack of consensus regarding the most significant independent predictors for BCVI, it has been shown that the greater the number of risk factors, the greater the likelihood of an associated vascular injury [12]. Table 1 Authors institutional guidelines for identifying trauma patients at high risk for blunt cerebrovascular injury 1. Cervical spine fracture with A. Foramen transversarium involvement, or B. 30% or more subluxation, or C. Significant rotation or distraction mechanism (e.g. unilateral interfacetal dislocation, atlanto-occipital distraction, etc.) 2. Basilar skull fracture crossing carotid canal, foramen lacerum, or cavernous sinus 3. Severe facial fractures (LeFort II or III; naso-orbital ethmoid complex, facial smash) 4. Carotid or vertebral artery perivascular hematoma by CT or MD-CTA 5. Horner s syndrome 6. Glasgow Coma Score 6 at 24 h after initial assessment 7. Neurologic examination incongruent with brain imaging 8. Stroke or transient ischemic attack 9. Hanging attempt with cervical hematomas or cervical spine fractures CT, Computed tomography; MD-CTA, multidetector computed tomographic angiography. References: Lee et al. [16], Miller et al. [19], Kerwin et al. [20], Cothren et al. [21], Mirvis et al. [23] and Parbhoo et al. [24]. Many proposed predictors of BCVI can be readily identified during routine imaging of acutely injured patients [11,17,22]. In a retrospective review of 75 patients with BCVI, Biffl et al. [22] identified diffuse axonal injury (DAI) (Fig. 2), LeFort II or III fractures (Fig. 3), and petrous temporal bone fractures with carotid canal extension (Fig. 4) as independent risk factors for BCI. Retrospective studies conducted by Berne et al. [25] and Utter et al. [17] support the association of skull base fractures with BCVIs. BVI and cervical spine injury have been strongly associated in several studies [4,7,10,21,17,22,24,26]. In one study [4], 38 patients with BVI were 11 times more likely than not to have a cervical spine injury. In another study [21], 77% of patients with BVIs demonstrated cervical spine injuries, with most exhibiting either fracture involving a transverse foramen Fig. 2. A 40-year old with persistent low Glasgow Coma Score following motor vehicle collision. Diffusion weighted brain MRI (a) demonstrates corpus callosum restricted diffusion related to diffuse axonal injury (curved arrows). (b) MRA demonstrates distal cervical segment internal carotid artery (ICA) pseudoaneurysm (arrowheads).

4 6 C.W. Sliker, S.E. Mirvis / European Journal of Radiology 64 (2007) 3 14 Fig. 3. A 44-year old male imaged with WB-MDCT following motor cycle crash. 3D volume rendered image (a) of face demonstrates extensive facial fractures that include LeFort II and III patterns (arrowheads). Sagittal multiplanar reconstruction (MPR) (b) demonstrates characteristic short-tapered post-traumatic occlusion of ICA just distal to carotid bulb (arrow). (Fig. 5), subluxation or dislocation, or injury level at C1 to C3. Parbhoo et al. [24] report a high frequency of BVIs among those with unilateral facet dislocations, which is an association supported by the authors experience. Some significant clinical risk factors can be inferred by findings on routine imaging. For example, a very low ( 8) Glasgow Coma Score (GCS) [11,22] has been associated with BCVI. Severe traumatic brain injury revealed by cerebral imaging may suggest a low GCS, and then allow the radiologist to propose evaluation for BCVI. Similarly, a high ( 3) thoracic abbreviated injury score (AIS) [11] has been linked to BCVI (Fig. 6). Since imaging plays the primary role in determining Fig. 4. A 23-year old with skull base fractures following motor vehicle collision. Computed tomography (CT) (a) demonstrates displacement of fragments into carotid canal (arrowheads). (b) Angiogram demonstrates petrous ICA (curved arrow) and cavernous ICA (arrow) pseudoaneurysms.

