The term dissection refers primarily to an elevation or separation of the intimal lining of an artery from the subjacent media and, less frequently, to separation of the media from the adventitia. Dissection is usually accompanied by hemorrhage into the arterial wall, which creates, as demonstrated in the first image below, a blind pouch or (uncommonly) a parallel subintimal second channel. The second image demonstrates the angiographic characteristics of a chronic subadventitial dissection of the right internal carotid artery [1, 2, 3]
Recognizing a dissection early is essential, because prompt anticoagulant and/or antiplatelet therapy and endovascular repair greatly minimize the patient's risk of infarction, neurologic disability, and death. However, carotid and vertebral dissections are still underrecognized despite their distinct clinical and radiologic manifestations.
Types of dissection
In most patients, the pathogenesis of arterial dissection is usually multifactorial. In general, such dissections can be categorized as traumatic or spontaneous. 
Traumatic dissection is the result of either external mechanical injury, such as a penetrating or blunt trauma, or trivial trauma that is related to a movement or abrupt change in head position. [5, 6] Examples of such movements include rapid turning of the head, [5, 6] flexion or extension of the head and neck,  and strenuous physical exertion. [6, 7]
Spontaneous dissections are those in which no definitive precipitating factor is recognized. However, spontaneous dissections may be associated with predisposing factors, such as connective tissue disorders, systemic hypertension, smoking, diabetes mellitus, a history of cerebral infarction, hyperlipidemia, cerebral and abdominal aortic aneurysms, use of oral contraceptives, and a family history of arterial dissection.
Nearly all the imaging modalities have been used to demonstrate cervicocephalic arterial dissection. Although conventional catheter angiography has been the criterion standard for the diagnosis of arterial dissection, increasing reliance has been placed on reconstructed cross-sectional imaging techniques, including computed tomography (CT) scanning and magnetic resonance angiography (MRA), because of the inherently noninvasive nature of these modalities. [8, 9, 10, 11]
The advantages of magnetic resonance imaging (MRI) make this the preferred method for the initial screening and evaluation of patients with suspected arterial dissection. MRI is the least invasive method for the follow-up of dissected cervicocephalic arteries.
A study by Bonati et al indicates that diffusion-weighted imaging (DWI) scans can differentiate cerebral ischemias caused by spontaneous cervical artery dissection from those resulting from patent foramen ovale (PFO). DWI was carried out in 93 patients with acute ischemic stroke, and 33 strokes were found to be caused by carotid or vertebral artery dissection and 60 of them by PFO.
In the study, the frequency with which multiple lesions appeared on DWI scans was greater in patients with a dissection-related stroke than in those whose stroke was PFO related (70% of patients vs 43%, respectively). Moreover, lesions in the dissection group were larger than those in the PFO group (median diameter of largest lesion, 50 mm vs 23 mm, respectively). 
Plain radiography is the least sensitive of all the imaging modalities for detecting carotid or vertebral artery dissection, yet plain radiographs may provide clues that suggest arterial injury. Penetrating injuries to the middle third of the neck in either the anteroposterior or lateral projections should raise the suspicion for carotid artery injury. Fractures of the lateral elements of the cervical spine, especially those fractures that extend through the foramina transversaria, should prompt evaluation for injury to the vertebral artery.
The degree of confidence in plain radiography is low.
CT scanning provides a noninvasive method for the diagnosis and follow-up of arterial dissections. Helical (spiral) CT scanning and multidetector-row dynamic CT (MDCT) scanning enable rapid data acquisition over a large area during peak contrast enhancement. The data can then be displayed by using a maximum intensity projection (MIP) algorithm to obtain CT angiographic images.
The most reliable criterion for diagnosis of acute arterial dissection is probably an enhancing, narrowed, eccentric lumen in association with enlargement of the overall diameter of the dissected artery from mural thickening (compared with the contralateral side).  The eccentric rim of mural thickening due to the acute intramural hematoma often appears darker than the adjacent muscle on CT scans and usually does not enhance.
From axial source images, CT angiography can provide exquisite 3-dimensional (3D) reconstructed images that are as thin as 0.6 mm. Although the vertebral artery is easily identified at the cervical levels above C2 and below C6, the status of the vertebral arteries can be difficult to assess between C2 and C6 as they course through the foramina transversaria. Extensive postprocessing can yield very diagnostic 3D images. Axial source images are sensitive to luminal abnormalities at all levels, including the foramen transversaria. 
Other findings on CT scans may include a pseudoaneurysm, an intimal flap with a double lumen, and arterial occlusion.
