First published online August 2, 2006
British Journal of Radiology (2006) 79, 962-967
© 2006 British Institute of Radiology
doi: 10.1259/bjr/23219572
Acquisition of MR perfusion images and contrast-enhanced MR angiography in acute ischaemic stroke patients: which procedure should be done first?
C W Ryu, MD1,
D H Lee, MD2,
H S Kim, MD2,
J H Lee, MD2,
C G Choi, MD2,
S J Kim, MD2 and
D C Suh, MD2
1 Department of Radiology, East-West Neomedical Center, Kyunghee University College of Medicine, Seoul, 2 Department of Radiology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea
Correspondence: Deok Hee Lee, Department of Radiology, Asan Medical Center, University of Ulsan College of Medicine, 388-1 Poongnap-2dong, Songpa-gu, Seoul, 138-736, Korea. E-mail: dhlee{at}amc.seoul.kr
 |
Abstract
|
|---|
Multimodal MRI for acute ischaemic stroke usually includes perfusion imaging (PI) and contrast-enhanced neck MR angiography (CE-MRA), as well as diffusion-weighted imaging and T2* weighted imaging. Because both PI and CE-MRA require the infusion of contrast medium, the likelihood exists that one study may conflict with the other due to the accumulation of previously injected contrast medium. The purpose of this study is to determine the appropriate order of PI and CE-MRA in this multimodal MRI protocol for evaluation of acute ischaemic stroke. We studied 35 patients with acute ischaemic stroke in the unilateral middle cerebral artery territory. 17 patients underwent CE-MRA following PI (group A) and 18 patients underwent PI following CE-MRA (group B). For qualitative analysis of the CE-MRA and colour-coded maps of the PI, two independent observers graded the image quality. Interobserver agreement was assessed using kappa statistics, and we assessed the statistical differences of imaging quality between groups A and B using the Mann-Whitney U-test). For the quantitative analysis of PI, two parameters – the maximum change in the transverse relaxation rate (
R2max) and the relative signal drop (S/S0) – were calculated from the time–signal intensity curve of an unaffected middle cerebral artery territory, and we compared the differences in the parameters of group A and B (t-test). Interobserver agreements for CE-MRA and PI were good. In the qualitative analysis of CE-MRA and PI, no significant difference was observed between groups A and B. In the quantitative analysis of PI, there were no relevant differences in R2max and S/S0 between the two groups. In simultaneous CE-MRA and PI, there was no deterioration of diagnostic imaging quality with regard to the order of the two post-contrast sequences. They can be performed according to the preference of each institution.
|
Introduction
|
|---|
Multimodal MRI, which includes diffusion-weighted imaging (DWI), perfusion imaging (PI), time-of-flight MR angiography (TOF-MRA) and contrast-enhanced neck MRA (CE-MRA), is very useful for evaluating the individual stroke mechanism and for therapeutic decision making. These techniques provide both anatomical and physiological information, such as the distribution of ischaemic lesions, patterns of haemodynamic compromise, and the status of the head and neck arteries [1–3]. Perfusion MRI, which uses a dynamic susceptibility-weighted bolus-tracking method with gadolinium chelate, is especially ease to use in the clinical setting [4].
MRI protocols at many institutes include both PI and CE-MRA, which require contrast injection for each scan. There are concerns that the accumulation of paramagnetic contrast media might spoil the image quality of subsequent studies. Time delay for the clearance of the contrast media between each sequence should be avoided, since a delay in a therapeutic decision may be hazardous to patients because of a narrower time window for therapeutic intervention. Curtailment of the imaging acquisition time is more important than obtaining a higher quality image in the MRI protocol for the acute stroke.
