High-Resolution Magnetic Resonance Imaging of Cervicocranial Artery Dissection: Imaging Features Associated with Stroke

Abstract

Background and Purpose:

We aimed to systematically investigate the characteristics of Cervicocranial artery dissection (CCAD) on high resolution magnetic resonance imaging (HRMRI) that are associated with acute ischemic stroke.

Methods:

Patients with CCAD were recruited and divided into stroke and non-stroke groups. The lesion location, the presence of a double lumen, intimal flap, intramural hematoma, pseudoaneurysm, irregular surface, intraluminal thrombus and other quantitative parameters of each dissected segment were reviewed. Multiple logistic regression was used to examine the association between imaging features of CCAD and ischemic stroke.

Results:

A total of 145 affected vessels from 118 patients with CCAD were analyzed. Anterior circulation, intramural hematoma, irregular surface, intraluminal thrombus and severe stenosis (>70%) on HRMRI were more prevalent in CCAD patient with stroke (54.4% vs. 36.4%; P =0.030, 96.2% vs. 84.8%; P =0.017, 74.7% vs. 37.9%; P < 0.001, 44.3% vs. 4.5%; P < 0.001, and 54.4% vs. 31.8%; P=0.008, respectively). In multivariable logistic regression analysis, the presence of irregular surface and intraluminal thrombus on imaging were independently associated with acute ischemic stroke in CCAD with odds ratios of 4.29 (95% confidence interval, 1.61-11.46, P = 0.004 and 7.48 (95% confidence interval, 1.64-34.07, P = 0.009).

Conclusions:

The current findings supported that the presence of irregular surface and intraluminal thrombus were related to stroke occurrence in patients with CCAD. HRMRI might give insights into pathogenesis of ischemic stroke in CCAD. It may be useful for individual prediction of ischemic stroke early in CCAD.

Introduction

Cervicocranial Artery Dissection (CCAD) is an important etiology of ischemic stroke in young patients¹. Population-based studies have reported that more than 50% of patients with CCAD presented with stroke or transient ischemic attack^2,3. Most strokes occur in the first few weeks after dissection⁴. Early detection of high-risk imaging characteristics of CCAD may be useful to aid in the preventive treatment of patients with CCAD without stroke but at higher risk. Several studies have investigated the relationship between demographic characteristics, risk factors and stroke in CCAD^5-7. The stroke subtypes of CCAD had been inferred to be embolization other than hemodynamic impairment and artery-to-artery embolism is the major mechanism⁸. Intimal damage and microthrombus formation may be associated with artery-to-artery embolism. A recent study has investigated the radiologic characteristics associated with ischemic stroke in patients suspected of CCAD using fat-suppressed T1WI and T1-weighted flow-suppressed magnetization-prepared rapid acquisition gradient-recalled echo (MPRAGE) sequences⁹. They showed that intramural hematoma (IMH) detection significantly contributed to stroke in patients with suspected CCAD. We hypothesized that there would be identifiable imaging features that would be associated with stroke.

Multiple modalities (CT, MRI, and echocardiography) have been used to complement each other to facilitate the diagnosis of CCAD^10,11. Nevertheless, definite diagnosis is often difficult because of the low sensitivity of these radiological methods for various signs, such as IMH, intimal flap, or double lumen¹¹. Recently, high resolution magnetic resonance imaging (HRMRI) has been used in the evaluation of patients with CCAD^12-15. HRMRI provides direct visualization both of vessel wall abnormalities as well as an intraluminal thrombus¹⁶. The potential contribution of individual risk factors and dissection imaging findings on HRMRI to stroke has not been well evaluated. Thus, the aim of this study was to investigate the imaging features that are associated to ischemic stroke in patients with CCAD.

Methods

Data availability

Anonymized data not published within this article will be shared by request from any qualified investigator.

Standard protocol approvals, registrations, and patient consents

The study was approved by the local Ethics Committee. Written informed consent was obtained from all patients in this study.

