| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Neuroradiology |
1 From the Mallinckrodt Institute of Radiology, Washington University School of Medicine, 510 S Kingshighway Blvd, St Louis, MO 63110. From the 2000 RSNA scientific assembly. Received February 10, 2000; revision requested April 5; final revision received October 10; accepted November 1. Address correspondence to P.M. (e-mail: mukherjeep@mir.wustl.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Twelve patients with the clinical features and conventional magnetic resonance (MR) imaging findings of RPLS underwent diffusion-tensor echo-planar MR imaging. The isotropic diffusion coefficient (
) and diffusion anisotropy (A
) were measured in posterior regions of diffusion abnormality and in anterior areas of normal-appearing brain.
RESULTS: Across all 12 subjects, the mean of (1.09 ± 0.13 [SD]) x 10-3 mm2/sec in affected posterior regions was 26% greater than its value of (0.87 ± 0.07) x 10-3 mm2/sec in normal-appearing anterior regions. The mean A of 0.15 ± 0.03 in posterior regions was 35% less than its value of 0.23 ± 0.02 in anterior regions (t11 = 9.58; P < .001). There was a significant inverse correlation between and A in posterior regions (r = -0.67; P < .018) but not in anterior regions (r = -0.12; P = .719). A follow-up study performed in one patient after resolution of symptoms documented reversal of elevated isotropic diffusion and at least partial recovery of anisotropy loss.
CONCLUSION: The increased magnitude of brain water diffusion characteristic of RPLS is accompanied by reduced A. The magnitudes of these two effects are correlated and may be reversible. These observations support the proposal that vasogenic edema due to cerebrovascular autoregulatory dysfunction is the underlying pathophysiologic mechanism in uncomplicated RPLS.
Index terms: Anisotropy • Brain, diffusion, 10.12144, 10.92 • Brain, diseases, 10.872 • Brain, gray matter, 10.92 • Brain, MR, 10.121413, 10.121416, 10.12144 • Brain, white matter, 10.92 • Diffusion tensor • Magnetic resonance (MR), diffusion study, 10.12144, 10.92 • Magnetic resonance (MR), rapid imaging, 10.121413
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Two pathophysiologic mechanisms for RPLS have been proposed (2). One postulates cerebral vasospasm with resulting ischemia within the involved territories (13–15), whereas the other posits a breakdown in cerebrovascular autoregulation with ensuing interstitial extravasation of fluid (3,16,17). Diffusion MR imaging can be used to discriminate between these two possibilities, as the cytotoxic edema of cerebral ischemia demonstrates decreased water mobility, whereas vasogenic edema due to cerebrovascular autoregulatory dysfunction results in increased water mobility. Results of case reports (4,9,18) and a small case series (19) of diffusion-weighted MR imaging in RPLS have shown elevated apparent diffusion coefficients in the involved brain regions. This evidence of increased water diffusion favors the cerebrovascular dysregulation mechanism. However, to our knowledge, no prior study has used diffusion-tensor MR imaging in the evaluation of RPLS, and no information regarding possible alterations of diffusion anisotropy (A) in this disorder has been available. Diffusion-tensor imaging generates a complete mathematic description of water diffusion at each imaging voxel, from which rotationally invariant parameters quantifying its overall magnitude (apparent diffusion coefficient) and directionality (anisotropy) can be derived.