We have found only a few reports concerning the influence of a pre-load contrast agent on a subsequent perfusion MR image [5, 6]. No study has looked at simultaneous performance of PI and CE-MRA, and no study has examined the ideal order of multimodal MRI in the diagnosis of acute stroke when a protocol including two post-contrast sequences is performed. Therefore, we compared the image quality of PI following CE-MRA and CE-MRA following PI in patients with acute stroke in order to ascertain the optimum order of PI and CE-MRA in a multimodal MRI protocol for patients with acute ischaemic stroke.
|
Materials and methods
|
|---|
Acute stroke MRI protocol
The acute stroke MRI protocol at our institute includes a T2 weighted image (T2WI), a DWI, three-dimensional TOF MRA of the circle of Willis, a T2* weighted image (T2*WI), PI and CE-MRA. MR images were obtained using a 1.5 T MR system (Signa CVi; GE Medical Systems, Milwaukee, WI) with neurovascular phased array coils. Total acquisition time for the acute stroke MRI protocol was 11 min and 4 s. Details of the protocol are described in Table 1
.
The PI sequence was acquired after administration of 15 ml of gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany) using a bolus technique (4 ml s–1) followed by a saline flush (20 ml at the same flow rate). The parameters were 10 slices, 10 mm thickness, 0 interslice gap, 60 dynamic scans, a 2000 ms repetition time (TR) and 60 ms echo time (TE). To post-process the perfusion MR data, the source data were transferred to a workstation computer to generate time–signal intensity curves. Changes in the transverse relaxation rate (R2) can be calculated from the signal intensity using the following equation: R2 = –ln (S/S0)/TE, where ln is the natural logarithm, S is signal intensity, S0 is baseline signal intensity and TE is echo time [7]. Mean transit time (MTT) and relative cerebral blood volume (rCBV) maps were created using post-processing software (Functool 2.5; GE Medical Systems, Milwaukee, WI). MTT maps were obtained as colour-coded images from the measured time–signal intensity curve on a pixel-by-pixel basis. rCBV maps were also generated as colour-coded images by numerical integration of the area under a fitted curve on a pixel-by-pixel basis [8, 9].
CE-MRA was performed to cover from the aortic arch to the intracranial arteries. A total of 15 ml of contrast media was injected (2 ml s–1). A tracker volume was placed in the common carotid artery on the sagittal scout image. Three-dimensional fast spoiled gradient-echo imaging was then performed with 2 s delay after the automated bolus-detection algorithm (SmartPrep technique) triggered the arrival of the contrast bolus into the tracking region [10]. The imaging parameters were as follows: 6.9/2.2 = TR/TE; flip angle 30°; single coronal slab thickness 80 mm; effective slice thickness 1 mm; field of view 320 mm; single excitation; matrix 256x256; sequence order; 512 zero-filled interpolation; and without half-Fourier technique. The angiographic images were reconstructed using a maximum intensity projection (MIP) algorithm. One set of 15 MIP images was obtained by rotation of the stacked images along a vertical axis.
Patients
In a prospective design, following the acquisition of all non-contrast imaging sequences, we made it a rule to perform a CE-MRA after PI between November 2003 and June 2004 (group A) and to perform a PI following CE-MRA between July 2004 and December 2004 (group B). There was no time gap or pause between the two scans.
Inclusion criteria were acute hemispheric ischaemic stroke; admission within 12 h after the onset of symptoms; unilateral middle cerebral artery (MCA) territorial infarction on DWI; proximal vessel (internal carotid artery or MCA stem) occlusion suspected on TOF-MRA; and the finding of a normal contralateral cerebral hemisphere on DWI, T2*WI, T2WI and TOF-MRA. Any case was excluded if there was a poor image quality caused by motion artefact, or no significant signal drop in the signal-to-time curve of PI reflecting inappropriate bolus transmission of the contrast material in the global cerebral circulation. 35 consecutive patients were included in this study and, of these, 17 patients (M:F = 8:9, median 60.0 years) underwent CE-MRA following PI (group A) and 18 patients (M:F = 10:8, median 64.5 years) underwent PI following CE-MRA (group B). Informed consent for the acute stroke MRI was obtained from each patient's next of kin.
MRI analysis
Two radiologists reviewed the image quality of each PI and CE-MRA in a blinded fashion, which meant that the readers did not have any information on the order of the acquisitions during the analysis. All images were reviewed on the PACS system, and all reviewers were permitted to choose the optimal window width and level setting.
