(Stroke. 2001;32:925.)
© 2001 American Heart Association, Inc.
Original Contributions |
Direct, Longitudinal Comparison of 1H and 23Na MRI After Transient Focal Cerebral Ischemia
From the Neuroscience Program (S.-P.L.) and Departments of Biostatistics (J.P.M.), Internal Medicine (J.J.H.A.), Pediatric Neurology (J.J.N.), and Radiology (J.J.H.A., J.J.N.), Washington University School of Medicine, St Louis, Mo, and the Department of Chemistry (S.-K.S., J.J.H.A.), Washington University, St Louis, Mo.
Correspondence to Jeffrey J. Neil, Biomedical MR Laboratory, Washington University School of Medicine, 4525 Scott Ave, Room 2313, St Louis, MO 63110. E-mail neil{at}wuchem.wustl.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Background and Purpose—23Na MRI may offer new insight into the evaluation of tissue injury. We performed a direct, longitudinal, morphological comparison of 1H T2 relaxation, 1H apparent diffusion coefficient (ADC), 23Na content, and histopathology after cerebral ischemia to address the hypotheses that (a) 23Na MRI is unique in comparison to 1H MRI, and (b) accumulation of 23Na is an unambiguous marker for dead tissue.
Methods—Rats underwent 30 minutes of focal ischemia. MRIs of 1H T2, 1H ADC, and 23Na content were acquired from 12 hours up to 1, 2, or 14 days after reperfusion. On excision, brains were stained with triphenyltetrazolium chloride (TTC).
Results—In all cases, the region of abnormality increased in size for 2 days. On day 5, both 1H T2 and ADC temporarily appeared normal despite the presence of TTC-defined infarction. By comparison, the volume of tissue exhibiting abnormally intense 23Na signal mirrored the TTC-defined infarct at all time points.
Conclusions—Regions of high 23Na content correlate well with the TTC-defined infarct and may be a quantitative in vivo marker for dead tissue. In contrast, the dynamics of the 1H T2 and ADC make it difficult to interpret these images without additional information because they may appear normal despite infarction. Neither type of 1H image delineates dead tissue, and none of these methods predicts the potential infarct size at early time points.
Key Words: animal models • magnetic resonance imaging • sodium • stroke, experimental • rats
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Imaging studies of stroke often test new techniques for their ability to (a) reveal the location and extent of ischemic injury at early time points and (b) quantify the degree of damage that has occurred or may subsequently develop. In the simplest terms, this involves distinguishing normal from dead or injured tissue, usually defined histologically. In the field of MRI, characteristic changes in 1H T2 relaxation and 1H apparent diffusion coefficient (ADC) have been correlated with tissue viability after ischemic stroke.1 2 3 4 5 Elevated 1H T2 is considered the MR gold standard for infarction, and decreased 1H ADC is considered an early marker of ischemic injury. However, the biophysical processes underlying the changes in T2 and ADC are incompletely understood. In addition, the temporal evolution of these signals can be dynamic, requiring one to account for variables such as the duration, depth, and elapsed time since ischemia to make the proper interpretation.5 6
All cells actively exclude sodium cations from the
intracellular space. On cell death, Na+ can
accumulate inside cells, resulting in an increase in bulk tissue
Na+ content (given a source for additional
Na+). Assuming an
intracellular-to-extracellular volume ratio of
4:17 and corresponding
Na+ concentrations of 10 and 140
mmol/L, bulk Na+ concentration in rat brain
is normally 
40 mmol/L. After cell death, bulk
Na+ concentration can approach extracellular
levels, potentially resulting in an increase of >200%. In this study,
we used a rat model of focal cerebral ischemia and performed a
direct, longitudinal, morphological comparison between estimated
1H T2 maps, estimated
1H ADC maps, 23Na
images, and triphenyltetrazolium chloride
(TTC)-stained sections. The hypotheses we tested were (a) that the
abnormalities detected by 23Na MRI are
unique in comparison to the estimated 1H T2
and ADC maps and (b) that the accumulation of
Na+ is an unambiguous marker for dead
tissue.