Participants

Between September 2013 and September 2018, patients with recent CCAD diagnosed by imaging techniques in our neurology department were recruited. All patients had neurological symptoms such as dizziness, headache, vomiting, sensory or motor disorders, transient ischemic attack or ischemic stroke. The indication for doing HRMRI was suspected of dissection by carotid artery ultrasound or carotid CT angiography. Head-neck combined HRMRI was performed within 28 days after symptom onset. Diffusion weighted imaging (DWI) was preformed within 7 days. Patients with DWI lesions but without clinical symptoms have been excluded. A diagnosis of CCAD was made if any pathognomonic signs of dissection (double lumen, intimal flap, intramural hematoma and pseudoaneurysm) were identified on imaging methods such as CT angiography, MR angiography, carotid artery ultrasound, digital subtraction angiography and HRMRI. Ischemic stroke status was determined by focal onset of neurological symptoms and positive DWI. The following types of patients were excluded: 1) dissection accompanied by other cerebral vasculopathies (e.g. Moyamoya disease, vasculitis, fibromuscular dysplasia, intracranial or carotid arteriosclerosis with ≥ 50% stenosis, identifiable plaque component of a large lipid-rich necrotic core or intraplaque hemorrhage¹⁷; 2) evidence of cardioembolism; 3) previous strokes or transient ischemic attacks; 4) history of transluminal intervention; 5) patients with renal dysfunction (estimated glomerular filtration rate assessed by creatinine clearance<60ml/min/1.73m²). The clinical characteristics including sex, age, history of trauma, recent infection, smoking, alcohol use, migraine, hypertension, hyperlipidemia, multiple lesions, and the time interval between symptom onset to HRMRI were recorded of each patient.

MRI protocol

All patients underwent HRMRI using a 3-tesla MRI scanner (Magnetom Verio, Skyra, Prisma, Siemens, Erlangen, Germany) with a 32-channel integrated head/neck coil. HRMRI was acquired by using a precontrast and postcontrast 3D turbo spin echo technique known as T1w-SPACE (Sampling Perfection with Application-optimized Contrast using different flip angle Evolutions)^18,19. The parameters were as follows: repetition time = 900ms; echo time = 14ms; FOV = 230 mm; matrix = 288×384; slices number = 224; slice thickness = 0.6 mm; voxel size = 0.6×0.6×0.6 mm³; acquisition time = 7 minutes. Post-contrast HRMRI was performed 5 minutes after the injection of single-dose (0.1 mmol per kilogram of bodyweight) gadolinium-based contrast agent (Magnevist; Schering, Berlin, Germany).

Image Interpretation

All DWI images were evaluated by the consensus of two neuroradiologists for the determination of a recent infarction²⁰. The carotid artery (ICA), vertebral artery (VA) and basilar artery (BA), anterior cerebral artery (ACA), middle cerebral artery (MCA), posterior cerebral artery (PCA) were included in the analysis. Dissected arteries were considered to be the cause of the stroke if the hyperintense lesions on DWI were detected in the vascular territory supplied by the branches distal to the site of the dissection. Dissected vessels without DWI abnormalities within the vascular territory were classified into non-stroke group including patients with transient ischemic attack.

HRMRI images were interpreted separately by two experienced neuroradiologists who were blinded to the DWI and clinical information using commercial software (Osirix MD; Pixmeo SARL, Switzerland; and Vessel Analysis; Surway Star Technology, China). All Source images and curved multiplanar reformation images were used for assessment. The sites of dissected vessels (intracranial or extracranial, anterior or posterior, single or multiple) on HRMRI were recorded. The presence of a double lumen, intimal flap, intramural hematoma, pseudoaneurysm, irregular surface, intraluminal thrombus and other quantitative parameters of each dissected segment was reviewed²¹. The double lumen sign was defined as blood flow that was divided into a true and a false lumen²². An intimal flap was considered as a curvilinear and isointense line crossing the flow void lumen or between a hyperintense haematoma that extended to the sidewall^15,23. Intramural hematoma was defined as crescent-shaped thickening of the arterial wall that was isointense/hyperintense on pre-contrast images^11,16,24. Pseudoaneurysm was defined as the diameter of aneurysmal vessels dilated over 1.5 times than the normal distal artery²⁵. Irregular surface was defined as a discontinuity of the juxtaluminal surface of the intramural hematoma. Intraluminal thrombus was defined as a hyperintense filling within the lumen on pre-contrast images, as well as an area of intraluminal contrast enhancement on post-contrast images^16,26(Figure 1). The lesion length was measured on curved planar reformation images of HRMRI. The stenosis degree at the most stenotic site was defined as (1-lesion lumen diameter/reference lumen diameter) × 100%. The stenosis degree was graded as: ≤ 49%, 50%~69% and ≥ 70%, according to the North American Symptomatic Carotid Endarterectomy Trial (NASCET) method of measurement²⁷. The characteristics of signal intensity (SI) of intramural hematoma (homogeneous or heterogeneous) and contrast enhancement ratio of intramural hematoma (SIpost-contrast / SIpre-contrast 100%) were measured. Any discrepancies between the reviewers were resolved by consensus. After an interval of 2 months, one reader reassessed all HRMRI studies.

Figure 1.