For qualitative analysis of the CE-MRA, each observer graded the imaging quality and vascular contrast in the following two sites: the extracranial arteries (from the aortic arch to the cervical segment of the internal carotid artery (ICA) and the extradural vertebral arteries) and the intracranial arteries (from the petrous ICA to the distal branch of the anterior, middle and posterior cerebral artery; intradural vertebral arteries; and the basilar trunk), using a three-point scoring system (Table 2).
View this table:
[in this window]
[in a new window]
|
Table 2. Description of the 3-point score used for the qualitative analysis of contrast enhanced MR angiography (CE-MRA)
|
|
Quantitative analysis of the PI was performed by a radiologist by evaluation of the maximum change in the transverse relaxation rate (R2max) and the relative signal drop (S/S0) in a round region of interest (ROI) in the temporal lobe (territory of the middle cerebral artery) of the unaffected side [11]. The ROI size was fixed in 120–160 pixels, and included cortex and white matter. R2max and the relative signal drop were calculated as follows: S0 is an average signal value in a round ROI on representative pre-contrast images and St is an average signal value in a round ROI with maximum signal decrease (peak post-contrast image). S0 and St were taken directly from the measured signal curve. R2max and the relative signal drop were calculated according to these two equations:
For the qualitative evaluation of PI, each observer rated the image quality of two parameter (rCBV and MTT) maps separately as either "good", "acceptable", or "unacceptable".
"Good" was defined as a parameter map which clearly demarcates the ischaemic zone and anatomy. (The clear demarcation of an ischaemic zone means that a larger or same size area containing the referent infarct area on DWI shows significant distinction from the surrounding normal perfusion area on a colour-coded parameter map, and clear demarcation of the anatomy means that basal ganglia, insular ribbon and internal capsule are clearly delineated from the surrounding tissue on rCBV map). "Acceptable" was defined as a parameter map which clearly demarcates the ischaemic zone, but delineation of the anatomy is unclear. A "poor" grade meant that the ischaemic zone not clearly demarcated on the parameter map.
Statistical analysis
The interobserver reliability of quality scoring of CE-MRA between the two observers was assessed using the pair-wise 
(kappa) statistics and was interpreted as described by Landis and Koch (<2.0, poor; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, good; 0.81–1.0, very good) [12]. Each of the quality scores of the two independent observers was used for the statistical analysis. Differences in quality scores of extracranial arteries and intracranial arteries on the CE-MRA between the two groups were tested for statistical significance using a Mann-Whitney U-test. The quality scores for the parameter maps of PI were statistically analysed using the same method mentioned above.
For the statistical analysis of R2max and the relative signal drop, a t-test was performed to evaluate the significant statistical difference between the two groups. A p-value of 0.05 or less was considered statistically significant.
|
Results
|
|---|
Of the 35 cases, CE-MRA demonstrated that 6 cases involved occlusion of the proximal ICA; three had stenosis of the proximal ICA; 14 had occlusion of the main trunk of middle cerebral artery; three had stenosis of the main trunk of MCA; five had occlusion of the M2 segment of the MCA; and four had an indistinct arterial occlusion level.
Interobserver agreements for intracranial and extracranial arteries on CE-MRA were good ( = 0.737, 0.641). There were no significant differences in the qualitative grade of CE-MRA between group A and group B (intracranial arteries (observer 1: p = 0.9867, observer 2: p = 0.6291); extracranial arteries (observer 1: p = 0.8016, observer 2: p = 0.1011)). Of the 17 intracranial artery images on CE-MRA in group A, 12 were rated good or acceptable by both observer 1 and observer 2. In the 18 images of group B, 12 were rated good or acceptable by observer 1 and 14 by observer 2. Of the 17 cases of CE-MRA images of extracranial arteries in group A, 16 were rated good or acceptable by observer 1 and 17 by observer 2. Of the 18 cases in group B, 16 were rated good or acceptable by observer 1 and 15 by observer 2 (Table 3).
