Download Pathology-guided MR analysis of acute and chronic experimental

JOURNAL OF MAGNETIC RESONANCE IMAGING 22:180 –188 (2005)
Original Research
Pathology-Guided MR Analysis of Acute and
Chronic Experimental Allergic Encephalomyelitis
Spinal Cord Lesions at 1.5T
Lisa L. Cook, PhD,1 Paula J. Foster, PhD,2 and Stephen J. Karlik, PhD1– 4*
Purpose: To directly correlate spinal cord pathology of
guinea pigs with experimental allergic encephalomyelitis
(EAE) to the MRI data obtained at 1.5T.
Materials and Methods: Spinal cords from EAE animals
were imaged in vivo with the following MRI sequences: T2FSE, PD-FSE, fluid-attenuated inversion recovery (FLAIR)FSE, T2-CSE, T1-CSE, T1-CSE ! gadolinium-DTPA (GdDTPA), PD-CSE, and short-tau inversion recovery (STIR)-FSE.
The spinal cords were removed and the lesions with specific
pathological compositions were identified by histological analysis. Regions of interest (ROIs) were drawn on the corresponding MR images, and signal-to-noise ratios (SNRs) were measured for each MR sequence and compared with controls.
Results: The receiver operating characteristic (ROC) analysis of STIR-FSE and PD-CSE was able to differentiate
tissue that contained cellular infiltrates with a high degree
of accuracy. The SNRs of T2-FSE, STIR-FSE, T2-CSE, PDCSE, and T1-CSE ! Gd-DTPA were elevated in lesions that
contained cellular infiltrates alone, whereas the SNRs of
PD-CSE and T1-CSE ! Gd-DTPA were reduced in demyelinated lesions that also contained inflammation.
Conclusion: The SNR difference between the two lesion
groups suggests that the combination of STIR-FSE, PDCSE, and T1-CSE ! Gd-DTPA sequences may be useful for
differentiating inflammatory lesions containing demyelination from lesions with inflammation alone.
Key Words: MRI; STIR; FLAIR; EAE; demyelination; inflammation
J. Magn. Reson. Imaging 2005;22:180 –188.
© 2005 Wiley-Liss, Inc.
THE ABILITY OF MRI to discriminate between white
and gray matter within the central nervous system
1
Department of Physiology and Pharmacology, University of Western
Ontario, London, Canada.
2
Imaging Research Laboratory, Robarts Research Institute, London,
Canada.
3
Department of Diagnostic Radiology, University of Western Ontario,
London, Canada.
4
Department of Pathology, University of Western Ontario, London, Canada.
*Address reprint requests to: S.J.K., Department of Pathology, Dental
Science Building 4035, University of Western Ontario, 1151 Richmond
Street, London, Ontario N6A 5C1, Canada. E-mail: [email protected]
Contract grant sponsor: Multiple Sclerosis Society of Canada.
Received 23 September 2004; Accepted 3 May 2005.
DOI 10.1002/jmri.20368
Published online in Wiley InterScience (www.interscience.wiley.com).
© 2005 Wiley-Liss, Inc.
(CNS) has made it an ideal method for the diagnosis and
monitoring of multiple sclerosis (MS) (1). The majority of
clinical trials in which new treatments for MS were
tested have utilized this in vivo tool to measure treatment efficacy (2). It is well accepted that T2-weighted
images are ideal for determining lesion load volume,
whereas contrast-enhanced T1-weighted images are
used to identify active lesions (3).
Several MRI studies examining MS have focused on
developing techniques that improve contrast or generate new types of contrast (4 –13). For instance, inversion recovery sequences, such as short-tau inversion
recovery (STIR) and fluid attenuated inversion recovery
(FLAIR), have been reported to provide improved lesion–
brain contrast by suppressing the signal from fat or
cerebrospinal fluid (CSF), respectively (4 – 8). Similarly,
fast spin-echo (FSE) sequences have been shown to be
comparable (if not superior) to conventional spin echo
(CSE) for detecting MS lesions (6,9 –10). The time saved
during the FSE sequence can yield higher-resolution
scans or a greater signal-to-noise ratio (SNR) in the
resulting images. Although these new techniques have
been successful in detecting more lesions, the correlations with clinical disability are modest or poor
(5,10,11). The limited specificity of the MR techniques
to the underlying pathology may contribute in part to
this incongruity.
The present in vivo study was performed to compare
multiple MR sequences at 1.5T with histologically-confirmed lesions within an animal model of MS, experimental allergic encephalomyelitis (EAE), and to determine whether the different MR sequences are sensitive
to specific pathologies. In this study we used the histopathology within the spinal cord to guide our measurements from the various MR image contrasts acquired in
vivo.
