Brain (1996), 119, 715-722
Assessment of lesion pathology in multiple
sclerosis using quantitative MRI morphometry and
magnetic resonance spectroscopy
P. M. Matthews,1-2E. Pioro,'*S. Narayanan,1 N. De Stefano,1 L. Fu,1 G. Francis,1 J. Antel,1
C. Wolfson3 and D. L. Arnold1
1
Department of Neurology and Neurosurgery, Montreal
Neurological Institute, the department of Human Genetics
and the ^Department of Epidemiology, McGill University,
Montreal, Quebec, Canada
Correspondence to: Dr P. M. Matthews, Department of
Clinical Neurology, Radcliffe Infirmary, Woodstock Road,
Oxford 0X2 6HE, UK
* Present address: Department of Neurology, Cleveland
Clinic, Cleveland, Ohio, USA
Summary
Quantitative measurement of MRI-defined brain lesions can
provide an index of the extent and activity of disease in
multiple sclerosis patients. However, the relationships
between these indices and clinical features are not wellunderstood. Heterogeneity of the pathological changes
underlying MRI lesions may be an important factor
determining the correlation between MRI lesion volumes
and clinical measures. Recent studies have suggested that
with magnetic resonance spectroscopy (MRS), it may be
possible to define chemical changes that better reflect the
pathological changes in multiple sclerosis. Here we report
results of combined quantitative brain T2-weighted MRI
lesion volume and proton MRS examinations that demonstrate heterogeneity of the chemical pathology underlying
brain lesions in patients selected on the basis of similar
clinical disability but differing with respect to the presence
or absence of clinical relapses. We examined 29 patients
with disease characterized by either clear relapses with at
least partial remissions (RR) or secondary, chronic
progression after an earlier history of a more relapsing
and remitting course (SP). Total hemispheric lesion volume
was greater (P < 0.04) in the RR (32.5±20.9 cm3) than in
the SP (16.2±9.0 cm3) patients, despite the longer duration
of disease in the latter group. Central brain N-acetyl
aspartate .creatine (NAA:Cr) ratios were reduced relative to
normal controls (4.0±0.3, n = 19) by similar amounts in the
two patient groups (RR, 3.1+0.5; SP, 3.2±0.4; P < 0.0001).
The ratio lesion volume:(NAA:Cr) was greater for the RR
group (11.7±9.3 cm3) than for the SP group (5.4±3.3 cm3,
P < 0.05), implying a greater average degree of axonal loss
per unit lesion volume defined by MRI for subjects in the SP
group or, alternatively, a greater proportion of lesions without
axonal damage or loss in the RR group. Our results emphasize
a limitation of using T2-weighted MRI lesion volume alone
and suggest that combined analysis of MR-based chemical
and imaging data might allow improved non-invasive
assessment of lesion pathology in order to better understand
its relationship to clinical features of multiple sclerosis.
Keywords: multiple sclerosis; magnetic resonance spectroscopy; MRI; pathology; /V-acetyl aspartate
Abbreviations: Cr = creatine; EDSS = extended disability status score: MRS = magnetic resonance spectroscopy;
NAA = N-acetyl aspartate; RR = relapsing-remitting multiple sclerosis; SP = secondary progressive multiple sclerosis
Introduction
Multiple sclerosis is a disease with inflammatory and
demyelinating pathology. A major problem in clinical studies
is assessment of the extent of the pathological process in
individual patients. Clinical measures of the extent of disease
are attractive as they directly assess the functional impact on
© Oxford University Press 1996
the patient, but they suffer from major limitations because
of bias for motor impairments, rater subjectivity, insensitivity
to changes with time, and non-linearity (Hohol et al., 1995).
Thus, there is considerable interest in developing non-invasive
measures of pathological changes in multiple sclerosis. A
716
P. M. Matthews et al.
particular focus of such work is to define better the
relationship between brain lesion pathology and clinical
course.
