Download Diagnostic Imaging of Patients in a Memory Clinic: Comparison of

ORIGINAL RESEARCH
䡲 NEURORADIOLOGY
Note: This copy is for your personal non-commercial use only. To order presentation-ready
copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
Diagnostic Imaging of Patients
in a Memory Clinic: Comparison
of MR Imaging and 64 –Detector
Row CT1
Mike P. Wattjes, MD
Wouter J. P. Henneman, MD
Wiesje M. van der Flier, PhD
Oscar de Vries, MD, PhD
Frank Träber, PhD
Jeroen J. G. Geurts, PhD
Philip Scheltens, MD, PhD
Hugo Vrenken, PhD
Frederik Barkhof, MD, PhD
Purpose:
To investigate the assessment of global cortical atrophy
(GCA), medial temporal lobe atrophy (MTA), and white
matter changes (WMCs) in patients screened at a memory
clinic with a 64 – detector row computed tomography (CT)
brain protocol, in comparison with the reference standard, magnetic resonance (MR) imaging.
Materials and
Methods:
The study protocol was approved by the local institutional
review board. Written informed consent was obtained from
all participants. Thirty patients (21 men, nine women; median age, 62 years) who presented to a memory clinic
underwent 64 – detector row CT and multisequence MR
imaging of the brain on the same day. Three readers
blinded to the clinical diagnosis assessed the resultant
images. Images were presented in random order and
scored for GCA, MTA, and WMC with published visual
rating scales. Intermodality agreement between CT and
MR imaging (intrareader agreement across both modalities), expressed by weighted ␬ analysis, and interobserver
agreement within each modality between readers (Kendall
W test) were assessed.
Results:
Overall, excellent intraobserver agreement between CT
and MR imaging was observed for GCA (mean ␬, 0.83) and
MTA (mean ␬, 0.88 and 0.86 on the left and right sides of
the brain, respectively). There was substantial overall
agreement concerning WMC (mean ␬, 0.79). For all three
tested scales, interobserver variability was low and comparable for CT and MR imaging.
Conclusion:
Use of 64 – detector row brain CT yields reliable information that is comparable with that obtained with MR imaging. Thus, multidetector row CT is a suitable diagnostic
imaging tool in a memory clinic setting.
娀 RSNA, 2009
1
From the Departments of Radiology (M.P.W., W.J.P.H.,
J.J.G.G., H.V., F.B.), Neurology (W.M.v.d.F., P.S.), Internal
Medicine (O.d.V.), Pathology (J.J.G.G.), and Physics and
Medical Technology (H.V.), VU University Medical Center,
De Boelelaan 1117, 1081 HV Amsterdam, the Netherlands; and Department of Radiology, University of Bonn,
Bonn, Germany (F.T.). Received December 19, 2008; revision requested February 25, 2009; revision received
March 31; accepted April 8; final version accepted April
27. Address correspondence to M.P.W. (e-mail:
[email protected] ).
姝 RSNA, 2009
174
radiology.rsna.org ▪ Radiology: Volume 253: Number 1—October 2009
NEURORADIOLOGY: Diagnostic Imaging of Patients in a Memory Clinic
S
tructural neuroimaging has become an important tool in the initial
work-up of patients suspected of
having dementia and has been incorporated into guidelines on the diagnosis of
dementia (1). The role of neuroimaging
extends beyond the detection of potentially treatable causes of dementia—such
as tumors, hematomas, or hydrocephalus—to include the assessment of specific
neurodegenerative findings, particularly
cerebral atrophy patterns and vascular
abnormalities (2–4).
The presence of vascular abnormalities is essential for a diagnosis of probable
vascular dementia (5). One of the vascular abnormalities is the presence of vascular white matter changes (WMCs).
Several rating scales, most of which were
developed for use with magnetic resonance (MR) imaging, have been established for use in the assessment of WMC
(6). Fazekas et al developed a simple and
robust scale that is used frequently and
can be applied to MR images, even those
with lower image quality (7).
Assessments of global cortical atrophy (GCA) and medial temporal lobe atrophy (MTA) are of special diagnostic
value. MTA can be used to support a diagnosis of Alzheimer disease; however, it
may occur in other diseases that cause
dementia and in the normal aging brain
(1–3,8,9). The assessment of GCA and
MTA is based on visual assessment with
Advances in Knowledge
䡲 Visual rating of global cortical atrophy (GCA) and medial temporal
lobe atrophy (MTA) on 64 – detector row computed tomography
(CT) images is comparable with
that on magnetic resonance (MR)
images and characterized by high
intermodality agreement.
䡲 Clinically relevant white matter
damage can be assessed equally
well with 64 – detector row CT
and MR imaging.