MATERIALS AND METHODS
Animal Model and EAE Inductions
EAE in the Hartley guinea pig produces lesions with
some characteristics similar to those seen in MS.
Perivascular cuffing of mononuclear cells and macrophages commences 10 days post-immunization (dpi)
and is continuous throughout the disease. The number
of inflammatory cells increases around the blood vessels and eventually infiltrates the neural parenchyma
180
Pathology-Guided MRI
181
Table 1
MR Sequences Parameters at 1.5T
a,b
T2-FSE
PD-FSEa
FLAIR-FSEa
STIR-FSEb
T2-CSEa,b
PD-CSEb
T1-CSEa
Gd-DTPAa,b
TR
TE
TI
NEX
ETL
Time
4700
2400
8000
2288
2000
2000
800
800
112
15
119
60
80
20
30
30
—
—
2000
110
—
—
—
—
3
3
2
4
1
1
2
2
15
7
7
11
–
—
—
—
4:18
4:38
9:04
11:11
8:56
8:56
6:52
6:52
a
Protocol 1.
Protocol.
TR " repetition time, TE " echo time, TI " inversion time (all in msec), NEX " number of excitations, ETL " echo train length, Time " length
of acquisition (min:sec).
b
(myelitis), resulting in demyelination. During the acute
phase of the disease (10 –25 dpi) the lesions are characterized primarily by cellular infiltrates. The chronic
phase occurs subsequently to 25 dpi, and during this
time the inflamed lesions can also be demyelinated.
However, in contrast to MS, in which demyelinated lesions can be observed with little or no cellular infiltrates, the EAE lesions retain an inflammatory cell component.
EAE was induced by nuchal intradermal injection of
0.2 mL of a 1:1 mixture of homogenized isologous CNS
tissue (in saline) and complete Freund’s adjuvant (CFA)
(Difco, Detroit, MI) with the addition of 10 mg of inactivated Mycobacterium tuberculosis (Difco, Detroit, MI)
per mL of CFA in adult female Harley guinea pigs
(Charles River Canada, St. Constant, PQ) weighing approximately 200 g. Food and water were provided ad
libitum in a controlled light environment. The animals
were assessed daily for signs of EAE (14). Age-matched,
nonimmunized female Hartley guinea pigs served as
controls.
MRI
All imaging was performed at 1.5T on a GE Signa Horizon using a custom-made surface RF coil. For the imaging sessions, 21 animals (11 EAE and 10 non-EAE)
were sedated with 1 mL/kg ketamine:xylazine (10:1
ratio; 115 mg/mL and 20 mg/mL, respectively). To obtain histological variability within the spinal cord, five
EAE animals were imaged during the acute phase of the
disease (15 dpi) and the remaining six animals were
imaged during the chronic phase (50 dpi). To aid in the
localization of the imaged area, a plastic catheter tube
was inserted into the muscle adjacent to the spinal cord
and an oil-based vitamin (1 cm) was adhered to the
dorsal surface of the animal. A total of 16 contiguous
axial imaging slices (3 mm thick) were acquired for a
4.8-cm length of the lumbar spinal cord, beginning at
the most caudal aspect. A sagittal localizer SE T1weighted sequence (TR/TE " 800/30 msec) was used to
select the image locations. The protocol was approved
by the animal use subcommittee of our institution and
conformed with the guidelines of the Canadian Council
for Animal Care.
MR Sequences
The guinea pigs were imaged with one of two imaging
protocols. The first protocol was used to image five EAE
animals (two during the acute phase and three during
the chronic phase) and consisted of T2-FSE, PD-FSE,
FLAIR-FSE, T2-CSE, T1-CSE, and T1-CSE performed
after a 0.2-mL intercardiac injection of gadolinium (Gd)
DTPA (T1-CSE ! Gd-DTPA). The second protocol was
used to imaged six EAE animals (three acute and three
chronic) and consisted of STIR-FSE, T2-FSE, PD-CSE,
T2-CSE, T1-CSE ! Gd-DTPA (Table 1). For each sequence, the following imaging parameters were used:
FOV " 7 cm, image matrix " 512 # 512, 16 slices, slice
thickness " 3 mm, and voxel size " 0.28 mm # 0.14
mm # 3 mm (0.12 mm3). Representative images from
the different MR sequences are shown in Fig. 1.
Whole Spinal Cord MR Analysis
The MR slices from each MR contrast were coregistered
using ImageJ (National Institutes of Health, USA) and
for each type of contrast the signal intensity was measured from a region of interest (ROI) that outlined the
entire spinal cord in each of the 16 slices. The ROIs
ranged from approximately 9.4 to 9.9 mm2. To account
for MR signal variability between the different imaging
sessions, the SNR was calculated (mean signal intensity of spinal cord divided by the standard deviation
(SD) of the air). The mean SNR of each MR sequence was
calculated for each spinal cord.
