Quantitative Magnetic Resonance Angiography in Internal Carotid Artery Occlusion with Primary Collateral Pathway.

BACKGROUND AND
PURPOSE: Quantitative magnetic resonance angiography (Q-MRA) enables direct measurement of volume flow rate (VFR) of intracranial arteries. We aimed to evaluate the collateral flows in internal carotid artery (ICA) occlusion with primary collateral pathway via circle of Willis using Q-MRA, and to compare them between patients who recently developed ipsilateral symptomatic ischemia and those who did not.
METHODS: Between 2012 and 2014, 505 patients underwent Q-MRA in our institution. Among these, 33 patients who had unilateral ICA occlusion with primary collateral pathway were identified, and grouped into asymptomatic patients, stable patients with chronic infarction, and symptomatic patients with acute/subacute infarction. Mean VFR (mVFR) in intracranial arteries was measured and compared between the patients' groups. Kruskal-Wallis test was used for statistical analysis.
RESULTS: Six patients were asymptomatic, fifteen with chronic infarction were stable, and twelve with acute/subacute infarction were symptomatic. The mVFR of ipsilateral middle cerebral artery in symptomatic patients was significantly lower than those in stable or asymptomatic patients (73.7±45.6 mL/min vs. 119.9±36.1 mL/min vs. 121.8±42.0 mL/min; P = 0.04). Total sum of the mVFR of ipsilateral anterior, middle, and posterior cerebral arteries was significantly lower in symptomatic patients than those in other groups (229.3 ± 51.3 mL/min vs. 282.0±68.6 mL/min vs. 314.0±44.4 mL/min; P = 0.02).
CONCLUSIONS: Q-MRA could be helpful to demonstrate the difference in the degree of primary collateral flow in ICA occlusion between the patients with recent symptomatic ischemia and those without.

Introduction

Acute ischemic stroke by occlusion of internal carotid artery (ICA) is known to have relatively high rate of ischemic progression and recurrence [1]. However, ICA occlusion can have variable clinical courses – from no symptom to severe neurologic deficit – according to the degree of collateral flow [2]. Thus, the assessment of collateral flow in ICA occlusion is of great clinical significance.

The collateral circulation in ICA occlusion is constituted with primary collateral pathways via circle of Willis and secondary pathways via ophthalmic artery and leptomeningeal vessels [3]. In particular, the presence of ipsilateral middle cerebral artery (MCA) flow via primary collateral pathway has been known to protect cerebral tissue from the progression of ischemic insult [4,5]. Accordingly, the presence of MCA flow has proven that it may reduce initial stroke severity and be associated with early favorable outcome [6].

Recently, the advent of quantitative magnetic resonance angiography (Q-MRA) has enabled direct measurement of aforementioned intracranial flow [7,8]. However, only a limited number of studies have quantitatively assessed the degree of collateral flow in ICA occlusion. Most recently, Rutgers et al. [9] have measured quantitative volume flow via circle of Willis using Q-MRA, along with collateral flow via ophthalmic artery using transcranial Doppler, and that via leptomeningeal anastomosis using conventional angiography. However, the authors reached the conclusion that recurrent ischemic stroke with symptomatic carotid artery occlusion was associated with increased collateral flow through posterior communicating artery, which appeared to be conflicting finding against those from previous studies [5,6,10].

In this study, we performed direct measurement of intracranial flow in unilateral ICA occlusion with primary Willisian collateral using Q-MRA. We used noninvasive optimal vascular analysis (NOVA; Vassol, Chicago, IL, USA) to calculate the mean volume flow rate (mVFR) of the intracranial arteries. The purpose of this study was to compare the mVFR of intracranial arteries between the symptomatic patients who recently developed cerebral ischemia ipsilateral to the occluded ICA and the stable or the asymptomatic patients.

Methods

Subjects

Between July, 2012 and September, 2014, 505 consecutive patients underwent Q-MRA in our institution as the diagnostic work-up or the follow-up study for the presumed or the alleged large-vessel cerebrovascular disease. Among these, 35 patients (29 men, 6 women; mean age, 64.8 ± 13.2 years; range, 23-81 years) who had occlusion of unilateral ICA with primary collateral pathways via circle of Willis were identified.