Interobserver agreement for PI was good ( = 0.663). There were no significant differences between groups A and B in terms of PI image quality (observer 1: p = 0.4278, observer 2: p = 0.4310). The image quality of PI was rated good or acceptable in 16 of the 17 cases in group A by both observers. Of the 18 PI images of group B, both observers rated 16 as good or acceptable (Table 4).
In the quantitative analysis of PI, the mean value of R2max of group A was 98.1x10–4±38.2x10–4 and that of group B was 96.6x10–4±36.7 x10–4 (Figure 1). The mean difference of the R2max between the two groups was 1.0x10–4 and the standard error difference was 12.7x10–4. No significant differences were observed between the two groups in R2max (p = 0.907). The mean value of the relative signal drop of group A was 43.1±12.8% and that of group B was 40.6±12.3% (Figure 2). The mean difference of the relative signal drop between the two groups was 2.490, and the standard error difference was 4.248. There was no significant difference between the two groups in the relative signal drop (p = 0.561).
View larger version (7K):
[in this window]
[in a new window]
|
Figure 2. Data comparison graph of the relative signal drop. The small blank boxes denote the value of the relative signal drop for each patient. There is no significant difference in the distribution of the relative signal drop between groups A and B.
|
|
|
Discussion
|
|---|
In this study, there was no difference between the imaging quality of CE-MRAs obtained before or after the acquisition of PI, nor was the image quality of PIs influenced by the order of image acquisition. In other words, this study shows that in the evaluation of acute stroke patients, neither of the contrast-using MR studies, PI or CE-MRA, were significantly affected by the previously administered contrast media.
After intravenous bolus injection of gadolinium, the high concentrations of gadolinium are typically present only during the first passage of contrast agent through the capillary bed and the recirculation phenomena rapidly dilutes the passing gadolinium bolus [14]. Gadolinium chelate is compartmentalized in the vessels of the brain due to the blood–brain barrier (BBB), without diffusion into the extravascular extracellular space [15]. Therefore, a pre-dose of gadolinium for CE-MRA may not be strong enough to influence the time–signal intensity curve of the subsequent PI. There have been some reports about the influence of pre-loaded gadolinium on the PI [5, 6]. When there is a disruption of the BBB, gadolinium leaks out of the vasculature into the extravascular extracellular space, resulting in enhanced T1 relaxation effects. Signal increases that result from shortening T1 compete with the susceptibility-induced signal decreases. Therefore, contrast agent leakage can lead to an underestimation of blood volume and flow to the lesion [16]. In the evaluation of brain tumours, the addition of a loading dose of contrast agent several minutes prior to PI could reduce the confounding effects of BBB leakage on the accuracy of the perfusion map by partial saturation of the tissue T1 [5, 6]. The T1 leakage effect need not be considered in our study, because a BBB break is not usually presented at the acute phase of ischaemic stroke. In this study, there were no qualitative differences in diagnosing the ischaemic zone between the PI maps with and without pre-loaded gadolinium for diagnosing the ischaemic zone. We believe that the parameters on PI would not be influenced by the concentration of pre-bolus gadolinium, because the relaxation effect of gadolinium is limited to the intravascular compartment of the brain with an intact BBB, which represents only 2–4% of the tissue, and the relaxation effect has a limited contrast-to-noise ratio [4, 17].
To our knowledge, this is the first study examining the influence of pre-loaded gadolinium chelate on a subsequent CE-MRA. Because the venous enhancement is related to the short time (8 s) after bolus contrast injection in vein [18], previously injected gadolinium for PI may not show venous contamination during a subsequent CE-MRA. Although the T1 signal of soft tissue or vein on post-contrast MRI is expected to be higher than that on pre-contrast MRI, maximum intensity projection (MIP) reconstruction would be helpful to accentuate the intra-arterial gadolinium signal by contrast with the less bright voxels of dilute contrast media in soft tissue or veins.