Tissue Preparation
Immediately following the imaging session, the anesthetized animal was euthanized with an overdose of
sodium pentobarbital (euthanyl forte). The imaged portion of the spinal column, as identified by the two external markers, was removed and immediately place in
10% buffered formalin. To prevent tissue distortion and
shrinkage artifacts, the spinal column was fixed in 10%
buffered formalin prior to dissection of the cord from
the surrounding vertebrae. The 4.8-cm spinal cord portion was embedded in paraffin and sectioned into 5-mm
blocks. From each paraffin block, two 5-$m axial sections were sectioned every millimeter. Two slides were
182
Cook et al.
Figure 1. Representative control images of MR sequences: (a) FLAIR-FSE,
(b) PD-FSE, (c) T2-FSE, (d) T2-CSE, (e)
T1-CSE, (f) T1-CSE ! Gd-DTPA (0.1
mmol/kg), (g) STIR-FSE, and (h) PDCSE.
stained with either hematoxylin-eosin (H&E) to assess
tissue morphology and inflammation, or solochrome-Rcyanin (ScR), to evaluate the amount of demyelination.
A total of 96 histological slides were prepared for each
animal (two at each anatomical location).
Lesion Classification and MR Measurement
The H&E- and ScR-stained spinal cord sections were
examined for the presence of inflammatory cells and
demyelination using light microscopy. When specific
pathological features were consistent on three consecutive histological sections (representing one 3-mm MR
image slice location), the spinal cord tissue was categorized into one of three lesion groups: 1) tissue that
contained no inflammatory cells or demyelination (microscopically the tissue showed a disorganized appearance, with vacuolization consistent with widespread
edema; this tissue was referred to as normal-appearing
white matter (NAWM); 2) tissue that contained perivascular and parenchymal inflammatory cells (cellular in-
filtrates); and 3) tissue that contained both cellular
infiltrates and complete demyelination (Fig. 2). In this
model we rarely observed partial demyelination except
at the edge of the lesion, which was not included in the
ROIs. Once a lesion was located, an ROI was carefully
drawn on the same area corresponding to coregistered
MR slices to include only neuroparenchyma and not the
surrounding CSF. The signal intensities for each MR
contrast were measured for the lesion. Individual lesion
ROIs ranged from 2.1 to 2.5 mm2. The SNR for each
image contrast was calculated for each identified lesion.
Statistical Analysis
The MR images were considered as independent samples since there were no significant correlations in
terms of signal intensity among the 16 imaging locations for individual spinal cords. Separate analyses of
variance (ANOVAs) were performed to determine
whether there were any significant differences between
the SNRs measured in EAE and control spinal cord/
Pathology-Guided MRI
183
Figure 2. Pathological classification
of spinal cord lesions. Tissue sections
stained with H&E (a, c, e, and g) and
ScR (b, d, f, and h). a and b: Non-EAE
control. c and d: Edematous white
matter. e and f: Tissue that contains
cellular infiltrates. g and h: Tissue
that contains extensive cellular infiltration and is completely demyelinated. Arrows are perivascular cuffs of
inflammatory cells. Arrowheads are
parenchymal infiltration of inflammatory cells (myelitis). gm " gray matter,
wm " white matter, dm "
demyelination.
lesions for each MR sequence. A receiver operating
characteristic (ROC) curve analysis was performed to
evaluate the accuracy (areas under the curve, Az) of
each MR contrast for each lesion categorization. All
analyses were performed using the Statistical Package
for Social Sciences (Student v10.0; SPSS, Chicago, IL,
USA).
RESULTS
When the whole cord ROIs were examined, the mean
SNRs of T2-FSE (P % 0.005), STIR-FSE (P % 0.05),
T2-CSE (P % 0.005), and T1-CSE ! GD-DTPA (P % 0.01)
images were significantly higher in the 16 axial spinal
cord sections for the EAE animals compared to the
controls. The mean SNRs of the FLAIR-FSE (P % 0.001)
and T1-CSE (P % 0.01) images were significantly lower
than control values, and the SNR of the whole spinal
cord in the PD-weighted images was not significantly
different from that of the controls (Table 2).
Figure 3 illustrates examples of MR images with the
corresponding pathological changes from an axial slice
location in an acute (Fig. 3a) and chronic (Fig. 3b)
guinea pig spinal cord. There were 16 imaging locations
for each animal (3-mm slice thickness), and three
Table 2
Mean (& SE) SNR of 16 Axial Spinal Cord MR Image Slices From
Each EAE and Control Animals
T2-FSE
PD-FSE
FLAIR-FSE
STIR-FSE
T2-CSE
PD-CSE
T1-CSE
T1-CSE ! Gd-DTPA
*P % 0.05, one-way ANOVA.