The identified patients were classified into three groups according to their clinical features. First group was composed of asymptomatic patients without any cerebral infarction (group 1). Second group consisted with stable patients with chronic infarction, which was present in the ipsilateral cerebral hemisphere to the occluded ICA (group 2). We defined stable clinical status as follows: a) no newly developed or progressed neurologic deficit for at least 6 months before the performance of Q-MRA, and b) no radiologic evidence of newly developed ischemic lesion on brain imaging. Symptomatic patients with ipsilateral acute/subacute infarction were grouped into the third group (group 3): accordingly, these patients had acute ischemic event within 1 month before Q-MRA. During this process, 2 patients in the second group were excluded for the reasons detailed here: one patient was diagnosed with Takayasu arteritis later, and the other patient showed progression of neurologic symptom regardless of the development of new infarction. Regarding the latter patient, the reason of the patient’s symptom progression was not clear.

Clinical information was obtained by reviewing the electronic medical records. Demographic data including age, sex, and medical history was acquired. The institutional review board approved this study. Informed consent requirements were waived by the board.

Imaging protocol and analysis

The protocol of Q-MRA was detailed in a previous study [11]. Q-MRA was performed with on 3T MRI scanner (Achieva, Philips Healthcare, Best, the Netherlands), and the mVFR was calculated from NOVA software on a separate workstation.

A standard axial 3-dimensional time-of-flight (TOF) MRA of intracranial and cervical arteries was obtained (Figure 1A). Then, the acquired images were transmitted to the workstation to reconstruct 3D surface-rendered vessel images (Figure 1B) [12]. After determining optimal perpendicular scan plane and setting the baseline coordinates following the method from the previous study [12], retrospectively gated, fast 2-dimensional phase-contrast sequence was performed using the parameters as follows: repetition time, 10-15 ms; echo time, 407 ms; flip angle, 15°; number of excitation, 1; field-of-view, 250 × 250 mm2; acquisition matrix, 256 × 128 mm2; slice thickness, 4 mm for intracranial arteries and 5 mm for cervical arteries. Velocity encoding was automatically adjusted by NOVA software.

Figure 1.

Q-MRA images of a 47-year-old male patient in group 1. (A) Left ICA is occluded, but patent ipsilateral MCA flow from posterior communicating arterial collateral is demonstrated on TOF-MRA. (B) The section plane (yellow) shows the location of the measurement of left MCA flow. (C) The overall mVFR of the intracranial and cervical arteries are shown in the report of NOVA software.

For the calculation of mVFR, the obtained phase-contrast images were transmitted to the workstation. A region-of-interest was automatically located on the acquired phase-contrast images and displayed on the 3D surface-rendered images for the vessel verification. The mVFR for each intracranial artery was calculated in milliliters per minute (mL/min), and reported (Figure 1C). The total scan time for the entire study was 20-30 minutes.

In addition, the following sequences were obtained according to the clinical need: diffusion-weighted imaging, T1- and T2-weighted turbo spin-echo imaging, Fluid-attenuated inversion recovery imaging, gradient recoiled echo T2*-weighed imaging, dynamic susceptibility perfusion magnetic resonance imaging (MRI), gadolinium-enhanced T1-weighted imaging, and/or contrast-enhanced MRA.

The degree of contralateral ICA stenosis was assessed according to the following criteria: mild, 0%-49%; moderate, 50%-69%; severe, 70%-99%; occlusion, 100%. To indirectly assess the volume of the infarction, MRI Alberta Stroke Program Early CT score (ASPECTS) was calculated for each patient [13].

mVFR calculation

The mVFR was calculated at anterior cerebral artery (ACA), MCA and posterior cerebral artery (PCA) in both ipsilateral and contralateral sides to the occluded ICA. Then, we calculated total ipsilateral and total contralateral cerebral circulation flow by adding up the mVFR from ACA, MCA, and PCA of the same side. We also calculated the ratio of the mVFR of ipsilateral MCA to that of contralateral side, and that of total ipsilateral cerebral circulation flow to contralateral side, respectively.

