Intracranial Vessel Wall Imaging at 7.0-T MRI
Background and Purpose—Conventional imaging methods cannot depict the vessel wall of intracranial arteries at sufficient resolutions. This hampers the evaluation of intracranial arterial disease. The aim of the present study was to develop a high-resolution MRI method to image intracranial vessel wall.
Methods—We developed a volumetric (3-dimensional) turbo spin-echo (TSE) sequence for intracranial vessel wall imaging at 7.0-T MRI. Inversion recovery was used to null cerebrospinal fluid to increase contrast with the vessel wall. Magnetization preparation was applied before inversion to improve signal-to-noise ratio. Seven healthy volunteers and 35 patients with ischemic stroke or transient ischemic attack underwent imaging to test the magnetization preparation inversion recovery TSE sequence. Gadolinium-based contrast agent (Gadobutrol, 0.1 mL/kg) was administered to assess possible lesion enhancement in the patients.
Results—The walls of intracranial arterial vessels could be visualized in all volunteers and patients with good contrast between wall, blood, and cerebrospinal fluid. The quality of the vessel wall depiction was independent of the vessel orientation relative to the plane of acquisition. In 21 of the 35 patients, a total number of 52 intracranial vessel wall lesions were identified. Eleven of the 52 lesions showed enhancement after contrast administration. Only 14 of the 52 lesions resulted in stenosis of the arterial lumen.
Conclusions—Intracranial vessel wall and its pathology can be depicted with the magnetization preparation inversion recovery TSE sequence at 7.0 T. The magnetization preparation inversion recovery TSE sequence will make it possible to study the role of intracranial arterial wall pathology in ischemic stroke.
Clinical Trial Registration Information—URL: http://www.trialregister.nl/trialreg/index.asp. Unique identifier: NTR2119.
For diagnosing intracranial arterial wall pathology, like atherosclerosis or Moyamoya disease, an imaging technique is needed to visualize the intracranial vessel wall. With conventional intra-arterial digital subtraction angiography or MRA methods, intracranial vessel wall pathology only becomes visible when it gives rise to luminal narrowing. However, it is known from extracranial atherosclerosis that because of remodeling of the arteries, lumen diameter can be maintained despite progressive wall thickening.1 Luminography-based methods therefore may underestimate the presence of intracranial arterial pathology.2–8 Another factor that complicates the detection of intracranial arterial pathology with current imaging methods is the small diameter of the intracranial arteries, which ranges from 2 to 3 mm proximally to <1 mm more distally.
An imaging method that is used for the detection of arterial vessel wall abnormalities also should preferably be able to depict the normal vessel wall. This facilitates the differentiation between healthy and abnormal vessels. With conventional MRI techniques only pathological conditions, such as enhancement of the intracranial vessel wall with vasculitis, can be detected.9–12 Only 5 studies to date have attempted to visualize both healthy and diseased intracranial vessel wall, all performed on 3.0-T MRI. They clearly showed abnormal intracranial vessel walls, but it proved difficult to depict the healthy vessel wall, mainly because of lack of contrast with surrounding tissues and cerebrospinal fluid.13–17
The aim of the present study was to develop an MRI method to image the vessel wall of intracranial arteries, also in the absence of disease. To allow for high-resolution imaging with sufficient sensitivity, MRI was performed at a magnetic field strength of 7.0 T. The developed MR sequence was tested in a pilot study with healthy volunteers and patients with an increased chance of having intracranial vessel wall pathology.
Materials and Methods
Intracranial Vessel Wall Sequence
Imaging was performed on a 7.0-T whole-body MRI scanner (Philips Healthcare) with a 16-channel receive coil and volume transmit/receive coil for transmission (Nova Medical). We developed a volumetric (3-dimensional) inversion recovery turbo spin-echo (TSE) sequence, in which the inversion pulse was used to null the cerebrospinal fluid for contrast with the vessel wall. Black blood was obtained because of flow between excitation and refocusing in the TSE train. A dedicated refocusing train with low and varying refocusing pulse angles was applied. The pulse angles for the train were calculated according to the approach described by Busse et al,18 which is optimized to obtain a constant signal response during the readout train and, hence, a sharp point-spread function, with low refocusing angles. The angles were computed, given the echo-spacing of 4.7 ms and train length of 116 echoes, and using a T1/T2 of 2500/50 ms, a minimum refocusing angle of 15 degrees, and a maximum angle of 90 degrees. Low refocusing angles lead to a high flow sensitivity, which contributed to the black blood appearance. Furthermore, the low refocusing angles yielded a limited specific absorption rate of the refocusing train. Because of the varying orientation of the vessels and branches of the circle of Willis, isotropic voxels were used to ensure quality of vessel wall depiction independent of vessel orientation. Finally, to improve the signal-to-noise ratio of the vessel wall, magnetization preparation was applied before the inversion pulse, leading to saturation of tissues with short T2 compared to the T2 of cerebrospinal fluid.19 This yields saturation recovery instead of inversion recovery for these tissues and, hence, more signal at the moment of acquisition.