In keeping with the premise of cerebrovascular autoregulatory dysfunction, we hypothesize the following regarding brain water diffusion in RPLS: (a) that the increases in apparent diffusion coefficient seen with RPLS will be accompanied by anisotropy loss, as the interstitial water accumulation associated with vasogenic edema should reduce the directionality of diffusion along white matter tracts; (b) that the magnitude of the changes in apparent diffusion coefficient and A will be correlated, since they both arise from the same underlying pathophysiologic process; (c) that the observed areas of increased apparent diffusion coefficient and anisotropy loss will be limited primarily to low-anisotropy structures such as gray matter and subcortical white matter, which is consistent with the observation of Shimony et al (20) that the magnitude of white matter anisotropy is directly related to the resistance of white matter to vasogenic edema; and (d) that the diffusion abnormalities in RPLS should be reversible with the resolution of clinical signs and symptoms, as well as reversal of the findings at conventional MR imaging. The purpose of this study was to report our findings with diffusion-tensor MR imaging in the clinical evaluation of RPLS and to test these four hypotheses.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
MR Imaging Protocol
All examinations were performed with 1.5-T system imagers (Magnetom Vision; Siemens, Erlangen, Germany) with circularly polarized head coils. The diffusion-tensor imaging protocol, which has been described in detail elsewhere (20), consisted of a single-shot multisection spin-echo echo-planar pulse sequence, with four tetrahedrally oriented diffusion gradients (b = 1,012.4 sec/mm2) and three orthogonally oriented diffusion gradients (b = 337.5 sec/mm2), with a 24 x 24-cm field of view, 5-mm section thickness, and 1-mm gap between sections. A reference intensity image (b = 0.0 sec/mm2) was also obtained. Fourteen transverse sections were acquired in 35 seconds at a 96 x 128-voxel matrix (2.50 x 1.88 x 5.00-mm voxels) interpolated to a 192 x 256-pixel matrix. All images were realigned in two dimensions to correct for image displacements and linear stretch and/or shear due to eddy currents (20). For each pixel, the elements of the diffusion tensor were derived from this combination of tetrahedral and perpendicular diffusion measurements.
The isotropic diffusion coefficient () and A index were computed from the diffusion tensor. measures the magnitude of diffusion in units of millimeters squared per second and was calculated as the arithmetic mean of the three diagonal elements of the diffusion tensor:
equals one-third of the trace of the diffusion tensor, where the trace is the sum of the eigenvalues of the diffusion tensor. has also been referred to as Dav (21) and is equivalent to the apparent diffusion coefficient in those studies in which the apparent diffusion coefficient was averaged over different encoding directions to remove anisotropy effects. is invariant to the coordinate system used to measure the diffusion tensor and thus is not affected by head orientation or the direction of white matter fibers.
The dimensionless anisotropy index A was calculated as follows (22):
is equivalent to the coefficient of variation of the eigenvalues of the diffusion tensor divided by 
2. It is directly proportional to the relative anisotropy measure defined by Basser and Pierpaoli (23), with a scaling factor of 1/2, which places A on an absolute scale from zero (no anisotropy) to one (complete anisotropy). Like , A is invariant to the measurement coordinate system. In addition to the quantitative and A images, an isotropic diffusion-weighted image was derived as the geometric mean of the four tetrahedrally oriented diffusion-weighted images obtained with the diffusion-tensor sequence (24,25). This directional averaging operation removes anisotropy effects from these images. The clinical neuroimaging protocol also included a fat-saturated turbo spin-echo fluid-attenuated inversion-recovery (FLAIR) sequence (9,999/119/2,309 [repetition time msec/echo time msec/inversion time msec], echo train length of seven), which was used in 11 of 12 subjects, and a gadolinium-enhanced spin-echo T1-weighted sequence (560/17), which was used in nine of 12 subjects. These conventional MR sequences were oriented along the plane parallel to the anterior commissure and posterior commissure, or AC-PC, line. These images were therefore not necessarily acquired in register with the diffusion-tensor images, which, for technical reasons, must be oriented in the transverse plane relative to the magnet bore.
Image Analysis
To investigate if the elevated isotropic diffusion found in affected posterior brain regions in patients with RPLS was also accompanied by reduced A, we compared the mean A values of posterior regions of interest (ROIs) encompassing the zone of abnormally increased with the mean A values of anterior ROIs in areas of normal-appearing brain. On the diffusion-tensor images, both anterior and posterior ROIs were defined on a single representative transverse section in each patient study (ANALYZEAVW; Mayo Foundation, Rochester, Minn). The representative transverse section chosen was that which showed the greatest confluent area of diffusion abnormality on images. Each ROI in posterior regions of diffusion abnormality on the image was defined with an automated threshold-based segmentation algorithm that included all contiguous pixels greater than a certain baseline value.