EAE
Control
28.31 (2.18)*
56.40 (4.66)
15.81 (1.03)*
36.81 (3.80)*
21.41 (1.28)*
46.39 (4.40)
34.76 (2.43)*
39.30 (2.46)*
19.47 (1.31)
50.23 (3.45)
27.70 (1.16)
24.13 (2.31)
13.90 (1.52)
36.47 (4.42)
46.93 (1.96)
29.23 (1.87)
184
Cook et al.
Figure 3. In vivo MRI and pathological findings. Representative MR images and distribution of pathological changes from (a)
acute and (b) chronic EAE animals. The MR images are labeled for each sequence: STIR-FSE, PD-CSE, T2-FSE, T2-CSE, and T1
post-Gd. The schematic diagrams depict the distribution of inflammation and demyelination on the three anatomical slides that
correlate with the imaging slice location.
pathological sections were prepared (at 1-mm intervals)
from the dissected cord for pathological comparison
with each imaging location. Intense foci can be observed in the STIR-FSE, T2-FSE, and T2-CSE images
from the acute animal (Fig. 3a). This corresponded to
extensive perivascular infiltration with limited demyelination. In the chronic animal that displayed extensive
parenchymal infiltrates and widespread demyelination
(Fig. 3b), the images were unremarkable, except for the
T1 Gd-DTPA image, which showed enhancement. These
MR changes were characteristic of the underlying
pathological changes as revealed by the individual lesion analysis.
From 176 imaging slice locations, a total of 153 individual lesion areas were identified in the EAE spinal
cords consisting of three varieties. Fifty-five were cate-
Pathology-Guided MRI
185
Figure 4. SNRs from the MR sequences for EAE spinal cords, lesion groups, and controls. T2-FSE, PD-FSE, FLAIR-FSE,
STIR-FSE, PD-CSE, T2-CSE, T1-CSE, and T1-CSE ! Gd-DTPA SNRs for con " non EAE controls (N " 192); NAWM " no cellular
infiltrates or demyelination (N " 55); cell " tissue that contains cellular infiltrates (perivascular cuffing, myelitis, and perivascular cuffing ! myelitis) (N " 47); celldm " tissue that contains cellular infiltrates and demyelination (perivascular cuffing and
demyelination, and perivascular cuffing, myelitis, and demyelination) (N " 51). The dashed horizontal line represents the mean
control value for each MR sequence. (* significantly different from control tissue at P % 0.05, ANOVA).
gorized as edematous (contained no inflammatory cells
or demyelination but had a disorganized vacuolar appearance consistent with edema, which we called
NAWM), 47 were tissues containing cellular infiltration
alone, and 51 were tissues containing both cellular
infiltration and demyelination. For the SNR analysis,
the composition of a lesion had to be consistent
throughout the three histological locations that defined
the imaging slice location.
An ANOVA followed by post-hoc analysis (Tukey, P %
0.05) shows that 1) lesions categorized as edematous
had higher SNRs compared to control tissue in STIRFSE images, but lower SNRs compared to control tissue
in FLAIR-FSE images; 2) lesions categorized as tissue
containing cellular infiltrates had higher SNRs compared to control tissue for T2-FSE, STIR-FSE, PD-CSE,
T2-CSE, and T1-CSE ! Gd-DTPA images; and 3) lesions categorized as tissue containing cellular infiltrates with demyelination had significantly decreased
SNRs compared to control tissue in PD-CSE and T1CSE ! Gd-DTPA images (Fig. 4 and Table 3).
ROC curves were used to determine whether lesion
composition could be accurately differentiated from
control tissue. This analysis showed that the sequences
tested were inadequate (Az % 70%) for differentiating
the different lesion types, with the exception of the
SNRs from STIR-FSE and PD-CSE, which were able to
differentiate tissue that contained cellular infiltrates
186
Cook et al.
Table 3
Relative SNR of MR Sequences Compared to Control Tissue for
Lesion Groups
T2-FSE
PD-FSE
FLAIR-FSE
STIR-FSE
T2-CSE
PD-CSE
T1-CSE
Gd-DTPA
Edema
(NAWM)
Cellular
infiltrates
Cellular
infiltrates !
demyelination
2
1
-
1
1
1
1
1
2
2
(1) " hyperintense, (-) " isointense, (2) " hypointense compared
to control.
from control tissues with a high degree of accuracy
(85% & 5% and 81% & 5%, respectively; Table 4).