High-resolution 23Na imaging is relatively slow (see Imaging section). To remove the potentially confounding effect of inadequate temporal resolution from our study, we used a stroke model in which the Na+ abnormality developed slowly relative to the time required for data acquisition. Tandem occlusion of the right middle cerebral artery (MCAO) and both common carotid arteries (CCAs) for 30 minutes has been shown to produce cortical infarctions that develop over several days8 (we refer to this as the "3-vessel MCAO"). On this time scale, despite the low 23Na MR sensitivity and low mean Na+ tissue concentration, we could readily acquire 23Na images at an adequate temporal and spatial resolution.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Surgery
Twenty-four male Long-Evans rats (Harlan Bioproducts), weighing 250 to 300 g, were subjected to 3-vessel MCAO.8 Anesthesia was induced by intraperitoneal injection of 400 mg/kg chloral hydrate, and fiberoptic probes (Luxtron Corp) were placed in the rectum and right nares for detection of body and head temperatures. A LabView "Virtual Instrument" (National Instruments) independently controlled temperature in both channels. Body temperature was maintained at 37°C with an electric heating pad, and head temperature was maintained with heat from a light bulb. Both CCAs were isolated through an incision in the neck. Next, under a dissecting microscope, the right MCA was visualized through a split in the temporalis muscle, and a burr hole was drilled at the junction of the zygomatic arch and squamous bone. A No. 1 Sundt AVM Micro-clip (Johnson and Johnson Professional, Inc)9 was placed on the distal MCA near the junction with the inferior cerebral vein, and microaneurysm clips (Roboz Surgical) were placed on both CCAs. Before occlusion, head and body temperatures remained the same without heating of the head, but after occlusion, head temperature dropped quickly. During ischemia, heat from the incandescent bulb was required to maintain head temperature, which we set at 36°C (Figure 1
). All clamps were removed after 30 minutes. During
this procedure, the site of MCAO was observed through the microscope to
visually verify refilling of the artery. All procedures were conducted
in accordance with protocols approved by the Washington University
Animal Studies Committee.
|
Imaging
Images were acquired at 4.7 T (Oxford Instruments and
Varian, Inc) with a 3-cm-diameter, 1H and
23Na double-tuned birdcage coil. The animals
were anesthetized with halothane (1.5% in
O2), and body temperature was maintained at
37°C with a heated water pad. The head was restrained within the
radiofrequency coil at the teeth and ears. Six animals were
killed at each end point of 0.5, 1, 2, or 14 days after reperfusion,
and the imaging was performed at each available time point of 0.5, 1,
2, 3, 5, 7, and 14 days.
23Na images were acquired with a 3D gradient echo sequence. Data for this experiment were acquired over a period of several months, during which time we optimized our acquisition parameters. Before optimization, 13 animals (of 24) were scanned with a voxel size of 0.55x0.80x2.5 mm3 (1.1 µL), with TE=2 ms, TR=40 ms, flip angle=65°, and bandwidth=50 kHz. Imaging time was 1 hour, 57 minutes for 125 repetitions. After optimization for maximum signal-to-noise per unit time, the remaining 11 animals were scanned with a voxel size of 0.55x0.55x2.5 mm3 (0.76 µL) with TE=5 ms, TR=20 ms, flip angle=45°, and bandwidth=8 kHz. Imaging time was 1 hour, 25 minutes for 125 repetitions. Image quality was comparable for the 2 sets of parameters. The relatively small voxel size (for 23Na MRI) kept volume estimation errors minimal.
To minimize scan time, 2-point estimates of 1H T2 and ADC were performed with a multislice, fast spin-echo sequence with an echo train length of 8. Ten coronal slices were acquired with an in-plane resolution of 0.27x0.27 mm2 and slice thickness of 2.5 mm (0.18 µL). The animal was not moved between the 23Na and 1H scans, and 1H images were aligned with the middle 10 coronal slabs from the 3D 23Na images (with 4x greater in-plane resolution). Two images, at TE=10 and 60 ms, and TR=2 seconds, were acquired to generate estimated T2 maps. Acquisition time was 4.3 minutes for 4 repetitions. Similarly, 2 images, at b=0 and 1193 s/mm2 and TE=36 ms and TR=2.2 seconds, were collected to generate estimated ADC maps. Diffusion encoding was along the axis of the animal’s body. Imaging time was 9.4 minutes for 8 repetitions.
Histopathology
On the predetermined day of euthanasia, the rats were
administered 400 mg/kg IP chloral hydrate and underwent
transcardiac perfusion with PBS. The brains were removed
and frozen in a cutting block with polyvinyl alcohol/polyethylene
glycol embedding medium (Sakura Finetek), after which they were cut
into 2-mm-thick coronal sections and incubated for 10 minutes at 37°C
in 2% TTC (Sigma) in PBS. Stained sections were placed on a Petri dish
such that the larger of the 2 surfaces was face down. Twenty-four–bit
color images of these surfaces were acquired at 300 dots per inch with
a flatbed scanner (Hewlett Packard).