Statistical analysis

Categorical variables are expressed as frequencies with percentages. Continuous variables are presented as mean values ± standard deviations. Demographic data between the patients’ groups were compared using Kruskal-Wallis tests and Fisher’s exact tests. MRI ASPECTS between the group of stable patients with chronic infarction and that of symptomatic patients with acute/subacute infarction was compared using Mann-Whitney test. Kruskal-Wallis test was also used to compare the calculated values of mVFR between 3 patients’ groups. If needed, further subgroup analysis was additionally performed. A P value less than 0.05 was considered statistically significant. The statistical analyses were performed using a statistical software package (SPSS, version 17.0; SPSS, Chicago, IL, USA).

Results

As a result, 6 patients were classified into group 1, 15 patients into group 2, and 12 patients into group 3. As for the group 3, 8 patients had hemodynamic dominant infarction (watershed zone infarction), and 4 patients had non-hemodynamic dominant infarction such as embolic infarction. Since we initially included patients with patent primary collateral pathways via anterior communicating and/or posterior communicating arterial flows, there was no significant difference of collateral pathways between the patients’ groups.

Table 1 demonstrates the patients’ clinical characteristics, degree of contralateral ICA stenosis, and MRI ASPECTS. There was no significant difference in the patients’ age, sex, and medical history related to the vascular ischemic risk factor. No patient had occlusion of contralateral ICA. MRI ASPECTS between the stable patients with chronic infarction and the symptomatic patients with acute/subacute infarction was not significantly different from each other.

Table 1.

Demographic and clinical findings of the subjects with unilateral internal carotid artery (ICA) occlusion

	Group 1[*] (n = 6)	Group 2[†] (n = 15)	Group 3[‡] (n = 12)	P value[§]
Men/women, n (%)	6/0 (100)	13/2 (86.7)	8/4 (66.7)	0.18
Age, mean ± standard deviation (yr)	63.5 ± 12.7	65.3 ± 13.8	65.8 ± 11.4	0.9
Hypertension, n (%)	2 (40)	8 (53.3)	9 (75)	0.22
Diabetes mellitus, n (%)	0 (0)	5 (33.3)	6 (50)	0.11
Heart disease, n (%)	3 (50)	2 (13.3)	0 (0)	0.12
Current smoking, n (%)	1 (20)	3 (20)	4 (33.3)	0.65
Hyperlipidemia, n (%)	3 (60)	6 (40)	5 (41.7)	0.91
Degree of contralateral
ICA stenosis, n (%)
0%-49%	6 (100)	14 (93.3)	12 (100)	0.54
50%-69%	0 (0)	0 (0)	0 (0)	-
70%-99%	0 (0)	1 (6.7)	0 (0)	0.54
100%	0 (0)	0 (0)	0 (0)	-
Alberta Stroke Program Early CT score (mean ± SD)	-	8.67 ± 1.91	8.33 ± 0.98	0.11

Group1, asymmetric patients;

Group 2, stable patients with chronic infarction;

Group 3, symptomatic patients with acute/subacute infarction;

P value is calculated by Kruskal-Wallis test and Mann-Whitney test.

Quantitative mVFR

Table 2 summarizes the results of the quantitative mVFR in intracranial arteries. The mVFR of MCA ipsilateral to the occlusion of ICA was significantly lower in the symptomatic patients with acute/subacute infarction than those in the stable patients with chronic infarction or the asymptomatic patients (group 1, 121.8 ± 42.0 mL/min; group 2, 119.9 ± 36.1 mL/min; group 3, 73.7 ± 45.6 mL/min; P= 0.04). Total ipsilateral cerebral circulation flow – the sum of mVFR from ipsilateral ACA, MCA, and PCA – was also significantly lower in the acute symptomatic patients than those in other groups (group 1, 314.0 ± 44.4 mL/min; group 2, 282.0 ± 68.6 mL/min; group 3, 229.3 ± 51.3 mL/min; P= 0.02). There was no significant difference in above two values between the stable patients with chronic infarction and the asymptomatic patients.

Table 2.