Degree of confidence
The degree of confidence in CT scanning is high. However, a minor drawback of this modality is the necessity for contrast enhancement. Findings on contrast-enhanced CT scans are similar to those on MRIs.
Magnetic Resonance Imaging
In conjunction with MRA, MRI is an effective noninvasive method that is used to diagnose arterial dissection. [14, 15, 16, 17] These techniques offer excellent tissue contrast, soft-tissue delineation, and demonstration of arterial flow, without the need for contrast enhancement. As a result, direct visualization of the intramural hematoma, simultaneous examination of the carotid and vertebral arteries, precise determination of the extent of the dissection, and examination of the brain parenchyma for possible ischemia are possible. 
The advantages of MRA and MRI make them the preferred methods for the initial screening and evaluation of patients with suspected arterial dissection, and they are the least invasive methods for the follow-up of dissected cervicocephalic arteries. [10, 18, 19, 20]
The classic MRI finding of cervical arterial dissection is a somewhat eccentric periluminal rim of intramural hematoma. This is probably best achieved with a T1-weighted technique or, occasionally, with a T2-weighted, fat-suppression technique.  The subacute intramural hematoma contains intracellular or extracellular methemoglobin and appears as a rim of hyperintensity on fat-suppressed, axial T1-weighted images, whereas the patent lumen appears as a flow void. If the predominant component of the hematoma is extracellular methemoglobin, the intramural hematoma can appear as a hyperintensity on fat-suppressed, T2-weighted images.
Fat suppression is important to confidently differentiate the periarterial fat from the hyperintense intramural hematoma. However, the age of the hematoma is an important consideration with regard to the signal intensity characteristics on the MRIs.  Early in the initial few days of the dissection, the rim of an acute intramural hematoma consists of deoxyhemoglobin and may not appear as a hyperintensity; instead, the rim may appear isointense relative to the adjacent muscle.
One study reported that none of the MRIs obtained in patients who underwent imaging within 1 week of the onset of dissection showed an intramural hematoma. The subacute intramural hematoma that contains intracellular and extracellular methemoglobin remains hyperintense on T1- and T2-weighted images into the early chronic stage. This finding is apparent for a period of months, after which the MRI signal intensity returns to normal and usually corresponds with the attainment of a normal appearance in the affected artery on angiograms.
Other MRI findings include an identifiable intimal flap on proton density– or T2-weighted images, an overall increase in the diameter of the affected vessel compared with the normal side, a double lumen, and enhancement of the wall and the septum on contrast-enhanced images that are obtained with a 3D spoiled gradient-recalled acquisition in the steady state (SPGR).
On images that have been obtained with a long repetition time (TR), the intimal flap is seen as a thin, curvilinear, hypointense partition that separates the residual arterial lumen from the intramural signal of the hematoma. The increase in diameter of the dissected artery is the result of the intramural hematoma. This finding is particularly useful in the acute stage of the dissection, when the hematoma is isointense on T1- and T2-weighted images.
Contrast-enhanced SPGR images show the flow channels with different degrees of hyperintensity; these appearances are thought to represent the presence of true and false lumina with different flow velocities. However, on catheter angiograms, this MRI finding often corresponds to a string sign or occlusion rather than to the double-lumen appearance that is characteristic for arterial dissection.
MRA should be routinely performed as part of the MRI evaluation of dissection. Various MRA techniques have been used; these include 2D time-of-flight (TOF), 3D TOF, and phase-contrast techniques. [9, 15] A gadolinium-based contrast material can be intravenously administered with these techniques.
Each of these modalities has certain advantages. For example, with 2D TOF imaging, large arterial segments can be imaged in a relatively shorter time than with the other 2 techniques. 3D TOF and phase-contrast MRA not only overcome the complex flow-related signal-loss artifacts that are seen with the 2D TOF technique but also depict better spatial resolution than does the 2D technique. However, imaging of long arterial segments takes a considerable length of time.
On TOF MRA source images, the intramural hematoma is often seen as a periarterial rim with a signal intensity between the high signal intensity of flow in the patent lumen of the artery and that of the surrounding soft tissue. The relatively high signal intensity of the intramural thrombus is due to the T1 effects of methemoglobin. The identification of the hyperintensity that eccentrically surrounds the artery is best seen on the TOF images because this technique has the least background-signal suppression.
The overall appearance due to the intramural hematoma is an apparent increase in the external diameter of the artery compared with that of the contralateral side. Early, before the conversion to methemoglobin occurs, the intramural hematoma may be isointense on the TOF images; in this setting, thickening of the vessel and narrowing of the patent lumen suggest dissection.