Data Analysis
To segment the color images of the TTC-stained brain
sections, we used the semiautomated algorithm devised by Goldlust et
al,10 with the Metamorph
image analysis package (Universal Imaging Corp). Images in
Figures 2, A through C, are shown with their segmentation
results. TTC-defined infarct volume was normalized to total ipsilateral
cortical volume, measured manually with Adobe Photoshop (Adobe Systems
Inc).
|
MRI analysis was performed with functions written to
run within the Image Browser analysis package (Varian, Inc). T2
maps were estimated from the variable TE images with a 2-point,
voxel-by-voxel fit to the
equation
 | (1) |
 | (2) |
The 23Na images and calculated
ADC and T2 maps were then segmented. We first manually delineated the
ipsilateral and contralateral cerebral hemispheres. Our segmentation
algorithm then extracted the volume of "abnormal voxels" in 2
steps. First, the images were smoothed with a blurring filter, and the
algorithm compared homologous locations in the 2 hemispheres in search
of a seed point within the abnormality. The position yielding the
largest percent difference in intensity between the 2 hemispheres was
taken as the seed point. Next, a comparison of homologous voxels in the
unfiltered images was performed, starting from the seed point and
extending recursively in all directions. Voxels exhibiting a predefined
minimum percent difference between the 2 hemispheres were labeled as
"abnormal" and allowed the algorithm to continue. Otherwise, the
search was terminated. We used a cutoff of 10% for the
1H images and 40% for the
23Na images (see Discussion). Segmentation
results have been superimposed onto the sample images in
Figure 2, D through O
. We normalized the abnormal volumes to
total ipsilateral cortical volume, measured manually. Having determined
the spatial extent of the abnormality, we recorded both its volume
and the average percent difference between voxels within the
abnormality and their homologous counterparts in the contralateral
hemisphere
 | (3) |
Statistical Analysis
The purpose of our statistical analysis was
2-fold. Within each data type, we first compared each time point with
(a) the first scan (day 0.5), and (b) the previous scan. These
comparisons indicated (a) when significant differences from our
"close to baseline," 0.5-day measurements existed and (b) at what
time point significant changes occurred. For the TTC-stained sections,
we performed 1-way, between-group ANOVA, and for the MRIs we used
1-way, repeated-measures ANOVA with the "mixed procedure" of the
SAS System (SAS Institute, Inc). This procedure accounted for the
decreasing number of data points at progressively later time points as
the result of animal euthanasia for histological study.
We evaluated the differences between specific pairs of time points, and
the probability values were adjusted for multiple comparisons. A value
of P<0.05 was considered
statistically significant. The final analysis that we performed
was a comparison (paired t
test) of the abnormality volume from each MRI type with the TTC-defined
infarct volume at each time point of
euthanasia.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Figures 2
, 3, and 4 provide sample images and summary data
for the temporal evolution of the TTC-defined infarct and MRI
abnormalities. The 30 minute, 3-vessel MCAO produced infarcts with
consistent location, extent, and temporal development. However,
in our hands, the rate of infarct development was somewhat faster than
previously reported.8 Of 24
rats, 3 were excluded. One animal from the 14-day group lacked a
TTC-defined infarct. From the 2-day group, 1 animal died and 1
exhibited subcortical infarction. (These 3 animals were scanned with
nonoptimized 23Na imaging
parameters.) Because of a spectrometer malfunction, the
0.5-day 23Na image was not acquired for 1
rat from the 2-day group.
|
). Data shown are mean±SEM.
|
TTC-stained sections are shown in
Figure 2, A through C, and the evolution of TTC-defined
infarct volume is shown in
Figure 3A. Infarct size grew to 50% of the ipsilateral
cortex over the first 2 days. On day 0.5, 1 animal of 6 exhibited a
TTC-defined infarct. That fraction increased to half of the animals on
day 1 and all of the animals on day 2. Statistical analysis
revealed 2 salient results. First, days 2 and 14 were significantly
different from day 0.5 (P=0.002
and P=0.003) but not from each
other. Thus, infarct volume peaked by day 2. Second, the difference
between days 1 and 2 was statistically significant, indicating that the
majority of infarct growth occurred between these 2 time points with no
significant changes thereafter
(P=0.016).
1H T2 maps are shown in
Figure 2, D through G, and the evolution of the volume of
tissue with prolonged T2 is shown in
Figure 3B. Abnormality volume grew to 50% of the affected
cortex over the first 2 days. On day 0.5, 5 of the 21 animals exhibited
areas of increased T2. That fraction increased to 11 of the remaining
15 animals on day 1 and all remaining animals on day 2. Beyond 2 days,
T2 contrast in the core of the abnormality normalized, decreasing
abnormality volume considerably by day 5. T2 contrast began to return
by day 14. Statistical analysis indicated that days 1, 2, 3,
and 14 were significantly different from day 0.5
(P<0.0001,
P<0.0001,
P<0.0001, and
P=0.003). Changes in volume
occurred on days 1, 2, and 5
(P<0.0001,
P=0.0002, and
P=0.0004). Thus, the
abnormality grew the first 2 days, decreased in size between days 3 and
5, and reappeared by day 14. Finally,
Figure 3C shows the average percent difference, over time,
between abnormal voxels and their homologous counterparts in the
contralateral cortex. Statistical analysis showed that whenever
an abnormality was detected, the average change in T2 was fairly
constant, with an overall average percent difference for all time
points of +33% (increase from 50 to 66.5 ms).