5 C.W. Sliker, S.E. Mirvis / European Journal of Radiology 64 (2007) Fig. 5. A 22-year old male with vertebral artery injury induced brainstem stroke following motor vehicle collision. Admission cervical spine CT (a) demonstrates C6 left transverse process fractures involving foramen transversarium (black arrow) (b) and angiogram demonstrates associated abrupt occlusion of the proximal left vertebral artery (white arrow). Diffusion weighted (c) brain MRI performed 10 days later demonstrates rostral brain stem infarct (arrowhead). the thoracic AIS, if the radiologist diagnoses severe thoracic injuries, he/she may then initiate the appropriate diagnostic evaluation for BCVI. Through awareness of the full spectrum of risk factors for BCVI, the radiologist may be the first to consider its presence, thereby facilitating early diagnosis and treatment. It is important to note that a large number of patients with BCVI do not manifest any of the reported markers of injury. According to Biffl et al., approximately 20% of patients with BCVI exhibit no recognized risk factors, and injuries are initially clinically occult [10] (Fig. 7). Supporting Biffl s data, recent work by Stein et al. [42] found no major craniofacial or cervical spine injury in 27% of 85 patients diagnosed with BCVI. As a means to diagnose the large number patients with truly occult acute injuries and limit resultant stroke, some [8] have suggested the routine use of computed tomographic angiography (CTA) to simultaneously examine the cervical spine and cervical arterial system during assessment of acute multitrauma patients. 6. Location and types of vascular injury Blunt cerebrovascular injuries may occur anywhere along the carotid and vertebral artery distributions, and evaluation of both circulations from their origins through their intracranial watersheds is vital. However, coinciding with the frequent association with cervical spine injuries, most BVIs occur at the foraminal (i.e. V2) vertebral artery segments [26]. As predicted by the

6 8 C.W. Sliker, S.E. Mirvis / European Journal of Radiology 64 (2007) 3 14 Fig. 6. A 39-year old male scanned with WB-MDCT protocol following motor vehicle collision. 3D volume rendered image of the chest (a) and neck (b and c) demonstrate a traumatic aortic injury with pseudoaneurysm (arrowheads), a right ICA dissection (black arrow), and a right vertebral artery dissection (white arrow). dominant mechanism of injury, BCIs typically affect the internal carotid artery just below the skull base [C.W. Sliker, unpublished data], although a significant number occur at or above the skull base [12]. Notably, a large number of those with BCVI manifest more than one injured vessel (Figs. 1 and 6). Multiarterial injury, in various combinations, has been reported in 18 38% of cases [18,20,27,28]. Given the violent and complicated force vectors

7 C.W. Sliker, S.E. Mirvis / European Journal of Radiology 64 (2007) Fig. 7. A 17-year old with female with femur fracture (not shown) and ICA injury following motor vehicle collision. Admission WB-MDCT (not shown) demonstrated narrowed distal left ICA cervical segment. Coronal maximum intensity projection (MIP) computed tomography angiogram (CTA) image (a) acquired day #2 demonstrates left ICA pseudoaneurysm (white arrow) and normal adjacent lumen. Coronal MIP CTA image (b) acquired day #7 and angiogram acquired day #9 (c) demonstrate larger pseudoaneurysm (black arrow) and near occlusion of adjacent ICA lumen (white arrowheads). Angiogram acquired after endovascular stent placement (d) demonstrates normal ICA lumen (black arrowheads) and minimal pseudoaneurysm filling (curved arrow).