MRI is currently the most widely used method for
assessment of lesions in multiple sclerosis. Quantitative
measurement of total lesion volume has been utilized as an
index of total burden of disease. Results from two longitudinal
studies of multiple sclerosis patients have documented
progressive increases in total T2-weighted MRI lesion
volumes with time (Paty et al., 1993, 1994). However, the
relationship between lesion volume and clinical disability
appears to be variable, in general (Miller, 1994). It is known
that the pathology underlying the hyperintense lesion signals
measured on T2-weighted images from brains of multiple
sclerosis patients is heterogeneous (arising from oedema,
demyelination or gliosis) within the same individual and that
the relative distribution of these different changes in lesions
varies considerably from one patient to another, depending
on the duration and course of disease (Newcombe et al.,
1991). The failure to observe similar correlations between
lesion volumes and disability for different patient populations
thus may reflect the inherent lack of specificity of the
hyperintense changes in the MRI.
Pathological studies indicate that axons, as well as myelin,
are damaged in multiple sclerosis lesions, adding further to
the potential heterogeneity of lesion pathology (reviewed
by McDonald et al., 1994). Our group and others have
demonstrated decreases of NAA, a neuronal-specific
acetylated amino acid, in brains of multiple sclerosis patients
(Arnold et al., 1990, 1992, 1994; Matthews et al., 1991;
Miller et al., 1991; Van Hecke et al., 1991). Although the
focal decreases in brain NAA are reversible in some acute
lesions, irreversible decreases in brain NAA are seen in other
lesions (Arnold et al., 1992; De Stefano et al., 1993; Davie
etal., 1994). These stable abnormalities have been interpreted
to be a result of the secondary axonal damage that is associated
with chronic lesions (Barnes et al., 1991; McDonald et al.,
1992). Recently, we have presented preliminary data showing
significant, monotonic decreases in brain NAA over 18
months, consistent with expected gradually increasing
severity of disease (Arnold et al., 1994). These observations
suggested to us that brain NAA may be useful in combination
with T2-weighted MRI lesion volume as an index of disease
pathology and in defining the relationship between pathology
and clinical course in multiple sclerosis.
Our hypothesis is that lesions with identical appearances
may have significant differences in underlying pathology that
could account (at least in part) for differences in clinical
course. The specific goal of the present study was to test
whether there is pathological heterogeneity underlying T2weighted MRI lesions between two groups of multiple
sclerosis patients with similar levels of clinical disability, but
differing in clinical course. One group was selected to include
patients with recurrent relapses and partial remissions (RR)
and the second group of patients had chronically progressive
disease without discrete relapses with an earlier history of
relapsing and partially remitting disease (SP). We performed
quantitative brain MRI morphometry with separate
segmentation of ventricular and T2-weighted lesion volumes
and single voxel proton MRS observations on these patients.
Methods
This study was approved by the Montreal Neurological
Hospital Ethics Committee and informed consent was
obtained from all subjects who participated. Twenty-nine
patients with clinically definite multiple sclerosis were chosen
from the population followed in the Montreal Neurological
Hospital multiple sclerosis clinic. Patients were classified
according to clinical course as having either recurrent relapses
with partial remissions (RR) or secondary, chronic progressive
disease (without any discrete relapses in our group) after an
earlier history of relapsing and partially remitting disease
(SP). We attempted to match the two patient groups on the
basis of disability and to choose groups with disability levels
relevant to current clinical therapeutic trials. This had the
effect of selecting for patients with intermediate Kurtzke
scores. Prior to the MRI examination the patients had a
clinical evaluation for final scoring of disability. Normal
control subjects were hospital and laboratory workers in
good health.
MRI and MRS examinations were performed in the same
session using a Philips Gyroscan S15 (1.5 T, Philips Medical
Systems, Netherlands). An axial view was taken to ensure
that the interhemispheric fissure was aligned with the vertical
axis and then sagittal scout views were performed to identify
the anterior callosal-posterior callosal line (AC-PC line).
Multislice transverse dual spin-echo images were obtained
(TR 2100/TE 30,78/5.5 mm slice thickness, 0.5 mm interslice
distance, 4 excitations). A brick-shaped volume of interest
(48X70X28 mm3) for spectroscopy was chosen from these
images, oriented along the AC-PC line, and centred on the
corpus callosum so that it included the superior lateral
ventricular regions.
