Content uploaded by Horst Müller
Author content
All content in this area was uploaded by Horst Müller on Jul 07, 2015
Content may be subject to copyright.
Abstract The primary objective of this study was to
test hypotheses about the relationship between HMPAO-
SPECT findings and probable Alzheimer’s disease (DAT)
in a relatively large sample of patients diagnosed accord-
ing to DSM-III-R. SPECT patterns of 20 controls and
116 DAT patients were investigated. Left and right fron
tal,
temporal, parietal and occipital regions of the brain were
rated as showing a hypoperfusion or not. A wide variety
of patterns were found and these are described in detail
below. In DAT patients, temporal and/or parietal regions
were affected significantly more often (88%, p> 0.001)
than frontal and/or occipital regions (70%). A bilateral
temporoparietal pattern, which has been repeatedly re-
ported as typical for DAT, was observed in 48% of DAT
patients, but also in 25% of controls, and did not differen-
tiate significantly between these two groups (p> 0.05).
Conversely, more than three regions with hypoperfusion
were observed significantly more often in DAT patients
(48%, p< 0.01) than in controls (10%). In DAT patients,
the number of regions with hypoperfusion correlated sig-
nificantly with the score of the Mini Mental State Exami-
nation (MMSE, r= 0.33, p< 0.001). The frequency of at
least one hypoperfusion was approximately equal in left
and right hemispheres (77% vs. 73%, p= 0.2). The hy-
pothesis that cognitive decline in DAT starts in the tempo-
ral regions was tested in 14 SPECT patterns showing only
one region with hypoperfusion. In 12 of these patterns, a
temporal region was in fact affected (p< 0.001). Whereas
hypoperfusion in frontal areas was not accompanied by a
significantly lower MMSE than when only temporoparietal
regions were affected, MMSE scores were significantly
lower when occipital regions were affected in addition to
temporoparietal regions (p< 0.05). The clinical use of
SPECT findings was tested in discriminating analyses with
the MMSE and a delayed recall test as additional predictors
of DAT. Whereas the MMSE and the delayed recall test
differentiated significantly between DAT patients and con-
trols, SPECT findings yielded no further differentiation.
In conclusion, the theoretical and clinical implications of
SPECT findings and their relationships to other physio-
logi
cal
and psychological variables deserve further inves-
tigation.
Key words HMPAO-SPECT · Alzheimer’s disease ·
Hypoperfusion
Introduction
Probable dementia of the Alzheimer type (DAT) is asso-
ciated with changes in regional cerebral blood flow that
can be detected using single photon emission computer
tomography (SPECT) (Gemmel et al. 1987, Burns et al.
1989, Geany et al. 1990, Montaldi et al. 1990, Habert et
al. 1991). The value of SPECT findings in the diagnosis
of DAT is still a matter of debate. For instance, Poremba
(1993) argues that there are no DAT-specific biological
markers to date, whereas Volles (1991) notes that functional
imaging such as PET and SPECT may identify DAT even
in early stages of the disease when the CCT is still nor-
mal. The different views may be explained by the hetero-
geneity of findings which is itself probably due to differ-
ences in methods and samples. The samples especially vary
in the severity of dementia, and the sizes of the groups in-
vestigated often seem rather low. The finding probably
noted most often in DAT patients is bilateral temporopari-
etal hypoperfusion. This finding is confirmed by some
early PET studies (Frackowiak et al. 1981, McGeer et al.