8 10 C.W. Sliker, S.E. Mirvis / European Journal of Radiology 64 (2007) 3 14 Table 2 Biffl blunt cerebrovascular injury grading system Injury grade I II III IV V Description Luminal irregularity (i.e. intimal injury) or dissection with <25% luminal narrowing Dissection or intraluminal hematoma with 25% luminal narrowing, intraluminal thrombus, raised intimal flap, or hemodynamically insignificant AVF Pseudoaneurysm Occlusion Transection or hemodynamically significant AVF AVF, Arteriovenous fistula. Reference: Biffl et al. [27]. responsible for many cases of multitrauma, this is not a surprising injury characteristic. Given the propensity for multivessel injuries, care should be taken to thoroughly assess all four vascular distributions during diagnostic imaging, especially during angiography when one may be tempted to limit the evaluation to less than a complete four-vessel study in an attempt to avoid the potential consequences of progressive catheter manipulations. The forms of injuries identified in the carotid and vertebral arteries are similar to those encountered in arteries elsewhere in the body. Typical types of injury include intimal injury, spasm, dissection with intramural hematoma, raised intimal flap, intramural hematoma, pseudoaneurysm, occlusion, transection with active hemorrhage, and arteriovenous fistulae (AVF) [16,29,30]. Dissections may vary in degree of luminal compromise from minimal to near occlusion. Combinations of injury can be seen (e.g. pseudoaneurysm with dissection, raised intimal flap with intraluminal thrombus, etc.) in the same artery. With occlusion or high grade luminal narrowing, inadequate collateral flow and/or hypotension can lead to arterial insufficiency and ischemia, while intimal injury may promote platelet aggregation and embolization [12]. Pseudoaneurysms may promote platelet aggregation with distal embolization or compress the native vessel lumen [31] with resultant downstream underperfusion. Rarely, pseudoaneurysm rupture or transaction may lead to ischemia or hemodynamic compromise [12]. In an effort to the predict risk of stroke, Biffl et al. [29] devised a blunt carotid artery injury grading scale based on the anatomic features of injury. In a subsequent report [27], the scale (Table 2) was revised to include AVFs, and it was used to grade BVIs as well as BCIs. When BCVIs are reported with the Biffl scale, the highest grade injury identified determines the score for that particular vessel. At the authors institution, all injuries in a particular vessel are described, although only one may be used to determine overall grade of injury. In addition to providing some prognostic information, the grading scale is a useful reporting shorthand that is easily understood by the traumatologist. 7. Diagnosis of BCVI 7.1. Angiography Four-vessel cervicocerebral angiography is considered the reference standard for diagnosing BCVI [10,20].As an invasive test, angiography carries an inherent risk of serious complications, including stroke and hemorrhage. Yet, with experienced operators, routine angiographic assessment of trauma patients, who are generally young and without pre-existing vascular disease, is a safe procedure. For example, in one report by Cothren et al. [9], 727 blunt trauma patients were angiographically screened for BCVI, and there was one instance (0.1%) of angiographyrelated stroke and two cases (0.3%) of conservatively-managed puncture-site hematomas. In those in whom BCVI is strongly suspected clinically, the risks would seem to be tolerable. Still, in a screening population without hard clinical signs of BCVI, many may consider potential complications of angiography unacceptable given the relative rarity of these injuries. Moreover, angiography is also expensive and resource intensive Consequently, other diagnostic modalities, including Doppler ultrasound, magnetic resonance angiography (MRA), and computed tomographic angiography (CTA), are being increasingly used to diagnose BCVI Multidetector computed tomographic angiography scanning options Computed tomography (CT) is an important tool commonly used to evaluate multitrauma patients for head, spine, chest, abdomen, and pelvic injuries. With the advances in multidetector computed tomography (MDCT) technology, multidetector computed tomographic angiography (MD-CTA) is increasingly used to evaluate the arterial systems throughout the body, and, recent reports describe it use to diagnose BCVI [6,7,17,32,33]. With the use of MDCT, routine diagnosis of BCVI does not need to be limited only to scans targeted to the craniocervical region. Current MDCT scanners, used in concert with contrast bolus timing, allow continuous scanning through all regions of interest in the typical multitrauma patient. Utilization of whole-body MDCT scanning techniques permits simultaneous detection of craniocervical vascular, skeletal, thoracic, and organ injuries while forgoing the scanning time and excess radiation [34] inherent to traditional segmental