DISCUSSION
Previous studies examined EAE lesions (15–17) and
postmortem and biopsy MS tissue (18 –26), and attempted to determine the effect that lesion composition
has on the MR signal. Typically, in postmortem MR
studies of MS, a lesion is identified on the MR image
and then the tissue is dissected to determine the pathological cause of that lesion (18 –21,24,25). We reversed
this common scenario and utilized histological analysis
to specify where the MR signal was sampled. To our
knowledge, this is the first study conducted in vivo at
1.5T to directly examine how different MR sequences,
such as inversion recovery, FSE, and CSE are altered
by histologically-verified pathological changes.
MR analyses of MS lesions have shown decreased
signal intensities on FLAIR (5) and T1 (3), and increased
signal intensities on STIR (8), T2 (3), PD (3), and contrast-enhanced T1 (3) sequences. When we sampled the
signal from every axial image slice of the entire spinal
cord, the results were consistent with previous data.
Specifically, the T2, T1 (& contrast), and two inversion
recovery sequences were all sensitive to the presence of
disease within the spinal cord. These ROIs encompassed the entire cord, capturing normal tissue and
central canal CSF as well as tissue altered by the disease.
To assess the specificity of the sequences for identifying the underlying pathological changes, we identified
three predominant pathological changes. Lesions categorized as edematous in the EAE spinal cord contained
no inflammatory cells or demyelination. Traditionally,
in MR studies NAWM is defined as an area that appears
normal on conventional MR sequences (27–30). Numerous MR studies have examined the relevance of this
tissue in the progression of MS, but only a few have
examined a possible histopathological composition
(30 –32). NAWM within MS is characterized by reduced
myelin density, vascular hyalinization, and evidence of
blood– brain barrier (BBB) breakdown (31,32). In the
EAE spinal cord, in addition to demyelinated lesions,
NAWM has an increased number of blood vessels compared to controls (33) and can appear disorganized and
vacuolated. The results of the current analysis of these
edema lesions are consistent with previous reports (i.e.,
“normal” signal on the T2-, T1-, and PD-weighted MR
sequences). However, the STIR-FSE images showed an
increased SNR, and the FLAIR-FSE images had a decreased SNR compared to controls. These two sequences have been shown to detect more lesions than
other MR sequences (7,13) and MTR technique may
assist in differentiating this tissue from control (34,35).
The acute MS lesion consists primarily of cellular
infiltrates, such as T-cells and macrophages, although
a small amount of demyelination may also be present
(36). The lesions that contain only cellular infiltrates
within the EAE spinal cord are thus similar to acute MS
lesions. Serial MR scans of MS brains show that when a
new lesion emerges, it is first identified by enhancement on T1-weighted images. Subsequently, this lesion
will appear to enlarge on T2- and PD-weighted sequences (37). Studies that focused on the optic nerve of
MS patients showed that an increased signal on STIR
images represents a lesion associated with edema and
inflammation (38). We found an increased SNR for the
same MR sequences (T2-FSE, STIR-FSE, T2-CSE, PDCSE, and T1-CSE ! Gd-DTPA) in lesions that contained
cellular infiltrates. The SNR from STIR-FSE and PDCSE also had a high degree of accuracy for differentiating this type of lesion from controls (85% and 81%,
respectively). The sensitivity to the presence of increased cellularity and inflammation suggests that the
combination of these MR sequences could be used to
identify MS lesions that contain primarily inflammation.
It is generally accepted that demyelination ensues as
a consequence to infiltrating immune cells in MS tissue
(36). Demyelinated MS lesions were recently catego-
Table 4
Areas Under the Curve (Az) & SE (95% Confidence Intervals)
T2-FSE
PD-FSE
FLAIR-FSE
STIR-FSE
T2-CSE
PD-CSE
T1-CSE
Gd-DTPA
Edema (NAWM)
Cellular infiltrates
Cellular infiltrates ! demyelination
52% & 5 (43–60%)
60% & 7 (45–74%)
68% & 6 (57–80%)
55% & 5 (46–56%)
54% & 5 (44–63%)
50% & 6 (39–62%)
58% & 7 (52–71%)
52% & 5 (42–61%)
67% & 6 (55–78%)
60% & 8 (45–75%)
53% & 7 (40–66%)
85% & 5 (75–94%)
70% & 5 (59–79%)
81% & 5 (72–91%)
65% & 7 (52–78%)
65% & 6 (54–76%)
64% & 4 (56–72%)
60% & 5 (41–80%)
61% & 10 (42–80%)
65% & 4 (59–74%)
56% & 5 (47–64%)
65% & 4 (56–74%)
59% & 9 (42–76%)
56% & 5 (46–65%)
Pathology-Guided MRI
rized into four patterns based on their histological features (39), such that each pattern was uniquely characterized by the presence or absence of blood vessels,
myelin proteins, and cell types. The demyelinated lesions within the EAE spinal cord develop from perivascular cuffs of infiltrating immune cells, and therefore
their histological characteristics are most closely related to MS patterns I and II. The SNRs from the T1CSE ! Gd-DTPA and PD-CSE images were significantly
reduced compared to controls when the MR signals