Flow quantification of intracranial arteries using quantitative magnetic resonance angiography (Q-MRA)

	Group 1 (n=6)	Group 2 (n=15)	Group 3 (n=12)	P value
Mean volume flow rate (mVFR), mean ± standard deviation (mL/min)
Ipsilateral anterior cerebral artery (ACA)	75.0 ± 28.7	57.3 ± 24.3	67.5 ± 33.3	0.49
Contralateral ACA	73.7 ± 33.4	65.3 ± 27.4	58.0 ± 12.4	0.21
Ipsilateral middle cerebral artery (MCA)	121.8 ± 42.0	119.9 ± 36.1	73.7 ± 45.6	0.04[*]
Contralateral MCA	167.3 ± 45.0	163.3 ± 50.5	170.2 ± 45.5	0.9
Ipsilateral posterior cerebral artery (PCA)	117.2 ± 48.9	104.8 ± 38.7	88.2 ± 49.9	0.48
Contralateral PCA	79.0 ± 14.6	77.1 ± 26.0	68.1 ± 29.8	0.59
Contralateral internal carotid artery (ICA)	350.0 ± 112.6	360.8 ± 85.8	349.5 ± 110.6	0.82
Ipsilateral ACA+MCA+PCA	314.0 ± 44.4	282.0 ± 68.6	229.3 ± 51.3	0.02[*]
The ratio of mVFR (%)
Ipsilateral to contralateral MCA	74.8 ± 24.3	75.7 ± 18.0	46.2 ± 26.8	0.02[*]
Ipsilateral to contralateral
ACA+MCA+PCA	99.7 ± 11.5	97.5 ± 35.8	79.6 ± 16.6	0.04[*]

Group1, asymptomatic patients; Group 2, stable patients with chronic infarction; Group 3; symptomatic patients with acute/subacute infarction. P value is calculated by Kruskal-Wallis test.

P value less than 0.05.

The ratio of mVFR of ipsilateral MCA to that of contralateral side was significantly different between the patients’ groups (group 1, 74.8 ± 24.3%; group 2, 75.7 ± 18.0%; 46.2 ± 26.8%; P= 0.02). In subgroup analysis, the value was significantly lower in symptomatic patients with acute/subacute infarction than in stable patients with chronic infarction (P< 0.01). The value in symptomatic patients was also lower than asymptomatic patients, but did not reach the statistical significance (P= 0.07).

The ratio of total ipsilateral cerebral circulation flow to total contralateral cerebral circulation flow was significantly different between the patients’ groups (group 1, 99.7 ± 11.5%; group 2, 97.5 ± 35.8%; group 3, 79.6 ± 16.6%; P= 0.04). In subgroup analysis, the difference between group 1 and 3 was statistically significant (P= 0.01). The value from group 3 was lower than that from group 2, but it did not reach the statistical significance (P= 0.09).

Representative Cases

Figure 2 presents the representative images from two patients from group 2 and group 3, respectively. A 67-year-old male in group 2 showed about 31 mm sized chronic infarction with tissue loss in right corona radiata (Figure 2A). He showed no neurologic symptom relevant to this lesion at the time of imaging. The other patient in group 3 was an 80-year-old male who visited the hospital with sudden onset right hand weakness. On diffusion-weighted imaging, about 33 mm sized acute infarction was noted in left corona radiata (Figure 2B). On TOF MRA, patent ipsilateral primary collateral flow via circle of Willis was equally verified in both patients (Figure 2C and D). However, when we performed Q-MRA and measured collateral flow in intracranial arteries, the mVFR of ipsilateral MCA (94 mL/min vs. 65 mL/min), the mVFR of total ipsilateral cerebral circulation (244 mL/min vs. 199 mL/min), the ratio of mVFR of ipsilateral MCA to that of contralateral side (95.9% vs. 50.0%), and the ratio of total ipsilateral cerebral circulation flow to total contralateral cerebral circulation flow (102.5% vs. 89.2%) was all lower in the patient from group 3 than the patient from group 2.

Figure 2.

(A) T2-weighted imaging and fluid-attenuated inversion recovery images of a 67-year-old male patient. About 3.1 cm sized chronic infarction is noted in right anterior corona radiata. (B) Diffusion-weighted imaging and Apparent diffusion coefficient map of an 80-year-old male. He had sudden onset right hand weakness at the time of imaging. There is about 3.3 cm sized acute infarction in left anterior corona radiata. (C) TOF MRA of the 67-year-old male with chronic infarction. Patent flow of MCA ipsilateral to the occluded ICA is visualized. The mVFR of the ipsilateral MCA was calculated as 94 mL/min. (D) The 80-year-old male with acute infarction in the territory in left MCA also shows intact flow of ipsilateral MCA to the occluded ICA on TOF MRA. However, the mVFR of this MCA was calculated as 65 mL/min.