Diagram of the Characteristics of CCAD and corresponding typical HRMRI images. CCAD=Cervicocranial Artery Dissection, HRMRI=High resolution MRI.

Statistical Analysis

All continuous data were described as the mean value ± standard deviations and categorical data were expressed as frequencies (percentage). Categorical variables were analyzed using a Chi-square test and continuous variables were compared using a t-test or Mann–Whitney U test between the two groups. Variables were entered into multivariate analysis model if they were P ≤ 0.1 on the univariate analysis. Multivariate logistic regression analysis with the method of enter stepwise was conducted to determine independent predictors of the stroke. A value of P < 0.05 was considered to indicate statistical significance. The intra-observer and inter-observer agreement in qualitative CCAD characteristics were evaluated by using Cohen’s kappa coefficient, while quantitative CCAD characteristics measurement were evaluated by intraclass correlation coefficient. Multiple collinearity of variables were diagnosed by Tolerance < 0.1 or variance inflation factor (VIF) > 10. All statistical analysis was performed with SPSS 20.0 (IBM, USA).

Results

Clinical characteristics

A total of 118 patients with CCAD (80 men; mean age, 40.3 years ± 10.1) were recruited, including 71 patients with stroke and 47 patients without stroke but with neurological symptoms. All Clinical characteristics of patients were presented in Table 1.

Table 1:

Clinical Characteristics of 145 Dissecting Arteries

	Stroke group n=79	Non-stroke group n=66	P value
Age, mean (SD), y	40.4±9.7	40.3±10.6	0.936
Male patients, No. (%)	64 (81.0)	31 (47.0)	< 0.001
BMI, mean(SD), kg/m2,	24.1±3.3	24.1±3.2	0.970
Risk factors, No. (%)
History of trauma	26 (32.9)	28 (42.4)	0.238
Recent infection	11 (13.9)	10 (15.2)	0.834
Smoking	33 (41.8)	23 (34.8)	0.394
Alcohol use	35 (43.8)	23 (34.8)	0.274
Migraine	11 (13.9)	8 (12.1)	0.749
Hypertension	17 (21.5)	13 (19.7)	0.787
Hyperlipidemia	27 (34.2)	22 (33.3)	0.915
NIHSS, Median(interquartile range)	3 (1,8)	0 (0,0)	0.000
Onset to HRMRI time, Median (interquartile range)	11 (6, 20)	14 (11, 20)	0.079

Abbreviations: National Institute of Health stroke scale; HRMRI, high resolution magnetic resonance imaging

Location of CCAD Lesions

A total of 145 dissecting arteries were included, with multiple vessels involved in 21 (17.8%) patients. Sixty-six (45.5%) of the 145 lesions were found in the ICA, seventy-one (49.0%) in the VA, six (4.1%) in the BA, one (0.7%) in MCA, and one (0.7%) in PCA. Seventy-nine (54.5%) affected vessels were classified as in the stroke group and 66 (45.5%) were in the non-stroke group. The dissected arteries in stroke group were more likely to be located in the anterior circulation than that in non-stroke group (54.4% vs. 36.4%; P = 0.030).

Intra- and interobserver agreement on HRMRI Measurements

For the intrareader agreement in the identification of the presence of double lumen, intimal flap, intramural hematoma, pseudoaneurysm, irregular surface, intraluminal thrombus and heterogeneous signal intensity of intramural hematoma, the К values were 0.91, 0.87, 0.86, 0.85, 0.96, 0.81, and 0.95, respectively. The interobserver reproducibility for the presence of double lumen, intimal flap, intramural hematoma, irregular surface, intraluminal thrombus, and heterogeneous signal intensity of intramural hematoma had К values of 0.91, 0.89, 0.90, 0.79, 0.95, and 0.91, respectively. The intraclass correlation coefficient of length and contrast enhancement ratio of intramural hematoma was 0.93 and 0.84, with all P < 0.05 respectively.

HRMRI Characteristics of CCAD

An intramural hematoma, irregular surface and intraluminal thrombus on HRMRI were more frequent in the stroke group, compared to the non-stroke group (96.2% vs. 84.8%; P = 0.017, 74.7% vs. 37.9%; P < 0.001, and 44.3% vs. 4.5%; P < 0.001, respectively). Severe stenosis (stenosis degree > 70%) was more frequent in the stroke group than in the non-stroke group (54.4% vs. 31.8%; P = 0.008). However, other HRMRI features, including the presence of a double lumen, an intimal flap, the length of dissecting artery, heterogeneous signal intensity and enhancement of intramural hematoma were not significantly different between the two groups (All P > 0.05). Details of the HRMRI features of the two groups were described in Table 2.