Other findings with MRA may include the complete absence of flow-related enhancement of the artery, which indicates occlusion of the artery and the presence of pseudoaneurysms. On MRI, pseudoaneurysms appear as an outpouching, with a signal intensity similar to that of the parent artery.
Gadolinium-based contrast agents (gadopentetate dimeglumine [Magnevist], gadobenate dimeglumine [MultiHance], gadodiamide [Omniscan], gadoversetamide [OptiMARK], gadoteridol [ProHance]) have recently been linked to the development of nephrogenic systemic fibrosis (NSF) or nephrogenic fibrosing dermopathy (NFD). For more information, see the eMedicine topic Nephrogenic Systemic Fibrosis. The disease has occurred in patients with moderate to end-stage renal disease after being given a gadolinium-based contrast agent to enhance MRI or MRA scans.
As of late December 2006, the FDA had received reports of 90 such cases. Worldwide, over 200 cases have been reported, according to the FDA. NSF/NFD is a debilitating and sometimes fatal disease. Characteristics include red or dark patches on the skin; burning, itching, swelling, hardening, and tightening of the skin; yellow spots on the whites of the eyes; joint stiffness with trouble moving or straightening the arms, hands, legs, or feet; pain deep in the hip bones or ribs; and muscle weakness. For more information, see Medscape.
Degree of confidence
The degree of confidence in MRI is high. The sensitivity of MRI for demonstrating the intramural hematomas is variably described as 22-93.5% in patients with internal carotid artery dissections. This variability among different studies is probably partly related to the timing from the date of the dissection to the use of fat-suppression imaging.
Most studies demonstrate a lower rate of detection with vertebral artery dissection than with carotid artery dissections; the demonstration of intramural hematomas in intracranially dissected arteries with MRI is similarly poor, as this technique is neither sensitive nor specific for detecting intracranial dissections. Studies have indicated MRI has less sensitivity in the detection of vertebral dissections than in that of extracranial carotid dissections.
Although MRI is sensitive for the detection of intramural hematomas, an additional thrombotic occlusion that is distal to the narrowed segment caused by dissection or slow flow can result in high signal intensity similar to that of the hematoma. Severe extracranial atherosclerotic carotid stenosis or occlusion can also produce a rim of abnormal signal called the partial flow-void effect. However, the rim signal is hyperintense on T2-weighted images but not on T1-weighted images. In addition, flow-related enhancement in the most caudal section of an imaging series is another possible cause of a false-positive MRI diagnosis. This typically occurs when saturation pulses are not applied inferior to the imaging volume.
Duplex ultrasonography is another noninvasive method that can be used for the initial screening and diagnosis of midcervical arterial dissection. However, this modality is less reliable than catheter angiography or MRA. Ultrasonography can be used only to screen the midcervical segments of the carotid or vertebral arteries; the proximal, distal cervical, and intracranial portions of the arteries cannot be adequately imaged with this technique. [21, 22]
The most specific ultrasonographic finding of arterial dissection is a double lumen that is separated by an echogenic intimal flap. An intramural hematoma can also be demonstrated as an eccentric echogenicity that surrounds a relatively narrowed arterial lumen.
Many arterial dissections involve the more distal segments of the arteries that cannot be imaged with ultrasonography, and in many cases, the diagnosis is then based on a discrepancy between the abnormal Doppler ultrasonographic signal and the ultrasonographic image. Abnormal Doppler ultrasonographic findings that have been described in carotid artery dissection include decreased velocities in the carotid bulb with either high resistance due to arterial stenosis or a biphasic pattern that suggests an occlusion. 
Color flow Doppler ultrasonographic imaging may best demonstrate the true and false lumina.  Findings of vertebral arterial dissection on Doppler sonograms include the absence of arterial flow or low blood velocities in the dissected artery, often with a compensatory increased blood flow in the contralateral vertebral artery. [25, 26] The sonographer may have difficulty in imaging the intraforaminal segments of the vertebral artery. Of note, the Doppler ultrasonographic findings are nonspecific, and the diagnosis of dissection is suggested only in the appropriate clinical setting. [9, 25]
Degree of confidence
The degree of confidence is high but only in the midcervical region, where the carotid artery is readily visualized. In the cervical vertebral arteries, the quality of duplex sonograms is fair, but they are considered poor for depicting the intrathoracic segments of the carotid and vertebral arteries.
Duplex ultrasonography is ineffective intracranially. Transcranial Doppler (TCD) ultrasonography is relatively sensitive to changes in the intracranial arterial velocities. Although an elevated intracranial flow velocity is not specific for intracranial arterial dissection, this condition is included in the differential diagnosis, which also includes atherosclerosis and vasospasm.
