AJR 2000; 174:853-857
© American Roentgen Ray Society
Cerebral Infarctions
Evaluation with Single-Axis Versus Trace Diffusion-Weighted MR Imaging
M. Castillo1,
S. K. Mukherji,
D. Isaacs and
J. K. Smith
1
All authors: Department of Radiology, University of North Carolina School of
Medicine, Campus Box 7510, Chapel Hill, NC 27599-7510
Received July 1, 1999;
accepted after revision August 13, 1999.
Address correspondence to M. Castillo.
Abstract
OBJECTIVE. Our purpose was to determine the usefulness of
single-axis diffusion-weighted imaging versus trace diffusion-weighted imaging
in the evaluation of cerebral infarctions.
SUBJECTS AND METHODS. Twenty-six patients harboring 34 infarctions
were examined using single-axis and trace diffusion-weighted imaging within 48
hr of the onset of symptoms. Two neuroradiologists who were not aware of the
clinical findings reviewed all images obtained with both techniques and noted
the following: type of infarction (small [<15 mm] versus territorial),
location of infarction, presence of infarction (see only on single-axis
images, seen only on trace images, seen on both), lesion conspicuity (better
on single-axis images, better on trace images, or equal on both), and lesion
size (larger on single-axis images, larger on trace images, or equal on both).
Differences in opinion were resolved by consensus.
RESULTS. Of the 18 small and 16 territorial infarctions, all were
identified on both single-axis and trace imaging. Lesion conspicuity was
judged to be slightly better on trace images for both types of infarctions.
Lesion size was judged to be larger on single-axis images for territorial
infarctions.
CONCLUSION. Both single-axis and trace diffusion-weighted imaging
showed all small and territorial cerebral infarctions. Both types of
infarctions were slightly larger on single-axis images but this did not affect
correct interpretation in any case. The single-axis technique provided
sufficient information for the diagnosis of cerebral infarction in our
clinical settings.
Introduction
Since its introduction, diffusion-weighted MR imaging has become the
diagnostic imaging method of choice for patients suspected of harboring acute
cerebral infarctions
[1,2,3,4].
The simplest method by which to obtain these studies is applying a diffusion
gradient in only one direction
[5]. This technique is
sensitive for the detection of acute infarctions but is limited by
shine-through contributions from spin-density and T2 as well as anisotropy
artifacts [6]. These artifacts
may be eliminated by obtaining trace diffusion images and apparent diffusion
coefficient maps. Trace imaging requires a slightly longer acquisition time
(thus increasing the risk of motion-induced image degradation) and apparent
diffusion coefficient maps are time-consuming, because they require the use of
three or more diffusion-gradient values and generally off-line
postprocessing.
In this study, we sought to determine the utility of single-axis versus
trace diffusion-weighted imaging in the evaluation of acute cerebral
infarctions.
Subjects and Methods
Twenty-six consecutive patients with acute cerebral infarctions underwent
brain MR imaging. All MR studies were obtained within 48 hr of the onset of
stroke symptoms. All MR studies were obtained on two 1.5-T units with
echoplanar capabilities. All patients also underwent conventional MR imaging
sequences, which were not used for this study. Multisection single-shot
spin-echo echoplanar diffusion-weighted images were obtained. The baseline set
of images was obtained with a TR of 0.8 msec, a TE of 123 msec, one
excitation, and a b value of 30 sec/mm2. The first set of
diffusion-weighted images was obtained using parameters identical to those of
the baseline images but applying a high-strength diffusion gradient (b = 1000
sec/mm2) in only the x-axis (frequency encoding, left to
right) direction (single-axis images). The second set of trace-weighted
(isotropic) images (TR/TE, 56.61/139) was acquired by playing out a diffusion
gradient for which bxx = byy = bzz = 333 and bxy = bxz = byz = 0. This results
in signal intensity images that incorporate the information of all three
diffusion directions (phase, read, and slice) into a single image. This was
done as ST = (SR x SP x SS)1/3, where for each pixel,
SR is the signal intensity of readout-direction diffusion sensitization, SP is
the signal intensity of phase-direction diffusion sensitization, and SS is the
signal intensity of slice-direction diffusion sensitization. The time of
acquisition was 4.6 sec for the low b value (baseline) and single-axis images
and 16 sec for the trace images. All images were obtained in the transverse
orientation with a field of view of 230 mm and a matrix of 128 x 200.