Proton MR spectra were obtained using a 90-180-180
(PRESS) sequence for volume selection with T,-nulling of
water (TR = 2 s, TE = 272 ms, 256 averages) as previously
described (Arnold et al., 1994). Resonance intensities were
determined from peak areas in the Fourier-transformed
spectra, relative to a baseline interpolated independently for
each of the resonances from mean noise intensities on either
side. Area measurements are reproducible to within ~5%.
Chemical shifts are expressed relative to NAA at 2.0 p.p.m.
Prior to semi-automated measurement of lesion or
ventricle volumes, images were transformed into a standard
brain (stereotaxic) space using manual homologous landmark
matching between the MRI volume and an average (n > 300)
MRI brain volume that is coextensive with the Talairach
atlas (Talairach and Tournoux, 1988; Evans et al., 1992).
Lesion classification was performed using locally developed
software offering the ability to toggle between the proton
density and T2-weighted images (to allow easier
Combined MRI and MRS in multiple sclerosis
Fig. 1 Axial section from a typical patient scan illustrating results
of semi-automated segmentation. The total brain, ventricular and
lesion volumes are outlined on the basis of operator defined
contrast thresholds. Volume measurements reported in the paper
are those of the segmented volumes after mapping of the image
into the standard brain space.
discrimination between grey matter and CSF). Semiautomatic outlining tools facilitated definition of lesion
volumes (Kamber, 1991). To use them, the operator
establishes a contrast threshold for the lesion edge, based on
a linear map of signal intensities in a central slice through
the lesion. The edge is then automatically traced in threedimensions to define the lesion volume (Fig. 1). This task is
performed for each lesion independently. Volumes are then
computed in cubic centimetres and total brain lesion volumes
for each patient are determined. Reproducibility of lesion
volume measurements made in this way is greater than with
manual tracing methods (our unpublished observations). Use
of data in stereotaxic space allows meaningful comparisons of
morphometric data from different individuals, who generally
vary in absolute brain dimensions.
Statistical analysis utilized non-parametric methods with
Spearman correlations or the Kruskal-Wallis t test, as
indicated.
Results
The 29 multiple sclerosis patients studied had durations of
disease ranging from 3 to 33 years (Table 1). Eleven were
clinically classified as RR and 18 as SP. There was a
significantly longer mean duration of disease for SP patients
717
(19.6±6.3 years) than for the RR group (8.7±5.4 years,
P < 0.0001). Extended disability status score (EDSS) ranged
from 3 to 7. The mean EDSS score for the RR patients of
4.8 was lower than that for the SP group (mean EDSS=6.2,
P < 0.006).
A significant difference in total hemispheric lesion volume
was found between patients in the two clinical subgroups:
the mean lesion volume was greater in RR (32.5±20.9 cm3)
than in SP patients (16.2±9.0 cm3, P < 0.04), despite the
longer mean duration of disease in the SP patients and the
lower mean EDSS in the RR patients. No correlation between
lesion volume and EDSS was apparent (Spearman
coefficients: total group, r = 0.20; RR patients only, r =
0.28; SP patients only, r = 0.20, P > 0.5 for all). Brain
ventricular volumes were increased for the multiple sclerosis
patients relative to controls (patients, 31 ±23 cm3, range 9119 cm3; controls, 18±6 cm3, range 9-29 cm3; P < 0.009).
No significant difference in ventricular volumes was found
between patients in the RR and SP groups.
Brain proton MRS study with a large central brain volume
of interest (Fig. 2) showed that the NAA:Cr ratio was
reduced by 18-42% for the multiple sclerosis patients (mean
NAA:Cr = 3.1 ±0.4) relative to controls (mean NAA:Cr =
4.0±0.3, n= 18, P < 0.0001). The decrease in NAA:Cr for
patients in the two clinical groups was similar (RR, 3.0±0.5,
P< 0.0001 compared with normal controls; SP, 3.2±0.4,
P < 0.0001). A strong correlation between NAA:Cr and total
brain lesion volumes was found for RR multiple sclerosis
patients (Spearman coefficient, r = -0.78, P < 0.03) (Fig. 3).
A similar correlation was not found with data from the SP
patients (Spearman coefficient, r = -0.12, P > 0.5), but both
ranges of values for correlation for this group were smaller:
RR brain lesion volume range, 12.2-75.0 cm3 with a range
for NAA:Cr of 2.3-4.0; SP brain lesion volume range, 7.632.3 cm3 with a NAA:Cr range of 2.9-3.9.