Table 2.

HRMRI Features of the Two Groups

	Total n=145	Stroke group n=79	Non-stroke group n=66	P value
Anterior circulation, No. (%)	67(46.2)	43 (54.4)	24 (36.4)	0.030
Extracranial, No. (%)	117(80.7)	65 (82.3)	52 (78.8)	0.596
Double lumen, No. (%)	25(17.2)	10 (12.7)	15 (22.7)	0.110
Intimal flap, No. (%)	49(33.8)	24 (30.4)	25 (37.9)	0.342
Intramural hematoma, No. (%)	132(91.0)	76 (96.2)	56 (84.8)	0.017
Pseudoaneurysm, No. (%)	27(18.6)	9 (11.4)	18 (27.3)	0.014
Irregular surface, No. (%)	84(57.9)	59 (74.7)	25 (37.9)	< 0.001
Intraluminal thrombus, No. (%)	38(26.2)	35 (44.3)	3 (4.5)	< 0.001
Length, mean (SD), cm	4.4(2.8)	4.5 (2.6)	4.3 (3.1)	0.767
Stenosis degree
≤ 49%	52(35.9)	20 (25.3)	32 (48.5)	0.008
50%~69%	29(20.0)	16 (20.3)	13 (19.7)
70%~100%	64(44.1)	43 (54.4)	21 (31.8)
Heterogeneous signal of intramural hematoma, No. (%)	85(57.0)	50 (63.3)	35 (53.0)	0.212
Enhancement of intramural hematoma, mean (SD), %	120(35)	118 (29)	122 (43)	0.603

Abbreviations: HRMRI, high resolution magnetic resonance imaging;

The Association of HRMRI Features with Stroke

Intramural hematoma, pseudoaneurysm, irregular surface, intraluminal thrombus, degree of stenosis as well as anterior circulation location showed significant associations with ischemic stroke in the univariate analysis (P < 0.05), then all variables with P < 0.1 were used as input variables for the multivariate logistic regression analysis. Irregular surface and intraluminal thrombus remained significant in the multivariate analysis (P < 0.05), with odds ratios of 4.29 (95% confidence interval, 1.61-11.46) and 7.48 (95% confidence interval, 1.64-34.07) (Table 3 and Figure 2, 3). There was no interaction between intramural hematoma, pseudoaneurysm, irregular surface, intraluminal thrombus and degree of stenosis in regression analysis with the Tolerance was all > 0.1 and VIF < 10. Colinear diagnostic features were shown in Supplemental Table I.

Table 3.

HRMRI Features of the Two Groups

	Univariate logistic regression			Multivariate logistic regression
	OR	95% CI	P value	OR	95% CI	P value
Anterior circulation	2.090	1.071, 4.081	0.031*	1.837	0.788, 4.284	0.159
Extracranial	1.250	0.547, 2.854	0.596
Double lumen	0.493	0.205, 1.186	0.114
Intimal flap	0.716	0.359, 1.428	0.342
Intramural hematoma	3.348	0.998, 11.229	0.050*	4.045	0.790, 20.725	0.094
Pseudoaneurysm	0.343	0.142, 0.827	0.017*	0.485	0.136, 1.736	0.266
Irregular surface	4.838	2.378, 9.843	<0.001*	4.289	1.605, 11.462	0.004
Intraluminal thrombus	16.705	4.832, 57.746	<0.001*	7.476	1.640, 34.074	0.009
Length	1.018	0.906, 1.144	0.765
Stenosis degree			0.010*			0.572
50%~69%	1.969	0.784, 4.945	0.149	1.251	0.403, 3.881	0.698
70%~100%	3.276	1.525, 7.037	0.002	0.687	0.227, 2.077	0.506
Heterogeneous signal of intramural hematoma	1.527	0.785, 2.971	0.212
Enhancement of intramural hematoma	0.654	0.220, 1.948	0.446

Abbreviations: HRMRI, high resolution magnetic resonance imaging; OR, odds ratio; CI, confidence interval.

Figure 2.

Comparison of patients with CCAD with or without ischemic stroke. (A) Diffusion weighted imaging (DWI) showed acute ischemic stroke in left basal ganglia and centrum ovale (arrows) of 48 years of male patient. Curved planar reformation of HRMRI demonstrated intramural hematoma (arrows) inner the vessel wall of C1 segment of left internal carotid artery (ICA) and distal intraluminal thrombus (dashed arrows) on the vessel wall of C2 and C4 segments of the ICA. Stretched curved planar reformation HRMRI image depicted irregular surface and intraluminal thrombus. (B) DWI showed no ischemic stroke in a 37 years old female patient. Curved planar reformation of HRMRI image depicted a smooth vessel wall surface (arrows) without intraluminal thrombus. Stretched curved planar reformation of right ICA demonstrated the intramural hematoma and normal distal vessel wall.