Nuclear medicine has no role in the workup of cervical arterial dissection. Brain single-photon emission CT (SPECT) scanning and positron emission tomography (PET) scanning may be useful for assessing the effect and distribution of cerebral and cerebellar hypoperfusion that result from an arterial dissection.
Catheter-directed digital subtraction angiography (DSA) remains the criterion standard for the diagnosis of cervicocephalic arterial dissection. The most common finding seen in approximately 65% of patients with subintimal arterial dissection is a relatively smooth or slightly irregular, tapered or spiraled luminal narrowing of the dissected segment, or the string sign. The degree of luminal narrowing can range from slight stenosis to severe narrowing to complete arterial occlusion.
Other angiographic findings include the double-lumen or double-barrel sign, which is most specific, and the less-specific pearl-and-string sign. The double-lumen sign demonstrates a patent false lumen or the accumulation of blood beneath an intimal flap, whereas the pearl-and-string sign is associated with stenosis and proximal dilatation. (The choice of descriptors for the double-lumen sign is poor at best and often confused with the string-of-pearls sign that is used as a descriptor for fibromuscular dysplasia.)
The double-lumen sign is seen angiographically when the distal false lumen recommunicates with the true lumen, allowing antegrade flow through the false channel. This most frequently occurs with subadventitial dissections. The recommunication of a small false lumen along the posterior margin of the left internal carotid artery with the native lumen is seen in the angiogram below.
Angiography also depicts pseudoaneurysms and distal-branch occlusions due to emboli. Although an intimal flap or the double-lumen sign is a specific finding for arterial dissection, it is seen in less than 10% of patients. Similarly, contrast stasis and thrombus in a dissected artery or emboli in the distal branches are seen in only about 10% of patients. The resolution of stenoses on follow-up angiography is also considered a reliable, albeit late, confirmatory finding of vertebrobasilar or carotid artery dissections.
Pseudoaneurysms are seen in up to 25-35% of subadventitial dissections. These dilatations commonly appear along the postbulbar segment of the internal carotid artery and are typically oval shaped, extend parallel to the artery, vary greatly in size, and are contained by the carotid sheath. Most pseudoaneurysms remain stable when observed with serial angiography; however, as with distal emboli, they can increase in size and rupture. Rupture of a midcervical pseudoaneurysm may result in the formation of a carotid-cutaneous or carotid-pharyngeal fistula.
Although conventional angiography has consistently been the best method to demonstrate changes in the vascular wall related to fibromuscular dysplasia, stenosis, and patent pseudoaneurysms, this technique has limitations. However, if the lumen of the artery is unaffected or minimally affected, as in a subadventitial dissection, or if a false lumen or pseudoaneurysm becomes thrombotic, dissection may not be reliably detected. Because angiography is an invasive procedure, the associated risks and complications should be taken into account when a diagnostic workup is planned.
If anticoagulation therapy is administered to a patient, follow-up imaging should be performed a few months after the event to reevaluate the injured area.  The angiogram below is from a patient who underwent successful treatment with aspirin and clopidogrel.
Degree of confidence
There is a high degree of confidence in angiographic findings.
Other pathologic entities can resemble cervical arterial dissection—in particular, atherosclerotic disease. The appearance of the lesion site and the site itself, as well as the presence of calcifications, can help to distinguish between these 2 conditions. Carotid artery dissection typically involves the postbulbar segment rather than the carotid bifurcation, which is the common site for atherosclerosis. Vertebral artery dissection tends to occur at the points of entry into and exit from the foramina transversaria—most commonly, at the C1-2 and C6-7 levels or at the intradural segment, rather than at the origin of the vertebral artery, where atherosclerosis is most common.
A long segment of narrowing with or without proximal dilatation is seen with dissections, whereas there is focal involvement in atherosclerosis. In addition, classic symptoms (eg, oculosympathetic paresis) and other angiographic findings (eg, pseudoaneurysm) further help to distinguish the 2 entities.
Nondissected type I fibromuscular dysplasia can also simulate cervical arterial dissection when the appearance is one of luminal narrowing rather than the typical beaded appearance. However, arterial dissection is a common finding in other forms of fibromuscular dysplasia. Again, the clinical features and the associated findings, especially pseudoaneurysm, can help to make the diagnosis.
Vasospasm can also mimic arterial dissection, and distinguishing this entity from dissection is particularly difficult after patient trauma. However, if vasospasm is related to catheter placement, the condition typically resolves in a short time, whereas a dissected segment requires weeks to months to resolve.