scanning techniques. In the authors trauma center, targeted neck MD-CTA and whole-body MDCT have been performed for over 3 years. Currently, both 16- and 40-detector row scanners are in use, though the 40-detector row scanner is preferred for most wholebody and neck MD-CTA scans (Table 3). In most multitrauma patients, the whole-body scanning protocol is routinely used. The targeted neck MD-CTA protocol is utilized in those cases where clinical interest is directed entirely to the cranicocervical arterial system, in addition to cases where the patient is considered at risk for harboring a BCVI and the whole-body MDCT provided suboptimal or equivocal evaluation of the neck vessels. With both protocols, advanced post-processing techniques (e.g. curved multiplanar reconstructions, volume rendering, etc.) are used liberally on an as needed basis Multidetector computed tomographic angiography diagnostic accuracy The relatively few studies that have investigated the accuracy of CTA when used to diagnose BCVI are limited by small cohorts

9 C.W. Sliker, S.E. Mirvis / European Journal of Radiology 64 (2007) Table 3 A 40-detector row MDCT a scanning protocols for evaluating craniocervical arterial system Parameter Whole-body MDCT protocol Neck MD-CTA protocol Patient positioning Arms extended superiorly over head Arms at his/her sides Detector configuration and collimation 40 mm mm 40 mm mm Pitch Rotation time 0.5 s 0.5 s kv ma Area scanned Circle of Willis to lung bases (or ischial tuberosities) Vertex to aortic arch Contrast Iohexol (350 mg I/mL) Iohexol (300 mg I/mL) Contrast volume 100 ml contrast, 50 ml saline push 100 ml contrast, 50 ml saline push Contrast injection rate 6 ml/s for 60 ml, then 4 ml/s for 40 ml 4 ml/s (contrast followed by saline push) (contrast followed by saline push) Scan trigger threshold 90 HU b Time to peak cervical ICA enhancement + 2 s c Axial images (thickness interval) 2 mm 2mm 1mm 0.5 mm Sagittal and coronal MPR images (thickness interval) 2 mm 2mm 2mm 2mm Sagittal and coronal MIP images (thickness interval) n/a 5 mm 2mm Thin section axial images to independent workstation (thickness interval) 2mm 1mm 1mm 0.5 mm n/a, MIP images through neck not part of routine whole-body MDCT protocol. MDCT, Multidetector computed tomography; MD-CTA, multidetector computed tomographic angiography; MPR, multiplanar reconstruction; MIP, maximum intensity projection. a Philips Brilliance-40 scanner (Philips Medical Systems, Cleveland, OH, USA). b Bolus Pro triggering system (Philips Medical Systems) with region of interest at the distal ascending aorta. c Based on test injection of 15 ml of contrast at 4 ml/s with region of interest placed over internal carotid artery at C4 level. of injured patient and varied study designs. Compared to the reference standard, angiography, reports of single-slice spiral CTA yielded disappointing results; sensitivities for BCI and BVI were 47 68% [10,19] and 53% [10], respectively. Recently, studies investigating the accuracy of MD-CTA have emerged. In 2006, Biffl et al. [32] and Berne et al. [6] reported their experience screening 331 and 435 patients, respectively, for BCVI with 16- slice MD-CTA. In both studies, injuries (combined total of 42 patients) were confirmed with angiography while those patients with normal (i.e. no BCVI) MD-CTA results were followed clinically for signs of neurovascular complications. Neither group reported delayed neurological complications of BCVI in those with normal MD-CTA examinations, though neither followed normal MD-CTA examinations with angiography. Therefore, in both reports, true accuracy of 16-slice MD-CTA for detecting BCVI was not determined. In a prospective study that also appeared in 2006, Eastman et al. [33] evaluated 146 at-risk patients with both 16-slice MD-CTA and angiography, and they determined 16-slice MD-CTA to have a sensitivity of 97.7% and specificity of 100% when used to diagnose BCVI. Despite these impressive results, the BCVI cohort consisted of only 43 patients [33], which is a small number when studying an uncommon entity; this is a limitation shared by the other two studies. Moreover, Eastman et al. [33] report only 46 injuries in the 43 patients diagnosed with BCVI, which is far from the previously reported multivessel injury rate of 18 32% in patients with BCVI [18,20,28], Although the results of this study are promising, the small cohort of injured patients and the atypically small number of patients with multiple injured vessels suggest the data may not reflect the true accuracy of 16-slice MD-CTA when used to evaluate to the blunt trauma population at large for BCVI. Controversies not withstanding, the authors feel that MD-CTA (16-slice or greater) plays a valuable role in evaluating patients with suspected BCVI, but, until proven by a large (preferably multicenter and prospective) trial, the authors do not think that MD-CTA should be considered diagnostically equivalent to angiography Magnetic resonance angiography Few studies have