Dynamic Magnetic Resonance Angiography Provides Collateral Circulation and Hemodynamic Information in Acute Ischemic Stroke
Background and Purpose—Contrary to usual static vascular imaging techniques, contrast-enhanced dynamic magnetic resonance angiography (dMRA) enables dynamic study of cerebral vessels. We evaluated dMRA ability to assess arterial occlusion, cerebral hemodynamics, and collateral circulation in acute ischemic stroke.
Methods—Twenty-five acute ischemic stroke patients with proximal anterior circulation occlusion underwent dMRA on a 3T scanner within 12 hours of symptoms onset. Diffusion weighted imaging, Tmax6 s lesion volumes and hypoperfusion intensity ratio as volume of Tmax>6 s/volume of Tmax>10 s were measured. Site and grade of occlusion (Thrombolysis in Myocardial Infarction criteria) were evaluated on time-of-flight MRA and dMRA. Leptomeningeal collaterality (American Society of Interventional and Therapeutic Neuroradiology/Society of Interventional Radiology [ASITN/SIR] Scale) and asymmetries in venous clearance were assessed exclusively on dMRA. Collateral filling was dichotomized into incomplete (ASITN/SIR 0–2) or complete (ASITN/SIR 3–4).
Results—On dMRA, site of occlusion was M1 in 21 patients, tandem internal carotid artery/M1 in 2 and tandem internal carotid artery/terminal internal carotid artery in 2 patients. Three tandem occlusions were not detected on time-of-flight–MRA. All patients had Thrombolysis in Myocardial Infarction 0 to 1 on time-of-flight–MRA, but three of them had Thrombolysis in Myocardial Infarction 2 on dMRA. Complete collateral filling (n=12, 48%) was associated with smaller diffusion weighted imaging lesion volume (P=0.039), smaller hypoperfused volume (P=0.018), and lower hypoperfusion intensity ratio (P=0.006). Patients with symmetrical clearance of transverse sinuses (52%) were more likely to have complete collateral filling (P=0.015).
Conclusions—As a fast, direct, feasible, noninvasive, and reliable method to assess site of occlusion, collateral circulation and hemodynamic alterations, dMRA provides profound insights in acute stroke.
Location of arterial occlusion, arterial circulation beyond primary occlusive lesion, and functionality of leptomeningeal collaterals predict arterial revascularization and clinical outcome in patients with acute ischemic stroke (AIS).1 Therefore, availability of data on cerebral hemodynamics is essential before undergoing an endovascular therapy.
In contrast to the usual static vascular imaging techniques as computed tomographic angiography or time-of-flight–magnetic resonance angiography (TOF–MRA), dynamic contrast-enhanced MRA (dMRA) assesses blood flow over time providing hemodynamic information. We evaluated dMRA ability to assess arterial occlusion, cerebral hemodynamics, and collateral circulation in AIS patients with anterior circulation large vessel occlusion.
Subjects and Methods
We studied consecutive patients with AIS admitted to our emergency department from August 2013 to November 2014 who met the following criteria: (1) wake-up stroke, symptoms onset >4.5 hours or patients in whom intravenous recombinant tissue-type plasminogen activator given within 4.5 hours failed; (2) anterior circulation proximal occlusion defined as Thrombolysis in Myocardial Infarction 0 to 1 on TOF–MRA; and (3) 3T magnetic resonance imaging ≤12 hours of onset. In addition to the conventional sequences of a standard acute stroke protocol, a dMRA sequence was acquired (Methods section in the online-only Data Supplement). Digital subtraction angiography (DSA) was performed only in patients who underwent therapeutical thrombectomy.
Two readers blinded to the results of the other radiological data, evaluated full-coverage maximum intensity projection of dMRA, reaching a consensus about discrepant interpretations. We measured infarct core volume on diffusion weighted imaging (apparent diffusion coefficient<600×10−6 mm2/s) and hypoperfusion volume on Tmax>6 s maps. Hypoperfusion intensity ratio was calculated as: volume Tmax>6 s/volume Tmax>10 s.2 Site of occlusion was evaluated on TOF–MRA, dMRA, and DSA. Thrombolysis in Myocardial Infarction grades were evaluated on TOF–MRA and dMRA, whereas modified Thrombolysis in Cerebral Infarction grades were assessed on dMRA and DSA. Anatomic variations, transverse sinus clearance, circle of Willis functionality, and secondary collateralization by using the American Society of Interventional and Therapeutic Neuroradiology/Society of Interventional Radiology (ASITN/SIR) Scale (Figure 1) were assessed on dMRA. ASITN/SIR scores were later dichotomized as incomplete (ASITN/SIR 0–2) and complete (ASITN/SIR 3–4).
Continuous variables were expressed as mean (SD) or median (quartiles) and compared using Student t test or the Mann–Whitney test as appropriate. Categorical variables were compared using χ2 test or Fisher test as appropriate.
Twenty-five patients were studied (mean age 65 years, 16 (64%) women, median baseline National Institutes of Health Stroke Scale International 17). Twelve patients underwent DSA.
