CC BY-NC-ND
0Accepted Manuscript
High Resolution Three Dimensional Intracranial Arterial Wall Imaging at 3T Using T1 Weighted SPACE
Lei Zhang, Na Zhang, Jun Wu, Lijuan Zhang, Yanyan Huang, Xin Liu, Yiu-Cho Chung
PII: S0730-725X(15)00151-4
DOI: doi: 10.1016/j.mri.2015.06.006
Reference: MRI 8371
To appear in:
Magnetic Resonance Imaging
Received date: Revised date: Accepted date:
20 May 2014 7 May 2015 20 June 2015
Please cite this article as: Zhang Lei, Zhang Na, Wu Jun, Zhang Lijuan, Huang Yanyan, Liu Xin, Chung Yiu-Cho, High Resolution Three Dimensional Intracranial Arterial Wall Imaging at 3T Using T1 Weighted SPACE, Magnetic Resonance Imaging (2015), doi: 10.1016/j.mri.2015.06.006
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
High Resolution Three Dimensional Intracranial Arterial Wall Imaging at 3T Using T1 Weighted SPACE
•1 1 2 1 2
Lei Zhang , Na Zhang , Jun Wu , Lijuan Zhang , Yanyan Huang ,
Xin Liu1, Yiu-Cho Chung1,
1Shenzhen key laboratory for MRI, Paul C. Lauterbur Research Center for Biomedical Imaging Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China Department of Neurology, Beijing University Shenzhen Hospital, Shenzhen, China
Please send correspondence to: Yiu-Cho Chung, Ph.D., Room B-614, 1068 Xueyuan Blvd, University Town of Shenzhen, Nanshan, Xili, Shenzhen, China, Email: yc.chung@siat.ac.cn Phone: +86-755-86392285 Fax: +86-75586392299
Running title: 3D high resolution vessel wall imaging for ICA
Abstract
Objective: To study the effect of imaging parameters on the contrast of T1 weighted SPACE (Sampling Perfection with Application optimized Contrast using different angle Evolutions, a 3D TSE variant) at 3T for high resolution imaging of intracranial plaques before contrast and with post-gadolinium induced enhancement, and evaluate its relevance to patients with intracranial atherosclerosis.
Materials and Methods: Optimized parameters giving good T1 contrast between intracranial vessel wall and cerebrospinal fluid (CSF) within a specific scan time and reasonable coverage were found by simulation and validated in a healthy volunteer study. Based on the results, a clinical protocol covering the three major intracranial arteries (middle cerebral arteries, MCA, basilar arteries, BA and petrous internal carotid arteries, ICA) was developed. It was applied in ten patients diagnosed with intracranial arterial lesions. The accuracy of the technique in depicting vessel lumen was assessed by comparison to contrast enhanced MR angiography. The contrast enhancement ratios of the vessel wall/plaque identified were analyzed. Results: Simulation and volunteer study showed that using T1w-SPACE, good T1 contrast between vessel wall and CSF occurred at TR of around 1000ms using an echo train length of 21 within 10 minutes at an isotropic spatial resolution of 0.5mm. In the 10 patients, 24 plaques were identified in the various segments of the intracranial arterial system of which eight appeared normal on MR angiography. Post contrast enhancement ratio of these plaques varied from 0% up to 156%.
Conclusions: T1w-SPACE provides good T1 contrast between intracranial arterial wall and CSF with high resolution and good coverage within a clinically acceptable scan time. It can 2 / 30
depict plaques pre- and post-contrast along the vessels surrounded by CSF in the intracranial arterial system, and would be a useful tool in the clinical assessment of intracranial arterial diseases.
Key words: intracranial arterial wall; intracranial plaque; T1 weighted SPACE; plaque enhancement
1. Introduction
Cerebrovascular disease is a leading cause of death in the world [1]. Autopsy studies found that intracranial plaques and stenoses are highly prevalent in fatal stroke [2]. Conventional luminographic techniques such as CT angiography, X-ray angiography and magnetic resonance angiography (MRA) can depict luminal stenosis but lack specificity because different pathologies may result in similar luminal defects. Arterial wall imaging by MRI has been shown useful in identifying reasons leading to the vessel wall stenosis [3].
MR imaging of intracranial arterial wall is challenging because of its small dimension, tortuous courses and the cerebrospinal fluid (CSF) surrounding the vessels. Two dimensional turbo spin echo techniques (2D TSE) [4-6] achieve good in-plane resolution at the expenses of poor resolution in the slice direction and limited spatial coverage. The thick slices in 2D images lead to over-estimation of arterial wall thickness [7] too. Three-dimensional (3D) isotropic imaging of intracranial vessels is highly desirable.
Recently, Qiao et al. [8] used VISTA for intracranial arterial wall imaging at 3T and achieved an isotropic resolution of 0.5mm in 8 minutes or, with reduced SNR, 0.4mm in 7.6 minutes. Attempts to use FLAIR pulse to suppress CSF were unsuccessful. The method thus had no provision to suppress/reduce signal from CSF surrounding the intracranial arteries. Magnetization prepared inversion recovery (MPIR) 3D TSE has been used for multi-contrast imaging of intracranial wall at 7T [9]. The MPIR pulse suppresses CSF around the intracranial arterial walls and improves visualization. The technique achieved an isotropic resolution of 0.8mm and took about 11-12 minutes. The spatial resolution is lower than that of [8], and the need for an ultra-high field system limits the routine clinical use of the technique.