assessed the diagnostic accuracy of magnetic resonance angiography (MRA) relative to angiography when used to detect BCVI. In one study by Biffl et al. [10], 16 patients at risk for BCVI were examined with both MRA and angiography. Angiography confirmed BCVI in three patients, and the collective sensitivity and specificity of MRA for BCVI were 75 and 67%, respectively. In a slightly larger study conducted by Miller et al. [19] 21 at-risk patients were examined with both MRA and angiography. Among the four patients with BCI, the sensitivity of MRA was 50% and its specificity was 100%. Vertebral artery injuries were found in 17 patients, and the sensitivity and specificity of MRA were 47 and 97%, respectively. Though its sensitivity is limited, the use MRA to diagnose BCVI has several potential advantages relative to other diagnostic options. Among them, the examination does not utilize ionizing radiation or require intravenous contrast. However, in most centers, magnetic resonance (MR) imaging is not as rapidly accessible as CT. In addition to the time and personnel required to transport the patient to the MR suite, which is remote from patient care areas in many centers, the MR environment itself is not conducive to the management of the acutely injured patient. Consequently, even if future well-designed studies demonstrate that contemporary MRA techniques have diagnostic accuracies

10 12 C.W. Sliker, S.E. Mirvis / European Journal of Radiology 64 (2007) 3 14 comparable or exceeding those of CTA, the authors feel that the physical limitations of MRA should limited it use in favor of MD-CTA (16 detector row or greater) or angiography, with the use of MRA relegated to instances when the other two diagnostic options are unavailable, their results are equivocal, or when the use of iodinated contrast material and/or ionizing radiation is contraindicated Sonography Although operator dependent, ultrasound is a portable, inexpensive tool in common use by traumatologists, and Duplex Doppler techniques are established means for diagnosing and monitoring non-traumatic cerebrovascular disease. Therefore, ultrasound is an attractive option for diagnosing BCVI. However, since the intracranial internal carotid arteries and most of the vertebral arteries are encompassed by bony canals, the ability of ultrasound to detect potentially significant injuries (e.g. intimal flaps, pseudoaneurysms) without associated flow disturbances can be limited. The few studies that have discussed the sensitivity of Doppler ultrasound for detecting BCVI report disappointing results. In one early study [35], Doppler ultrasound diagnosed 14 patients with BCI within the neck for a sensitivity of 86%, although it missed two injuries near the skull base. In a more recent study, Mutze et al. [8] detected BCVI in 0.9% of 1471 patients screened with Doppler ultrasound with sensitivity of only 38.5%. The eight patients in whom Doppler ultrasound missed injury came to clinical attention as result of cerebrovascular ischemia, supporting the notion that injuries missed by sonography are potentially significant. Given its limitations, Doppler ultrasound should only be utilized when the use of other imaging techniques is impossible. 8. Injury evolution Both blunt carotid and vertebral arteries tend to be dynamic lesions. That is, over the course of days or weeks, the severity of both treated and untreated lesion may either progress or regress. In their review of 114 carotid artery and 65 vertebral artery injuries followed with angiography 7 10 days after sustaining injury, Biffl et al. [27] noted that grade I and II injuries frequently healed by initial follow-up, and therapy could then be discontinued in 57% of grade I and 8% of grade II lesions. Conversely, 8% of grade I lesions and 43% of grade II lesions progressed, leading to more aggressive therapeutic interventions. Ninety-three percent of grade III injuries and 82% of grade IV injuries were unchanged, although in the authors experience, these lesions, especially carotid injuries, change with greater frequency. Notably, Biffl et al. noted no significant difference in injury progression or regression among patient populations receiving systemic anticoagulation, antiplatelet therapy, or no medical therapy [27]. At the authors trauma center, blunt cerebrovascular injuries are followed with the imaging modality that most satisfactorily characterized it during the initial work-up; this is usually an MDCT angiogram. The first follow-up imaging is typically performed between 24 and 72 h after initial diagnosis. During this brief interval, BCVI frequently change enough to alter management, which is particularly important given the relative frequency of coexisting injuries (e.g. unstable cervical spine injury requiring surgery, major visceral injury, etc.) that may complicate treatment with anti-thrombotic agents. A second follow-up examination is usually obtained within 1 week to 1 month (Fig. 7a and b). Beyond this point, the timing of repeat imaging is generally based on injury morphology and response to management. Uncommonly, rapid morphologic changes render differentiation between injury progression and vascular remodeling associated with healing difficult, and, in these cases, more frequent follow-up imaging can be especially useful. 