Some institutes omit CE-MRA to avoid any potential influence on the quality of a following PI, to avoid the potential influence of pre-loaded gadolinium from the previous PI, or to shorten the acquisition time of MRI. Although TOF-MRA can often help identify the causative arterial occlusion in acute ischaemic stroke, TOF-MRA of the intracranial arteries is not sufficient for evaluation of the entire vasculature of the neck and brain. In our study, a steno-occlusive lesion of the carotid artery in nine cases (26%) was the cause of acute stroke in unilateral MCA territorial infarction. Since CE-MRA can display a tremendous field of view from the aortic arch to the cerebral branches using a single bolus of contrast medium [19], it is considered useful for looking for occlusion in both the intracranial arteries and the collateral pathways, and also to evaluate atherosclerotic pathology of the carotid arteries [20, 21]. Because the pathology of supra-aortic vessels (carotid artery, vertebral artery) is a significant cause of acute stroke, it is desirable to evaluate the supra-aortic arteries as part of the routine imaging protocol in acute stroke.
A small test injection of contrast material with a rapid two-dimensional MR pulse sequence is one of the other methods for the optimization of imaging delay time before CE-MRA [22]. Although this method has good reliability, it has limitations such as associated overhead costs of performing and interpreting the test bolus acquisition and the consumption of the scan time for test injection. PI prior to CE-MRA could provide a very good bolus timing technique by analysing the signal curve in the region of a large artery. This enables scan triggering at the optimal moment with maximal arterial opacification. Doing PI first may avoid the requirement of test injection bolus for setting up a bolus triggering sequence prior to CE-MRA.
There were limitations to our study. First, we could not include a quantitative analysis of the MRA because the source images in some cases had been lost. Second, in the analysis of the data, we did not consider the influence of other patient factors such as cardiac output or patient body weight, etc. Third, intraindividual evaluation was not performed with either imaging technique. Fourth, the sample sizes of PI parameters with negative results were too small for quantitative analysis.
For comparative evaluation of the MR perfusion images in this study, the qualitative evaluation is emphasised rather than the accuracy of the measurement of perfusion parameters. During the data acquisition period of in this study, the perfusion parameter maps of all cases were automatically reformatted on a workstation computer directly after acquisition of the acute stroke MRI, and they were used with other sequences to quickly make decisions about therapeutic interventions. Therefore, the qualitative evaluation of PI in this study focused primarily on emergent clinical applications.
|
Conclusion
|
|---|
There was no reciprocal influence upon diagnostic imaging quality with regard to the order of the two post-contrast sequences; PI and CE-MRA. When the multimodal stroke MRI series including PI and CE-MRA is performed, it is recommended that, regardless of the order of these two post-contrast sequences, all image sequences should be acquired quickly. During simultaneous CE-MRA and PI, we found no deterioration of diagnostic imaging quality related to the order of the two post-contrast sequences. They can be performed according to the preference of each institute.
This work was partly supported by grants from the Korean Ministry of National Health and Welfare (03-PJ1-PG1-CH06-0001).
Received for publication March 22, 2006.
Revision received May 8, 2006.
Accepted for publication May 12, 2006.
|
References
|
|---|
- Szabo K, Gass A, Hennerici MG. Diffusion and perfusion MRI for the assessment of carotid atherosclerosis. Neuroimaging Clin N Am 2002;12:381–90.[CrossRef][Medline]
- Sunshine JL. CT, MR imaging, and MR angiography in the evaluation of patients with acute stroke. J Vasc Interv Radiol 2004;15:S47–55.[Medline]