Regions containing cerebrospinal fluid, such as cerebral sulci and ventricles, were first excluded from consideration by the automated segmentation algorithm with a manually drawn limiting trace. All such posterior ROIs were located in the posterior parietal lobes or occipital lobes. In cases of discontinuity among the regions of diffusion abnormality on the representative transverse section, more than one ROI was defined, and the results from these ROIs were combined to represent the posterior ROI. No more than two spatially separated ROIs were required to encompass the region of visible abnormality for the representative transverse section in any MR examination. The baseline value used for the threshold-based segmentation was set to the mean of the anterior ROI in areas of normal-appearing brain.
Each anterior ROI was manually traced on the representative image in regions of normal-appearing gray matter and subcortical and deep white matter of the frontal lobe or anterior temporal lobe that were homologous to the corresponding posterior ROI. The anterior ROIs were also drawn to be similar in area to the posterior ROI encompassing the region of diffusion abnormality. In one case of posterior leukoencephalopathy that was complicated by acute cerebral ischemia, a third ROI delineating the region of cytotoxic edema was defined on the image, by using the automated segmentation procedure, as all contiguous pixels less than the mean of the anterior ROI.
The mean and SD of the pixel values were computed for each ROI in each of the 12 subjects. The Shapiro-Wilk W test for normality was used to establish that the distribution of ROI means of and A did not deviate significantly from a Gaussian distribution. The mean A values in the posterior ROIs were compared with those of the control anterior ROIs across all 12 patients by using the paired two-population two-tailed t test. To examine whether the measured changes in and A were correlated across subjects, the linear correlation coefficient between the ROI means of and A was computed. Statistical inference testing was performed on the correlation coefficient by using the t test.
To investigate the hypothesis that low-anisotropy structures would be preferentially affected in RPLS, the distribution of areas of abnormally increased was cataloged on images in all 12 patient studies. The tissue types involved by the lesions, with regard to gray matter or white matter, were also classified. This image analysis was limited to the cerebral hemispheres, because coverage of infratentorial brain regions with the diffusion-tensor sequence was incomplete in several subjects. White matter was categorized as commissural, projectional, or association, according to the classification scheme used by Shimony et al (20). Gray matter is known to have lower anisotropy than does white matter. Association fiber tracts, composing the subcortical and deep white matter of the cerebral hemispheres, have lower anisotropy than do projectional tracts, such as the internal capsule, which in turn have lower anisotropy than do commissural tracts, such as the corpus callosum (20). All images were evaluated by means of consensus of the two authors, a board-certified radiologist (P.M.) and a board-certified attending neuroradiologist with a certificate of added qualification in neuroradiology (R.C.M.).
Follow-up MR images obtained after clinical recovery were available in one of the 12 subjects (Fig 1). To establish quantitatively whether the regions of increased isotropic diffusion and decreased A seen with RPLS are reversible, we manually traced an ROI on the follow-up image (Fig 1, part g) in the same area as the posterior region of diffusion abnormality on the initial image (Fig 1, part c). The posterior ROI generated for the initial study could not be repeated for the follow-up study owing to differences in the section plane between examinations because of a change in patient head orientation relative to the magnet bore. The pixel values of and A from the posterior ROI of the follow-up examination were then compared with those from the posterior and anterior ROIs of the initial study by using a two-population two-tailed t test.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
and reduced anisotropy on diffusion-tensor images (Fig 1). The distribution of hyperintense lesions on images (Table 2) was almost identical to that on FLAIR images, but the lesions were more conspicuous with FLAIR images. There was preferential involvement of the subcortical white matter in all 12 cases, with extension of the abnormality into cortical gray matter in 10 of the subjects. In three subjects, some involvement of deep white matter tracts was also seen. However, projectional white matter tracts, such as the internal capsule, and commissural white matter tracts, such as the corpus callosum, were always spared. A follow-up examination after treatment was available in one patient, which documented reversal of the posterior T2-weighted signal intensity abnormalities, as well as the regions of elevated (Fig 1). There was also partial recovery of the reduced A in the affected posterior areas.
