Helical CT for the Diagnosis of Extracranial Internal Carotid Artery Dissection
Background and Purpose We attempted to evaluate the sensitivity of helical CT for the diagnosis of extracranial internal carotid artery (ICA) dissection.
Methods Sixteen consecutive patients with 18 angiographically confirmed extracranial ICA dissections were studied with a helical CT protocol with large-volume acquisition and thin axial slice reconstructions. A control group including normal and atherosclerotic ICAs was formed for comparison, and a blind interpretation of CT images was made by two observers. We evaluated the presence of stenosis, eccentric lumen, mural thickening, aneurysm, occlusion, and annular contrast enhancement. When the artery seemed to be occluded, we measured the external diameter of the ICA (1) on the occluded side, at its upper portion and most enlarged level, (2) at its lower portion, beyond the bulb, and (3) on the contralateral side, at its upper portion.
Results Interobserver agreement was good except for the presence of annular contrast enhancement. In the stenotic dissection group (n=12), the presence of a narrowed eccentric lumen at the upper portion of the ICA on axial CT images was classified correctly in all cases (sensitivity, 100%; specificity, 100%). An arterial wall thickening was seen in all cases of dissection but also in three cases of the control group. In the occlusive dissection group (n=6), the enlargement of the dissected artery was the best criterion (sensitivity, 100%, specificity, 100%) for occlusive dissection.
Conclusions Helical CT seems to be a reliable method for evaluating extracranial ICA dissection. The analysis of the residual arterial lumen and the measurement of the external diameter of the carotid artery were the best criteria for the diagnosis. Further studies with larger groups are required to determine whether ICA dissections might be diagnosed using helical CT as a first procedure.
Cervical artery dissection accounts for up to 10% to 20% of ischemic stroke in young adults.1 Angiography remains the gold standard for evaluating the presence and extent of dissection,2 3 but this technique carries a risk of complications4 and does not analyze the arterial wall. Previous studies have suggested the use of MRI and Doppler sonography for the diagnosis of cervical artery dissection.5 6 7 Some reports have described the CT appearances of dissection8 9 10 : narrowed eccentric lumen, mural thickening, and thin annular contrast enhancement. More recently, Zuber et al10 evaluated the value of the dynamic incremental CT scan in cervical artery dissections. These studies report signs suggestive of cervical artery dissection but did not evaluate the specificity, sensitivity, and reliability of such findings. Moreover, they did not examine the imaging findings independently from angiography because CT technique was directed by angiographic abnormalities10 or clinical findings.8 All examinations were performed with conventional CT using either a dynamic CT technique (axial images at a single level of the neck during contrast administration) or a dynamic incremental CT technique (dynamic CT scan with table incrementation providing multilevel slices).
Helical acquisition CT is a new technique combining continuous gantry rotation with simultaneous displacement of the examination table throughout the acquisition.11 This technique allows data acquisition of a large volume in a short time. When intravenous contrast material is used and timed correctly, helical CT scans can be obtained at the peak of arterial enhancement.12 13 Thin reconstructions of images in the axial plane provide excellent visualization of both arterial lumen and arterial wall.13 The value of this technique in the diagnosis of internal carotid artery (ICA) atherosclerosis has already been evaluated.14 The aim of this study was to assess the value of helical CT imaging for the diagnosis of ICA dissection.
Subjects and Methods
The study was conducted over a 15-month period (July 1994 through October 1995) in 16 consecutive patients (7 men and 9 women; mean age, 44 years [range, 23 to 53 years]) with extracranial ICA dissection on conventional angiography. All patients underwent helical CT scan. Clinical presentations consisted of symptoms of acute cerebral ischemia associated with headache, ipsilateral neck pain, or both (11 cases), Horner’s syndrome (3 cases), and lower cranial nerve palsies (2 cases). Conventional angiography was performed within 6 days after clinical onset in 15 patients and 3 months after the initial event in 1. Digital subtraction angiograms were obtained on a Philips Integris V3000 through a femoral approach. Arch aortograms followed by selective carotid and vertebral angiograms were performed in the anterior and lateral projections with cervical and head views. The contrast dose for the entire study was approximately 150 mL (Omnipaque 300, Nycomed). The diagnosis of dissection was based on the following angiographic criteria: (1) irregular or smooth stenosis involving the upper portion of cervical ICA, (2) tapered occlusion beginning distal to the bulb, and (3) saccular pseudoaneurysm. Because angiographic findings are not specific to occlusive dissection, etiologic workup15 was performed, including blood tests (sedimentation rate, fibrinogen, packed cell volume, platelet count, and coagulation tests), electrocardiography, 24-hour continuous electrocardiography monitoring, color and duplex Doppler sonography of the cervical arteries, and transesophageal echocardiography. We considered the diagnosis of dissection as probable when potential cardiac sources of cerebral embolism and atherosclerotic disease were excluded. Eighteen dissections of the ICA were found in 16 patients. In 6 cases, the ICA was completely occluded with a tapered lumen beyond the bulb. In 9 cases, the angiography showed a stenotic segment with a long tapered lumen, and in the remaining 3 cases, there was a narrowed and irregular short segment of the upper ICA associated with an adjacent pseudoaneurysm. The narrowing was smooth in 3 arteries and irregular and eccentric in 8 other arteries. Dissection extended beyond the skull base in only 1 case.