Good T1 contrast is essential in intracranial arterial wall imaging. Early studies showed that T1-weighted imaging of intracranial arterial wall allows identification of various pathologies
of intracranial artery diseases such as atherosclerosis [10]. A recent study found that enhanced intracranial atherosclerotic plaques correlate with plaque instability and may serve as a sign of inflammation [4]. Good T1 contrast also helps reduce CSF signal surrounding the intracranial arteries and improve delineation of intracranial arterial wall boundaries.
In this study, 3D high resolution T1 weighted SPACE (T1w-SPACE) was proposed for intracranial arterial wall imaging. We looked into the effect of various imaging parameters of T1w-SPACE on the CNR between intracranial arterial vessels and CSF based on spatial coverage and scan time constraints. The technique was then evaluated in selected subjects for its clinical relevance.
1. Materials and Methods
2. 1 Theory and Simulation
SPACE [11,12] is considered for intracranial arterial wall imaging because it has high sampling efficiency and good blood suppression properties. T1w-SPACE [13] has been used in vessel wall imaging of various vascular territories [14,15]. The SPACE sequence assumes the grey matter signal (T1/T2 = 940ms/100ms at 1.5T) to follow a tissue-specific prescribed signal evolution [11,12] which consists of a rapid exponential decay, followed by a flat portion, and then another exponential decay for the rest of the echo train. Based on the echo train length and echo spacing selected, the flip angles for the refocusing pulses are then computed using the formulation described in [16,17]. The resulting flip angles of the refocusing pulses would vary throughout the echo train.
Let f() be the signal of T1w-SPACE, then:
signal = f (TR, ETL, T1, T 2) (1)
where T1 and T2 are the MR parameters of the tissue of interest. For 3D imaging, the noise of the images is inversely proportional to the square root of slice number (SL) [18], thus: 5 / 30
SNR x signal-sfSL
In clinical imaging, scan time T and spatial coverage SL need to be carefully balanced. Using Cartesian sampling, the various imaging parameters of T1w-SPACE are related to T and SL
Here, NEX is the number of averages; PE is the number of phase encoding lines. For given PE and NEX, Eq. (1), (2) and (3) can be combined to give:
Hence, by choosing TR and ETL as independent variables and T as a constraint, imaging parameters for optimal CNR between two tissues (each with different T1/T2 values) can be found.
In the simulation, the dependence of CNR between intracranial arterial wall and CSF on TR and ETL at a given scan time was examined in the following way: ETL was varied from 17 to 65. For each ETL selected, the corresponding refocusing flip angle series were computed. Using Bloch equation simulation (MATLAB, MathWork, MA, USA) and Eq. (4), the CNR between arterial wall and CSF at different TR was computed. Preliminary experiment found that the flip-down pulse in [13] reduces image SNR and hence was not used in this study. In the simulation, the scan time T was set to 10 min based on clinical consideration. The relaxation parameters for arterial wall at 3T are unavailable from existing literature and were estimated using the method in [20,21]. The T1 value for the arterial wall was 1200ms, while that of CSF was 4300ms [22] at 3T. Assuming T2 values to be field strength independent, the T2 value for vessel wall would be 50ms [21] while its value for CSF would be 2200ms [23]. Note that the T1 and T2 values at 3T for white matter are 1084ms and 69ms respectively [24]. Other parameters used in the simulation were: echo time = 25ms, echo spacing = 5ms, NEX = 2, phase encoding lines (PE) vary between 310 and 330, depending on the ETL used.
T = TR ■ NEX ■ SL ■ PE / ETL
Simulation results of the relationship between SL and TR (see Eq. (3)) were also plotted to show the effect of TR on spatial coverage for a given scan time.
2.2 Volunteer study:
An IRB approved healthy volunteer study was used to validate the simulation results. Six healthy volunteers with informed consents were recruited. A 3T MRI system (Magnetom TIM Trio, Siemens, Germany) equipped with a 32-channel head coil was used (it has superior SNR and lower g-factor compared to a 12-channel head coil [25]). A localizer scan was first used to identify the intracranial vascular tree. High resolution 3D time of flight (TOF) was then used to obtain bright blood angiography of the intracranial arteries (see Table 1, protocol A). After that, T1w-SPACE images covering the middle cerebral arteries (MCA) were acquired with the following parameters: FOV = 165mm*165mm; matrix size = 330^320; voxel size = 0.5mm*0.52mm*0.5mm; bandwidth = 460Hz/pixel; TE = 25ms; ETL =33; 2 averages; scan time T = 10 minutes. TR was varied from 500ms (60 partitions) to 1880ms (16 partitions). No parallel imaging was used.