|
characteristic of acute cerebral infarction, as illustrated in Figure 2. In the case presented in Figure 3 (patient 5), there was hyperintensity that localized to regions of T2 prolongation with elevated on diffusion-weighted images, which is consistent with the T2 shine-through phenomenon. In the third case (patient 12), a small focus of markedly increased diffusion-weighted signal intensity abnormality was apparent in the right caudate nucleus, with much less conspicuous T2-weighted abnormality. There was no difference in between the right caudate nucleus and the normal-appearing left caudate nucleus. This was thought to represent a small subacute infarct, in which values had pseudonormalized during the 5-day interval between the patient’s initial presentation and the MR examination.
|
|
and A within regions of posterior diffusion abnormality and anterior regions of normal-appearing brain. The mean area of these posterior ROIs in the representative transverse section was 16.5 cm2 ± 10.9 (SD), with a range of 6.0–43.6 cm2. The for the posterior ROI was greater than that for the anterior ROI in each of the 12 subjects, as would be expected from the manner in which the posterior ROIs were defined (see Materials and Methods and subsequent Discussion). The mean of the posterior ROIs across all 12 subjects was (1.09 ± 0.13) x 10-3 mm2/sec, which was 26% greater than the mean of (0.87 ± 0.07) x 10-3 mm2/sec for the anterior ROIs. The A of the posterior ROI was less than that of the anterior ROI in each of the 12 patients. The mean A for the posterior ROIs was 0.15 ± 0.03, which was 35% less than the corresponding value of 0.23 ± 0.02 for the anterior ROIs, a statistically significant difference (t11 = 9.58; P < .001).
|
and A for the posterior ROIs were negatively correlated (r = -0.67). This inverse relationship was statistically significant (P < .018). No statistically significant correlation was found between and A for the anterior ROIs (r = -0.12; P = .719).
In addition to comparing mean values from multiple ROIs across patients, the relationship between and A can also be examined by evaluating individual pixel values within a single ROI. Figure 4 presents pixel data from the anterior ROI, posterior ROI, and an additional ROI defined in the area of cytotoxic edema in the left posterior parietal lobe for the case illustrated in Figure 2. Compared with the pixels in the region of normal-appearing brain in the left frontal lobe, many of the pixels in the zone of posterior vasogenic edema in the right parietal lobe are shifted toward higher values and lower anisotropy levels. The pixels with the greatest elevation in are the ones that show the lowest levels of anisotropy. This was not the case in the region of cytotoxic edema in the left posterior parietal lobe. values were shifted to lower values, as compared with those in the region of normal-appearing brain, with relative preservation of the range of anisotropy values throughout the ROI.
|
|
|
elevation at the initial pretreatment MR examination, as well as at least partial reversal of the abnormally reduced anisotropy (Fig 1). The mean of the posterior ROI of the posttreatment MR examination, encompassing the zone of formerly elevated diffusion, was (0.974 ± 0.20) x 10-3 mm2/sec, compared with the mean of (1.37 ± 0.33) x 10-3 mm2/sec in the posterior ROI and the mean of (0.968 ± 0.25) x 10-3 mm2/sec in the control anterior ROI of the pretreatment MR examination. The mean A in the posterior ROI of the posttreatment study was 0.19 ± 0.13, compared with the mean A of 0.12 ± 0.06 in the posterior ROI and the mean A of 0.23 ± 0.14 in the control anterior ROI of the pretreatment study. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Results of prior reports (4,9,18,19) of quantitative diffusion MR imaging in RPLS, comprising a total of 13 cases, have consistently shown elevations of apparent diffusion coefficient in posterior regions of T2-weighted signal intensity abnormality. This agrees with our results, as the mean of (1.09 ± 0.13) x 10-3 mm2/sec measured herein for the posterior ROIs in the 12 patients is greater than the normal brain apparent diffusion coefficient range of 0.70 x 10-3 mm2/sec to 0.88 x 10-3 mm2/sec found in studies of adult volunteers (26–28). The average of (0.87 ± 0.07) x 10-3 mm2/sec that we obtained in anterior regions of normal-appearing brain is within this normal range. It is also similar to the mean apparent diffusion coefficient of (0.80 ± 0.05) x 10-3 mm2/sec found by Schwartz et al (19) in areas of normal-appearing brain in seven patients with RPLS.