1H ADC maps are shown in
Figure 2, H through K, and the evolution of the volume of
tissue with abnormal ADC is shown in
Figure 3D. Considering decreases in ADC first, abnormality
volume grew the first 2 days to 50% of the ipsilateral cortex. On day
0.5, 15 of 21 animals had regions of decreased ADC, and by day 1, this
was true of all remaining animals. Abnormality volume peaked on day 2,
after which the ADC in the abnormal region normalized and ultimately
evolved into increased ADC. Statistical analysis showed that
days 1 and 2 were significantly different from day 0.5
(P=0.013 and
P<0.001). It is important to
point out, however, that in contrast to TTC, T2, and
23Na (as we shall see), there was a
nonnegligible abnormality volume on day 0.5. Changes in abnormality
volume occurred on days 1, 2, and 3
(P=0.013,
P=0.002, and
P=0.002), consistent
with the impression that abnormality volume peaks transiently.
Figure 3E shows the average percent difference between
abnormal voxels and their contralateral counterparts. A difference was
found between days 0.5 and 1
(P=0.043). Otherwise, the
difference in ADC between voxels within the abnormality and their
homologous counterparts was fairly constant, with an average percent
difference for all time points of -23% (decrease from
0.65x10-3 to
0.5x10-3
mm2/s).
A similar analysis was performed for regions of abnormally high ADC. No abnormal volumes were detected until day 7. Statistical analysis showed that days 7 and 14 were different from day 0.5 (P=0.0003 and P<0.0001), with the only change in volume found on day 7 (P=0.028). Thus, the region of abnormally high ADC developed between days 5 and 7 and persisted through the second week.
Last, 23Na images are shown in
Figure 2, L through O, and the evolution of the volume of
tissue with abnormally intense 23Na signal
is shown in
Figure 3F. Abnormality volume grew over the first 2 days to
include 50% of the affected cortex. On day 0.5, 8 of 21 animals
exhibited volumes of elevated 23Na. That
fraction increased to 12 of the remaining 15 animals on day 1 and all
remaining animals on day 2. The 23Na
abnormality persisted for the length of our study. Statistical
analysis showed that all time points from day 1 to 14 were
different from day 0.5
(P=0.002,
P<0.0001,
P<0.0001,
P<0.0001,
P<0.0001, and
P=0.0004), but changes in
volume occurred only on days 1 and 2
(P=0.002 and
P=0.0004). Thus, the
abnormality volume grew for 2 days, then remained stable.
Figure 3G shows the temporal evolution of the average
percent difference between abnormal voxels and their contralateral
counterparts. Differences were found between day 0.5 and days 2, 3, and
5 (P=0.003,
P=0.018,
P=0.032), indicating a peak in
contrast between 2 and 5 days that corresponded with the peak in the
volume of abnormality and the peak in ipsilateral cortical volume
(Figure 4). The average percent difference for all time
points was +213%.
The data from
Figure 3, A, B, D, and F, are superimposed in
Figure 5. At the time points at which both TTC and T2 data
were acquired (days 0.5, 1, 2, and 14), no statistically significant
differences between the abnormality volumes were detected. However,
assuming that tissue can never reverse its state of TTC staining, there
was clearly a deviation on days 5 and 7. Comparing the volumes of
abnormally low ADC with the TTC-defined infarct, we found significant
differences on days 0.5, 1, and 14
(P=0.0009,
P=0.02, and
P=0.003). The volume with low
ADC was larger than the TTC-defined infarct size at the first 2 time
points but still below its maximum volume. These 2 data types matched
on day 2, and on day 14, there was no volume of abnormally low ADC
despite the presence of infarct. Comparing the TTC-defined infarct with
volumes of elevated ADC revealed a significant difference on day 2
(P=0.028). The infarct reached
maturity at that time point, but no regions of elevated ADC had
developed. However, by day 14, the volume of elevated ADC grew to match
the infarct volume. Finally, no difference was found between the volume
of abnormally elevated 23Na signal intensity
and the TTC-defined infarct. Also, the shapes of the
23Na MRI and TTC time courses were very
similar.