2.3 Protocol development
Based on the simulation results, the set of imaging parameters that gave the optimal CNR were shown in Table 1 (protocol B). To meet the need for spatial coverage in clinical application (it should cover the middle cerebral arteries, MCA, basilar artery, BA, and petrous internal carotid artery, ICA), several changes were made (longer ETL, improved spatial coverage, use of parallel imaging, see Protocol C). By slightly reducing image SNR, the imaging volume of the new protocol would cover the three major intracranial arteries comfortably in 10 minutes at an isotropic spatial resolution of 0.5mm.
2.4 Patient study:
The patient study was IRB approved. Ten patients (2 female, age from 36-71 years old; average age: 54.4 years) were randomly selected from a group of patients who had previously undergone intracranial artery examination, have symptoms of stroke, and were diagnosed with intracranial arterial stenosis of varying degree based on MRA or CTA examinations performed earlier. Informed consents were obtained from the patients for this study. These patients were scanned in the following way: Localizer was performed to locate the intracranial arteries. High resolution 3DTOF was then acquired using protocol A, followed by T1w-SPACE (protocol C). Contrast enhanced MR angiography (ceMRA) was performed (protocol D) after intravenous injection of Gd-DTPA (BeiLu pharmaceutical Co., Ltd., Beijing, China) at a dose of 0.2mmol/kg and an injection rate of about 3-4ml/s. The same T1w-SPACE protocol was repeated post-contrast 1-2 minutes after contrast administration. The imaging slabs in all three protocols were carefully positioned to ensure that the MCA, BA, and petrous ICA were covered.
2.5 Image Analysis
2.5.1 Volunteer study - SNR and CNR
In the volunteer study, SNR and CNR measurements performed directly on the thin arterial
walls would be biased due to the small ROIs and the associated partial volume effect. Since
the T1 and T2 values of vessel wall and white matter are very close, white matter was used
instead of vessel wall in the signal measurement when validating the simulation.
For each image set, white matter signals from three contiguous slices around the center of the
slab where signal appears homogenous were measured and averaged. The same was true for 8 / 30
CSF signals selected from the two ventricles. Noise was given by the standard deviation of the image background. Simulation data were linearly scaled to facilitate comparison with experimental results.
2.5.2 Patient study A. Lumen area measurement
For each patient, pre-contrast T1w-SPACE images were used to identify vessel stenoses and their locations. Pre-contrast T1w-SPACE and ceMRA images were then co-registered on a 3D post-processing workstation (Syngo Fusion, Siemens, Germany). After registration, multi-planar reconstruction (MPR) images perpendicular to the long axis of the diseased vessels were generated at the location of each stenosis from the two datasets. Based on the T1w-SPACE images, plaques were identified and cross-sectional images at these sites were generated from both T1w-SPACE and ceMRA images (throughout this study, a plaques was defined as a site with vessel wall thickening or enhancement after contrast injection). In addition, two sites from the apparent normal segments of MCA, BA and petrous ICA were selected, and cross-sectional images were obtained from T1w-SPACE and ceMRA images for lumen area comparison. Lumen area measurement was performed on these images using post-processing software (VesselMass, Medis Specials, Netherlands) on a workstation. Two readers (L.Z. and N.Z., both with more than 4 years of MRI experience) performed the measurements on ceMRA and T1w-SPACE independently. Evaluation of ceMRA and SPACE images were repeated at least two weeks or longer to minimize recall bias. Both readers were unaware of each other's results and patients' clinical information.
Statistical analyses were performed using SPSS (SPSS Inc., version 17, Chicago, IL). Pearson correlation was used to assess the agreement of the measurements of lumen area obtained from pre-contrast Tlw-SPACE and ceMRA. Scatter plot was used to show the relation between lumen areas measured by pre-contrast Tlw-SPACE and ceMRA for each reader. Intraclass correlation coefficient (ICC) was used to evaluate inter-observer variability for both ceMRA and Tlw-SPACE.
B. Contrast enhancement
For each patient, 3D images from pre-contrast Tlw-SPACE and post-contrast Tlw-SPACE were registered to each other (Syngo Fusion, Siemens, Germany) for signal enhancement ratio measurement. After registration, locations where signal appeared enhanced in the post-contrast Tlw-SPACE were visually identified. The cross-sections of the vessels at these locations were generated using multi-planar reconstructions (MPR). ROI was drawn at the signal enhancement region on the post-contrast Tlw-SPACE image and the mean signal was obtained (Splaque post). The same measurement was performed on the corresponding pre-contrast Tlw-SPACE image (Splaque pre). The signal intensities of the plaque were normalized by the signal intensities of the nearby grey matter (an area of around 30mm2) in the pre-contrast (SGM pre) and post-contrast (SGM post) images. Contrast enhancement ratio (CE) of the plaque
was determined by [26]. :
f \ S • S
plaquepost GM pre ^
S • S
^ plaquepre GM post j
l00% (5)
3. Results:
3.1 Simulation
Figure 1(a) shows the refocusing flip angles series for ETL = 33 and 60, and the corresponding signal evolutions of vessel wall (VW) and CSF. The flip angles decreased rapidly to establish a pseudo steady state for grey matter at the start of the echo train. The contrast between arterial wall and CSF was high for short ETL, and this contrast was reversed towards the end of the echo train. When ETL increased, more echoes had reversed contrast.