were obtained from this lesion group. The decreased
SNR in T1-CSE ! Gd-DTPA images suggests the presence of a non-enhancing lesion. These lesions, which
are common in MS, evolve from ring-enhancing lesions
and are believed to reflect a chronic area with axonal
damage and demyelination (40). The reduced SNR from
PD-CSE images would also suggest an area that contains severe tissue destruction. The PD-weighted signal
intensity acquired at 4.7T was found to be inversely
related to the amount of demyelination within the MS
spinal cord. However, the tissue examined was formalin-fixed prior to imaging (21), and the fixation process
removes water from tissue and reduces the sensitivity
of detecting lesions (19). Our in vivo results indicate
that the SNR from PD-CSE was decreased in areas
where demyelination and extensive tissue damage existed. This observation can be compared with that of
Bot et al (41), who examined formalin-fixed cords from
end-stage MS patients at 4.7T. Although the expected
prolongation of the T1 and T2 in lesions was observed,
demyelinated lesions without an inflammatory component were identified as hyperintense on intermediate
density-weighted images (SE: TR " 3000, TE " 15). We
observed increased intensity on PD-CSE when the lesions were inflamed, and a decrease when demyelination accompanied the cellular changes. This disparity
may be attributable to differences in lesion composition
or the tissue fixation method, field strength, and sequence used.
The use of both FSE and CSE to obtain T2- and
PD-weighted images allowed a direct comparison of the
two pulse sequences in this study. T2-FSE and T2-CSE
showed similar differences within the lesion groups.
Both had an increase SNR in tissue that contained
cellular infiltration, but did not have different SNRs
when compared to controls with the other lesion
groups. T2-FSE and T2-CSE had comparable accuracies as regards the ROC curves for the different lesions
groups as well. In terms of sensitivity to the underlying
pathology, one pulse sequence was no better than the
other. However, the shorter time needed to acquire the
FSE sequences may be more comfortable for the patient. The same cannot be said for the PD-CSE and
PD-FSE sequences. Not only did the SNR from PD-FSE
fail to differentiate the mean EAE lesions from controls,
there was no differentiation of any of the lesion groups
from control tissue or from one another. Furthermore,
the ROC accuracies of PD-FSE for differentiating EAE
tissue and lesion groups from controls were poor (all
below 65%). The SNR from PD-CSE, however, was elevated in tissue that contained cellular infiltrates, and
decreased once the lesions were demyelinated. In addition, the SNR from PD-CSE differentiated lesions that
187
contained cellular infiltrates from control tissue with
high accuracy. The SNR of the PD-FSE had a greater
variability within the lesion groups, which prevented in
vivo characterization of these lesions.
There are three limitations that should be considered
in relation to these data: First, we limited this analysis
to three well-defined tissue pathological changes:
NAWM (as defined here for guinea pig EAE), perivascular infiltration with myelitis, and myelitis and demyelination. This model is not equivalent to the sclerotic
demyelinated plaque seen in MS patients. Second, despite our best efforts to exclude surrounding tissue,
even the small ROIs used in this study had the potential
to consist of a mixture of the specific pathological
changes and either NAWM or unaffected white matter.
Third, although the TI-Gd DTPA data revealed a characteristic pattern of SNR change for the 47 lesions identified, there were instances in which an acute animal
had little enhancement (with inflammation) and a
chronic animal showed significant enhancement (with
demyelination; see Fig. 3 for examples). The possibility
exists that BBB permeability detected by Gd-DTPA permeation is related more to the expression of permeability factor(s) than to the pathological constituents of a
lesion.
In this study we have attempted to provide a histological validation for eight different MR sequences acquired in vivo. The novel approach we used, however,
left us with the challenge of comparing our results with
those obtained from postmortem MS tissue, and MR
studies with no pathological correlate. For obvious reasons, the methodology we utilized could not be replicated in an MS patient. However, it is important to
reiterate the high specificity of STIR-FSE and PD-CSE
to the presence of inflammation, and the reversal of
signal intensities of T1-CSE ! Gd-DTPA and PD-CSE
when demyelination was present in an inflamed lesion.
Even though the lesions we examined represent a specific pathological component that is common in MS
tissue, the data suggest that these MR sequences (T1CSE ! Gd-DTPA, PD-CSE, and STIR-FSE) may be useful for distinguishing an acute MS lesion from a chronic
demyelinated area.
REFERENCES
1. Kappos L, Stadt D, Ratzka M, et al. Magnetic resonance imaging in
the evaluation of treatment in multiple sclerosis. Neuroradiology
1988;30:299 –302.