Heterogeneity of MRI lesions in the two patient groups
was suggested by significant differences in the relationship
of total brain T2-weighted lesion volume and NAA:Cr. The
mean ratio of lesion volume:NAA:Cr was greater for RR
patients (11.7±9.3 cm3) than for SP patients (5.4±3.3 cm3,
P < 0.05). A similar difference in the relationship of lesion
volume to NAA:Cr was found between the two groups
when lesion volume just within the volume of interest was
considered [RR (lesion volume in the volume of interest)/
(NAA:Cr) = 2.2± 1.0 cm3: SP, 1.5±1.0 cm3; P < 0.05].
The hypothesis that ventriculomegaly in multiple sclerosis
indirectly reflects neuronal loss was tested by correlating
NAA:Cr and ventricular size. A correlation between NAA:Cr
and ventricular size was not apparent for the full study group.
Comparison of the five patients with the highest NAA:Cr
(mean NAA:Cr, 3.6±0.3; mean ventricular size, 21.3±8.4
cm3) with the five patients with the lowest NAA:Cr (mean
NAA:Cr, 2.6±0.3; mean ventricular size, 46.9±41.1 cm3)
suggested a trend towards increased ventricular size with
lower NAA:Cr, although the difference between the two
groups did not reach statistical significance (P > 0.2).
718
P. M. Matthews et al.
Table 1 Patient characteristics and quantitative MRI morphometry and MRS results
Patient
no.
Sex
Relapsing-remitting
1
F
2
M
3
M
4
F
5
F
6
M
7
M
M
8
F
9
10
M
11
F
Secondary progressive
12
M
13
M
14
F
15
M
16
M
17
M
18
M
19
F
20
F
F
21
22
F
23
F
24
F
25
M
27
M
28
M
M
29
Controls
Age
(years)
Duration of
symptoms
(years)
EDSS
Total brain
lesion
volume (cm3)
Lesion volume
in VOI
(cm3)
Ventricle
size
(cm3)
NAA:Cr
36
29
26
45
26
40
31
28
39
30
34
9
10
7
7
7
7
3
3
23
9
11
4.0
4.5
5.0
3.5
4.0
3.0
5.5
6.0
5.0
6.5
6
44.1
30.6
14.4
75.0
15.5
12.2
12.3
33.5
56.2
46.2
17.5
8.9
6.8
6.3
9.0
3.8
2.7
4.4
9.3
NA
10.2
6.4
37.7
30.2
18.6
119.4
26.5
14.2
9.0
18.2
46.6
31.6
18.9
3.24
2.80
3.99
2.35
3.42
3.29
3.25
3.02
2.32
3.01
3.13
45
57
43
56
40
53
55
44
52
51
47
49
45
44
58
32
58
13
21
22
28
21
19
22
21
23
11
17
16
17
14
27
9
33
6.0
6.5
6.5
6.5
6.0
6.0
6.5
6.0
6.0
6.0
6.0
5.0
6.0
7.0
4.0
7.0
6.5
11.2
28.2
14.7
6.8
15.3
10.0
20.3
12.2
12.2
22.5
34.7
1.6
32.2
17.7
7.4
20.8
15.0
2.7
9.6
5.0
2.0
5.4
3.3
3.4
3.3
2.0
9.7
5.7
0.9
10.7
4.6
1.7
3.0
3.4
17.4
33.2
22.5
66.8
19.7
12.8
26.6
13.9
66.3
21.1
39.2
11.6
24.2
15.7
12.8
53.7
47.2
18.0 ± 6.0
3.22
3.59
2.97
3.18
2.71
3.26
2.32
3.45
3.46
3.59
2.85
3.04
2.66
2.78
3.32
3.40
3.88
3.98 ± 0.33
All volumes are measured in standard space. VOI = volume of interest; NA = data not available.