Helical CT scans were obtained on a Siemens Somatom Plus S system. The examination was performed within 3 to 19 days (mean, 11 days) after clinical onset in 15 patients and 3 months after onset in 1. Patients were placed in the supine position with the head tilted back as far as possible to avoid dental hardware. A lateral scanogram was acquired with the shoulders placed as low as possible. We performed scans in 5-mm sections without contrast injection from C2 to C6 to determine the level of carotid bifurcations. Continuous data were acquired during a scan time of 32 seconds and started at the level of carotid bifurcations. Patients were instructed to breathe quietly without swallowing during the scanning period. A total volume of 90 mL of nonionic contrast material (Omnipaque 350, Nycomed) was injected at 3 mL/s into an antecubital vein with a 20-second scan delay after the start of the contrast bolus, a 2-mm collimation, and a table speed of 3 mm/s (total coverage, 90 mm). Axial images were reconstructed at 1-mm increments (total number of sections, 90) using low-order interpolation and small field of view. The time required for analyzing all slices was approximately 15 minutes per patient.
A comparative study of the ICA images on helical CT was performed for those consecutive patients requiring angiography for atherosclerotic carotid artery disease (n=14) or vertebrobasilar infarcts (n=8). A total of 44 nondissected carotid arteries were included as a control group. In this group, angiography showed a normal ICA in 22 cases, atherosclerotic stenosis at the origin of the ICA in 15 cases (mild and moderate stenosis, n=6; severe stenosis, n=9), and atherosclerotic occlusion in the remaining 7 arteries. All CT examinations were performed with the same protocol and separately interpreted in a fully randomized order by two independent examiners (X.L. and A.S.). As described in previous reports,8 9 10 we evaluated the presence of stenosis, eccentric lumen, mural thickening, aneurysm, occlusion, annular contrast enhancement (thin hyperdensity surrounding the arterial wall), and target picture as described by Dal Pozzo et al8 (narrowed eccentric lumen surrounded by a mural thickening, itself surrounded by a thin annular enhancement). When the ICA was judged to be occluded, we performed several measurements with the use of a large zoom and a computer caliper to separate atherosclerosis from dissection. The ICA external diameter was measured on the occluded side (1) at its lower portion, immediately beyond the bulb, (2) at its upper portion in the last 2 cm below the petrous bone, and (3) on the contralateral side, at its upper level. Measurements were made at the most enlarged level in cases with nonuniform diameter. To take into account possible intersubject variability of ICA diameter, two additional ratios were calculated: the ipsilateral/contralateral ratio (100×upper ipsilateral carotid diameter/upper contralateral carotid diameter) and the upper/lower ratio (100×upper ipsilateral carotid diameter/lower ipsilateral carotid diameter). A value higher than 100 indicated an ipsilateral or upper ICA enlargement.
The first step of the statistical analysis evaluated the interobserver agreement. For the presence of stenosis, eccentric lumen, mural thickening, aneurysm, occlusion, and annular enhancement, the κ statistic was used.16 This test takes into account the proportion of agreement expected by chance; κ values between 0.4 and 0.8 suggest a moderate to substantial agreement and values higher than 0.8 an excellent agreement. The interobserver agreement for the measurement of ICA diameter was assessed by the method of limits of agreement.17 This test uses the interobserver difference for each diameter and provides an index of difference between each observer for 95% of observations (95% limits of agreement).
The second step consisted of a group comparison. In cases without occlusion, frequencies of stenosis, eccentric lumen, aneurysm, target picture, and mural thickening were compared with Fisher’s exact test.18 In cases with occlusion, ICA diameters and ratios were compared with the Mann-Whitney U test.17
The scan volume of 90 mm was sufficient to cover the whole ICA in each case, from the bifurcation to the petrous portion of the ICA. The position of internal and external carotid arteries was accurately determined on axial images by both examiners in all cases. The total volume of contrast material injected during 32 seconds with 20-second delay led to a poor opacification of the carotid bifurcation in 5 dissected arteries but optimal contrast enhancement of the upper portion of the ICA, providing a good imaging of the arterial lumen at the level of dissection. In 3 other carotid arteries, the opacification of the lumen at its upper portion was moderate but did not prevent the interpretation. Artifacts due to dental hardware were a minor problem to analyze and measure the ICA in 6 cases. The interpretation was made in these cases on the underlying nonartifacted sections.