All patient information and alphanumeric data were removed from the studies.
Both single-axis and trace images were photographed with the same window width
and center settings.
Three separate sets of images (baseline, single-axis, and trace images)
were presented for interpretation to two neuroradiologists. Each image set was
interpreted separately. Criteria for the diagnosis of acute infarction were a
focal high signal intensity on diffusion-weighted images not thought to
represent an artifact caused by magnetic susceptibility and to be in
accordance with the clinical symptoms. We then categorized each infarct as
being small (<15 mm in greatest dimension) or involving the territory of a
major cerebral artery and recorded them according to location. The presence of
an infarct was noted as appearing in the single-axis images, trace images, or
both. The subjective conspicuity of the infarcts was noted as being more
obvious on the single-axis images, more obvious on the trace images, or
equally obvious on both. The size of the infarcts was noted as being larger on
the single-axis images, larger on the trace images, or equal on both. The data
were reviewed and any differences in scoring were resolved by consensus
between the two observers. The results of the data collected were then
analyzed.
Results
Twenty-six patients (12 males and 14 females) who harbored 34 infarctions
comprised our study population. Patients ranged in age from 3 months to 85
years. There were 18 small infarctions (occipital lobes, n = 6;
frontal lobes, n = 5; basal ganglia, n = 4; brainstem,
n = 2; and corpus callosum, n = 1). There were 16
territorial infarctions (predominantly temporal, n = 12;
predominantly occipital, n = 3; and predominantly frontal, n
= 1). Using the independent scores, both observers identified all infarctions
in both the single-axis and the trace images (thus, specificity and
sensitivity are equal). Because of a lack of randomization and a relatively
small sample size, the data analysis is presented as simple tabular
comparisons. Using the consensus data with regard to the conspicuity of the
infarctions, both techniques were judged to be equal in 20 infarctions (small,
n = 10; territorial, n = 10) (Fig.
1A,1B),
better on single-axis images in seven instances (small, n = 3;
territorial, n = 4) (Fig.
2A,2B),
and better on trace imaging in seven infarctions (small, n = 6;
territorial, n = 1) (Fig.
3A,3B)
(Table 1). Using the consensus
data with regard to size of the infarcts, both techniques showed infarcts to
be of a similar size in 25 instances (Fig.
1A,1B),
larger on the single-axis images in six instances (small, n = 2;
territorial, n = 4) (Fig.
2A,2B),
and larger on the trace images in three instances (all small infarctions)
(Figs.
4A,4B
and
5A,5B)
(Table 2).
View larger version (118K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2A. —Posterior cerebral artery territory infarction in 75-year-old woman
shown on diffusion-weighted MR images (window width, 522 H; window center, 181
H). Single-axis diffusion-weighted axial MR image shows infarction
(arrows).
|
|
View larger version (123K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2B. —Posterior cerebral artery territory infarction in 75-year-old woman
shown on diffusion-weighted MR images (window width, 522 H; window center, 181
H). Trace diffusion-weighted MR image at same level as A shows
infarction (arrows) to be less conspicuous and slightly smaller than
that on A.
|
|
View larger version (172K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A. —Small infarction in 78-year-old man shown on diffusion-weighted MR
images (window width, 500 H; window center, 187 H). Single-axis
diffusion-weighted axial MR image shows small infarction (arrow) in
splenium of corpus callosum. Separating infarct from normal brightness of
corpus callosum resulting from anisotropy artifact is difficult. Second small
infarction is in left thalamus.
|
|
View larger version (172K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B. —Small infarction in 78-year-old man shown on diffusion-weighted MR
images (window width, 500 H; window center, 187 H). Trace diffusion-weighted
MR image at same level as A shows infarction (arrow) to be
more conspicuous but equal in size. Increased conspicuity of this infarct is
caused by significant decreased anisotropy artifact on this trace image.