Discussion
Assessment of the extent of disease by in vivo imaging
techniques should provide further insight into evolution of
the disease process in multiple sclerosis patients and provide
an additional outcome measure in therapeutic trials. Total
lesion volume has been proposed as one measure and
progressive increases in brain T2-weighted MRI lesion
volume with time have been shown (Paty et al., 1993, 1994;
Filippi etai, 1994; Miller. 1994). However, while correlations
between changes in T2-weighted MRI lesion load and clinical
disability have been shown, they are not strong (Khoury
et al., 1994; Miller, 1994; Filippi et al., 1995). The potential
limitations of use of this lesion volume as an independent
index of the extent of disease are emphasized further by our
observation that total brain lesion volumes were lower in a
SP group than in a group of RR patients, despite the more
than 2.5-fold longer duration of disease and greater mean
clinical disability in the SP group. Pathological heterogeneity
of the T2-weighted MRI lesions may play a significant role
in determining the relationship between lesion volume and
disability, clinical course and disease duration. In fact,
anatomical, cellular, and chemical heterogeneity of multiple
sclerosis lesions has been demonstrated by descriptions of
(i) variable pathology of lesions even within the same patient,
(ii) lesion-to-lesion, time-dependent blood-brain barrier
permeability differences, (iii) variable patterns of changes in
lesion sizes with time, and (iv) lesion heterogeneity in
diffusion imaging, magnetization transfer and MR
relaxometry studies (Kermode et al., 1990; Barnes et al.,
1991; Thompson et al., 1991; Dousset et al., 1992; McDonald
et al., 1992; Christiansen et al., 1993).
The decreased brain NAA:Cr ratio found in the multiple
sclerosis patients in this study is consistent with previous
work, which has established that decreases in the ratio are a
consequence primarily of decreased amounts of NAA, a
compound found only in neurons in the mature brain (Arnold
etal., 1990, 1992, 1994; Matthews et al., 1991; Miller et al.,
1991; Van Hecke et al., 1991). This has been interpreted
generally to be a consequence of secondary axonal damage
associated with demyelination. Axonal loss distinguishes
what is presumed to be an irreversible aspect of lesion
evolution (McDonald, 1994). Although reversible decreases
in NAA can be seen in acute lesions that diminish in size
(Arnold et al., 1992; De Stefano et al., 1993; Davie et al..
Combined MR! and MRS in multiple sclerosis
719
(A)
(C)
NAA
Cho
iCho
LA
LA
3.5
(P-P-m.)
0.8
3.5
(P-P-m.)
0.8
Fig. 2 (A) Sagittal MR scout view illustrating positioning of the MRS volume of interest with centering on the corpus callosum. (B)
Axial view of the same MRS volume of interest. (C) Brain proton MR spectra from central brain volume of interest (as illustrated in A
and B) of a normal control and a SP multiple sclerosis patient with ventriculomegaly. From right to left, resonances from NAA, Cr and
choline (Cho) are seen. The prominent lactate (LA) resonance at 1.2 ppm arises from lactate in the CSF. An increased lactate signal is a
common finding associated with ventriculomegaly, reflecting the increased proportion of the spectroscopic volume of interest including
CSF. As CSF does not contain significant amounts of NAA, Cr or Cho, the relative intensities of these resonances reflect parenchymal
concentrations.
1994), decreases in NAA are relatively stable in chronic
lesions. Decreases of NAA also are found in normal-appearing
white matter (Arnold et al., 1992; Davie et a/., 1994;
Husted, 1994), likely reflecting axonal loss due to Wallerian
degeneration from sites of damage in focal lesions. By
choosing a volume in which a single large brain volume
is centred on the corpus callosum (a volume in which
cortical axons converge) for sampling of NAA signal,
our measurements provide, therefore, an index for
characterization of brain axonal loss in a volume of brain
considerably greater than the dimensions of the volume
of interest.
720
P. M. Matthews et al.
4-S-i (A)
3.53-
2.52-
1.5
c
10 20 30 40 50 60 70 80
4.5-1 (B)
•
4"
•
3.5-
• -•
1h
3- •
•
•
_ •
"
•
2.5-
•
•
-
—
~t-—
i
•
2-
()
5
10 15 20 25 30 35 40
Lesion volume (cm3)
Fig. 3 Correlation between total brain lesion volume and
NAA:Cr in the central volume of interest for patients with either
a clinical course marked by recurrent relapses and partial or
complete remissions (RR) (A) or a secondary stage of progressive
clinical deterioration without well-defined relapses (SP) (B). The
upper plot illustrates the significant correlation found with the RR
group (Spearman coefficient, r = -0.78, P < 0.02).