1986). However, unilateral hypometabolism may also oc-
cur, especially in early DAT (Salmon and Franck 1989)
H. Müller · H.-J. Möller · A. Stippel · M. Fric ·
F. Grünwald · G. Laux · E. Klemm · H.-J. Biersack
SPECT patterns in probable Alzheimer’s disease
Eur Arch Psychiatry Clin Neurosci (1999) 249:190–196 © Steinkopff Verlag 1999
Received: 3 December 1998 / Accepted: 1 April 1999
ORIGINAL PAPER
H. Müller · A. Stippel
Psychiatric Department, Bonn University, Germany
M. Fric · G. Laux
Gabersee Regional Hospital, Wasserburg/Inn, Germany
F. Grünwald · E. Klemm · H.-J. Biersack
Department of Nuclear Medicine, Bonn University, Germany
Prof. H.-J. Möller (쾷)
Psychiatric Hospital, Nussbaumstr. 7, D-80336 Munich, Germany
Tel.: +49-89-5160-5501, Fax: +49-89-5160-5522
and bilateral temporoparietal hypoperfusion is also found
in Parkinson’s disease with dementia (Kuhl et al. 1984,
Holman et al. 1992, Salmon et al. 1993), vascular demen-
tia (Kuwabara et al. 1990), and Creutzfeld-Jakob disease
(Friedland et al. 1984). Recent studies show that bilateral
temporoparietal hypoperfusion (with or without additional
defects) characterises approximately 50–65% of DAT pa-
tients (Holman et al. 1992, Horn et al. 1995). Thus, the
sensitivity of this hypoperfusion pattern seems low, but
the specificity seems high, if Parkinson’s disease with de-
mentia can be excluded. Moreover, it seems unlikely that
a certain pattern of hypoperfusion is characteristic for all
stages of DAT.
Eberling et al. (1992) suggested – as many authors have
done previously – that the temporal lobes are the first neo-
cortical regions affected in DAT. Eberling’s group ob-
served temporal changes in mild and moderate DAT,
whereas decline was only seen in other regions in moder-
ate DAT. Wolfe et al. (1995) reported that the temporal lobe
regional cerebral blood flow ratio – but not parietal or
frontal ratios – predicts cognitive decline over a longitu-
dinal follow-up. In criticism, it must be mentioned that
Wolfe et al. (1995) used the ratio of region of interest to
occipital cortex, and assumed the occipital cortex to be
“relatively unaffected” by DAT. This assumption seems
questionable, at least at later stages of DAT. Nevertheless,
current evidence suggests that the role of the temporal
lobes in the first stages of DAT deserves further investiga-
tion.
Besides the inherent methodological difficulties already
mentioned, the evaluation of previous findings is compli-
cated by the fact that most studies describe the observed
perfusion patterns incompletely. The purpose of the cur-
rent study is to evaluate hypoperfusion patterns in a rela-
tively large sample of DAT patients, to describe them ex-
haustively and to test them in comparison to SPECT pat-
terns of a control group of patients diagnosed as “normal”.
To our knowledge, this is the first study with more than
100 DAT patients. A detailed evaluation of hypoperfusion
patterns that show only one region with hypoperfusion
(“isolated patterns”) describes the beginning of cortical
deterioration in DAT. A predominant affection of tempo-
ral lobes in these isolated patterns was expected. A critical
evaluation of the clinical value of SPECT patterns in the
diagnosis of DAT leads to the question whether SPECT
findings supply additional information compared to other
well-established diagnostic techniques. We, therefore, per-
formed discriminating analyses that tested whether SPECT
findings enhance discrimination between controls and DAT
patients beyond the level reached by evaluation of the
Mini Mental State Examination (MMSE, Folstein et al.
1975) and two delayed verbal memory test. Delayed mem-
ory tests are known to be sensitive indicators of even mild
DAT (Welsh et al. 1992).
Method
Diagnosis
First of all, thorough psychiatric and neurological examinations were
conducted. Furthermore, the MMSE (Folstein et al. 1975), Hachin-
ski Scale (Hachinski et al. 1975) and an extensive neuropsycho-
logical test battery were performed. DSM-III-R criteria (American
Psychiatric Association 1987) were employed for diagnosis. Cause
s
of dementia other than DAT were excluded by laboratory-chemical
analyses, CCT/MRT, EEG, and extracranial and transcranial Dop-
pler sonography. Patients with a Hachinski score greater than 4 or
more than two lacunal infarctions in CCT or MRT were excluded
under a probable diagnosis of a vascular or a mixed-type dementia.