- Lee DH, Kang DW, Ahn JS, Choi CG, Kim SJ, Suh DC. Imaging of the ischemic penumbra in acute stroke. Korean J Radiol 2005;6:64–74.[Medline]
- Grandin CB. Assessment of brain perfusion with MRI: methodology and application to acute stroke. Neuroradiology 2003;45:755–66.[CrossRef][Medline]
- Quarles CC, Ward BD, Schmainda KM. Improving the reliability of obtaining tumor hemodynamic parameters in the presence of contrast agent extravasation. Magn Reson Med 2005;53:1307–16.[CrossRef][Medline]
- Schmainda KM, Rand SD, Joseph AM, Lund R, Ward BD, Pathak AP, et al, Characterization of a first-pass gradient-echo spin-echo method to predict brain tumor grade and angiogenesis. AJNR Am J Neuroradiol 2004;25:1524–32.[Abstract/Free Full Text]
- Weisskoff RM, Zuo CS, Boxermann JL, Rosen BR. Microscopic susceptibility variation and transverse relaxation: theory and experiment. Magn Reson Med 1994;31:601–10.[Medline]
- Cho SK, Na DG, Ryoo JW, Roh HG, Moon CH, Byun HS, et al. Perfusion MR imaging: clinical utility for the differential diagnosis of various brain tumors. Korean J Radiol 2002;3:171–9.[Medline]
- Sorensen AG, Copen WA, Ostergaard L, Buonanno FS, Gonzalez RG, Rordorf G, et al. Hyperacute stroke: simultaneous measurement of relative cerebral blood volume, relative cerebral blood flow, and mean tissue transit time. Radiology 1999;210:519–27.[Abstract/Free Full Text]
- Ho VB, Foo TK. Optimization of gadolinium-enhanced magnetic resonance angiography using an automated bolus-detection algorithm (MR SmartPrep). Original investigation. Invest Radiol 1998;33:515–23.[CrossRef][Medline]
- Tombach B, Benner T, Reimer P, Schuierer G, Fallenberg EM, Geens V, et al. Do highly concentrated gadolinium chelates improve MR brain perfusion imaging? Intraindividually controlled randomized crossover concentration comparison study of 0.5 versus 1.0 mol/L gadobutrol. Radiology 2003;226:880–8.[Abstract/Free Full Text]
- Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159–74.[CrossRef][Medline]
- Huisman TA, Sorensen AG. Perfusion-weighted magnetic resonance imaging of the brain: techniques and application in children. Eur Radiol 2004;14:59–72.[CrossRef][Medline]
- Barbier EL, Lamalle L, Decorps M. Methodology of brain perfusion imaging. J Magn Reson Imaging 2001;13:496–520.[CrossRef][Medline]
- Aronen HJ, Glass J, Pardo FS, Belliveau JW, Gruber ML, Buchbinder BR, et al. Echo-planar MR cerebral blood volume mapping of gliomas. Clinical utility. Acta Radiol 1995;36:520–8.[Medline]
- Edelman RR, Mattle HP, Atkinson DJ, Hill T, Finn JP, Mayman C, et al. Cerebral blood flow: assessment with dynamic contrast-enhanced T2*-weighted MR imaging at 1.5 T. Radiology 1990;176:211–20.[Abstract/Free Full Text]
- Abe O, Aoki S, Shirouzu I, Kunimatsu A, Hayashi N, Masumoto T, et al. MR imaging of ischemic penumbra. Eur J Radiol 2003;46:67–78.[CrossRef][Medline]
- Jung HW, Chang KH, Choi DS, Han MH, Han MC. Contrast-enhanced MR angiography for the diagnosis of intracranial vascular disease: optimal dose of gadopentetate dimeglumine. AJR Am J Roentgenol 1995;165:1251–5.[Abstract/Free Full Text]
- Leclerc X, Gauvrit JY, Nicol L, Pruvo JP. Contrast-enhanced MR angiography of the craniocervical vessels: a review. Neuroradiology 1999;41:867–74.[CrossRef][Medline]
- Cosottini M, Pingitore A, Puglioli M, Michelassi MC, Lupi G, Abbruzzese A, et al. Contrast-enhanced three-dimensional magnetic resonance angiography of atherosclerotic internal carotid stenosis as the noninvasive imaging modality in revascularization decision making. Stroke 2003;34:660–4.[Abstract/Free Full Text]
- Okumura A, Araki Y, Nishimura Y, Iwama T, Kaku Y, Furuichi M, et al. The clinical utility of contrast-enhanced 3D MR angiography for cerebrovascular disease. Neurol Res 2001;23:767–71.[CrossRef][Medline]
- Earls JP, Rofsky NM, DeCorato DR, Krinsky GA, Weinreb JC. Breath-hold single-dose gadolinium-enhanced three-dimensional MR aortography: usefulness of a timing examination and MR power injector. Radiology 1996;201:705–10.[Abstract/Free Full Text]
Top
Abstract
Introduction