2. Barkhof F, Filippi M, Miller DH, Tofts P, Kappos L, Thompson AJ.
Strategies for optimizing MRI techniques aimed at monitoring disease activity in multiple sclerosis treatment trials. J Neurol 1997;
244:76 – 84.
3. Miller DH, Grossman RI, Reingold SC, McFarland HF. The role of
magnetic resonance techniques in understanding and managing
multiple sclerosis. Brain 1998;121:3–24.
4. Campi A, Pontesilli S, Gerevini S, Scotti G. Comparison of MRI
pulse sequences for investigation of lesions of the cervical spinal
cord. Neuroradiology 2000;42:669 – 675.
5. Rovaris M, Comi G, Rocca MA, et al. Relevance of hypointense
lesions on fast fluid-attenuated inversion recovery MR images as a
marker of disease severity in cases of multiple sclerosis. AJNR Am J
Neuroradiol 1999;20:813– 820.
6. Yousry TA, Filippi M, Becker C, Horsfield MA, Voltz R. Comparison
of MR pulse sequences in the detection of multiple sclerosis lesions.
AJNR Am J Neuroradiol 1997;18:959 –963.
188
7. Rocca MA, Mastronardo G, Horsfield MA, et al. Comparison of three
MR sequences for the detection of cervical cord lesions in patients
with multiple sclerosis. AJNR Am J Neuroradiol 1999;20:1710 –
1716.
8. Thorpe JW, MacManus DG, Kendall BE, et al. Short tau inversion
recovery fast spin-echo (fast STIR) imaging of the spinal cord in
multiple sclerosis. Magn Reson Imaging 1994;12:983–989.
9. Patola WB, Coulter BA, Chipperfield PM, Lingawi SS. A comparison
of conventional spin-echo and fast spin-echo in the detection of
multiple sclerosis. J Magn Reson Imaging 2001;13:657– 667.
10. Kidd D, Thorpe JW, Thompson AJ, et al. Spinal cord MRI using
multi-array coils and fast spin echo. II. Findings in multiple sclerosis. Neurology 1993;43:2632–2637.
11. Gawne-Cain ML, O’Riordan JI, Coles A, Newell B, Thompson AJ,
Miller DH. MRI lesion volume measurement in multiple sclerosis
and its correlation with disability: a comparison of fast fluid attenuated inversion recovery (fFLAIR) and spin echo sequences. J Neurol Neurosurg Psychiatry 1998;64:197–203.
12. Bot JC, Barkhof F, Lycklama a Nijeholt GJ, et al. Comparison of a
conventional cardiac-triggered dual spin-echo and a fast STIR sequence in detection of spinal cord lesions in multiple sclerosis. Eur
Radiol 2000;10:753–758.
13. Stevenson VL, Gawne-Cain ML, Barker GJ, Thompson AJ, Miller
DH. Imaging of the spinal cord and brain in multiple sclerosis: a
comparative study between fast FLAIR and fast spin echo. J Neurol
1997;244:119 –124.
14. Karlik SJ, Grant EA, Lee D, Noseworthy JH. Gadolinium enhancement in acute and chronic-progressive experimental allergic encephalomyelitis in the guinea pig. Magn Reson Med 1993;30:326 –
331.
15. Jordan EK, McFarland HI, Lewis BK, et al. Serial MR imaging of
experimental autoimmune encephalomyelitis induced by human
white matter or by chimeric myelin-basic and proteolipid protein in
the common marmoset. AJNR Am J Neuroradiol 1999;20:965–976.
16. Hart BA, Bauer J, Muller HJ, et al. Histopathological characterization of magnetic resonance imaging— detectable brain white matter
lesions in a primate model of multiple sclerosis: a correlative study
in the experimental autoimmune encephalomyelitis model in common marmosets (Callithrix jacchus). Am J Pathol 1998;153:649 –
663.
17. Duckers HJ, Muller HJ, Verhaagen J, Nicolay K, Gispen WH. Longitudinal in vivo magnetic resonance imaging studies in experimental allergic encephalomyelitis: effect of a neurotrophic treatment on cortical lesion development. Neuroscience 1997;77:1163–
1173.
18. Mottershead JP, Schmierer K, Clemence M, et al. High field MRI
correlates of myelin content and axonal density in multiple sclerosis—a post-mortem study of the spinal cord. J Neurol 2003;250:
1293–1301.
19. Schmierer K, Scaravilli F, Barker GJ, Gordon R, MacManus DG,
Miller DH. Stereotactic co-registration of magnetic resonance imaging and histopathology in post-mortem multiple sclerosis brain.
Neuropathol Appl Neurobiol 2003;29:596 – 601.
20. De Groot CJ, Bergers E, Kamphorst W, et al. Post-mortem MRIguided sampling of multiple sclerosis brain lesions: increased yield
of active demyelinating and (p)reactive lesions. Brain 2001;124:
1635–1645.