In the present study, the posterior ROIs would be expected to have greater mean values than the anterior ROIs, given that each posterior ROI was defined as contiguous pixels with values greater than the mean of the corresponding anterior ROI. However, if there were no difference in average between the normal-appearing anterior region and the affected posterior region, then it would not have been possible to define an extensive confluent ROI within the posterior brain region by using the automated segmentation algorithm with a minimum threshold set to the mean of the anterior region. That this was possible in every one of the 12 subjects confirmed quantitatively that the areas of visually apparent posterior hyperintensity in the images represented real elevations in isotropic diffusion.
Schwartz et al (19) report a mean apparent diffusion coefficient of (1.36 ± 0.14) x 10-3 mm2/sec in the abnormal brain regions in their subjects, which is greater than our measurements in posterior ROIs. This discrepancy is almost certainly due to methodologic differences in the definition of ROIs. Schwartz et al (19) used small ROIs that were manually positioned, likely in the areas of greatest diffusion abnormality. Since posterior regions of abnormal isotropic diffusion in RPLS are inhomogeneous (Fig 1, part c), their method may yield higher values than the automated procedure used herein, which defined larger ROIs that also contain areas of more modestly elevated isotropic diffusion. We think that the more inclusive ROIs used in this study are representative of the full extent of diffusion abnormality in RPLS.
We extend the results of earlier studies (4,9,18,19) of diffusion-weighted imaging in RPLS by establishing that the known increases in isotropic diffusion in affected posterior brain regions are also accompanied by large reductions in A, by an average of 35%, as compared with anterior areas of normal-appearing brain. Since there is no anisotropy difference between homologous anterior and posterior regions of the cerebral hemispheres in the normal brain (20), this anisotropy loss can be attributed to the pathophysiologic process. We also demonstrate that the magnitude of the changes in and A are inversely correlated, as would be expected if the same underlying mechanism were giving rise to both the increased overall rate of water diffusion and its loss of directionality. No statistically significant relationship between and A was found in the anterior areas of normal-appearing brain.
The results of this study confirm that the areas of elevated and reduced A are restricted largely to gray matter and association white matter, which have the lowest baseline levels of anisotropy in the brain. In addition, this study provides an example of reversal of the increase and A loss in a patient with RPLS who had a follow-up examination after her symptoms resolved with antihypertensive therapy.
The increase in isotropic diffusion, the decline in anisotropy, the confinement of these diffusion changes to zones of normally low anisotropy, and their reversibility all support the hypothesis of vasogenic edema as the pathophysiologic mechanism responsible for the imaging attributes of uncomplicated RPLS. The interstitial accumulation of water that occurs with vasogenic edema leads to an increase in extracellular volume and in free water diffusing capacity, reflected by an increase in . Extravasation of fluid in the spaces between myelinated axons, as well as between fiber bundles, also reduces the directionality of water diffusion in white matter.
Shimony et al (20) have pointed out that the magnitude of white matter anisotropy, as measured with the A index also used herein, directly correlates with the known resistance of white matter to the spread of vasogenic edema. High-anisotropy white matter, such as commissural and projectional fiber tracts, is relatively impervious to vasogenic edema, whereas the low-anisotropy association fibers found in subcortical white matter are more susceptible. Hence, rotationally invariant indices such as A appear to be sensitive to the microstructural properties of white matter, such as degree of myelination and fiber packing density (23,26), that account for its pathophysiologic response to edema.