Conspicuity of infarct in left thalamus was judged to be equal on both
studies, although it appears slightly larger on trace image.
|
|
View larger version (131K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4A. —Small infarctions in 51-year-old man shown on diffusion-weighted MR
images (window width, 350 H; window center, 152 H). Single-axis
diffusion-weighted MR image shows infarctions in pons.
|
|
View larger version (142K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4B. —Small infarctions in 51-year-old man shown on diffusion-weighted MR
images (window width, 350 H; window center, 152 H). Trace diffusion-weighted
MR image at same level as A shows infarcts to appear slightly larger
but judged to have similar conspicuity.
|
|
View larger version (111K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5A. —Small infarction in 39-year-old woman shown on diffusion-weighted MR
images (window width, 360 H; window center, 160 H). Single-axis
diffusion-weighted MR image shows infarction (arrow) in right
cerebral peduncle.
|
|
View larger version (129K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5B. —Small infarction in 39-year-old woman shown on diffusion-weighted MR
images (window width, 360 H; window center, 160 H). Trace diffusion-weighted
MR image at same level as A shows infarction (arrow) to appear
more conspicuous and larger.
|
|
Discussion
Diffusion-weighted images show cerebral infarctions within minutes of their
onset and have rapidly become the mainstay in the diagnostic evaluation of
stroke patients [7]. Normal
random motion of water molecules, particularly in the extracellular space,
results in greater diffusion and signal-intensity loss. Restriction of water
movement in areas of infarction results in less diffusion and increased signal
intensity with respect to surrounding normal tissues. In the normal brain,
motion of water molecules may be normally restricted in one direction by the
organization of the surrounding tissues. Anisotropy refers to a property of
tissues that affects the translational movement of water molecules
[8]. In the brain, anisotropy
occurs predominantly in the white matter tracts
[9]. If a white matter tract is
oriented along the axis in which a diffusion gradient is applied, high signal
intensity may be normally seen in the images and should not be mistaken for a
lesion. This potential problem exists when diffusion-weighted images are
obtained using a diffusion gradient applied in only one direction. A solution
to this problem is to produce images that portray the average diffusion in
three orthogonal directions (trace or isotropic imaging). Another option is to
generate apparent diffusion coefficient maps. These maps need at least three
different b values and, in most current equipment, need to be generated
off-line. Therefore, coefficient maps may not be practical from a clinical
standpoint. A third option is to apply very high diffusion values (b 
3000
sec/mm2)—which require the use of gradients of 40 mT or
more—that are not available in all scanners. Thus from a clinical
standpoint, single-axis or trace diffusion-weighted images are two practical
methods for the examination of stroke patients. On our MR scanner, single-axis
diffusion-weighted images may be obtained in less than 5 sec but these images
suffer from anisotropy artifacts and, although trace imaging eliminates these
artifacts, trace imaging requires three times as long to acquire. In stroke
patients, a short acquisition time is desired to eliminate motion
artifacts.
Different processing methods for data from diffusion-weighted studies have
recently received attention. Chong et al.
[10] compared the results of
simple three orthogonal-axis diffusion-weighted images with isotropic-weighted
images, diffusion trace images, and diffusion trace-weighted images (apparent
diffusion coefficient maps). The single-axis images of Chong et al. were
presented as different sets for each of the three directions and the presence
of an abnormality in only one of these sets was considered an infarct. These
researchers found that these single-axis images had the highest accuracy and a
high specificity, closely followed by the isotropic studies. Chong et al.
concluded that diffusion trace-weighted images and apparent diffusion
coefficient maps were not as effective in delineating lesions and that these
techniques are needed predominantly for research protocols for which
quantification is desired. Chong et al. stated that they could not evaluate
if"only a single direction of diffusion sensitivity may be adequate to
detect these lesions."
In another study, trace images were found to be inferior to single-axis
images in the detection of brainstem infarctions (Britt PM et al., presented
at the American Society of Neuroradiology Meeting, May 1999). These authors
argued that the accentuation of normal anisotropy may be the only sign in
early brainstem infarctions. They did not study infarctions occurring in the
cerebral hemispheres. Two of our patients harbored small infarctions in the
brainstem and these infarctions were better visualized and appeared larger on
the trace images than on the single-axis studies.