For this magnetization preparation inversion recovery (MPIR) TSE sequence, the following scan parameters were used: field of view 220×180×13 mm3 in transverse orientation, acquired resolution 0.8×0.8×0.8 mm3 (0.5 μL), echo train length (TSE factor) of 116 (including 4 start-up cycles), repetition time 6050 ms, inversion time 1770 ms, echo time 23 ms, magnetization preparation mixing time 250 ms, no SENSE was applied, and 2 averages were acquired to avoid free induction decay (FID) artifacts. The scan duration was ≈12 minutes.
For possible depiction of lesion activity, we administered 0.1 mL/kg of a gadolinium-containing contrast agent (Gadobutrol; Gadovist 1.0 mmol/mL; Bayer Schering Pharma) to all patients and obtained vessel wall information before and ≈5 minutes after contrast administration.
For confirmation of the observed vessels seen on the MPIR-TSE images, a 3-dimensional time-of-flight MRA by means of fast field-echo sequence was added to the scan protocol with the following parameters: field of view 180×180×110 mm3 in transverse orientation, acquired resolution 0.4×0.5×0.6 mm3, repetition time 22 ms, echo time 2.5 ms, flip angle 25 degrees, receiver bandwidth 202 Hz/pixel, and scan duration ≈10 minutes.
Offline Postprocessing and Assessment
MPIR-TSE images were analyzed on an offline workstation (Philips) by 2 observers who sought to identify the vessel walls of the major arteries of the circle of Willis and its branches. The 3-dimensional format with isotropic resolution allowed multiple reformatting depending on local vessel orientation, which was used for assessment of both vessel wall and possible vessel wall lesions. Time of flight MRA data were used to identify the observed vessels on the MPIR-TSE images. For assessment of contrast enhancement of the (local) vessel wall, the postgadolinium MPIR-TSE scans were registered to the precontrast scans using rigid mutual information registration20 with partial volume interpolation (32 bins; sampling factors: x=4, y=4, z=1). After coregistering, precontrast and postcontrast registered images were subtracted to obtain only those areas with contrast enhancement. After obtaining the subtracted images, both precontrast and postcontrast scans as well as the subtracted images were put in a viewing format that allowed for direct comparison between enhancing areas and concomitant vessel wall anatomy. To see whether normal contrast enhancement had taken place, the infundibulum was assessed for enhancement on the subtracted images.
Our sequence was tested on healthy volunteers and patients with a high chance of intracranial vessel wall pathology. This prospective study was approved by the Institutional Review Board of the University Medical Center Utrecht, the Netherlands. All subjects gave written informed consent. Healthy volunteers without known cardiovascular disease were recruited by advertisement posters in the University Medical Center Utrecht in November and December 2009. Hereafter, all consecutive patients presenting with arterial ischemic stroke or transient ischemic attack (TIA) of the anterior cerebral circulation at the neurology ward of the University Medical Center Utrecht between December 2009 and January 2011 were screened for inclusion in the present study. Patients who were unable to endure the MRI examination because of their clinical condition, healthy volunteers and patients with contraindications for 7.0-T MRI (claustrophobia, metal objects in or on the body), and patients with a known allergic reaction to gadolinium-containing contrast agents or impaired renal function were excluded.
Seven healthy volunteers (3 males; mean age, 30 years; range, 20–55 years) without known cardiovascular disease were included in the study and scanned with the MPIR-TSE sequence. Vessel wall of the distal internal cerebral artery, basilar artery, M1 segment of the middle cerebral artery, A1 segment of the anterior cerebral artery, and P1 segment of the posterior cerebral artery could be visualized along their complete trajectories in all volunteers (Supplemental Figure I, http://stroke.ahajournals.org). Also, in most subjects the vessel wall of the smaller A2, M2, and P2 branches of the anterior cerebral artery, middle cerebral artery, and posterior cerebral artery, respectively, could be identified and tracked along their course on the MPIR-TSE images. It was difficult to distinguish the vessel wall of these smaller branches when it was immediately adjacent to brain parenchyma (Supplemental Figure I). The quality of the vessel wall depiction was independent of the vessel orientation. No signs of gross artifacts from insufficient flow spoiling within the TSE train were observed.