䡲 In the assessment of GCA, MTA,
and white matter changes
(WMCs), 64 – detector row CT
shows a similar low interobserver
variability when compared with
MR imaging in a dementia setting.
scales, considering the width of the sulci
and the volume of the gyri (10,11) and
widened choroid fissures, widening of the
temporal horn, and volume loss of the
hippocampus (8,12,13).
Although MR imaging is the preferred
imaging modality in the diagnostic
work-up of patients suspected of having
dementia, many patients cannot undergo
MR imaging because of safety concerns
(eg, pacemaker), claustrophobia, or agitation. Moreover, in some hospitals, MR
imaging is not available to these patients.
Multi– detector row computed tomography (CT) could be an acceptable alternative for use in patients who cannot undergo MR imaging.
MR imaging is superior to CT in the
evaluation of patients suspected of having
or who definitely have dementia when
GCA, MTA, and WMC are assessed,
leading to better sensitivity, specificity,
and diagnostic accuracy (14–17). However, the studies in which this finding was
observed were performed with older CT
scanners. To our knowledge, no one has
investigated whether MR imaging is also
superior to multi– detector row CT (up to
64 detector rows) with higher spatial resolution and better opportunities in terms
of multiplanar reconstructions.
The purpose of this study was to investigate the assessment of GCA, MTA,
and WMC in patients screened at a memory clinic with a 64 – detector row CT
brain protocol, in comparison with the
reference standard, MR imaging.
Materials and Methods
Study Design and Patients
Between September 2007 and June 2008,
we performed a prospective intraindividual
comparative study in patients who were
Implication for Patient Care
䡲 Sixty-four– detector row CT yields
sufficient information in the evaluation of patients in a memory
clinic setting in terms of the assessment of GCA, MTA, and
WMCs and can be used in patients who cannot undergo MR
imaging.
Radiology: Volume 253: Number 1—October 2009 ▪ radiology.rsna.org
Wattjes et al
suspected of having or who definitely had
dementia. The inclusion criteria were cognitive or behavioral complaints that were
suggestive of dementia. The exclusion criteria were pregnancy and inability to undergo
MR imaging, CT, or both. The study protocol was approved by our local institutional
review board (VU University Medical Center). Written informed consent was obtained from all patients before they entered
the study during their first visit to our memory clinic and after the study had been fully
explained to them.
Thirty patients (median age, 62 years;
age range, 44–85 years), 21 of whom were
men (median age, 58 years; age range,
44 – 82 years) and nine of whom were
women (median age, 65 years; age range,
51–85 years), presented to the outpatient
memory clinic of our institution with memory deficiencies, as reported at clinical history taking. These patients were consecutively selected and included by one neurologist (P.S., 20 years experience). None of
the primarily selected patients were excluded. All patients underwent a routine
multisequence MR examination as part of
the diagnostic work-up and were asked to
undergo an additional CT examination on
the same day prior to or after the scheduled
MR examination.
MR Protocol
MR examinations were performed with a
clinical whole-body MR system with a
field strength of 1.5 T (Sonata; Siemens,
Published online before print
10.1148/radiol.2531082262
Radiology 2009; 253:174 –183
Abbreviations:
GCA ⫽ global cortical atrophy
MTA ⫽ medial temporal lobe atrophy
WMC ⫽ white matter change
Author contributions:
Guarantors of integrity of entire study, M.P.W., P.S., F.B.;
study concepts/study design or data acquisition or data
analysis/interpretation, all authors; manuscript drafting or
manuscript revision for important intellectual content, all
authors; manuscript final version approval, all authors;
literature research, M.P.W.; clinical studies, M.P.W.,
W.J.P.H., P.S., H.V., F.B.; statistical analysis, M.P.W.,
W.M.v.d.F., O.d.V., F.T.; and manuscript editing, all
authors
Authors stated no financial relationship to disclose.
175
NEURORADIOLOGY: Diagnostic Imaging of Patients in a Memory Clinic
Erlangen, Germany) and use of an eightchannel head coil and the following pulse
sequences: (a) a coronal three-dimensional T1-weighted magnetization-prepared rapid gradient-echo sequence (repetition time msec/echo time msec/inversion time msec, 2700/5/950; voxel size,
1 ⫻ 1 ⫻ 1.5 mm; one signal acquired) with
multiplanar reconstructions (voxel size, 1 ⫻
1 ⫻ 3 mm) in transverse, sagittal, and
oblique coronal orientations perpendicular
to the hippocampus (Fig 1); (b) a twodimensional transverse T2-weighted fast
spin-echo sequence (repetition time msec
/echo time msec, 4590/114; echo train
length, 15; voxel size, 0.5 ⫻ 0.5 ⫻ 5
mm; two signals acquired); and (c) a
two-dimensional transverse fluid-attenuated inversion-recovery sequence
(9000/108/2500; voxel size, 1 ⫻ 1 ⫻ 5
mm; echo train length, 21; one signal
acquired).