9. Treatment Though treatment of some BCVI may be impossible or fruitless due to early extensive stroke or other contraindications (e.g. intracranial hemorrhage, spinal cord injury) [36], multiple authors have suggested that treatment with either medical therapy or endovascular techniques can either prevent or limit BCVI-related neurologic deficits [2 5,18,19,37,38]. To this date, the optimum treatment regimen has yet to be determined, but the current trend is towards medical therapy as option of choice. Many consider anticoagulation first line therapy [2 5,9,18,19,29], with antiplatelet therapy reserved for those with contraindications to anticoagulation. In one study conducted by Miller et al. [5], timely treatment with either anticoagulation or antiplatelet therapy lowered stroke incidence to 6% in patients diagnosed with BCI and to 0% in those with BVI (compared to 50% for both untreated BCI and BVI). Antiplatelet has been proposed by others [37] to exhibit efficacy comparable to anticoagulation while reducing the risk of hemorrhagic complications. At this time, the superiority of one approach over the other is unproven. Of special interest to radiologists, endovascular repair with stent placement has been used to treat BCIs when lesions are refractory to medical therapy [31,38] (Fig. 7c and d), medical therapy is contraindicated [38], or either angiography or perfusion imaging demonstrates evidence of cerebral hypoperfusion [38]. In a study reported by Cohen et al. [38], 10 patients with post-traumatic cervical internal carotid artery (ICA) dissections were treated with endovascular stent placement. There were no procedure-related complications or stent occlusions, and at a mean clinical follow-up time of 16 months four patients demonstrated no deterioration in neurologic status while six patients improved. In contrast, Cothren et al. [39] reported a 21% complication rate (three strokes and one subclavian artery dissection) and an occlusion rate of 45% among 23 patients with post-traumatic cervical ICA pseudoaneurysms treated with stents. In a cohort of 18 patients who received only medical therapy, none suffered stroke or progressive deficit after initiation of treatment, and one suffered carotid artery occlusion despite anticoagulation. Interestingly, this report contradicts an earlier one from the same institution, in which Coldwell et al. [31] described uncomplicated endovascular treatment of cervical ICA pseudoaneurysms in 14 patients in whom a mean clinical follow-up of 16 months revealed no evidence of stroke or delayed stent

11 C.W. Sliker, S.E. Mirvis / European Journal of Radiology 64 (2007) complication. It must be noted that in Cohen s study, all patients received antiplatelet therapy with aspirin and clopidogrel [38], while those suffering stent occlusion in Cothren s study only received anticoagulation [39], thereby suggesting that the antiplatelet therapy either alone or in combination with the stents contributed to the more promising results reported by Cohen. Clearly, the role and methods of endovascular repair of BCI remain areas in need of continued investigation. Endovascular therapy of vertebral artery injuries is also a diagnostic option. Although emergent coil embolization of vertebral artery transections is an appropriate approach [40], the merits of a comparable approach to less severe injuries remain incompletely explored. Since Biffl et al. [27] suggest the highest rate of BVI-related stroke (38%) occurs with grade II injuries, and up to 75% of traumatic vertebral artery occlusions can recanalize [24], coil embolization of occluded vessels has been proposed [41] to prevent reconstitution and the risk of distal thromboembolization. While this approach may pose no additional harm to the patient, its long term merits measured against those of medical therapy and the immediate risks of an endovascular procedure are unproven. Although dissections and pseudoaneurysms can be addressed with endovascular occlusion, this should be done with caution with documentation of adequate collateral flow through the contralateral vertebral artery and circle of Willis established. 