The reversibility of the changes in and A is also consistent with vasogenic edema and excludes a host of other irreversible pathophysiologic mechanisms. To our knowledge, recovery of anisotropy loss has never been reported before. This finding has implications for studies in which diffusion anisotropy is used as a gauge of the structural integrity of white matter. Results of investigations in which diffusion-tensor imaging was used have revealed white matter anisotropy loss in disorders such as multiple sclerosis (29,30), age-related leukoaraiosis (31,32), cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (33), and wallerian degeneration and gliosis due to evolving infarction (34–36). Implicit in these studies is the idea that reductions in anisotropy can be used as a marker for irreversible white matter injury. However, caution should be exercised in interpreting changes in anisotropy, given the observation herein that marked anisotropy loss can be produced by vasogenic edema and is reversible with resolution of the edema. This is especially true for disorders in which vasogenic edema is a known part of the pathophysiologic process, such as evolving infarction.
Sorensen et al (37) demonstrated a small reduction in white matter anisotropy during the acute phase of cerebral infarction. This anisotropy loss is not as large as the mean 35% decrease in A encountered in the patients with RPLS in our study. In a case in which RPLS was complicated by acute ischemia, the anisotropy loss was much more apparent in the region of vasogenic edema than in the region of cytotoxic edema (Figs 2, 4). In addition, cytotoxic edema does not evince the same preference for low-anisotropy structures that vasogenic edema does. In a recent diffusion-tensor imaging study (38) of acute and early subacute cerebral infarction, greater reductions in were found in white matter than in gray matter. Hence, cytotoxic edema differs from vasogenic edema in its effects on both the magnitude of water diffusion and its anisotropy.
Diffusion-weighted imaging has been shown to be the most accurate test for acute stroke (39). However, hyperintensity on diffusion-weighted images can sometimes be observed in uncomplicated RPLS owing to T2 prolongation in areas of vasogenic edema. The three cases in this series that manifested regions of increased diffusion-weighted signal intensity highlight the difficulty of interpreting this finding in the setting of RPLS. Quantitative diffusion measurements, in the context of the time between the onset of the patient’s symptoms and the MR examination, can aid in arriving at a definitive diagnosis (4,18).
In one case of hyperintensity on diffusion-weighted images (patient 3), the finding of reduced with relative preservation of anisotropy suggested the diagnosis of acute cerebral ischemia (Fig 2). One differential consideration for this finding is restricted diffusion secondary to prolonged seizure activity, which has recently been reported in one case of RPLS (40). However, the persistent neurologic deficits in patient 3 were more consistent with infarction. This finding was confirmed at follow-up conventional MR imaging 5 days after the initial examination, which demonstrated petechial gray matter hemorrhages and leptomeningeal gadolinium enhancement in the left temporal lobe, which are characteristic of subacute infarct.
In a second case of increased diffusion-weighted signal intensity (patient 12), the finding of apparently normal values within the affected region of the head of the right caudate nucleus 5 days after onset of symptoms allowed the diagnosis of early subacute infarct with pseudonormalization of . In the third case (patient 5), elevated values and strikingly reduced A within 5 hours after symptom onset implied that the observed diffusion-weighted signal intensity increase was merely the result of T2 shine-through in a case of uncomplicated RPLS (Fig 3). It should be noted that these same changes in isotropic diffusion and anisotropy, with residual hyperintensity on diffusion-weighted images, could also be encountered the first few days after cerebral infarction (34–36,41). Hence, clinical history with regard to the time of onset of symptoms may be important in distinguishing acute vasogenic edema from late subacute ischemic stroke.