CT Protocol
CT scanning of the brain was performed
with a 64 – detector row clinical CT scanner (Sensation 64; Siemens) and use of a
dementia imaging protocol (380 mAs,
120 kV, 0.6-mm collimation, pitch of
1.15). The effective radiation dose was
1.5 mSv. Images were presented and
viewed in the brain window (window center, 35 HU; window width, 80 HU).
Oblique coronal, sagittal, and transverse
reconstructions (section thickness, 3
mm; in-plane spatial resolution, 0.4 ⫻ 0.4
mm) were made in an orientation similar
to that of the corresponding MR images
(Fig 1).
Image Analyses
All obtained images were rated separately by three readers with different
levels of experience with MR imaging
and CT (reader 1, M.P.W. [7 years experience]; reader 2, F.B. [20 years experience]; reader 3, W.J.P.H. [2 years
experience]). CT and MR images were
assessed for GCA, MTA, and WMC.
We used the five-point visual rating
scale (scores ranged from 0 to 4) described by Scheltens et al (8,12) to assess MTA. To assess GCA, we used a
four-point rating scale (scores ranged
from 0 to 3) based on the rating scale
for analysis of regional atrophy estab176
Wattjes et al
Figure 1
Figure 1: Sagittal planning images for (a) MR imaging and (b) CT serve as examples of the oblique coronal
reconstructions perpendicular to the hippocampus axis. White lines indicate the exact anatomic position of c
and d. (c) MR imaging and (d) CT oblique coronal reconstructions perpendicular to the hippocampus axis do
not show any relevant differences in terms of repositioning.
Figure 2
Figure 2: Oblique coronal (a) CT and (b) T1-weighted MR images in a 58-year-old man with right-sided variant
of frontotemporal lobe degeneration. All three readers identified focal atrophy of the right temporal lobe (arrow).
radiology.rsna.org ▪ Radiology: Volume 253: Number 1—October 2009
NEURORADIOLOGY: Diagnostic Imaging of Patients in a Memory Clinic
lished by Pasquier et al (11). The
amount of subcortical WMC was scored
with the four-point rating scale (scores
ranged from 0 to 3) described by Fazekas et al (7). For MR images, GCA and
WMC analyses were based on axial twodimensional fluid-attenuated inversionrecovery images, while MTA analysis
was based on coronal reconstructions of
the T1-weighted magnetization-pre-
pared rapid gradient-echo sequence
performed perpendicular to the hippocampus axis, as described previously
(8,12). On CT images, GCA and WMC
analyses were based on the axial reconstruction, whereas MTA analysis was
based on the oblique coronal reconstructions. The images from different
patients obtained with the two imaging
modalities were presented in a random
Figure 3
Wattjes et al
order. All three readers were blinded to
the patient’s clinical presentation and
findings of additional investigations. In
addition, each reader separately recorded and documented incidental findings.
Statistical Analysis
The intermodality agreement between
CT and MR imaging (intrareader agreement across both modalities) was calculated separately for all three rating scales
and for each reader by using weighted ␬
analysis based on the raw scores. For
weighted ␬ values, the degree of agreement was defined according to the
method of Landis and Koch (18), as follows: no agreement, the ␬ value was less
than 0.00; slight agreement, the ␬ value
ranged from 0.00 to 0.20; fair agreement,
the ␬ value ranged from 0.21 to 0.40;
moderate agreement, the ␬ value ranged
from 0.41 to 0.60; substantial agreement,
the ␬ value ranged from 0.61 to 0.80; and
excellent agreement, the ␬ value ranged
from 0.81 to 1.00.
Interobserver variability was assessed for CT and MR imaging with the
Kendall W test. Assessment of systematic
trends in terms of higher or lower scores
obtained with one modality (CT or MR
imaging) was performed with the nonparametric test of marginal homogeneity
and included power analysis.
All statistical calculations were performed with SPSS software (version 14.0;
SPSS, Chicago, Ill). In addition, power
analysis was performed with a computer
program (G*Power, version 3; Institute
for Experimental Psychology, University
of Düsseldorf, Düsseldorf, Germany)
(19), taking into account the Spearman
correlation coefficients between the
matched samples.
Results
Figure 3: Examples of incidental findings in the patient group that were seen on both CT and MR images. (a) Transverse T2-weighted MR image in an 85-year-old woman shows a mass lesion (arrow) indicative of mengioma.