The solid lines in Figure 1(b) show the simulation results of the contrast between vessel wall and CSF for a 10 min protocol. In the figure, (1) optimal ETL was about 21 beyond which CNR decreased with increasing ETL; (2) optimal TR at ETL=21 was about 1000ms; (3) the CNR curves did not have a pronounced peak - that is, the CNR decreased for using a TR slightly different from the optimal value was very modest; (4) the CNR decrease was also modest when ETL was slightly increased from its optimal value. Eq. (3) was plotted as dashed lines in Figure 1(b). The figure showed that the slice number decreased with increasing TR, and increased with increasing ETL. Minor change in TR could increase spatial coverage more effectively at short TR than at long TR.
From the simulation results, it was also found that within a 10 minute scan time, increasing ETL (and hence sampling efficiency) from 21 to 35 would improves spatial coverage to 30 slices, and the SNR penalty was only 10.2%.
3.2 Volunteer Study
Figure 2(a) shows how signals of white matter and CSF, and CNR between the two tissues evolve with TR from the healthy volunteer study after taking the number of slices into account. The experimental results matched well with simulation. The SNR and CNR variances observed were most likely due to the coil sensitivity variation in the imaging slices 11 / 30
among different volunteers. The figure also shows that the signal behavior of vessel wall tracks that of white matter very well in simulation. Hence, imaging parameters optimized for white matter to CSF contrast would apply equally well to arterial wall to CSF contrast.
Figure 2(b) shows visually how the contrast between MCA vessel and CSF varies with TR from one healthy volunteer at the same window level. As expected, the SNR improved when TR increased. Beyond a TR of ll60ms (the optimal TR for ETL = 33 in this case), the signals of white matter/arterial wall and CSF both increased, but the contrast between the two decreased.
In the simulation, white matter was used instead of vessel wall in SNR and CNR measurement because the Tl and T2 values of the two tissues are very close. Experimental results in Figure 2 showed that this approximation can be justified.
3.3 Patient Study
All patients underwent the examination successfully. In all patients, the imaging area covered the three main intracranial arteries (MCA, BA and petrous ICA) and the V4 segment of the vertebral arteries.
A. Lumen area measurement
24 plaques on l6 vessels were identified from the Tlw-SPACE images. The lumen areas of 20 plaques were included in the study. Four plaques which occluded the vessels totally were excluded. 28 sectional images were obtained from the remaining l4 normal vessels. A total of
48 lumen areas were included for comparison. Figure 3 shows the correlation of the lumen area measurement between pre-contrast T1w-SPACE and ceMRA by reader 1. The correlation coefficients (R) for readers 1 and 2 were 0.96 and 0.92 (P<0.001 in both cases) respectively. Inter-observer ICC for lumen area measurements of T1w-SPACE and ceMRA were 0.88 and 0.83 for the two readers respectively. The results suggested that blood signal in pre-contrast T1w-SPACE was well suppressed, and this allowed for accurate observation of the lumen.
B. Contrast enhancement
Image quality of T1w-SPACE was consistently good among the patients. The vessel walls of the major arteries were well depicted in all cases. In total, 24 plaques (11 at MCA, 8 at ICA, 2 at BA, 3 at vertebral artery) were identified in the ten patients. Among the 24 plaques, 8 of them did not show luminal narrowing in MRA (in either 3DTOF or ceMRA). Three plaques had diffuse wall thickening in T1w-SPACE (see figure 5b). Five other plaques missed in MRA had area stenosis less than 30% (measurement based on T1w-SPACE images using the method in [27]). The contrast enhancement ratio of these plaques varied from 0% (no enhancement) to 156%.
Figure 4 shows one clinical case where the ceMRA images did not show apparent vessel narrowing along the supraclinoid segment of left ICA. However, eccentric narrowing of the vessel was clearly seen in the pre-contrast and post-contrast T1w-SPACE images.
The importance of spatial coverage in detecting multiple arterial wall abnormalities is demonstrated in Figure 5. The MIP of the ceMRA images showed severe narrowing of the left ICA and occlusion of the left MCA. Close examination of the pre- and post-contrast T1w-SPACE images found:
1. Complete occlusion of the Ml segment of the left MCA in all three image sets (Figure 5(c) dashed inserts). The wall of the vessel in that vicinity was also thickened and enhanced (Figure 5(c), solid inserts). The attenuation of the CSF signal around this MCA segment greatly facilitated the depiction of the arterial wall and its enhancement;
2. Diffuse thickening of the arterial wall at the Ml segment of the right MCA. The artery appeared normal in ceMRA (see Figure 5(b)-i);
3. Narrowing of the petrous ICA and the reduced signal in that vessel in ceMRA. The pre- and post-contrast Tlw-SPACE revealed that the vessel segment had circumferential wall thickening (Figure 5(d-ii) and 5(d-iii)). The vessel showed no stenosis in ceMRA. The residual blood signal in Tlw-SPACE images suggested that blood flow at that vessel segment was slow, resulting in reduced signal dephasing in Tlw-SPACE.