The interobserver agreement was excellent for the evaluation of the presence of stenosis, eccentric lumen, mural thickening, aneurysm, and occlusion (all κ=1.00), but it was moderate (κ=0.5) for the presence of a thin annular contrast enhancement. This sign was only observed in the dissection group after consensus reading (n=4). The interobserver agreement for the measurement of ICA diameter was judged as good (95% range for agreement of −0.82 to 0.63). A consensus reading of the external diameter measurement was performed in two cases because of a difference of more than 10% between the two examiners.
In cases with ICA stenosis, the location of maximum stenosis differed between the dissected and atherosclerotic groups: it was higher at the lower portion in the atherosclerotic group and maximum at the upper portion in the dissected group. The presence of an eccentric lumen at the upper portion of the ICA was found in all cases of stenotic dissection but never in atherosclerotic stenosis (Table 1⇓ and Fig 1⇓). An aneurysm was only found in 3 cases of dissection (Fig 2⇓). The target picture was highly specific but was seen in only 4 cases (Fig 3⇓). Arterial wall thickening was encountered in all cases of stenotic dissection but was also seen in 3 cases of the control group (Fig 4⇓).
In cases with ICA occlusion (Fig 5⇓), group comparison showed that the upper ipsilateral diameter, the upper/lower ratio, and the ipsilateral/contralateral ratio were higher in the dissection group, and this result was not attributable to change in either the lower ipsilateral diameter of occluded ICA or in the upper contralateral diameter. The examination of range values (Table 2⇓) showed no overlap between groups for the upper ipsilateral carotid diameter and upper/lower ratio. Thus, a threshold value might be determined to classify all cases with occlusion within the present sample19 (upper ipsilateral carotid diameter, cutoff value between 4.7 and 5.5 mm; upper/lower ratio, cutoff value between 102% and 167%). Conversely, 1 case with bilateral dissection leading to occlusion on one side had an ipsilateral/contralateral ratio close to those observed in the atherosclerosis group.
Our study showed that stenosis, eccentric lumen, mural thickening, occlusion, and aneurysm are well evaluated with helical CT, whereas a thin annular contrast enhancement is a less reliable sign. All ICA dissections with stenosis were characterized by a narrowed eccentric lumen at the upper portion of the ICA, providing a very sensitive and specific sign of dissection. Conversely, the typical target picture was not a sensitive criterion, and the presence of arterial wall thickening was not specific to dissection.
The CT appearance of carotid dissection has already been reported in previous studies,8 9 10 but the usefulness of helical CT for the diagnosis of ICA dissection has not been previously estimated. In the study of Petro et al,9 the arterial wall thickening was the best criterion of dissection on conventional CT. However, we have observed 3 cases of mural thickening in the nondissected group, and this result is consistent with ultrasonographic findings showing increased thickness of arterial wall in patients with physical inactivity, abdominal adiposity, abnormal glucose metabolism, or atherosclerotic disease.20 The typical target picture8 was present in only 4 cases of our study but was frequently found in the study of Zuber et al.10 This discrepancy could be explained by the poor reliability of the detection of the thin annular contrast enhancement. The peripheral hyperdensity is probably due to the contrast enhancement of the vasa vasorum in the adventitial layer.10 With the use of helical CT technique, the shortened scanning time and the thin slice thickness might lead to a low contrast level in the arterial wall and consecutively to a disagreement on the peripheral hyperdensity analysis. The increase in the external diameter of the artery was the best indicator of dissection in the study of Levy et al,6 which used MR angiography. Consistently, we found that the enlargement of the upper ICA was higher in occlusion related to dissection. However, this sign might be not specific to dissection because cervical arteries may be enlarged in other conditions, such as arterial hypertension, cardiac dysfunction, or fibromuscular dysplasia. Thus, we used the comparison between the external diameter of the carotid artery at the site of the dissection and beyond the bulb. This upper/lower ratio allowed the correct classification of all cases of ICA occlusion within the present study and might be less sensitive to other causes of ICA enlargement. The comparison between the external diameter of the dissected artery and the contralateral side may be misleading in patients with bilateral dissection, and this situation appears to be not uncommon in our study, as well as in a previous study.2 Further studies are required to determine the best criterion to classify various mechanisms of ICA occlusion.