(b) Corresponding CT image in the same patient shows this same lesion (arrow). (c) Transverse fluid-attenuated inversion-recovery MR image in a 66-year-old man shows end-stage atrophy of the cerebellum (arrow).
(d) Corresponding CT image in the same patient shows the same end-stage atrophy of the cerebellum (arrow).
Radiology: Volume 253: Number 1—October 2009 ▪ radiology.rsna.org
Clinical Features
The clinical diagnoses after the neurologic and radiologic evaluation of the
included patients were as follows: probable Alzheimer disease (n ⫽ 14), mild
cognitive impairment (n ⫽ 2), frontotemporal lobe degeneration (n ⫽ 3),
subjective memory complaints (n ⫽ 6),
177
NEURORADIOLOGY: Diagnostic Imaging of Patients in a Memory Clinic
Table 1
Table 2
Interobserver Agreement of All
Readers Regarding MR and CT
Examinations
Finding
GCA
MTA
Left side
Right side
WMC
Wattjes et al
Intermodality Agreement between CT and MR Imaging
Finding
MR Imaging
CT
0.84
0.79
0.82
0.83
0.92
0.84
0.88
0.89
GCA
MTA
Left side
Right side
WMC
Reader 1
Reader 2
Reader 3
Mean
0.91
0.80
0.79
0.83
0.91
0.92
0.79
0.94
0.86
0.78
0.80
0.80
0.81
0.88
0.86
0.79
Note.—Data are weighted ␬ values.
Note.—Data are Kendall W values of the three readers.
Table 3
temporal lobe epilepsy (n ⫽ 1), and dementia with Lewy bodies (n ⫽ 1). In
three patients, a final diagnosis could not be
established. An example of the contribution
of MR imaging and CT to the clinical diagnosis is shown in Figure 2.
Incidental Imaging Findings
Incidental imaging findings were observed in four patients and included an
arachnoid cyst, a meningioma, end-stage
atrophy of the cerebellum, and infarction
of the cerebellum. All three readers recognized all incidental findings on MR and
CT images (Fig 3). Furthermore, in three
patients, the atrophy pattern of frontotemporal lobe degeneration could be seen
equally well on CT and MR images.
Interobserver Agreement
The results of the interobserver agreement of all three readers are summarized
in Table 1. Overall, the interobserver
agreement was high within each modality
regarding all variables. Almost identical
values of interobserver agreement could
be observed for MR imaging and CT.
GCA Scores
The mean GCA score for all readers was
0.9 for both CT and MR imaging (range,
0 –3). Different GCA scores for CT and
MR imaging were obtained in three patients by reader 1 and in eight patients
each by readers 2 and 3, leading to
weighted ␬ values for agreement between
MR imaging and CT of 0.91, 0.80, and
0.79 for observers 1, 2, and 3, respectively (Table 2). The detailed analysis of
discordant results between MR imaging
and CT concerning GCA is given in Table 3.
Overall, there was no systematic trend of
178
Discordant Results Concerning GCA Assessment
GCA Score
MR imaging vs CT
1 vs 0
2 vs 1
3 vs 2
CT vs MR imaging
0 vs 1
1 vs 2
2 vs 3
Reader 1 (n ⫽ 3)
Reader 2 (n ⫽ 8)
Reader 3 (n ⫽ 8)
1
0
1
0
2
1
1
0
2
1
0
1
6
3
2
1
8
4
4
0
0
0
0
0
Note.—Data are numbers of patients.
higher scores being obtained with CT or
MR imaging. Higher scores at CT or MR
imaging were significant for one of the three
readers (reader 1, P ⬎ .99 and power ⫽
0.12; reader 2, P ⫽ .289 and power ⫽ 0.35;
reader 3, P ⫽ .008 and power ⫽ 0.9). Examples of corresponding CT and MR images in terms of different stages of GCA are
presented in Figure 4.
MTA Scores
The mean MTA scores were 1.1 (range,
0 – 4) for both sides of the brain on CT
images and 1.2 (range, 0 – 4) for the left
side of the brain and 1.1 (range, 0 – 4)
for the right side of the brain on MR
images.