In our patients, we found that the infarctions were equally detected by
both single-axis (we chose the x-axis because of the technical
restrictions of our equipment) and trace images. In addition, both techniques
were similar with regard to lesion conspicuity (20/34) and lesion size
(25/34). The conspicuity of all infarcts (regardless of their type) was also
judged to be equal with both techniques. In seven patients, infarcts were
better seen in single-axis images, whereas infarcts were better seen on trace
imaging in seven patients. If one categorizes these infarcts according to
size, trace imaging was slightly better than single-axis imaging in assessing
conspicuity (nine small versus five territorial infarcts, respectively). With
respect to lesion size, six lesions were larger on the single-axis diffusion
weighted images, whereas three lesions were larger on the trace imaging. This
finding implies that on the single-axis diffusion-weighted images there is
some contribution from spin density and T2 shine-through artifact, which we
believe did not affect our interpretation of the studies. If one judges the
lesions according to size only, the trace images performed similar to the
single-axis images with respect to small infarctions (two versus three
infarctions, respectively). Four territorial infarctions were seen better on
single-axis imaging than on trace imaging, again implying that some
shine-through artifact contributes to make infarcts slightly larger on the
former sequence. Again, the shine-through contribution from spin-density and
T2 did not alter our image interpretation.
In conclusion, we found that both single-axis and trace diffusion-weighted
images were able to show all small and territorial cerebral infarctions.
Single-axis diffusion-weighted imaging is faster to perform than trace imaging
(and thus less susceptible to motion degradation) and we believe that the
information single-axis diffusion-weighted imaging provides is sufficient for
diagnosis of cerebral infarction in clinical settings.
Acknowledgments
We thank David Richardson (School of Public Health, University of North
Carolina, Chapel Hill, NC) for his help.
References
-
Warach S, Chien D, Li W, Ronthal M, Edelman RR. Fast magnetic
resonance diffusion-weighted imaging of acute human stroke.
Neurology
1992;42:1717-1723[Abstract/Free Full Text]
-
Kohno K, Ohta S, Kohno K, Kumon Y, Sakaki S, Okujima S. Early
detection of cerebral ischemic lesion using diffusion-weighted MRI. J
Comput Assist Tomogr
1995;19:982-986[Medline]
-
Sorensen AG, Buonanno FS, Gonzalez RG, et al. Hyperacute stroke:
evaluation with combined multisection diffusion-weighted and hemodynamically
weighted echo-planar MR imaging. Radiology
1996;199:391-401[Abstract/Free Full Text]
-
Gonzalez RG, Schaefer PW, Buonanno FS, et al. Diffusion-weighted MR
imaging: diagnostic accuracy in patients imaged within 6 hours of stroke
symptom onset. Radiology
1999;210:155-162[Abstract/Free Full Text]
-
Patel MR, Siewert B, Warach S, Edelman RR. Diffusion and perfusion
imaging techniques. Magn Reson Imaging Clin N Am
1995;3:425-438[Medline]
-
Gray L, MacFall J. Overview of diffusion imaging. Magn Reson
Imaging Clin N Am
1998;6:125-138[Medline]
-
Mosely ME, Cohen Y, Mintorovitch J, et al. Early detection of
regional cerebral ischemia in cats: comparison of diffusion- and T2-weighted
MRI and spectroscopy. Magn Reson Med
1990;14:330-346[Medline]
-
Hajnal J, Doran M, Hall A, et al. MR imaging of anisotropically
restricted diffusion of water in the nervous system: technical, anatomic, and
pathologic considerations. J Comput Assist Tomogr
1991;15:1-18[Medline]
-
Doran M, Hajnal J, Van Bruggen N, King M, Young I, Bydder G. Normal
and abnormal white matter in the nervous system: technical, anatomic, and
pathologic considerations. J Comput Assist Tomogr
1990;14:865-873[Medline]
-
Chong J, Lu D, Aragao F, et al. Diffusion-weighted MR of acute
cerebral infarction: comparison of data processing methods.
AJNR
1998;19:1733-1739[Abstract]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
|
|
 
N. Evangelou, D. Konz, M. M. Esiri, S. Smith, J. Palace, and P. M. Matthews
Regional axonal loss in the corpus callosum correlates with cerebral white matter lesion volume and distribution in multiple sclerosis
Brain,
September 1, 2000;
123(9):
1845 - 1849.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