HMPAO-SPECT
Interictal brain SPECT imaging was performed 30 to 120 min after
injection of 555-740 MBq
99m
Tc-HMPAO. Four different types
of SPECT acquisition systems were used: three different rotating
gamma cameras: APEX 409 (Elscint, single head, A); DYNA-SCAN
(Picker, dual-head, B); Genesys (ADAC, single-head, C). With an
acquisition time of 20–30 min, 60 (A) or 64 (B and C) projections
were acquired with a 64*64 matrix. A ring system has been used
since 1992: CERASPECT system (Digital Scintigraphic Inc., D).
The system works with three collimators rotating within a single
crystal system iodide ring detector and acquires three views from
three angles simultaneously. The acquisition time was 30 min. One
hundred and twenty projections were acquired uisng a 512*64 ma-
trix. Coronal, sagittal and transaxial (parallel to the orbitomeatel
line) slices were calculated from the original transaxial slices and
summed in order to obtain two (A/B/C) or four (D) pixel thick
slices. Additionally, “thin” slices (one pixel thick, parallel to the
long axis of the temporal lobes) were calculated for evaluation of
the temporal lobes. Two experienced nuclear medicine physicians
who were blind to diagnosis and neuropsychological test results
rated the findings. Eight regions of interest were defined: left and
right frontal, temporal, parietal and occipital lobes, and it was eval-
uated whether these regions showed a hypoperfusion or not.
Assessment of cognitive performance
Although many neuropsychological tests were performed in this
study, this article is restricted to the evaluation of the MMSE and
two delayed word tests. Delayed word recall has been proven to be
extremely sensitive in the detection of mild cases of DAT (Welsh
et al. 1992). Two tests of delayed word recall of varying difficulty
were employed in our sample. Thirty-five patients who seemed less
handicapped at an initial interview heard and recalled a 15-item
word list five times. A second 15-item list was then given and re-
called. Afterwards, patients were asked to recall the first word list
without additional presentation. After a 30 min interval, the first
word list had to be recalled a seventh time and the number of cor-
rectly reproduced words was scored (delayed word recall of 15 items,
DWR15). Fifty-nine patients who seemed more handicapped un-
derwent a similar procedure, but with an eight-item word list. This
list was presented and recalled five times and then had to be re-
called after a 30 min interval (DWR8). Although it is a method-
ological weakness of this study that not both delayed recall tests
(or at least the same one) were employed for all patients, the use of
different lists and methods seemed clinically useful at the time of
testing in order to adjust the testing procedure according to the
severity of impairment.
Hypotheses
Review of the literature led to formulation of the following hypothe-
ses: 1. A bilateral temporoparietal pattern of hypoperfusion should
191
192
significantly differentiate between DAT patients and controls. 2. In
DAT patients, temporal and parietal cortex should be affected more
frequently than frontal and occipital cortex, and the frequency of
affection of left and right hemispheres should be approximately
equal.
3. In DAT patients, more regions of the brain should show
hypoperfusion than in controls. 4. If SPECT shows only one re-
gion with hypoperfusion, then this region is expected to be tempo-
ral since it is assumed that this is the first cortical area affected. 5.
If SPECT is clinically useful, then the discrimination between DAT
patients and controls should improve after inclusion of SPECT find-
ings in discriminating analyses with MMSE, DWR8 and DWR15
as predictors of diagnosis.
Sample
The sample was recruited in the memory clinic of the psychiatric
department of Bonn university hospital. One hundred and thirty-
six patients had undergone a HMPAO-SPECT and were diagnosed
according to the criteria of DSM-III-R as suffering from DAT
(116) or not suffering from dementia (20) (Table 1). The controls
of the latter group, however, reported at least some cognitive com-
plaints – typically memory problems – which were the reason for
their examination at the memory clinic.