Another interesting case showing the importance of Tl contrast in intracranial arteries imaging is shown in Figure 6. In this patient, ceMRA showed the irregularities of the vertebral arteries at the right and a severe stenosis on the left (Figure 6a). Pre-contrast Tlw-SPACE showed that there was diffuse atherosclerosis along the right artery and a tight stenosis on the left side. Upon contrast injection, the fibrous cap of the plaque was enhanced by 80.9% in Tlw-SPACE while the signal at its core was enhanced by only ll.5%. This pattern was consistent with previous observations using 2D TSE [28] and seemed to suggest the presence of lipid core in the plaque.
4. Discussion
This study looked into the properties of Tlw-SPACE and its relevance to 3D high resolution imaging of intracranial arteries with and without use of gadolinium contrast. Based on its
properties, a clinically useful protocol that could cover the MCA, BA and petrous ICA in 10 minutes was developed and evaluated in patients diagnosed with intracranial arterial lesions.
The T1w-SPACE technique proposed here improves upon the two existing three dimensional approaches for intracranial arterial wall imaging. Our technique uses a short TR to attenuate CSF. Compared to MPIR 3D TSE [29], T1w-SPACE reduces the significant SNR penalty associated with the MPIR preparation pulse and could therefore achieve a higher spatial resolution than [29] even at 3T. Long TR in VISTA favors SNR of both the arterial wall and CSF but introduces some PD contrast between vessel wall and CSF (Figure 2(b)) [30]. Meanwhile, the short TR in T1w-SPACE attenuates CSF signal and improves delineation of vessel wall of MCA surrounded by CSF. T1w-SPACE achieves similar spatial resolution as VISTA with a longer scan time. A short ETL in T1w-SPACE also helps improve SNR and CNR, which is much needed for high spatial resolution imaging. More work would be needed to identify the effectiveness of the two techniques in identifying gadolinium induced T1 contrast in intracranial walls/plaques.
In all the images, vessel wall to CSF contrast was better around the BA segment than in the MCA. We found in the volunteer study that the SNR of CSF around the basilar arteries was only about 30% - 40% of that around the MCA, probably due to the motion induced signal dephasing of CSF around the basilar artery region [31]. To avoid this effect on the CNR measurements in the volunteer study, signal measurements were all performed around the MCA region where the CSF signal was least influenced by flow. Hence, the T1 contrast (or 15 / 30
CNR) predicted in the simulation was the "worst case" scenario: for instance, basilar arteries immersed in CSF were usually well delineated with no difficulty.
The close correlation of luminal area measured using pre-contrast Tlw-SPACE and ceMRA in the patient study showed that blood suppression for the major branches of the intracranial arteries was good among patients in Tlw-SPACE. In Tlw-SPACE, blood nulling is caused by flow induced spin dephasing and blood signal would be attenuated by more than 75% for intracranial blood flow velocity exceeding 5cm/s [32]. As the cerebral blood flow velocity is about 20cm/s - 80cm/s in healthy subjects [33], blood signal at the BA, petrous ICA and major segments of MCA would be dephased with no need for increased TE or blood suppression pulse in this application.
The clinical Tlw-SPACE protocol used (Protocol C in Table l) was designed around the simulation results and volunteer experiments. As shown in Fig l(b), CNR between vessel wall and CSF is optimal at ETL=2l and TR=l000ms, but it would cover only a l0mm slab (20 partitions, each with thickness of 0.5mm). To ensure the imaging slab includes the three major intracranial arteries, a more than 30mm thick slab is required without significantly affecting scan time and SNR. Based on the observation of Figure l(b), three changes were made to Protocol B:
l. ETL was increased from 2l to 35 (sampling efficiency increased by more than 50%), resulting in a slight compromise in SNR of about l0.2% but increased the number of slices to 30;
2. Parallel imaging using GRAPPA rate 2 was used to increase the number of slices covered while keeping scan time constant. As a result, the number of slices increased from 30 to 60 while SNR remained practically unchanged.
3. TR was slightly reduced to around 938ms, representing a further 2.1% SNR reduction. The final protocol covered 64 slices in 10 minutes;
This protocol reduced the overall SNR by about 12% but covered the three major intracranial arteries and the V4 segment of the vertebral arteries comfortably. The clinical protocol in this study is by no means unique. For instance, partial Fourier may be used to reduce scan time or increase spatial coverage. The simulation results provide a useful starting point for protocol development based on specific clinical needs.
The voxel volume from 2D TSE commonly used in intracranial artery imaging [6] is comparable to T1w-SPACE in this study (~0.125mm3). However, the partial volume effect [7] in 2D TSE along the slice direction (especially in the torturous vessels) would hinder the detection of lesion morphology. 3D TSE techniques have isotropic spatial resolution and ideal voxel profile, and are more favorable in the depiction of plaques along the vessel wall. Interpolation can help reduce partial volume effect and improve vessel wall delineation further [34]. Though the spatial resolution from T1w-SPACE (as well as other 3D techniques) is still insufficient for vessel wall thickness quantification, the close correlation between T1w-SPACE and ceMRA in measuring luminal area showed that T1w-SPACE can reliably detect plaque along the vessel wall. 17 / 30
In this small patient study, 8 plaques identified from post-contrast Tlw-SPACE images were missed by MR angiography. Arterial expansive remodeling is one reason for this discrepancy (Figure 4). Another reason is that diffuse wall thickening cannot be detected by MR lumenographic techniques (Figure 5b). The existence of plaque undetected by MRA coincides with the observation of an autopsy study [2]. Tlw-SPACE would be useful in identifying plaques missed in MRA. More study would be needed.