Figure 6 is a scatterplot of TTC-defined infarct volume
versus the volume of tissue with abnormally elevated
23Na signal intensity for each animal at the
time point of euthanasia. Eight animals were killed before the
detection of any abnormality (5 from the 0.5-day group and 3 from the
1-day group) and consequently fell on the origin. Linear regression
yielded a slope of 1.0,
y-intercept of 2.4%, and
r=0.95, indicating a very
strong correlation.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We compared 1H T2 maps, 1H ADC maps, 23Na images, and TTC-stained sections in a rat model of focal cerebral ischemia to test the hypotheses (a) that the abnormalities reported by 23Na MRI are unique in comparison to 1H T2 and ADC and (b) that the accumulation of Na+ is an unambiguous marker for dead tissue. These hypotheses were based on the notion that frank loss of the cellular electrochemical gradient and subsequent equilibration of Na+ into the intracellular space is synonymous with cell death. Although we have not made independent measurements of tissue sodium concentration (TSC) to calibrate our images, the use of 23Na MRI for quantitative TSC mapping is well established.11 In particular, Thulborn et al12 have estimated that a TSC of
55 mmol/L is indicative of tissue
that can be preserved with reperfusion, whereas a TSC >70 mmol/L
is indicative of infarction.
23Na Images
Without Na+ delivery, total
Na+ content must remain constant.
Equilibration of Na+ into the
ischemic region depends on (a) the rate of diffusion and (b)
the distance to the source. Assuming flowing capillary blood to be an
infinite source, we must consider both reperfused and unperfused
tissue. For isotropic
diffusion,
r2/t=ADC
In other words, the mean-squared displacement
(r) per unit time equals the
ADC. In saline, the ADC for Na+ is 50% to
60% smaller than for
H2O.13
Given that the 1H ADC for
H2O in rat brain is roughly
1x10-3
mm2/s, and assuming that the proportionality
is maintained, we estimate the ADC for Na+
in vivo to be
0.55x10-3
mm2/s. In reperfused cortex, the average
distance between active capillaries is 50
µm,14 and it would take
1 second for Na+ to diffuse into the
intervening space. In unperfused cortex, the distance to the nearest
collateral vessel is 0.5 mm, and it would take 8 minutes for
Na+ to permeate the ischemic tissue.
In humans, the distance to collateral vessels could be 1 cm, in which
case the time for diffusion into the ischemic core would be 2
days. Thus, unlike our rat model, 23Na MRI
of human stroke must be interpreted with this potential time delay in
mind.
An alternative interpretation of the increased
23Na signal intensity is a change in MR
relaxation. Either a decrease in 23Na T1 or
an increase in 23Na T2 could also cause an
increase in 23Na signal intensity. However,
on the basis of published values for 23Na T1
and
T215 16 17 18 19 20
and our finding that T1 of 23Na is 60 ms
for both normal and infarcted brain at 4.7 T (unpublished observation,
2000), this effect is unlikely. An alternative approach to this
question is to measure 23Na signal intensity
after ischemia under circumstances in which total
Na+ content remains constant (ie, complete
global cerebral ischemia). In a gerbil model of global
ischemia, Allen et
al21 demonstrated a decrease
in signal intensity associated with injury. Similarly, we observed a
10% drop in signal intensity in rat after cardiac arrest (with the use
of the same relevant MRI parameters as described herein;
unpublished observation, 2000). Thus, changes in MR relaxation appear
more likely to account for a decrease in
23Na signal intensity. We thus attribute the
increased 23Na signal intensity to increased
spin density.
Estimated 1H T2 and
ADC Maps
Our 2-point estimates of 1H
T2 and ADC with the fast spin-echo pulse sequence were nonquantitative
relative measurements subject to measurement bias. However, this did
not pose a problem because we followed changes in contrast over time,
relative to the contralateral cortex. In addition, this method (a)
removed the effects of radiofrequency inhomogeneity and T1 relaxation
and (b) minimized scan time while producing results comparable to
values found in the literature. Previously published results report
normal T2 in rat brain to be 60 to 75 ms and normal ADC to be 0.6 to
0.7x10-3
mm2/s.5 22 23 24
We found T2 to be 50 ms and ADC to be 0.7 to
0.8x10-3
mm2/s. The discrepancy probably is due to a
combination of measurement bias and the fact that our measurements were
confined to the cortex.
One potential source of bias was the use of single-axis diffusion measurements. However, even though diffusion in the rat cortex is not isotropic, it has been reported that ADC maps sensitized along the body axis correlate well with both TTC-defined infarction and abnormalities defined by directionally averaged ADC values.25 26 There is also evidence that the difference in ADC between normal and infarcted tissue may be larger in this direction, resulting in greater sensitivity.27
Data Analysis
Regarding MRI, a potential source of bias was the
cutoffs chosen to define abnormal versus normal voxels. We arrived at
these values by plotting histograms of small cortical regions, as in
Figure 7. The voxel distributions were clearly bimodal, so
we simply chose cutoffs to approximate the width of the distribution of
normal voxels.