On the left side of the brain, discordant CT and MR imaging scores were
obtained in six patients by reader 1, five
patients by reader 2, and 13 patients by
reader 3. On the right side of the brain,
discordant CT and MR imaging scores
were observed in five patients by reader
1, eight patients by reader 2, and 13 patients by reader 3. There was no substantial trend for any reader in terms of
higher MTA sores being seen solely on CT
or MR images. The detailed results of
comparison of MTA scores on CT and
MR images are given in Table 4. Agreement between MR imaging and CT with
weighted ␬ statistics ranged from substantial to excellent on the left (mean ␬
value, 0.88; range, 0.80 – 0.94) and right
(mean ␬ value, 0.86; range, 0.80 – 0.92)
side of the brain (Table 2). No systematic
trend of obtaining higher or lower scores
with MR imaging or CT could be observed
(reader 1: P ⬎ .99 for both sides; power,
0.05 for the left side and 0.06 for the
right; reader 2: P ⫽ .375 for the left side
and P ⫽ .727 for the right; power, 0.18
for the left side and 0.09 for the right;
reader 3: P ⫽ .076 for the left side and
P ⫽ .58 for the right; power, 0.43 for the
left side and 0.12 for the right). Figure 5
shows examples of different MTA scores
on CT images and the corresponding MR
images.
WMC Scores
Mean WMC scores were 0.8 for CT and
1.0 for MR imaging (range, 0 –3). The
agreement between modalities ranged
from 0.78 to 0.81 (mean, 0.79) (Table
radiology.rsna.org ▪ Radiology: Volume 253: Number 1—October 2009
NEURORADIOLOGY: Diagnostic Imaging of Patients in a Memory Clinic
2). Discordant results between CT and
MR imaging were observed in nine patients by reader 1, 10 patients by reader
2, and eight patients by reader 3. Among
the patients with discordant results,
scores were higher on MR images than
on CT images in all nine (100%) patients for reviewer 1, eight (80%) of 10
patients for reviewer 2, and five (62%)
of eight patients for reviewer 3. The
number of higher WMC scores obtained
with MR imaging compared with that obtained with CT was significantly higher for
reader 1 (P ⫽ .004; power, 0.87) but not
for readers 2 (P ⫽ .092; power, 0.33) or 3
(P ⫽ .563; power, 0.13). The majority of
the patients with discrepant scores received a score of 1 (punctuate WMC) at
MR imaging and a score of 0 (no WMC)
at CT. A detailed analysis of the discor-
Wattjes et al
dant results regarding the WMC score is
given in Table 5. Figure 6 gives an overview of different stages of white matter
lesions on MR images and the corresponding CT images.
Discussion
Despite the availability of time-consuming
quantitative MR methods that can be
Figure 4
Figure 4: Different degrees of GCA on transverse (a, b, c) MR and (d, e, f) CT images obtained in the same patient. Note grade 0 GCA in a and d, grade 1 GCA in b and
e, and grade 2 GCA in c and f. Different degrees of GCA could be easily classified on CT images, with a high level of agreement when compared with the corresponding
MR images.
Radiology: Volume 253: Number 1—October 2009 ▪ radiology.rsna.org
179
NEURORADIOLOGY: Diagnostic Imaging of Patients in a Memory Clinic
Wattjes et al
Figure 5
Figure 5: Different degrees of MTA on oblique coronal, A–D, T1-weighted MR images and, E–H, corresponding CT images. All images were obtained in the same
patient. Examples of grade 0 (A, E), 1 (B, F), 2 (C, G), 3 (right side of D and H), and 4 (left side of D and H) MTA. Differences in repositioning might lead to imaging of different aspects of the medial temporal lobe, which can lead to different MTA assessments (right side of C and G).
used to precisely assess focal and global
atrophy in patients with dementia, these
techniques are more applicable for research purposes than for clinical use (20).
Thus, visual rating scales are still the most
frequently used method in the routine
memory clinic setting, and they enable
fast and accurate assessment of atrophy
with a good correlation to volumetric
measurements. Because of the increasing
number of patients with memory complaints suggestive of dementia, there are
a considerable number of patients who
cannot undergo MR imaging for any number of reasons (eg, limited imager availability, patients have a pacemaker, etc).
Thus, multidetector CT might be an alternative imaging modality that can be used
to examine these patients. However, atrophy assessment with use of visual rating
scales is based on MR findings, and it is
important to determine whether these rating scales might also be applicable to CT.
In contrast to the findings of previous
studies in which single-section CT techniques were used, we found substantial
agreement of the GCA and MTA assessments on CT images when compared
with those on MR images. Overall, excellent agreement was found for GCA, expressed by a mean weighted ␬ value of
180
Table 4
Discordant Results Concerning MTA Assessment
MTA Score
Left side of the brain
MR imaging vs CT
1 vs 0
2 vs 0
2 vs 1
3 vs 2
CT vs MR imaging
1 vs 0
2 vs 1
3 vs 2
Right side of the brain
MR imaging vs CT
1 vs 0
2 vs 1
3 vs 1
CT vs MR imaging
1 vs 0
2 vs 1
3 vs 2
Reader 1
Reader 2
Reader 3
6
3
2
0
1
0
3
2
1
0
5
2
1
1
0
3
2
1
0
5
1
0
0
1
0
4
2
2
0
8
3
3
0
0
5
2
2
1
13
10
6
1
2
1
3
0
1
2
13
8
6
1
1
5
1
3
1
Note.—Data are numbers of patients.