The observation that the lesion volume:(NAA:Cr) ratio
was lower in the SP than in RR patients suggests that either
the lesions or the surrounding normal-appearing white matter
in the two patient groups have different pathological
characteristics on average, despite indistinguishable MR
appearances on T2-weighted scans. The similar mean NAA:Cr
for the two groups suggests a similar overall extent of brain
axonal loss or damage. However, as the total lesion volume
was greater in RR than in SP patients, we conclude either
that the mean degree of axonal loss or damage per unit lesion
volume is higher in the SP group or that there is a greater
proportion of lesions in the brains of RR patients that are
not associated with axonal loss or damage. A progression of
pathological changes towards increased axon loss or damage
in lesions with greater duration of disease (as found with the
SP patients) may correlate with increasing proportions of
the so-called 'open' lesions denned in fixed, post-mortem
specimens and with an increasing proportion of lesions
showing bi-exponential MR T2 relaxation decay curves in
vivo (Barnes et al., 1991). In preliminary work we have begun
to explore whether MRS imaging (MRSI) with registration of
lesion volumes could be used to define the spatial variation
in chemical pathology across lesions and normal-appearing
white matter that is 'averaged' in the single voxel
measurements reported here (L. Fu, D. L. Arnold, N. De
Stefano and P. M. Matthews, unpublished observations).
An alternative explanation that could account for the
differences in lesion volume:(NAA:Cr) between the two
patient groups is that the lesions in the SP patients may have
a higher probability of being localized to the periventricular
region within the volume of interest than those of RR patients.
However, data in Table 1 do not show a significant difference
in the proportion of total lesion volume within the volume
of interest between the two groups of patients (RR,
0.26±0.09; SP, 0.29±0.10). More generally, comparison of
the relative proportion of lesions in the periventricular region
relative to regions of white matter more distant from the
ventricles has failed to demonstrate a significant difference
in lesion distribution between the two groups (S. Narayanan,
L. Fu and D. L. Arnold, unpublished observations).
Ventriculomegaly was initially explored as a measure of
neuronal damage or loss in neurodegenerative and cerebral
vascular diseases (Coffey et al., 1992; DeCarli et al., 1992;
Liu et al., 1992) and may be useful in multiple sclerosis
(Clark et al., 1992), as callosal atrophy and ventriculomegaly
are common radiological findings in advanced stages of the
disease. The quantitative segmentation techniques employed
here incidently allowed us to examine ventriculomegaly as
a possible indirect morphological measure of axonal loss.
There was not a clear relationship between neuronal loss
(as assessed from the NAA:Cr ratio) and ventriculomegaly,
in contrast to expectations based on previous observations in
primary neurodegenerative disease and correlations between
neuropsychological indices and ventricular size in multiple
sclerosis (Clark et al., 1992; DeCarli et al, 1992). We
speculate that the trend towards increased ventricular size
with decreased NAA:Cr reported here suggests that
ventriculomegaly in multiple sclerosis may be due, in pan,
to neuronal loss [consistent with the proportionality of
atrophy of the corpus callosum and functional measures of
interhemispheric transfer of information (e.g. Rao et al.,
1989)]. The lack of a strong correlation emphasizes that
other factors (e.g. oligodendroglial cell and process loss,
demyelination, astrocytic proliferation and consolidation of
parenchyma, and interstitial fluid and matrix changes) must
make dominant contributions to ventriculomegaly.
This report illustrates how chemical data from MRS and
morphometric data from MRI can be combined for an
improved description of multiple sclerosis pathology. In
principle, other MRI information reflecting additional
pathological characteristics of lesions (e.g. hypodensity on
T2-weighted images, altered magnetization transfer, or
gadolinium enhancement) could be collected during the same
examinations to define better the burden of disease. We
believe that simultaneous statistical analysis of data from
morphometry of MRIs collected with different sequences,
contrast studies, and MRS or MRSI prove useful for assessing
outcome in patients with multiple sclerosis in treatment
Combined MRI and MRS in multiple sclerosis
or natural history studies by providing multi-dimensional,
quantitative descriptions of the underlying pathology.