On dMRA, site of occlusion was M1 in 21 patients, tandem internal carotid artery/M1 in 2, and tandem internal carotid artery/terminal internal carotid artery in 2 patients. Three tandem occlusions were not detected on TOF–MRA. All patients had Thrombolysis in Myocardial Infarction 0 to 1 on TOF–MRA, but 3 of them had Thrombolysis in Myocardial Infarction 2 on dMRA (Figure 2; Table I and Movie in the online-only Data Supplement). Agreement between DSA and TOF–MRA was found in 9 of 12 cases and between DSA and dMRA in 11 of 12 cases. Modified thrombolysis in cerebral infarction score agreement between DSA and dMRA was found in 11 of 12 cases.
About circle of Willis, 6 patients had segment P1 hypoplasia, 2 had segment A1 hypoplasia, and none had fetal posterior cerebral artery, and 11 patients (44%) had a complete circle of Willis. Leptomeningeal collateral flow was classified as ASITN/SIR score 1 in 9 (36%) patients, score 2 in 4 (16%), score 3 in 7 (28%), and score 4 in 5 (20%) patients, thus collateral flow was incomplete in 13 (52%) and complete in 12 (48%) patients. Inter-reader agreement for ASITN/SIR was excellent (k=0.93). Complete circle of Willis was associated with complete leptomeningeal collateral filling (P=0.028). Vascular filling was faster through the anterior cerebral artery in 7 (28%) patients, faster through the posterior cerebral artery in 6 (24%) patients, and simultaneous in 12 (48%) patients. Complete leptomeningeal filling was not associated with any pattern of fast artery supply (P=0.676).
Patients with complete leptomeningeal filling had smaller diffusion weighted imaging lesion (12.7 versus 22.9 mL, P=0.039), smaller Tmax 6 s lesion (56.8 versus 135 mL, P=0.018), and lower hypoperfusion intensity ratio (2.9 versus 1.8, P=0.006) than those with incomplete filling. There was a negative correlation between the ASITN/SIR grades and both diffusion weighted imaging and Tmax 6 s lesion volumes, whereas ASITN/SIR grades correlated positively with hypoperfusion intensity ratio (Figure 3).
Ten patients (40%) had transverse sinus hypoplasia (2 right, 7 left, and 1 both). Twelve patients had asymmetrical clearance of tranverse sinuses. The ipsilateral sinus to the arterial occlusion cleared faster in 4 patients and slower in 8 patients. Asymmetrical clearance of transverse sinuses was not associated with transverse sinus hypoplasia (P=0.163). Patients with complete collateral filling had more frequently symmetrical clearance of transverse sinuses (66.7% versus 30.8%, P=0.015).
Dynamic MRA is a fast and reliable method to assess cerebral hemodynamics and collateral circulation in patients with AIS. The ASITN/SIR scale applied to dMRA highly correlated with infarct core and hypoperfusion intensity ratio highlighting that it might be used as a surrogate of collateral degree in conventional angiography.2
Furthermore, dMRA detected more incomplete occlusions and forward-flow through the thrombus than TOF–MRA. This finding is particularly interesting because a recent study proves that anterograde flow through the thrombus predicts early recanalization with intravenous recombinant tissue-type plasminogen activator.3 Finally, asymmetrical clearance of tranverse sinuses assessed by dMRA was associated with poor collateralization. The role of venous system in AIS is poorly understood and dMRA may be a useful tool for future research.
dMRA enables obtaining hemodynamic information in just 90 s. The acquired images do not need complex postprocessing methods and can be assessed using maximum intensity projection, a simple volume rendering method. Other MR techniques have been developed to study collaterals. Four-dimensional steady state–free precession MRA seems excellent for the assessment of collaterality through the Circle of Willis but does not show leptomeningeal collaterals.4 Perfusion-based collateral indexes2,5 and arterial spin-labeling6 sequences measure collaterality indirectly without giving anatomic information about occlusion site or main blood input through primary or secondary collateral circuits. By contrast, dMRA enables simple, noninvasive angiographic characterization with high inter-rater reproducibility for collaterality.
We acknowledge some important limitations. The number of patients studied was small and assessment of collaterals with dMRA was not compared against conventional angiography. However, dMRA may be used as a noninvasive sequence of multimodal MR in patients who are screened for endovascular therapy. The usefulness of dMRA in predicting favorable response to endovascular thrombectomy warrants validation in future clinical research.
Sources of Funding
Dr Hernández-Pérez is granted with a Río Hortega research contract (Carlos III Health Institute), cofinanced by the Germans Trias i Pujol Research Institute Foundation. Project partially supported by a grant from the Spanish Ministry of Health cofinanced by Carlos III Health Institute, Spanish Stroke Research Network INVICTUS (RD 12/0014/008).
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.115.010748/-/DC1.
- Received August 4, 2015.
- Revision received October 27, 2015.
- Accepted October 29, 2015.
- © 2015 American Heart Association, Inc.
- Bang OY,
- Saver JL,
- Kim SJ,
- Kim GM,
- Chung CS,
- Ovbiagele B,
- et al
- Olivot JM,
- Mlynash M,
- Inoue M,
- Marks MP,
- Wheeler HM,
- Kemp S,
- et al
- Ahn SH,
- d’Esterre CD,
- Qazi EM,
- Najm M,
- Rubiera M,
- Fainardi E,
- et al
- Lanzman RS,
- Kröpil P,
- Schmitt P,
- Bi X,
- Gliem M,
- Miese FR,
- et al
- Chng SM,
- Petersen ET,
- Zimine I,
- Sitoh YY,
- Lim CCT,
- Golay X.