0.83. For visual rating of MTA, agreement between CT and MR images from
the descriptive point of view seemed to be
lower and more variable when compared
with that for GCA. However, the inter-
modality agreement expressed by the
weighted ␬ value is substantial for a junior
reader and excellent for two senior readers. These apparently contradictory results between the descriptive statistics
radiology.rsna.org ▪ Radiology: Volume 253: Number 1—October 2009
NEURORADIOLOGY: Diagnostic Imaging of Patients in a Memory Clinic
and the weighted ␬ value are due to the
fact that we are dealing with a fivepoint scale for MTA and a four-point
scale for GCA. The weighted ␬ values
of all three readers lead us to believe
that intermodality agreement concerning the MTA rating seems to be
related to the experience of the
Wattjes et al
reader. Both readers with at least 7
years of experience showed excellent
agreement between CT and MR imaging, with ␬ values greater than 0.86,
whereas agreement was slightly lower
(but still substantial) for the lessexperienced reviewer. Thus, intermodality agreement can be considered
Table 5
Discordant Results Concerning WMC Assessment
WMC Score
MR imaging vs CT
1 vs 0
2 vs 1
2 vs 0
CT vs MR imaging
0 vs 1
2 vs 3
Reader 1 (n ⫽ 9)
Reader 2 (n ⫽ 10)
Reader 3 (n ⫽ 8)
9
7
2
0
0
0
0
8
7
0
1
2
2
0
5
3
1
1
3
1
2
Note.—Data are numbers of patients.
Figure 6
Figure 6: Different stages of WMC assessed on transverse, A–C, fluid-attenuated inversion-recovery MR
and, D–F, CT images in the same patients. WMC scores ranged from 1 (A and D) to 3 (C and F). MR imaging
shows higher sensitivity in the detection of subtle WMCs, especially small punctuate lesions (arrow in A).
However, some of those focal lesions can also be seen on CT images (arrow in D). Beginning confluent white
matter lesions (arrowhead in B and E) indicating a grade 2 WMC can almost equally be assessed on MR and
CT images. Clinically relevant confluent WMCs indicating a grade 3 WMC (C and F) can be equally detected
on MR and CT images. Also note the different stage of GCA.
Radiology: Volume 253: Number 1—October 2009 ▪ radiology.rsna.org
robust, even across readers with different degrees of experience.
Given the better sensitivity of MR
imaging in the detection of WMCs when
compared with that of single-section
CT, we expected to find similar results
for multi– detector row CT, thus leading
to rather moderate intermodality agreement between CT and MR imaging (21).
In fact, the intermodality agreement for
WMC reached almost excellent values,
with a mean weighted ␬ value of 0.8.
Almost all discordant results between
modalities were caused by higher scores
on MR images reflecting the higher sensitivity of MR imaging regarding WMC,
as described in a direct comparison between MR imaging and single-section
CT (21). Most of the discordant ratings
between CT and MR imaging dealt with
lower clinically irrelevant scores of 0
and 1. Agreement was actually good for
the higher clinically relevant scores of 2
and 3. None of our patients fulfilled the
criteria for vascular dementia; however, in clinical practice, it is unlikely
that the diagnosis of vascular dementia
based on a large amount of WMCs (involving more than 25% of the white
matter) according to the National Institute
of Neurological Disorders and Stroke
(NINDS) and Association Internationale
pour la Recherché et l’Enseignement en
Neurosciences (AIREN) criteria might be
missed on CT images when compared with
MR images because almost all of the discordant results concern patients with lower
WMC scores that are not relevant for the
diagnosis (22,23).
Similar to quantitative methods used
in the volumetric measurement of atrophy, volumetric assessment of WMCs
might enable more precise measurement
when compared with that achieved with
simple rating scales. However, these
time-consuming volumetric methods are
not applicable in the routine clinical setting. The modified Fazekas rating scale
used in this study is widely accepted and
yields good global assessment of WMCs.
It has been suggested in a large overview
of 26 rating scales for evaluation of WMC
on MR images that the simplicity of the
Fazekas scale might make it robust, even
for images of poorer quality (7). Thus,
this rating scale is probably the most ap181
NEURORADIOLOGY: Diagnostic Imaging of Patients in a Memory Clinic
propriate scale with which to evaluate
WMCs on CT images. Moreover, this rating scale has been successfully validated
with histopathologic analysis (24,25). It
has also been shown that simple rating
scales (such as the Fazekas score) are
comparable with complex measures of
WMC in terms of associations with clinical outcome measures (26,27).