Acknowledgements
We wish to thank Dr Alan Evans and his group in the
Neuroimaging Laboratory of the Montreal Neurological
Institute for development of segmentation software and our
colleagues in the Montreal Neurological Institute multiple
sclerosis Imaging Research Group for encouragement and
stimulating criticism. This project was supported by the MRC
(D.L.A., G.F., J.A. and C.W.) and the Multiple Sclerosis
Society of Canada (D.L.A., G.F., J.A., P.M.M., C.W. and
Alan Evans). P.M.M. is a Medical Research Council
clinician-scientist.
References
Arnold DL, Matthews PM, Francis G, Antel J. Proton magnetic
resonance spectroscopy of human brain in vivo in the evaluation of
multiple sclerosis: assessment of the load of disease. Magn Reson
Med 1990; 14: 154-9.
721
Dousset V, Grossman RI, Ramer KN, Schnall MD, Young LH,
Gonzalez-Scarano F, et al. Experimental allergic encephalomyelitis
and multiple sclerosis: lesion characterization with magnetization
transfer imaging [published erratum appears in Radiology 1992;
183: 878]. Radiology 1992; 182: 483-91.
Evans AC, Marrett S, Neelin P, Collins L, Worsley K, Dai W,
et al. Anatomical mapping of functional activation in stereotactic
coordinate space. Neuroimage 1992; 1: 43-53.
Filippi M, Horsfield MA, Morrissey SP, MacManus DG, Rudge P,
McDonald WI, et al. Quantitative brain MRI lesion load predicts
the course of clinically isolated syndromes suggestive of multiple
sclerosis. Neurology 1994; 44: 635^11.
Filippi M, Paty DW, Kappos L, Barkhof F, Compston DAS,
Thompson AJ, et al. Correlations between changes in disability and
T2-weighted brain MRI activity in multiple sclerosis: a follow-up
study. Neurology 1995; 45: 255-60.
Hohol MJ, Orav EJ, Weiner HL. Disease steps in multiple sclerosis:
a simple approach to evaluate disease progression. Neurology 1995;
45: 251-5.
Husted C. Contributions of neuroimaging to diagnosis and
monitoring of multiple sclerosis. [Review]. Curr Opi Neurol 1994;
7: 234-^1.
Arnold DL, Matthews PM, Francis GS, O'Connor J, Antel JP.
Proton magnetic resonance spectroscopic imaging for metabolic
characterization of demyelinating plaques. Ann Neurol 1992; 31:
235-41.
Kamber M. Automated detection of multiple sclerosis lesions in
magnetic resonance images of the human brain [Master's Thesis].
Montreal; Concordia University, 1991.
Arnold DL, Riess GT, Matthews PM, Francis GS, Collins DL,
Wolfson C, et al. Use of proton magnetic resonance spectroscopy
for monitoring disease progression in multiple sclerosis. Ann Neurol
1994; 36: 76-82.
Kermode AG, Thompson AJ, Tofts R MacManus DG, Kendall BE,
Kingsley DP, et al. Breakdown of the blood-brain barrier precedes
symptoms and other MRI signs of new lesions in multiple sclerosis.
Pathogenetic and clinical implications. Brain 1990; 113: 1477-89.
Barnes D, Munro PM, Youl BD, Prineas JW, McDonald WI.
The longstanding MS lesion. A quantitative MRI and electron
microscopic study. Brain 1991; 114: 1271-80.
Christiansen P, Gideon P, Thomsen C, Stubgaard M, Henriksen O,
Larsson HB. Increased water self-diffusion in chronic plaques and
in apparently normal white matter in patients with multiple sclerosis.
Acta Neurol Scand 1993; 87: 195-9.
Clark CM, James G, Li D, Oger J, Paty D, Klonoff H. Ventricular
size, cognitive function and depression in patients with multiple
sclerosis. Can J Neurol Sci 1992; 19: 352-6.
Coffey CE, Wilkinson WE, Parashos IA, Soady SA, Sullivan RJ,
Patterson LJ, et al. Quantitative cerebral anatomy of the aging
human brain: a cross-sectional study using magnetic resonance
imaging. Neurology 1992; 42: 527-36.