Between December 2009 and January 2011, 344 patients with ischemic infarct and 267 patients with TIA of the anterior cerebral circulation were admitted to the University Medical Center Utrecht, of whom 173 patients were excluded because of MRI contraindications and 226 patients were excluded because of their clinical status. In total, 35 patients (18 males; mean age, 59 years; range, 26–83 years) were included in the study, of whom 16 had experienced an acute ischemic stroke and 19 patients had experienced a TIA. All patients were scanned within 1 week after symptom onset. In 6 patients (17%) an underlying pathology other than atherosclerosis was diagnosed clinically (Moyamoya disease, reversible vasoconstriction syndrome, other vasculopathy of unknown cause, fibromuscular dysplasia, internal cerebral artery dissection [twice]). In 4 patients no cause for TIA or stroke was found. In 3 patients vessel wall MRI examinations were of insufficient quality for vessel wall assessment because of motion artifacts.
Of the 32 remaining patients in whom vessel wall could be assessed, 21 had identifiable vessel wall lesions on the MPIR-TSE sequence in major intracranial arteries (anterior cerebral artery, middle cerebral artery, or posterior cerebral artery with their respective major branches; Table), 10 of the 17 TIA patients (59%) and 11 of the 15 stroke patients (73%). Seven of these 21 patients had only 1 vessel wall lesion; in the remaining 14 patients, intracranial lesions were detected at multiple locations (Table); in 4 patients 2 lesions were found, 5 patients showed 3 lesions, and in 5 patients >3 lesions were found, totaling 52 lesions. Lesions most often consisted of a small focal or more elongated thickening of the arterial vessel wall (Supplemental Figure II, http://stroke.ahajournals.org), sometimes of the whole vessel circumference, and causing luminal stenosis as seen on conventional imaging (time-of-flight MRA) in only 14 of 52 lesions (Figures 1, 2). Image subtractions from precontrast and postcontrast MPIR-TSE scans showed enhancement in 11 of the 52 locations of an intracranial lesion (Table, Figures 2, 3, Supplemental Figure III, http://stroke.ahajournals.org). In 14 of the 21 patients with an intracranial lesion, a lesion was present in an artery of the flow territory in which the ischemic event had occurred.
In the present study we developed the 3-dimensional MPIR-TSE MR sequence to image intracranial arterial vessel wall of normal nondiseased intracranial arteries and possible pathological arteries. With the use of 7.0-T, we achieved high image resolution and sufficient sensitivity to allow identification of vessel wall of the major arteries of the circle of Willis in all subjects, including healthy controls.
Only a few studies have thus far succeeded in imaging the intracranial arterial vessel wall using MRI.13–17 Niizuma et al13 visualized middle cerebral artery vessel wall with 3.0 T in 3 patients but not in healthy controls. Swartz et al14 found enhancement patterns of different pathologies such as atherosclerosis and inflammation of the intracranial vessel wall with 3.0-T in a patient group with more complicated pathology, which could have reduced the chance of visualizing healthy vessel wall. Ryu et al15 found multicontrast-weighted black blood on 3.0-T MRI to have the potential for characterization of atherosclerotic plaques. Li et al16 and Xu et al17 differentiated wall thickening, arterial remodeling, and atherosclerotic plaque with their 3.0-T T2-weighted sequence in symptomatic and asymptomatic patients with intracranial middle cerebral artery stenosis. They used relatively thick slices (2 mm), which had to be planned perpendicular to the vessel, which makes it difficult to image the complete circle of Willis. None of these studies used cerebrospinal fluid suppression to improve contrast between cerebrospinal fluid and vessel wall and to avoid possible misinterpretation of cerebrospinal fluid signal around a vessel for signal from the vessel wall.
We have used 7.0-T field strength for further development of an intracranial vessel wall sequence; 7.0 T yields a higher signal-to-noise ratio compared to lower field strengths, but intracranial vessel wall imaging remains challenging. The vessels of the circle of Willis do not have a single orientation, which prohibits the use of thick slices perpendicular to the vessel orientation, as is normally performed in imaging of the carotid artery wall.22 Second, inhomogeneity in the radio-frequency transmit field at 7.0 T leads to a reduced flip angle in the occipital and temporal lobes of the brain, making small, more peripherally located arteries more difficult to identify. However, this was not a limiting factor for identifying the vessels of the circle of Willis. Third, no body transmit coil is available with 7.0-T systems, so nonselective inversion as used in dual-inversion black blood sequences cannot be used. In the MPIR-TSE sequence, we used the intrinsic flow sensitivity of TSE to obtain dark blood attributable to flow between excitation and refocusing in the TSE train.