The results of several studies showed
high interreader agreement of GCA,
MTA, and WMC on MR images; this finding was confirmed by the results of our
study (10–12). We extend those findings
and report similarly high values of interreader agreement for CT, even among
readers with different degrees of experience.
In our study, we observed a limited
number of incidental findings that were
equally detectable with CT and MR imaging. We have to assume that MR imaging will be superior to CT when it
comes to the specific information from
additional sequences for several specific
disease entities that can cause dementia, such as prion-linked dementias
(Creutzfeldt-Jacob disease), viral infections (herpes simplex encephalitis, human immunodeficiency virus), and certain disorders that affect the white matter (28,29). However, these entities
represent only a minority of the diseases that can cause dementia.
Although not formally tested, it is unlikely that any clinically relevant finding in
our sample would have been missed if the
imaging part of the diagnostic work-up
had been based only on CT. In addition to
comparable results for clinically relevant
differences between GCA, MTA, and
WMC scores, other findings that contributed to a diagnosis were recognized on
CT and MR images. Examples are cerebellar atrophy or disproportional atrophy
patterns of the frontal and temporal
lobes, such as those observed in patients
with frontotemporal lobe degeneration,
which are easy to detect on multidetector
CT images with multiplanar reconstructions (Fig 2).
A limitation of our study was the relatively small sample of patients. The clinical relevance of our results in terms of
the effect on the clinical diagnosis in patients was not formally tested and could
182
Wattjes et al
not be adequately tested because of the
small sample size. Multicenter studies
that include a larger number of patients
are needed to investigate whether there
might be differences in the diagnoses in patients in a memory clinic setting that are
dependent on the imaging modality used.
It also must be stressed that these
results probably are not simply due to the
fact that we used a 64 – detector row clinical CT scanner. In other words, the number of detector rows does not necessarily
correlate with the degree of intermodality
agreement. Given thin enough sections,
similar results could also be expected
with use of spiral CT scanners operating
with fewer detector rows.
In conclusion, although MR imaging
should be the preferred imaging modality due to its lack of ionization and
higher contrast resolution, 64 – detector
row CT is a suitable and accurate imaging method with which to evaluate GCA,
MTA, and WMCs in a memory clinic
setting. It can be considered a nearly
equivalent alternative to MR imaging in
patients who cannot undergo an MR examination.
Acknowledgments: We thank all participants
who agreed to take part in this study. We also
thank Patrick Schenkers for the excellent technical assistance, Dr Dirk L. Knol for statistical
advice, and Rolinka Romkes and Freek Gillissen
for their organizational support.
References
for research studies—report of the NINDSAIREN International Workshop. Neurology
1993;43:250 –260.
6. Scheltens P, Erkinjunti T, Leys D, et al.
White matter changes on CT and MRI: an
overview of visual rating scales. Eur Neurol
1998;39:80 – 89.
7. Fazekas F, Chawluk JB, Alavi A, Hurtig HI,
Zimmermann RA. MR signal abnormalities
at 1.5T in Alzheimer’s dementia and normal
ageing. AJR Am J Roentgenol 1987;149:
351–356.
8. Scheltens P, Leys D, Barkhof F, et al. Atrophy of medial temporal lobe on MRI in “probable” Alzheimer’s disease and normal ageing:
diagnostic value and neuropsychological correlates. J Neurol Neurosurg Psychiatry 1992;
55:967–972.
9. Jack CR Jr, Petersen RC, Xu YC, et al. Medial temporal lobe atrophy on MRI in normal
ageing and very mild Alzheimer’s disease.
Neurology 1997;49:786 –794.
10. Scheltens P, Pasquier F, Weerts JG, Barkhof
F, Leys D. Qualitative assessment of cerebral
atrophy on MRI: inter- and intra-observer
reproducibility in dementia and normal ageing. Eur Neurol 1997;37:95–99.
11. Pasquier F, Leys D, Weerts JG, MounierVehier F, Barkhof F, Scheltens P. Inter- and
intraobserver reproducibility of cerebral atrophy assessment on MRI scan with hemispheric infarcts. Eur Neurol 1996;36:268 –
272.
12. Scheltens P, Launer LJ, Barkhof F,
Weinstein HC, van Gool WA. Visual assessment of medial temporal lobe atrophy on
magnetic resonance imaging: interobserver
reliability. J Neurol 1995;242:557–560.
1. Knopman DS, DeKosky ST, Cummings JL,
et al. Practice parameter: diagnosis of dementia (an evidence-based review) report of
the quality standards subcommittee of the
American Academy of Neurology. Neurology
2001;56:1143–1153.