10. Conclusion Blunt cerebrovascular injuries are infrequently encountered, and a high index of suspicion coupled with understanding of injury mechanism and recognition of associated injuries facilitates early diagnosis and treatment of asymptomatic lesions. Based on available resources, each institution must determine in what setting and with what means will the evaluation for BCVI be initiated. Although its accuracy remains unproven, liberal screening with 16-slice (or higher) MD-CTA facilitates early diagnosis and treatment. In cases where MD-CTA results are suspicious, equivocal or normal in high-risk patients, or clearly positive in those whom endovascular therapy is an option, prompt follow-up with four-vessel cerebral angiography is recommended. Regardless of the diagnostic and therapeutic approaches utilized to address blunt cerebrovascular injuries, the authors advocate that institutions caring for blunt trauma patients adopt a coordinated and systematic approach in order efficiently utilize resources while optimizing the outcomes of those diagnosed with these uncommon but potentially devastating injuries. References [1] Carillo EH, Osborne DL, Spain DA, et al. Blunt carotid artery injuries: difficulties with the diagnosis prior to neurologic event. J Trauma 1999;46(6): [2] Fabian TC, Patton JH, Croce MA, et al. Blunt carotid injury: importance of early diagnosis and anticoagulant therapy. Ann Surg 1996;223(5): [3] Biffl WL, Moore EE, Ryu RK, et al. The unrecognized epidemic of blunt carotid arterial injuries: early diagnosis improves neurologic outcome. Ann Surg 1998;228(4): [4] Biffl WL, Moore EE, Elliott JP, et al. The devastating potential of blunt vertebral arterial injuries. Ann Surg 2000;231(5): [5] Miller PR, Fabian TC, Bee TK, et al. Blunt cerebrovascular injuries: diagnosis and treatment. J Trauma 2001;51(2): [6] Berne JD, Reuland KS, Villarreal DH, et al. Sixteen-slice multi-detector computed tomographic angiography improves the accuracy of screening for blunt cerebrovascular injury. J Trauma 2006;60(6): [7] Schneidereit NP, Simons R, Nicolaou S, et al. Utility of screening for blunt vascular neck injuries with computed tomographic angiography. J Trauma 2006;60(1): [8] Mutze S, Rademacher G, Matthes G, Hosten N, Stengel D. Blunt cerebrovascular injury in patients with blunt multiple trauma: diagnostic accuracy of duplex Doppler US and early CT angiography. Radiology 2005;237(3): [9] Cothren CC, Moore EE, Ray Jr EC, et al. Screening for blunt cerebrovascular injuries is cost-effective. Am J Surg 2005;190(6): [10] Biffl WL, Ray Jr CE, Moore EE, et al. Noninvasive diagnosis of blunt cerebrovascular injuries: a preliminary report. J Trauma 2002;53(5): [11] McKevitt EC, Kirkpatrick AW, Vertesi L, Granger R, Simon RK. Blunt vascular neck injuries: diagnosis and outcomes of extracranial vessel injury. J Trauma 2002;53(3): [12] Biffl WL, Moore EE, Offner PJ, Burch JM. Blunt carotid and vertebral arterial injuries. World J Surg 2001;25(8): [13] Yang WG, Chen CT, de Villa GH, Lai JP, Chen YR. Blunt internal carotid artery injury associated with facial fractures. Plast Reconstr Surg 2003;111(2): [14] Schneider RC, Crosby EC. Vascular insufficiency of brain stem and spinal cord in spinal trauma. Neurology 1959;9: [15] Carpenter S. Injury of neck as cause of vertebral artery thrombosis. J Neurosurg 1961;18: [16] Lee CL, Woodring JH, Walsh JW. Carotid and vertebral artery injury in survivors of atlanto-occipital dislocation: case reports and literature review. J Trauma 1991;31(3): [17] Utter GH, Hollingworth W, Hallam KD, Jarvik JG, Jurkovich GJ. Sixteenslice CT angiography in patients with suspected blunt carotid and vertebral artery injuries. J Am Coll Surg 2006;203(6): [18] Cothren CC, Moore EE, Biffl WL, et al. Anticoagulation is the gold standard therapy for blunt carotid injuries to reduce stroke rate. Arch Surg 2004;139(5): [19] Miller PR, Fabian TC, Croce MA, et al. Prospective screening for blunt cerebrovascular injuries: analysis of diagnostic modalities and outcomes. Ann Surg 2002;236(3): [20] Kerwin AJ, Bynoe RP, Murray J, et al. Liberalized screening for blunt carotid and vertebral artery injuries is justified. J Trauma 2001;51(2): [21] Cothren CC, Moore EE, Biffl WL, et al. Cervical spine fracture patterns predictive of blunt vertebral artery injury. J Trauma 2003;55(5): [22] Biffl WL, Moore EE, Offner PJ, et al. Optimizing screening for blunt cerebrovascular injuries. Am J Surg 1999;178(6): [23] Mirvis SE, Shanmuganathan S, Buell J, Rodriguez A. Use of spiral computed tomography for the assessment of blunt trauma patients with potential aortic injury. J Trauma 1998;45(5): [24] Parbhoo AH, Govender S, Corr P. Vertebral artery injury in cervical spine trauma. Injury 2001;32(7): [25] Berne JD, Norwood SH, McAuley CE, Vallina VL, Creath RG, McLarty J. The high morbidity of blunt cerebrovascular injury in an unscreened population: more evidence of the need for mandatory screening protocols. J Am Coll Surg 2001;192(3): [26] Torina PJ, Flanders AE, Carrino JA, et al. Incidence of vertebral artery thrombosis in cervical spine trauma: correlation with severity of spinal cord injury. AJNR Am J Neuroradiol 2005;26(10): [27] Biffl WL, Ray CE, Moore EE, et al. Treatment-related outcomes from blunt cerebrovascular injuries: importance of routine follow-up arteriography. Ann Surg 2002;235(5): [28] Berne JD, Norwood SH, McAuley CE, Villareal DH. Helical computed tomographic angiography: an excellent screening test for blunt cerebrovascular injury. J Trauma 2004;57(1):11 7.