In summary, diffusion-tensor imaging of RPLS revealed that the increase in the magnitude of water diffusion, confined largely to cortical gray matter and subcortical white matter, is accompanied by a reduction in diffusion anisotropy; these changes may be reversible and support the hypothesis of vasogenic edema due to cerebrovascular autoregulatory dysfunction as the underlying pathophysiologic mechanism. However, progression of cerebrovascular dysregulation may lead to blood-brain barrier breakdown, as manifested by abnormal gadolinium enhancement; to intracerebral hemorrhage; and to acute cerebral ischemia. Interpretation of diffusion-weighted images alone, without the benefit of quantitative diffusion information, can be problematic in differentiating uncomplicated RPLS from cerebral infarction. With continued improvements in MR hardware, such as the advent of 3-T imagers with stronger and faster diffusion gradients, as well as software, such as online tensor computation with image realignment, diffusion-tensor imaging may become more widely used in the detection and characterization of RPLS.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
= diffusion anisotropy,
= isotropic diffusion coefficient,
FLAIR = fluid-attenuated inversion recovery,
ROI = region of interest,
RPLS = reversible posterior leukoencephalopathy syndrome Author contributions: Guarantors of integrity of entire study, P.M., R.C.M.; study concepts and design, P.M., R.C.M.; literature research, P.M., R.C.M.; clinical studies, P.M.; data acquisition, P.M.; data analysis/interpretation, P.M., R.C.M.; statistical analysis, P.M., R.C.M.; manuscript preparation, P.M.; manuscript definition of intellectual content, editing, revision/review, and final version approval, P.M., R.C.M.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. L. Hollingsworth, D. P. Frush, J. Kurtzburg, and V. K. Prasad Pediatric Hematopoietic Stem Cell Transplantation and the Role of Imaging Radiology, August 1, 2008; 248(2): 348 - 365. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W.S. Bartynski Posterior Reversible Encephalopathy Syndrome, Part 1: Fundamental Imaging and Clinical Features AJNR Am. J. Neuroradiol., June 1, 2008; 29(6): 1036 - 1042. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W.S. Bartynski Posterior Reversible Encephalopathy Syndrome, Part 2: Controversies Surrounding Pathophysiology of Vasogenic Edema AJNR Am. J. Neuroradiol., June 1, 2008; 29(6): 1043 - 1049. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W.S. Bartynski, H.P. Tan, J.F. Boardman, R. Shapiro, and J.W. Marsh Posterior Reversible Encephalopathy Syndrome after Solid Organ Transplantation AJNR Am. J. Neuroradiol., May 1, 2008; 29(5): 924 - 930. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Mukherjee, J.I. Berman, S.W. Chung, C.P. Hess, and R.G. Henry Diffusion Tensor MR Imaging and Fiber Tractography: Theoretic Underpinnings AJNR Am. J. Neuroradiol., April 1, 2008; 29(4): 632 - 641. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Mak, B. P. L. Chan, I. B. Yeh, R. C. M. Ho, M. L. Boey, P. H. Feng, D. R. Koh, and B. K. C. Ong Neuropsychiatric lupus and reversible posterior leucoencephalopathy syndrome: a challenging clinical dilemma Rheumatology, March 1, 2008; 47(3): 256 - 262. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Leroux, J. Sellam, N. Costedoat-Chalumeau, D. Le Thi Huong, A. Combes, N. Tieulie, J. Haro |
/123; b = 1,006 sec/mm2) obtained with a widely used commercial protocol, with the diffusion gradient applied only in the section-select direction, demonstrates hyperintensity bilaterally in both frontal lobes (long arrows) and both posterior parietal lobes (short arrows). These findings could be interpreted as infarcts along the border zones between the territories of the anterior cerebral artery and middle cerebral artery, or as parasagittal infarcts secondary to superior sagittal sinus thrombosis. b, FLAIR image (9,999/119/2,309) exhibits T2-weighted signal intensity abnormalities in the same anterior cerebral artery and middle cerebral artery watershed distribution. c, Nonenhanced and d, gadolinium-enhanced T1-weighted images (560/17) show abnormal enhancement in the anterior cerebral artery and middle cerebral artery border zones. This parenchymal enhancement and the degree of increased T2-weighted signal intensity would be unusual for hyperacute stroke. e, Isotropic diffusion-weighted image (3,000/97.4; b = 1,012.4 sec/mm2) does not show the same abnormalities as does the single-axis diffusion-weighted image in a. This discrepancy results from the shorter echo time (97.4 vs 123.0 msec) and the absence of anisotropy on the isotropic diffusion-weighted image, which indicates that the hyperintensity on the diffusion-weighted image in a is the product of both T2 shine-through and anisotropy effects. f, 