Figure 3 (C, E, and G) shows that the average percent
difference for each type of image was consistently larger than
the chosen cutoff. From this, we can also conclude that although the
addition of more data points in our calculated images would have
decreased the amount of uncertainty, it is unlikely that this would
have altered the conclusions of our study.
|
Because ipsilateral cortical volume varied over time
(Figure 4), a method to account for edema was required. The
"indirect method," which has previously been used for ex vivo
studies, calculates infarct volume from measurements of unaffected
tissue: Abnormality volume equals contralateral hemisphere volume minus
the volume of unaffected tissue in the ipsilateral
hemisphere.28 This method
did not work well for our MRIs because (a) in vivo, swelling of the
ipsilateral cortex distorted the contralateral hemisphere, and (b)
swelling of the ipsilateral cortex did not always coincide with other
imaging abnormalities (on day 0.5 in particular), resulting in
"corrected" volumes less than zero. Thus, in lieu of this
procedure, we calculated the percentage of ipsilateral cortex that was
abnormal. This method accounted for edema as well as animal-to-animal
variability in size. Regardless of the method of analysis
chosen, the conclusions of this study remain the
same.
Conclusions
Figure 5 confirms that the temporal evolution of the
abnormality detected on 23Na MRI is
different from that detected by 1H T2 and
ADC maps. The similarity in temporal evolution of the
23Na MRI abnormality and the TTC-defined
infarct volume, taken in conjunction with the correlation between the
two for individual animals
(Figure 6), suggests that 23Na MRI
may be a reliable method for determining the volume of dead tissue in
vivo. However, we cannot unequivocally assert that increased
23Na signal intensity is an unambiguous
marker for dead tissue. Further analysis of the spatial
correlation between the TTC-defined and 23Na
MRI-defined abnormalities and the addition of various other stroke
models is required.
The dynamics of the 1H signals made them difficult to interpret without contextual information, the time after injury in particular. Similar animal studies have reported decreases in T2 contrast at late time points, but none demonstrated the biphasic behavior we observed, even when the time course extended beyond 2 weeks.23 However, transient, subacute normalization of T2 contrast has been reported in cases of human stroke as the "MR fogging effect"6 and has become the subject of further investigation in our laboratory. Because previous rat studies primarily used models of permanent ischemia, it is possible that reperfusion plays a role in this phenomenon.
In contrast to the simple notion that ADC should be low during the acute phase after ischemia, biphasic behavior of the ADC has been reported by several authors.4 5 29 In particular, Li et al5 have shown (in rats) that even though ADC drops immediately after the onset of ischemia, reperfusion after 30 minutes results in an immediate, full recovery of ADC followed by a subsequent secondary decrease roughly 12 hours later. Our results corroborate this finding. We did not obtain data during the period of ischemia, but we did observe a delayed appearance of regions of low ADC, which presumably corresponds with the secondary decline reported by Li et al. On the basis of our data, this secondary decline appears to be associated with the onset of cell death.
In conclusion, none of the imaging methods tested proved useful for predicting the final infarct volume before histopathological evidence of cell death. In other words, we had no early indication of the volume of tissue that was injured, still viable, but destined to die. Intraischemic measurement of ADC or perfusion-weighted imaging might have provided this information. However, we have shown that for time points up to 12 hours after a brief but severe episode of ischemia, 23Na content, 1H T2, ADC, TTC, and presumably perfusion-weighted imaging could all appear normal despite the subsequent development of infarction.
|
Acknowledgments |
|---|
This research was funded by National Institutes of Health grants R0-1NS35912 and F30-MH12173. The authors thank Mark P. Goldberg, MD, for use of the segmentation algorithm for TTC-stained brain sections; Charles H. Anderson, PhD, for assistance in designing the segmentation algorithm for our MR images; Dmitriy Yablonskiy, PhD, for many helpful discussions; and Charles S. Springer, PhD, for insightful comments regarding the 23Na NMR experiment.