Our results show the ability of the MPIR-TSE sequence to detect lesions and healthy intracranial vessel wall. Arteries surrounded by cerebrospinal fluid, such as the major branches of the circle of Willis, are best visible because of the cerebrospinal fluid suppression applied in the sequence. This is especially useful in older individuals because atrophy of the brain causes the arteries to be surrounded by more cerebrospinal fluid. The majority (66%) of the patients in our study had ≥1 lesions in major intracranial arteries. Only 27% of the lesions (14/52) caused a stenosis, as seen on conventional time-of-flight MRA images. Hence, the majority of the lesions probably would have gone undetected by luminography-based methods. Several patients showed vessel wall lesions in an arterial territory contralateral to the side of neurological symptoms. This indicates that the lesions are not always directly related to the ischemic event. The burden of intracranial lesions probably reflects the presence and severity of more generalized arterial wall pathology23 in a patient. We believe that the presented method has potential to examine the relationship between intracranial arterial vessel wall pathology and TIA or ischemic stroke, for instance, in clinical longitudinal studies. In this regard, the MPIR-TSE sequence also could be used complementary to the current diagnostic imaging techniques for identifying intracranial vessel wall pathology, like atherosclerosis or Moyamoya disease, in a minimally invasive way.
The current study has some limitations. First, we scanned only a limited number of healthy volunteers. By scanning more subjects with no known cardiovascular diseases, we will be able to better discern pathological vessel wall changes from normal vessel wall and facilitate the need for a reference standard regarding vessel wall thickness. In the older healthy volunteer, vessel wall was more clearly visible than in the younger volunteers (Supplemental Figure I). With the current data, we are not able to judge whether this is a consequence of normal aging of the vessel wall24 or asymptomatic vessel wall pathology. Second, for everyday clinical use, 7-T MRI is still hampered by strict safety rules, which currently make the MPIR-TSE sequence only applicable to a subgroup of patients. In our study, only some of all eligible patients could be scanned because of these safety rules. We expect that with technical improvements, such as an increase in the number of receive elements in a head coil, the signal-to-noise ratio at lower field strengths may improve to such a level that the resolution needed for vessel wall imaging with MPIR-TSE or other sequences may become feasible at 1.5-T and 3.0-T MRI field strengths in the future. Third, we used an MRI sequence with small coverage (ie, a transverse slice of 13 mm) and a relatively long acquisition time of ≈12 minutes. Because of the small coverage, small arterial branches of the large circle of Willis arteries often could not be tracked completely. Enlarging the coverage results in even longer acquisition times, prolonging the time subjects have to lie completely still. However, the majority of subjects tolerated the acquisition time well, with limited motion artifacts. Moreover, even with the restricted coverage the large intracerebral arteries arising from the circle of Willis could be assessed completely.
In conclusion, intracranial vessel wall and its pathology can be depicted with MPIR-TSE imaging on 7.0-T MRI in healthy volunteers and in ischemic stroke and TIA patients. This sequence makes it possible to study the role of intracranial arterial wall pathology in stroke and TIA in more detail.
Sources of Funding
This research was performed within the framework of CTMM, the Center for Translational Molecular Medicine (www.ctmm.nl), project PARISk (grant 01 C-202), and supported by the Netherlands Heart Foundation.
Fredy Visser is an employee of Philips Healthcare, Best, the Netherlands.
The online-only Data Supplement is available at http://stroke.ahajournals.org/cgi/content/full/STROKEAHA.110.620443/DC1.
- Received March 14, 2011.
- Accepted April 7, 2011.
- © 2011 American Heart Association, Inc.
- Stiel GM,
- Stiel LS,
- Schofer J,
- Donath K,
- Mathey DG
- Kiechl S,
- Willeit J
- Bash S,
- Villablanca JP,
- Jahan R,
- Duckwiler G,
- Tillis M,
- Kidwell C,
- et al
- Kasner SE,
- Chimowitz MI,
- Lynn MJ,
- Howlett-Smith H,
- Stern BJ,
- Hertzberg VS,
- et al
- Nguyen-Huynh MN,
- Wintermark M,
- English J,
- Lam J,
- Vittinghoff E,
- Smith WS,
- et al
- Klein IF,
- Lavallée PC,
- Touboul PJ,
- Schouman-Claeys E,
- Amarenco P
- Swartz RH,
- Bhuta SS,
- Farb RI,
- Agid R,
- Willinsky RA,
- Terbrugge KG,
- et al
- Homma S,
- Hirose N,
- Ishida H,
- Ishii T,
- Araki G