Discussion

The degree of intracranial collateral flows has been mainly evaluated using regional cerebral blood flow, which can be assessed by diverse modalities including technetium-99m-hexamethyl propyleneamineoxime single-photon emission computed tomography [14,15], perfusion computed tomography [16,17], and perfusion MRI such as dynamic susceptibility perfusion MRI [18] or arterial spin-labeling perfusion MRI [19]. Although cerebral blood flow may provide information about tissue-level collateral flow adequacy, it is incapable of quantitatively demonstrating collateral flow routes [8].

The direct measurement of volume flow via collateral pathways or intracranial arteries has not been commonly performed. A few studies in late 1900’s used transcranial Doppler to evaluate the velocity of ipsilateral MCA via primary collateral pathways in ICA occlusion [20,21]. These studies reported that diminished blood flow velocity less than approximate 30 cm/sec was associated with poor clinical prognosis [20,21]. However, transcranial Doppler is an operator-dependent, subjective study which cannot provide objective result of volume flow measurement [8].

Recently, the direct vascular flow measurement has become possible with the use of Q-MRA [7]. Q-MRA is a technique that uses TOF and phase-contrast MRI to visualize extracranial and intracranial vascular anatomy, and provide the direct measurement of volumetric blood flow within intracranial arteries [7,22]. The accuracy of the flow measurement using Q-MRA has been extensively tested with phantom blood flow simulators [11,23]. Many previous studies also have verified the reproducibility of flow measurement using Q-MRA, and intra-rater and inter-rater reliabilities of Q-MRA were also investigated [11,24-26]. In addition, the flow measurement from Q-MRA has proved to be correlated with several other methods such as H2(15)O positron emission tomography [26], ultrasound vector velocity technique [27], laser Doppler velocimetry and computational fluid dynamics [28].

Our study showed that Q-MRA using NOVA software could be a feasible tool for the direct measurement of the volume flow of intracranial arteries including primary collateral flow via circle of Willis in ICA occlusion. Moreover, we demonstrated that the mVFR of ipsilateral MCA and that of total ipsilateral cerebral circulation were reduced in the symptomatic patients with acute/subacute infarction than in the stable patients with chronic infarction or in the asymptomatic patients without any infarction. The ratio of mVFR in ipsilateral to contralateral MCA and that of total mVFR in ipsilateral cerebral circulation to contralateral side also decreased in the symptomatic group.

It has been well known that reduced primary Willisian collateral flow is related to the development of cerebral ischemia [5,10]. The North American Symptomatic Carotid Endarterectomy Trial study proved that increased collateral flow via anterior communicating artery or posterior communicating artery was associated with lower cerebral ischemic risks [4]. Our study has its merit in that, with the use of Q-MRA, we objectively verified the larger reduction in the amount of primary Willisian collateral flow in the symptomatic patients with ICA occlusion and ipsilateral recent infarction than in the neurologically stable or asymptomatic patients.

Our study has some limitations. First, this study was cross-sectional study with small number of patients. In this regard, there could be selection bias from small size of study population, and age correction of mVFR could not be achieved. However, the statistical analysis verified that the included patients showed similar clinical risk factors for stroke, which could give credibility to the comparison of mVFR between the patients’ groups. Still, to determine whether decreased collateral flow can really predict the future development of symptomatic ischemia in ICA occlusion, further follow-up prospective study with larger number of patients will be necessary. Second, we did not directly measure the volume of the infarcted cerebral tissue in the assessment of the collateral degree. It can be easily anticipated that the degree of collateral flow will be measured low in the larger infarcted area than in the smaller one. However, we compared the values of MRI ASPECTS between the symptomatic patients with acute/subacute infarction and the stable patients with chronic infarction, which could reflect the volume of the infarcted area [29]. Therefore, the difference in the infarction volume would have had little impact on the result of our study.

Conclusions

In conclusion, Q-MRA can demonstrate the difference in the degree of primary collateral flow via circle of Willis in unilateral ICA occlusion. The volume flow in MCA and in total cerebral circulation was significantly reduced in the patients with acute/subacute symptomatic ipsilateral cerebral infarction than in the patients without. Further follow-up study with prospective nature and the larger number of patients will be needed to solidify the result from this study.