Received October 17, 2000; revision received December 14, 2000; accepted December 21, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1. Knight RA, Dereski MO, Helpern JA, Ordidge RJ, Chopp M. Magnetic resonance imaging assessment of evolving focal cerebral ischemia: comparison with histopathology in rats. Stroke. 1994;25:1252–1261.[Abstract]
2. Mintorovitch J, Yang GY, Shimizu H, Kucharczyk J, Chan PH, Weinstein PR. Diffusion-weighted magnetic resonance imaging of acute focal cerebral ischemia: comparison of signal intensity with changes in brain water and Na+,K(+)-ATPase activity. J Cereb Blood Flow Metab. 1994;14:332–336.[Medline] [Order article via Infotrieve]
3. Hoehn-Berlage M, Eis M, Back T, Kohno K, Yamashita K. Changes of relaxation times (T1, T2) and apparent diffusion coefficient after permanent middle cerebral artery occlusion in the rat: temporal evolution, regional extent, and comparison with histology. Magn Reson Med. 1995;34:824–834.[Medline] [Order article via Infotrieve]
4.
Miyasaka N, Nagaoka
T, Kuroiwa T, Akimoto H, Haku T, Kubota T, Aso T. Histopathologic
correlates of temporal diffusion changes in a rat model of cerebral
hypoxia/ischemia. AJNR
Am J Neuroradiol. 2000;21:60–66.
5.
Li F, Liu KF, Silva
MD, Omae T, Sotak CH, Fenstermacher JD, Fisher M, Hsu CY, Lin W.
Transient and permanent resolution of ischemic lesions on
diffusion-weighted imaging after brief periods of focal
ischemia in rats: correlation with histopathology.
Stroke. 2000;31:946–954.
6. Asato R, Okumura R, Miki Y, Konishi J. MR T2-weighted image of subacute cerebral infarct can be isointense to the surrounding brain: MR fogging in cerebral infarct. MAGMA. 1994;2:539–544.
7.
Hansen AJ. Effect
of anoxia on ion distribution in the brain.
Physiol Rev. 1985;65:101–148.
8. Du C, Hu R, Csernansky CA, Hsu CY, Choi DW. Very delayed infarction after mild focal cerebral ischemia: a role for apoptosis? J Cereb Blood Flow Metab. 1996;16:195–201.[Medline] [Order article via Infotrieve]
9.
Buchan AM, Xue D,
Slivka A. A new model of temporary focal neocortical ischemia
in the rat. Stroke. 1992;23:273–279.
10.
Goldlust EJ,
Paczynski RP, He YY, Hsu CY, Goldberg MP. Automated measurement of
infarct size with scanned images of
triphenyltetrazolium chloride-stained rat
brains. Stroke. 1996;27:1657–1662.
11. Christensen JD, Barrere BJ, Boada FE, Vevea JM, Thulborn KR. Quantitative tissue sodium concentration mapping of normal rat brain. Magn Reson Med. 1996;36:83–89.[Medline] [Order article via Infotrieve]
12.
Thulborn KR,
Gindin TS, Davis D, Erb P. Comprehensive MR imaging protocol for stroke
management: tissue sodium concentration as a measure of tissue
viability in nonhuman primate studies and in clinical studies.
Radiology. 1999;213:156–166.
13. Burstein D, Gray ML, Hartman AL, Gipe R, Foy BD. Diffusion of small solutes in cartilage as measured by nuclear magnetic resonance (NMR) spectroscopy and imaging. J Orthop Res. 1993;11:465–478.[Medline] [Order article via Infotrieve]
14. Metzger HP, Bruggemann H, Plewnia A. The use of fluorescent-protein conjugates for staining brain capillaries in hypo- and hyperthermia (26–42 degrees). Microvasc Res. 1987;34:311–317.[Medline] [Order article via Infotrieve]
15. Hilal SK, Maudsley AA, Ra JB, Simon HE, Roschmann P, Wittekoek S, Cho ZH, Mun SK. In vivo NMR imaging of sodium-23 in the human head. J Comput Assist Tomogr. 1985;9:1–7.[Medline] [Order article via Infotrieve]
16. Lee SW, Hilal SK, Cho ZH. A multinuclear magnetic resonance imaging technique: simultaneous proton and sodium imaging. Magn Reson Imaging. 1986;4:343–350.[Medline] [Order article via Infotrieve]
17. Winkler SS, Thomasson DM, Sherwood K, Perman WH. Regional T2 and sodium concentration estimates in the normal human brain by sodium-23 MR imaging at 1.5 T. J Comput Assist Tomogr. 1989;13:561–566.[Medline] [Order article via Infotrieve]
18.
Perman WH, Turski
PA, Houston LW, Glover GH, Hayes CE. Methodology of in vivo human
sodium MR imaging at 1.5 T.
Radiology. 1986;160:811–820.