Recent studies on intracranial arterial wall imaging showed that enhanced intracranial atherosclerotic plaques correlate with plaque instability and might be a sign of inflammation [30,35]. Another study reported the relationship between patients with acute ischemic stroke and intracranial arterial wall/plaque enhancement [36]. The small patient study here demonstrated the ability of Tlw-SPACE to depict intracranial arterial wall and plaque enhancement post-contrast. The case from Figure 6 suggests that the Tl weighed technique may have the potential to identify lipid core in intracranial plaque (histology would be needed to verify this hypothesis). We believe that the good Tl contrast and 3D spatial coverage of Tlw-SPACE will be a good tool in the study of intracranial plaque composition and clinical relevance of arterial wall enhancement.
In conclusion, the contrast properties of Tlw-SPACE were study and optimized imaging parameters were found. The technique gives good Tl contrast, enough SNR and sufficient coverage needed to include the three major intracranial arteries in one scan in a reasonable 18 / 30
scan time. Initial experience with this technique shows that it will be a useful technique for the imaging of intracranial plaques.
Acknowledgements
This work is funded by grants No. 2013CB733800 / 2013CB733803, 81470077, 81301216, JCYJ20140417113430589, JSGG20141020103440414. We also thanked the reviewers for their constructive comments.
References
[1] Murphy SL, Xu JQ, KD K. Death: Preliminary Data for 2010. National Vital Statistics Report; 2012. p. 1-51
[2] Mazighi M, Labreuche J, Gongora-Rivera F, Duyckaerts C, Hauw JJ, Amarenco P. Autopsy prevalence of intracranial atherosclerosis in patients with fatal stroke. Stroke 2008;39:1142-7.
[3] Bodle JD, Feldmann E, Swartz RH, Rumboldt Z, Brown T, Turan TN. High-Resolution Magnetic Resonance Imaging An Emerging Tool for Evaluating Intracranial Arterial Disease. Stroke 2013;44:287-92.
[4] Swartz RH, Bhuta SS, Farb RI, Agid R, Willinsky RA, Terbrugge KG, et al. Intracranial arterial wall imaging using high-resolution 3-tesla contrast-enhanced MRI. Neurology 2009;72:627-34.
[5] Xu WH, Li ML, Gao S, Ni J, Zhou LX, Yao M, et al. In vivo high-resolution MR imaging of symptomatic and asymptomatic middle cerebral artery atherosclerotic stenosis. Atherosclerosis 2010;212:507-11.
[6] Ryu CW, Jahng GH, Kim EJ, Choi WS, Yang DM. High resolution wall and lumen MRI of the middle cerebral arteries at 3 tesla. Cerebrovasc Dis 2009;27:433-42.
[7] Antiga L, Wasserman BA, Steinman DA. On the overestimation of early wall thickening at the carotid bulb by black blood MRI, with implications for coronary and vulnerable plaque imaging. Magn Reson Med 2008;60:1020-8.
[8] Qiao Y, Steinman DA, Qin Q, Etesami M, Schär M, Astor BC, et al. Intracranial arterial wall imaging using three-dimensional high isotropic resolution black blood MRI at 3.0 Tesla. J Magn Reson Imaging 2011;34:22-30.
[9] Van der Kolk AG, Hendrikse J, Brundel M, Biessels GJ, Smit EJ, Visser F, et al. Multi-sequence whole-brain intracranial vessel wall imaging at 7.0 tesla. Eur Radiol 2013;23:2996-3004.
[10] Klein IF, Lavallee PC, Touboul PJ, Schouman-Claeys E, Amarenco P. In vivo middle cerebral artery plaque imaging by high-resolution MRI. Neurology 2006;67:327-9.
[11] Mugler JP 3rd, Kiefer B, JR. B. Three-dimensional T2-Weighted imaging of the brain 19 / 30
using very long spin-echo trains. In: Proceedings of the 8th Annual Meeting of ISMRM. Denver, CO, USA; 2000: p. 687.
[12] Mugler JP 3rd, JR. B. Efficient spatially selective single-slab 3D turbo-spin-echo imaging. In: Proceedings of the l2th Annual Meeting of ISMRM. Kyoto; 2004: p. 695.
[13] Park J, Mugler JP, 3rd, Horger W, Kiefer B. Optimized Tl-weighted contrast for single-slab 3D turbo spin-echo imaging with long echo trains: application to whole-brain imaging. Magn Reson Med 2007;58:982-92.
[ l4] Mihai G, Chung YC, Merchant A, Simonetti OP, Raj agopalan S. Tl -weighted-SPACE dark blood whole body magnetic resonance angiography (DB-WBMRA): initial experience. J Magn Reson Imaging 20l0;3l:502-9.