Zuber et al10 suggested the use of dynamic CT scan to demonstrate the presence of mural hematoma when angiographic findings are not conclusive, but they estimated that indications of CT scan are limited because CT needs to be directed by prior angiography to determine the level of the dissection. This limitation is the result of CT technology and may be resolved by helical CT, which allows the use of a volumetric acquisition. With conventional CT, the examination table does not move during the scan, and interscan delays are required to perform multilevel slices. A few slices only are acquired during the period of maximum contrast enhancement, and the examination of the cervical region requires thick sections or a large volume of contrast material or both.21 With the helical CT technique, the acquisition is performed in a very short time, and a small volume of contrast material provides a good contrast between the enhanced lumen and the surrounding structures. A large volume from C5 to the skull base is assessable with 2-mm-thick sections and 1-mm-thick increments. Thus, the helical CT technique provides a sensitive imaging of the whole cervical portion of the ICA and does not need to be directed by prior angiography.
Previous studies have suggested the usefulness of MRI for the diagnosis of carotid dissection5 6 because of a specific picture on the T1-weighted images associating a narrowed eccentric signal void surrounded by a semilunar signal hyperintensity (mural hematoma). Despite the important role of MRI in the management of ICA stenosis due to dissection, helical CT might provide a complementary contribution. In most cases, the mural hematoma in the subacute phase is better visualized on MRI because of its high-intensity signal on T1-weighted images. Conversely, if the diagnosis is suspected at a very early stage, within the first few days after the clinical onset, helical CT might be theoretically superior to MRI, since a fresh hematoma appears isointense or slightly hyperintense on MRI.22 However, this situation has never been encountered in our experience. Another important point is the diagnosis of occlusive dissection. MR findings may be nonspecific, showing a hypersignal covering all sections of the artery and leading to confusion between the mural hematoma and the thrombus in the true arterial lumen.10 With the use of helical CT technique, thin sections and thin increments provide accurate measurements of the external diameter of the ICA, which seems to be a reliable method for the diagnosis of occlusive-type dissection.
Our study suggests that helical CT may be an interesting tool for the diagnosis of ICA dissection, but it requires an evaluation with larger groups and a comparative study with MRI before its exact role is determined.
The authors gratefully thank F. Lemarchand for the photographic reproductions, T. Saint-Michel for technical assistance, C. Rose for help in preparing the manuscript, and the technical staff in the CT department.
- Received November 21, 1995.
- Revision received December 21, 1995.
- Accepted December 22, 1995.
- Copyright © 1996 by American Heart Association
Leys D, Moulin T, Stojkovic T, Begey S, Chavot D, for the DONALD Investigators. Follow-up of patients with history of cervical artery dissection. Cerebrovasc Dis. 1995;5:43-49.
Houser OW, Mokri B, Sundt TM, Baker HL, Reese DF. Spontaneous cervical cephalic arterial dissection and its residuum: angiographic spectrum. AJNR Am J Neuroradiol. 1984;5:27-34.
Bui LN, Brant-Zawadzki M, Verghese P, Gillan G. Magnetic resonance angiography of cervicocranial dissection. Stroke. 1993;24:126-131.
Steinke W, Rautenberg W, Schwartz A, Hennerici M. Noninvasive monitoring of internal carotid artery dissection. Stroke. 1994;25:998-1005.
Petro GR, Witwer GA, Cacayorin ED, Hodge CJ, Bredenberg CE, Jastremski MS, Kieffer SA. Spontaneous dissection of the cervical internal carotid artery: correlation of arteriography, CT, and pathology. AJNR Am J Neuroradiol. 1986;7:1053-1058.
Zuber M, Meary E, Meder JF, Mas JL. Magnetic resonance imaging and dynamic CT scan in cervical artery dissections. Stroke. 1994;25:576-581.
Leclerc X, Godefroy O, Pruvo JP, Leys D. Computed tomographic angiography for the evaluation of carotid artery stenosis. Stroke. 1995;26:1577-1581.
Mounier-Vehier F, Degaey I, Leclerc X, Leys D. Cerebellar border zone infarcts are often associated with presumed cardiac sources of ischaemic stroke. J Neurol Neurosurg Psychiatry. 1995;59:87-89.
Brennan P, Silman A. Statistical methods for assessing observer variability in clinical measures. BMJ. 1992;304:1491-1494.
Siegel S, Castellan JJ. Nonparametric Statistics for the Behavioral Sciences. 2nd ed. New York, NY: McGraw-Hill Publishing Co; 1988.
Folsom AR, Eckfeldt JH, Weitzman S, Ma J, Chambless LE, Barnes RW, Cram KB, Hutchinson RG, for the Atherosclerosis Risk in Communities (ARIC) Study Investigators. Relation of carotid artery wall thickness to diabetes mellitus, fasting glucose and insulin, body size, and physical activity. Stroke. 1994;25:66-73.
Kitanaka C, Tanaka JI, Kuwahara M, Teraoka A. Magnetic resonance imaging study of intracranial vertebrobasilar artery dissections. Stroke. 1994;25:571-575.