19. Bansal N, Germann MJ, Seshan V, Shires GTd, Malloy CR, Sherry AD. Thulium 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetrakis(methylene phosphonate) as a 23Na shift reagent for the in vivo rat liver. Biochemistry. 1993;32:5638–5643.[Medline] [Order article via Infotrieve]
20. Van Emous JG Van Echteld CJ. Changes of intracellular sodium T2 relaxation times during ischemia and reperfusion in isolated rat hearts. Magn Reson Med. 1998;40:679–683.[Medline] [Order article via Infotrieve]
21. Allen KL, Busza AL, Williams SR, Williams SC. Early changes in cerebral sodium distribution following ischaemia monitored by 23Na magnetic resonance imaging. Magn Reson Imaging. 1994;12:895–900.[Medline] [Order article via Infotrieve]
22.
Knight RA,
Ordidge RJ, Helpern JA, Chopp M, Rodolosi LC, Peck D. Temporal
evolution of ischemic damage in rat brain measured by proton
nuclear magnetic resonance imaging.
Stroke. 1991;22:802–808.
23. Verheul HB, Berkelbach van der Sprenkel JW, Tulleken CA, Tamminga KS, Nicolay K. Temporal evolution of focal cerebral ischemia in the rat assessed by T2- weighted and diffusion-weighted magnetic resonance imaging. Brain Topogr. 1992;5:171–176.[Medline] [Order article via Infotrieve]
24. Jiang Q, Zhang ZG, Chopp M, Helpern JA, Ordidge RJ, Garcia JH, Marchese BA, Qing ZX, Knight RA. Temporal evolution and spatial distribution of the diffusion constant of water in rat brain after transient middle cerebral artery occlusion. J Neurol Sci. 1993;120:123–130.[Medline] [Order article via Infotrieve]
25. Lythgoe MF, Busza AL, Calamante F, Sotak CH, King MD, Bingham AC, Williams SR, Gadian DG. Effects of diffusion anisotropy on lesion delineation in a rat model of cerebral ischemia. Magn Reson Med. 1997;38:662–668.[Medline] [Order article via Infotrieve]
26.
Minematsu K, Li
L, Fisher M, Sotak CH, Davis MA, Fiandaca MS. Diffusion-weighted
magnetic resonance imaging: rapid and quantitative detection of focal
brain ischemia.
Neurology. 1992;42:235–240.
27. Dardzinski BJ, Sotak CH, Fisher M, Hasegawa Y, Li L, Minematsu K. Apparent diffusion coefficient mapping of experimental focal cerebral ischemia using diffusion-weighted echo-planar imaging. Magn Reson Med. 1993;30:318–325.[Medline] [Order article via Infotrieve]
28. Swanson RA, Morton MT, Tsao-Wu G, Savalos RA, Davidson C, Sharp FR. A semiautomated method for measuring brain infarct volume. J Cereb Blood Flow Metab. 1990;10:290–293.[Medline] [Order article via Infotrieve]
29.
Dijkhuizen RM,
Knollema S, van der Worp HB, Ter Horst GJ, De Wildt DJ, Berkelbach van
der Sprenkel JW, Tulleken KA, Nicolay K. Dynamics of cerebral tissue
injury and perfusion after temporary hypoxia-ischemia
in the rat: evidence for region-specific sensitivity and delayed
damage. Stroke. 1998;29:695–704.
This article has been cited by other articles:
 |
|
 |
|
  E.A. Mellon, D.T. Pilkinton, C.M. Clark, M.A. Elliott, W.R. Witschey 2nd, A. Borthakur, and R. Reddy Sodium MR Imaging Detection of Mild Alzheimer Disease: Preliminary Study AJNR Am. J. Neuroradiol., May 1, 2009; 30(5): 978 - 984. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. C. Jones, A. Kharlamov, B. Yanovski, D. K. Kim, K. A. Easley, V. E. Yushmanov, S. K. Ziolko, and F. E. Boada Stroke Onset Time Using Sodium MRI in Rat Focal Cerebral Ischemia Stroke, March 1, 2006; 37(3): 883 - 888. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Chen, S.-K. Song, W. Liu, M. McLean, J. S. Allen, J. Tan, S. A. Wickline, and X. Yu Remodeling of cardiac fiber structure after infarction in rats quantified with diffusion tensor MRI Am J Physiol Heart Circ Physiol, August 7, 2003; 285(3): H946 - H954. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Wang, L. Xu, H. Wang, P. R. Young, M. Gaestel, and G. Z. Feuerstein Mitogen-activated Protein Kinase-activated Protein (MAPKAP) Kinase 2 Deficiency Protects Brain from Ischemic Injury in Mice J. Biol. Chem., November 8, 2002; 277(46): 43968 - 43972. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. C. McKinstry, J. H. Miller, A. Z. Snyder, A. Mathur, G. L. Schefft, C. R. Almli, J. S. Shimony, S. I. Shiran, and J. J. Neil A prospective, longitudinal diffusion tensor imaging study of brain injury in newborns Neurology, September 24, 2002; 59(6): 824 - 833. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||