[15] Chung YC, Shea S, Qiao Y, Simonetti OP, Wasserman BA. High Resolution 3D Intracranial Imaging at 3.0T. In: Proceedings of the l2th Annual Meeting of ISMRM. Stockholm, Sweden,; 20l0: p. 2255.
[16] Busse RF, Hariharan H, Vu A, Brittain JH. Fast spin echo sequences with very long echo trains: design of variable refocusing flip angle schedules and generation of clinical T2 contrast. Magn Reson Med 2006;55: l030-7.
[17] Hennig J, Weigel M, Scheffler K. Multiecho sequences with variable refocusing flip angles: optimization of signal behavior using smooth transitions between pseudo steady states (TRAPS). Magn Reson Med 2003;49:527-35.
[18] Haccke EM, Brown RW, MR T. Magnetic resonance imaging: physical principles and sequence design: New York: John Wiley & Sons; l999.
[20] Brown R, Nguyen TD, Spincemaille P, Cham MD, Choi G, Winchester PA, et al. Effect of blood flow on double inversion recovery vessel wall MRI of the peripheral arteries: quantitation with T2 mapping and comparison with flow-insensitive T2-prepared inversion recovery imaging. Magn Reson Med 20l0;63:736-44.
[21] Xie J, Bi X, Fan Z, Bhat H, Shah S, Zuehlsdorff S, et al. 3D flow-independent peripheral vessel wall imaging using T(2)-prepared phase-sensitive inversion-recovery steady-state free precession. J Magn Reson Imaging 20l0;32:399-408.
[22] Rooney WD, Johnson G, Li X, Cohen ER, Kim SG, Ugurbil K, et al. Magnetic field and tissue dependencies of human brain longitudinal lH2O relaxation in vivo. Magn Reson Med 2007;57:308-l8.
[23] Park J, Park S, Kim EY, Suh JS. Phase-sensitive, dual-acquisition, single-slab, 3D, turbo-spin-echo pulse sequence for simultaneous T2-weighted and fluid-attenuated whole-brain imaging. Magn Reson Med 20l0;63:l422-30.
[24] Stanisz GJ, Odrobina EE, Pun J, Escaravage M, Graham SJ, Bronskill MJ, et al. Tl, T2 relaxation and magnetization transfer in tissue at 3T. Magn Reson Med 2005;54:507-l2.
[25] Reiss-Zimmermann M, Gutberlet M, Kostler H, Fritzsch D, Hoffmann KT. Improvement of SNR and acquisition acceleration using a 32-channel head coil compared to a l2-channel head coil at 3T. Acta Radiol 20l3;54:702-8.
[26] Lou X, Ma N, Shen H, Shi K, Jiang W, Ma L. Noninvasive visualization of the basilar artery wall and branch ostia with high-resolution three-dimensional black-blood
sequence at 3 tesla. J Magn Reson Imaging 2014;39:911-6.
[27] Klein IF, Lavallee PC, Touboul PJ, Schouman-Claeys E, Amarenco P. In vivo middle cerebral artery plaque imaging by high-resolution MRI. Neurology 2006 25;67:327-9.
[28] Turan TN, Rumboldt Z, Brown TR. High-resolution MRI of basilar atherosclerosis: three-dimensional acquisition and FLAIR sequence. Brain Behav 2013;3:1-3.
[29] van der Kolk AG, Zwanenburg JJ, Brundel M, Biessels GJ, Visser F, Luijten PR, et al. Intracranial vessel wall imaging at 7.0-T MRI. Stroke 2011;42:2478-84.
[30] Qiao Y, Zeiler SR, Mirbagheri S, Leigh R, Urrutia V, Wityk R, et al. Intracranial Plaque Enhancement in Patients with Cerebrovascular Events on High-Spatial-Resolution MR Images. Radiology 2014;271:534-42.
[31] Busse RF. Flow Sensitivity of CPMG Sequences with Variable Flip Refocusing and Implications for CSF Signal Uniformity in 3D-FSE Imaging. In: Proceedings of the 14th Annual Meeting of ISMRM. Seattle, Washington, USA; 2006: p. 2430.
[32] Park J, Kim EY. Contrast-enhanced, three-dimensional, whole-brain, black-blood imaging: application to small brain metastases. Magn Reson Med 2010;63:553-61.
[33] Bammer R, Hope TA, Aksoy M, Alley MT. Time-resolved 3D quantitative flow MRI of the major intracranial vessels: initial experience and comparative evaluation at 1.5T and 3.0T in combination with parallel imaging. Magn Reson Med. 2007;57:127-40.
[34] Du YP, Parker DL, Davis WL, Cao G. Reduction of partial-volume artifacts with zero-filled interpolation in three-dimensional MR angiography. Journal of magnetic resonance imaging : J Magn Reson Imaging 1994;4:733-41.
[35] Vergouwen MD, Silver FL, Mandell DM, Mikulis DJ, Swartz RH. Eccentric narrowing and enhancement of symptomatic middle cerebral artery stenoses in patients with recent ischemic stroke. Arch Neurol 2011;68:338-42.