13. Wahlund LO, Julin P, Johansson SE,
Scheltens P. Visual rating and volumetry of
the medial temporal lobe on magnetic resonance imaging in dementia: a comparative
study. J Neurol Neurosurg Psychiatry 2000;
69:630 – 635.
2. Scheltens P, Fox N, Barkhof F, De Carli C.
Structural magnetic resonance imaging in
the practical assessment of dementia. Lancet
Neurol 2002;1:13–21.
14. De Leon MJ, George AE, Golomb J, et al.
Frequency of hippocampal formation atrophy in normal aging and Alzheimer disease.
Neurobiol Aging 1997;18:1–11.
3. Waldemar G, Dubois B, Emre M, et al. Recommendations for the diagnosis and management of Alzheimer’s disease and other
disorders associated with dementia: EFNS
guideline. Eur J Neurol 2007;14:e1– e26.
15. Frisoni GB. Structural imaging in the clinical
diagnosis of Alzheimer’s disease: problems
and tools. J Neurol Neurosurg Psychiatry
2001;70:711–718.
4. Dubois B, Feldman HH, Jacova C, et al. Research criteria for the diagnosis of Alzheimer’s
disease: revising the NINCDS-ADRDA. Lancet
Neurol 2007;6:734–746.
5. Roman GC, Tatemichi TK, Erkinjuntti T,
et al. Vascular dementia: diagnostic criteria
16. Frisoni GB, Geroldi C, Beltramello A, et al.
Radial width of the temporal horn: a sensitive measure in Alzheimer’s disease. AJNR
Am J Neuroradiol 2002;23:35– 47.
17. O’Brien JT, Metcalfe S, Swann A, et al. Medial temporal lobe width on CT scanning in
Alzheimer’s disease: comparison with vascu-
radiology.rsna.org ▪ Radiology: Volume 253: Number 1—October 2009
NEURORADIOLOGY: Diagnostic Imaging of Patients in a Memory Clinic
lar white matter lesions predict symptomatic
cerebrovascular disease in probable Alzheimer’s disease. Arch Neurol 1995;52:659 –
664.
lar dementia, depression and dementia with
Lewy bodies. Dement Geriatr Cogn Disord
2000;11:114 –118.
18. Landis JR, Koch GG. The measurement of
observer agreement for categorical data. Biometrics 1977;33:159 –174.
19. Faul F, Erdfelder E, Lang AG, Buchner A.
G*Power3: a flexible statistical power analysis program for the social, behavioral, and
biomedical sciences. Behav Res Methods
2007;39:175–191.
20. Karas G, Scheltens P, Rombouts S, et al.
Precuneus atrophy in early-onset Alzheimer’s disease: a morphometric structural
MRI study. Neuroradiology 2007;49:967–
976.
21. Lopez OL, Becker JT, Jungreis CA, et al.
Computed tomography— but not magnetic
resonance imaging—identified periventricu-
22. Van Straaten EC, Fazekas F, Rostrup E,
et al. Impact of white matter hyperintensities scoring method on correlations with
clinical data: the LADIS study. Stroke
2006;37:836 – 840.
23. Guermazi A, Miaux Y, Rovira-Cañellas A,
et al. Neuroradiological findings in vascular
dementia. Neuroradiology 2007;49:1–22.
24. Fazekas F, Kleinert R, Offenbacher H, et al.
The morphologic correlate of incidental
punctate white matter hyperintensities on
MR images. AJNR Am J Neuroradiol 1991;
12:915–921.
25. Fazekas F, Kleinert R, Offenbacher H, et al.
Pathologic correlates of incidental MRI white
Radiology: Volume 253: Number 1—October 2009 ▪ radiology.rsna.org
Wattjes et al
matter signal hyperintensities. Neurology
1993;43:1683–1689.
26. Gouw AA, van der Flier WM, van Straaten
EC, et al. Simple versus complex assessment
of white matter hyperintensities in relation
to physical performance and cognition: the
LADIS study. J Neurol 2006;253:1189 –
1196.
27. Inzitari D, Simoni M, Pracucci G, et al. Risk
of rapid global functional decline in elderly
patients with severe cerebral age-related
white matter changes: the LADIS study.
Arch Intern Med 2007;167:81– 88.
28. Tschampa HJ, Zerr I, Urbach H. Radiological
assessment of Creutzfeldt-Jacob disease.
Eur Radiol 2007;17:1200 –1211.
29. Achim CL, Wiley CA. Virus mediated dementias. In: Markesbery WR, ed. Neuropathology of dementing disorders. London,
England: Edward Arnold, 1998; 312–339.
183