21. Nijeholt GJ, Bergers E, Kamphorst W, et al. Post-mortem highresolution MRI of the spinal cord in multiple sclerosis: a correlative
study with conventional MRI, histopathology and clinical phenotype. Brain 2001;124:154 –166.
22. Barkhof F, Scheltens P, Kamphorst W. Pre-and post-mortem MR
imaging of unsuspected multiple sclerosis in a patient with Alzheimer’s disease. J Neurol Sci 1993;117:175–178.
Cook et al.
23. Macchi G, Cioffi RP. An in vivo and post mortem MRI study in
multiple sclerosis with pathological correlation. Ital J Neurol Sci
1992;13:97–103.
24. van Waesberghe JH, Kamphorst W, De Groot CJ, et al. Axonal loss
in multiple sclerosis lesions: magnetic resonance imaging insights
into substrates of disability. Ann Neurol 1999;46:747–754.
25. van Walderveen MA, Kamphorst W, Scheltens P, et al. Histopathologic correlate of hypointense lesions on T1-weighted spin-echo
MRI in multiple sclerosis. Neurology 1998;50:1282–1288.
26. Bruck W, Bitsch A, Kolenda H, Bruck Y, Stiefel M, Lassmann H.
Inflammatory central nervous system demyelination: correlation of
magnetic resonance imaging findings with lesion pathology. Ann
Neurol 1997;42:783–793.
27. Gallo A, Rovaris M, Riva R, et al. Diffusion-tensor magnetic resonance imaging detects normal-appearing white matter damage unrelated to short-term disease activity in patients at the earliest
clinical stage of multiple sclerosis. Arch Neurol 2005;62:803– 808.
28. Caramia F, Pantano P, Di Legge S, et al. A longitudinal study of MR
diffusion changes in normal appearing white matter of patients
with early multiple sclerosis. Magn Reson Imaging 2002;20:383–
388.
29. Whittall KP, MacKay AL, Li DK, Vavasour IM, Jones CK, Paty DW.
Normal-appearing white matter in multiple sclerosis has heterogeneous, diffusely prolonged T(2). Magn Reson Med 2002;47:403–
408.
30. Allen IV, McQuaid S, Mirakhur M, Nevin G. Pathological abnormalities in the normal-appearing white matter in multiple sclerosis.
Neurol Sci 2001;22:141–144.
31. Allen IV, McKeown S. A histological, histochemical and biochemical
study of the macroscopically normal white matter in multiple sclersois. J Neurol Sci 1979;41:81–91.
32. Gay D, Esiri M. Blood– brain barrier damage in acute multiple
sclerosis plaques. Brain 1991;114:557–572.
33. Kirk SL, Karlik SJ. VEGF and vascular changes in chronic inflammation. J Autoimmun 2003;21:353–363.
34. Ropele S, Strasser-Fuchs S, Augustin M, Stollberger R, Enzinger C,
Hartung HP, Fazekas F. A comparison of magnetization transfer
ratio, magnetization transfer rate, and the native relaxation time of
water protons related to relapsing-remitting multiple sclerosis.
AJNR Am J Neuroradiol 2000;21:1885–1891.
35. Ge Y, Grossman RI, Babb JS, He J, Mannon LJ. Dirty-appearing
white matter in multiple sclersois: volumetric MR imaging and
magnetization transfer ration histogram analysis. AJNR Am J Neuroradiol 2003;24:1935–1940.
36. Lassmann H. Neuropathology in multiple sclerosis: new concepts.
Mult Scler 1998;4:93–98.
37. Kermode AG, Thompson AJ, Tofts PS, et al. Breakdown of blood–
brain barrier precedes symptoms and other MRI signs of new lesions in multiple sclerosis. Pathogenetic and clinical implications.
Brain 1990;113:1477–1489.
38. Gass A, Moseley IF. The contribution of magnetic resonance imaging in the differential diagnosis of optic nerve damage. J Neurol Sci
2000;172:s17–s22.
39. Lucchinetti C, Bruck W, Parisi J, Scheithauer B, Rodriguez M,
Lassmann H. Heterogeneity of multiple sclerosis lesions: implications for the pathogenesis of demyelination. Ann Neurol 2000;47:
707–717.
40. Rovira A, Alonso J, Cucurella G, et al. Evolution of multiple sclerosis lesions on serial contrast-enhanced T1-weighted and magnetization transfer MR images. AJNR Am J Neuroradiol 1999;20:
1939 –1945.
41. Bot JC, Blezer EL, Kamphorst W, et al. The spinal cord in multiple
sclerosis: relationship of high-spatial-resolution quantitative MR
imaging findings to histopathological results. Radiology 2004;233:
531–535.