[36] Skarpathiotakis M, Mandell DM, Swartz RH, Tomlinson G, Mikulis DJ. Intracranial atherosclerotic plaque enhancement in patients with ischemic stroke. AJNR Am J Neuroradiol 2013;34:299-304.
Table legends:
Table 1: Protocols used in healthy volunteer and patient studies
Figure legends:
Figure 1: Simulated signal and contrast change between vessel wall (VW) and CSF in
T1w-SPACE for different imaging parameters. (a) The refocusing flip angle series for ETL =
33, TR = 938ms and ETL = 60, TR = 1000ms. The signal behaviors of vessel wall and CSF
along the echo train are also shown. (b) The contrast between vessel wall and CSF, the slice
number as a function of TR for 4 different ETLs (17, 21, 33, 65).
21 / 30
Figure 2: Dependence of signals and contrast from vessel wall, white matter and CSF on TR. (a) How the SNR and CNR between WM and CSF change with TR from simulation and experiment at a scan time of l0 minutes and ETL=33. The experimental results matched simulation results well. Highest CNR was achieved when TR = ll60ms, (b) the dependence of contrast (between WM and CSF) on TR in-vivo. The image corresponding to the optimal CNR was highlighted in dashed box.
Figure 3: Scatter plot of the lumen area comparison between ceMRA and pre-contrast Tlw-SPACE, Very good agreement between the two methods was found.
Figure 4. A patient with eccentric wall thickening at the supraclinoid segment of ICA. (a) ceMRA showed no apparent narrowing along the vessel. A cross section of the vessel showed a subtle irregularity on the vessel wall; (b) pre-contrast Tlw-SPACE showed the obvious eccentric thickening of the vessel wall due to a plaque, suggestive of expansive remodeling. The plaque was most obvious in the axial view - see insert; (c) contrast enhanced Tlw-SPACE showed localized (albeit subtle) enhancement of the vessel wall.
Figure 5: A patient with several vascular abnormalities in the intracranial arterial tree. (a) The ceMRA showed the stenosis of the left internal carotid artery segment (the arrow); (b) Tlw-SPACE acquired at pre- (ii) and post-contrast (iii) revealed diffuse wall thickening of an otherwise normal lumen as observed by ceMRA (i). The plaque was enhanced by 38.3%. (c) Complete occlusion of the Ml segment of the left MCA (the dotted line and the dotted insert). Wall thickening and contrast enhancement was observed at a nearby vessel (solid line and solid insert). The plaque was enhanced by 47.7%. Note that attenuation of CSF signal around the vessel help the depiction of the vessel wall. (d) The stenosed segment of the left internal carotid artery was found to have concentric wall thickening. The residual signal at the vessel lumen may be attributed to the slow blood flow at that vessel segment. The plaque was enhanced by ll0.4%.
Figure 6: A patient with a tight stenosis at the left vertebral artery. (a) ceMRA shows luminal irregularity along the right vertebral artery, and a stenosis at the left vertebral artery. (b) Plaques shown in pre-contrast Tlw-SPACE correspond to those shown in ceMRA. Note the eccentric plaque at the right vertebral artery (see dashed insert). (c) Post-contrast Tlw-SPACE shows the enhanced fibrous cap (dashed arrow). The plaque is isointense, suggestive of a lipid core (solid arrow).
•• FA:ETL=33 .......FA:ETL=60
—VW:ETl=33 - - VW:ETL=60 — CSF:ETL=33 - - CSF:ETL=60
0.1 8 s
-CNR: ETL=17 -CNR: ETl=21
- - Slices:ETL=17 - - Slices:ETL=21
-CNR: ETL=33 -CNR: ETL=65
Slices:ETL=33 - - Slices:ETL=65
RF pulse number
Figure 1
Figure 2
10 15 20
SPACE (mm2)
Figure 3
Figure 5
Table 1: Protocols used in healthy volunteer and patient studies
A B C D
3DTOF T1w-SPACE (optimal) T1w-SPACE (clinical) ceMRA
FOV (mm2) 180x137.8 165x162.9 165x162.9 ^ 221x151.9
Base matrix 512 320 320 448
PE* lines 314 315 326 262
Phase resolution 80% 100% 100% 85%
Partitions/Slices 28/slab, 5 slabs 20 64 88
Partition 78% 100% 100% 71%
resolution
TR / TE 24ms / 3.91ms 1000ms / 25ms 938ms / 25ms 3.57ms / 1.4ms
Flip angle 18o Variable Variable 21o
GRAPPA / 2 / 24 No / - 2 / 24 2 / 24
reference lines
Bandwidth/pixel 238 460 460 590
Echo train 1 21 35 1
Voxel size (mm3) 0.44x0.35x0.46 0.52x0.52x0.52 0.5x0.52x0.52 0.58x0.49x0.55
Partial Fourier PE:6/8; PA§:6/8 No No PE:6/8; PA§:6/8
Scan time (min) 6.3 10.0 10.0 24s
§ Partition encoding
Table legends:
Table 1: Protocols used in healthy volunteer and patient studies