Figure 3.

Odds ratio (OR) and 95% confidence interval (Cl) of stroke group vs. non-stroke group on the basis of multivariable logistic regression for each feature.

Discussion

In this study, we found a higher prevalence of intramural hematoma, irregular surface, intraluminal thrombus and severe stenosis in patients with CCAD with stroke. Multivariate analysis showed that the presence of irregular surface and intraluminal thrombus were independently associated with the occurrence of acute ischemic stroke. Therefore, it is understandable that these high risk imaging features may help to better understand the relationship between unstable CCAD, formation of blood clots, and embolic stroke.

CCAD implies a tear in the wall of the artery leading to the intrusion of blood within the layers of an arterial wall. An important pathologic feature of symptomatic CCAD is that the affected vessel wall becomes the nidus for distal embolization. Previous studies demonstrated that intraluminal contrast enhancement on HRMRI, representing intraluminal thrombus formation, is strongly correlated with ischemic symptoms in patients with spontaneous CCAD²⁶. In our study, a hyperintense signal was more likely to be found on the dissected vessel wall of stroke patients (44.3% vs 4.5%, P < 0.001), which represents fresh intraluminal thrombus. The findings of intraluminal thrombus on HRMRI in our study supports previous research findings that thromboembolism is the most common stroke mechanism in patients with CCAD⁸. Consistent with previous similar studies in carotid artery atherosclerosis, intraluminal thrombosis was a significant predictors of ischemic stroke (OR = 7.48; 95% confidence interval, 1.64-34.1, P = 0.009)²⁸.

An irregular surface is another high risk characteristic of CCAD associated with ischemic stroke observed in our study. Similarly, studies of carotid atherosclerotic plaques have revealed that an irregular surface is a risk factor for thrombosis formation and is associated with an increased risk of transient ischemic attack/stroke^29,30. The irregular surface of dissecting artery may represent intimal damage which may lead to microthrombus formation, consequently be correlated to ischemic stroke. In patients without stroke, we speculate this may due to the uneven fluctuation caused by different stages of intramural hematoma without intimal damage.

A double lumen and intimal flap are important imaging features for CCAD diagnosis, however, in our study they did not differ between the two groups. We found that patients with CCAD with stroke were more likely to have intramural hematoma, severe stenosis/occlusion. Intramural hematoma may represent a state in which blood flow tends to coagulated or accumulated inside the false lumen. Severe stenosis/occlusion as can be associated with hypoperfusion and a greater risk of thrombogenesis resulting in ischemic stroke³¹. However, our multivariate analysis demonstrated that the occurrence of stroke was not influenced by the degree of stenosis indicating that hemodynamic impairment may be a less important cause of stroke in patients with CCAD³².

CCAD tends to affect multiple arterial segments and develop simultaneously in intracranial and extracranial vessels^33,34. In our study, we found intraluminal thrombus usually forms in the distal segments of the dissected vessel. Conventional HRMRI has limited coverage and do not allow the visualization of both extra- and intracranial vessels^24,35-36. In the current study, we used 3D HRMRI technique with an integrated head-neck coil which provided larger vessel coverage and higher spatial resolution, enabling a more comprehensive evaluation of the extracranial and intracranial vessels. The complementary information with HRMRI of vessel wall and vessel lumen may facilitate individualized risk stratification of CCAD. However, whether patients with these high-risk features could benefit from anti-thrombotic prophylaxis needs to be further investigated in future studies.

This study has several limitations. First, because this is a cross-sectional study, additional prospective studies are still needed to determine the predictive power of these imaging characteristics in occurrence of ischemic stroke. Second, the low rate of intracranial dissection indicates the potentially selective bias of our study. Large-scaled studies are needed to investigate the relationship between pseudoaneurysms and subarachnoid hemorrhage. Third, HRMRI studies were performed within 28 days after symptom onset in this study. Recent follow-up studies of CCAD have shown that the dissection may recover in short-term^37,38. Thus, the dynamic nature of CCAD may affect the findings of high-risk features of dissected vessel wall. Finally, a separate validation cohort was absent due to the limited sample size in this single-center study.

In conclusion, irregular surface and an intraluminal thrombus were associated to stroke occurrence in patients with CCAD and may be useful for individual risk assessment early after clinical presentation.

Glossary:

CCAD

cervicocranial artery dissection

HRMRI

high resolution magnetic resonance imaging