12 14 C.W. Sliker, S.E. Mirvis / European Journal of Radiology 64 (2007) 3 14 [29] Biffl WL, Moore EE, Offner PJ, Brega KE, Franciose RJ, Burch JM. Blunt carotid arterial injuries: implications of a new grading scale. J Trauma 1999;47(5): [30] Nuñez Jr DB, Torres-Leon M, Munera F. Vascular injuries of the neck and thoracic inlet: helical CT-angiographic correlation. RadioGraphics 2004;24(4): [31] Coldwell DM, Novak Z, Ryu RK, et al. Treatment of posttraumatic internal carotid arterial pseudoaneurysm with endovascular stents. J Trauma 2000;48(3): [32] Biffl WL, Egglin T, Benedetto B, Gibbs F, Cioffi WG. Sixteen-slice computed tomographic angiography is a reliable noninvasive screening test for clinically significant blunt cerebrovascular injuries. J Trauma 2006;60(4): [33] Eastman AL, Chason DP, Perez CL, McAnulty AL, Minei JP. Computed tomographic angiography for the diagnosis of blunt cervical vascular injury: is it ready for primetime? J Trauma 2006;60(5): [34] Ptak T, Rhea JT, Novelline RA. Radiation dose is reduced with a single-pass whole-body multi-detector row CT trauma protocol compared with a conventional segmented method: initial experience. Radiology 2003;229(3): [35] Cogbill TH, Moore EE, Meissner M, et al. The spectrum of blunt injury to the carotid artery: a multicenter perspective. J Trauma 1994;37(3): [36] Mayberry JC, Brown CV, Mullins RJ, Velmahos GC. Blunt carotid artery injury: the futility of aggressive screening and diagnosis. Arch Surg 2004;139(6): [37] Wahl WL, Brandt MM, Thompson BG, Taheri PA, Greenfield LJ. Antiplatelet therapy: an alternative to heparin for blunt carotid injury. J Trauma 2002;52(5): [38] Cohen JE, Ben-Hur T, Rajz G, Umansky F, Gomori JM. Endovascular stent-assisted angioplasty in the management of traumatic internal carotid artery dissection. Stroke 2005;36(4):e45 7. [39] Cothren CC, Moore EE, Ray CE, et al. Carotid artery stents for blunt cerebrovascular injury: risks exceed benefits. Arch Surg 2005;140(5): [40] Atar E, Griton I, Barchar GN, Bartal G, Klugar Y, Belenky Y. Embolization of transected vertebral arteries in unstable trauma patients. Emerg Radiol 2005;11(5): [41] Veras LM, Pedraza-Gutierrez S, Castellanos J, Capellades J, Casamitjana J, Rovira-Canellas A. Vertebral artery occlusion after acute cervical spine trauma. Spine 2000;25(9): [42] Stein DM, Boswell S, Sliker CW, Lui F, Scalea TM. Blunt cerebrovascular injuries: does treatment really matter? Eastern Association for the Surgery of Trauma 20th Annual Scientific Assembly. Fort Myers, FL. January 8, 2007.