Davie C A, Hawkins CP, Barker GJ, Brennan A, Tofts PS, Miller
DH, et al. Serial proton magnetic resonance spectroscopy in acute
multiple sclerosis lesions. Brain 1994; 117: 49-58.
DeCarli C, Haxby JV, Gillette JA, Teichberg D, Rapoport SI,
Schapiro MB. Longitudinal changes in lateral ventricular volume
in patients with dementia of the Alzheimer type. Neurology 1992;
42: 2029-36.
De Stefano N, Francis G, Antel JP, Arnold DL. Reversible decreases
of N-acetylaspartate in the brain of patients with relapsing-remitting
multiple sclerosis. Proc Soc Magn Reson Med 1993; 1: 280.
Khoury SJ, Guttmann CRG, Orav EJ, Hohol MJ, Ahn SS, Hsu L,
et al. Longitudinal MRI in multiple sclerosis: correlation between
disability and lesion burden. Neurology 1994; 44: 2120-4.
Liu CK, Miller BL, Cummings JL, Mehringer CM, Goldberg MA,
Howng SL, et al. A quantitative MRI study of vascular dementia.
Neurology 1992; 42: 138^3.
Matthews PM, Francis G, Antel J, Arnold DL. Proton magnetic
resonance spectroscopy for metabolic characterization of plaques in
multiple sclerosis [published erratum appears in Neurology 1991;
41: 1828]. Neurology 1991; 41: 1251-6.
McDonald WI. Rachelle Fishman-Matthew Moore Lecture. The
pathological and clinical dynamics of multiple sclerosis. [Review].
J Neuropathol Exp Neurol 1994; 53: 338^13.
McDonald WI, Miller DH, Barnes D. The pathological evolution
of multiple sclerosis. [Review]. Neuropathol Appl Neurobiol 1992;
18: 319-34.
Miller DH. Magnetic resonance in monitoring the treatment of
multiple sclerosis. [Review]. Ann Neurol 1994; 36 Suppl: S9l-^l.
Miller DH, Austin SJ, Connelly A, Youl BD, Gadian DG, McDonald
WI. Proton magnetic resonance spectroscopy of an acute and chronic
lesion in multiple sclerosis [letter] [see comments]. Lancet 1991;
337: 58-9. Comment in: Lancet 1991; 337: 362.
Newcombe J, Hawkins CP, Henderson CL, Patel HA, Woodroofe
MN, Hayes GM, et al. Histopathology of multiple sclerosis lesions
722
P. M. Matthews et al.
detected by magnetic resonance imaging in unfixed postmortem
central nervous system tissue. Brain 1991; 114: 1013-23.
Paty DW, Li DK. Interferon beta-lb is effective in relapsingremitting multiple sclerosis. II. MRI analysis results of a multicenter,
randomized, double-blind, placebo-controlled trial. UBC MS/MRI
Study Group and the IFNB Multiple Sclerosis Study Group [see
comments]. Neurology 1993; 43: 662-7. Comment in: Neurology
1993; 641-3.
Paty DW, Li DKB, Oger J-L, Kastrukoff L, Koopmans R, Tanton
E, et al. Magnetic resonance imaging in the evaluation of clinical
trials in multiple sclerosis. Ann Neurol 1994; 36 Suppl: S956.
Rao SM, Leo GJ, Haughton VM, St. Aubin-Faubert P, Bernardin L.
Correlation of magnetic resonance imaging with neuropsychological
testing in multiple sclerosis. Neurology 1989; 39: 161-6.
Talairach J, Tournoux P. Co-planar stereotaxic atlas of the human
brain: 3-dimensional proportional system: an approach to cerebral
imaging. Stuttgart: G. Thieme, 1988.
Thompson AJ, Kermode AG, Wicks D, MacManus DG, Kendall
BE, Kingsley DP, et al. Major differences in the dynamics of
primary and secondary progressive multiple sclerosis. Ann Neurol
1991; 29: 53-62.
Van Hecke P, Marchal G, Johannik K, Demaerel P, Wilms G, Carton
H, et al. Human brain proton localized NMR spectroscopy in
multiple sclerosis. Magn Reson Med 1991; 18: 199-206.
Received July 11, 1995. Revised October 31, 1995.
Accepted December 29, 1995