Statistics
First of all, a complete description of the observed hypoperfusion
patterns in the 116 DAT patients and the 20 controls is given. Thus,
the data may be used to test hypotheses other than the ones tested
here or may be combined with data sets from other studies. Frequen-
cies of patterns were tested with the binomial test. Associations be-
tween patterns, and between patterns and diagnoses, were tested
with the Pearson chi
2
-test and sometimes described with phi val-
ues. MMSE scores in different groups were tested with the t-test
for independent samples and the predictive value of SPECT find-
ings was evaluated by discriminating analyses using MMSE and
delayed free recall as additional predictors. All statistical analyses
were performed with SAS 6.03 for personal computers.
Results
The following SPECT patterns were observed in the
116 DAT patients (Table 2). There were 16 patterns which
were significantly more frequent than would be expected
by chance. The multiple use of the binomial test in 50 pat-
terns infers the risk that some of the “significantly” frequent
patterns are significant by chance. Therefore, we would
only like to comment on the patterns with a frequency of
at least five before we draw some general conclusions. In
8.6% of the DAT patients, no hypoperfusion was detected.
In 6.9%, only the left temporal lobe seemed to be affected.
In 6.0%, all regions except the frontal ones showed hy-
poperfusion. In 4.3%, only the parietal and the occipital
Table 1 Sample description
DAT Control
N 116 20
Age (mean ± std. dev.) 66 ± 9 56 ± 10
Sex (female/male) 60/56 12/8
MMSE (mean ± std. dev.) 19.9 ± 5.4 28.8 ± 1.2
Table 2 Observed frequencies of SPECT patterns in 116 DAT pa-
tients
Pattern
a
Frequency Relative fre- p< 0.01
b
quency (%)
00000000 10 8.6 *
00001000 1 0.9
00001010 3 2.6 *
00001111 5 4.3 *
00010000 4 3.4 *
00010001 2 1.7
00010100 2 1.7
00011010 1 0.9
00011100 1 0.9
00011111 1 0.9
00100000 8 6.9 *
00100010 1 0.9
00101010 1 0.9
00101100 1 0.9
00101111 5 4.3 *
00110000 3 2.6 *
00110011 1 0.9
00110101 1 0.9
00111011 1 0.9
00111100 5 4.3 *
00111111 7 6.0 *
01000100 3 2.6 *
01010000 2 1.7
01010100 3 2.6 *
01010101 1 0.9
01011111 3 2.6 *
01100000 1 0.9
01110000 1 0.9
01110101 1 0.9
01111000 1 0.9
01111100 2 1.7
10000000 1 0.9
10100000 1 0.9
10100100 1 0.9
10100101 1 0.9
10101010 5 4.3 *
10101111 1 0.9
10110000 1 0.9
10111100 1 0.9
11000000 3 2.6 *
11001000 2 1.7
11001100 1 0.9
11001111 3 2.6 *
11010000 2 1.7
11011101 1 0.9
11100000 1 0.9
11110000 2 1.7
11111000 1 0.9
11111100 5 4.3 *
11111111 1 0.9
a
0 denotes normal perfusion, 1 hypoperfusion; the strings are in
the sequence left frontal lobe – right frontal lobe – left temporal
lobe – right temporal lobe – left parietal lobe – right parietal lobe –
left occipital lobe – right occipital lobe
b
binomial test with equal prior probability of each pattern =
1/256
regions were affected. In 4.3%, a similar pattern was ob-
served: all regions except the frontal ones and the right
temporal lobe seemed to be affected. In a further 4.3% of
the DAT patients, only the temporal and parietal regions
were affected. In a further 4.3%, all frontal, temporal and
parietal regions showed hypoperfusion, whereas the oc-
cipital regions were spared. Thus, the predominant impres-
sion from Table 2 is that there is a wide variety of SPECT
patterns in DAT. Of course, the patterns result from dif-
ferent stages of the disease and are not error free. How-
ever, it seems unlikely that even for different stages of
DAT “the” typical pattern arises. The frequency of invol
ve-
ment of left and right hemispheres was approximately
equal (77% and 73% respectively, p= 0.2, binomial test).
Table 3 shows the patterns observed in the 20 con-
trols.
Of course, the relative frequencies given in Table 3 are
rough estimates, with N being only 20. For this reason, we
do not want to comment on the patterns in detail but would
just like to mention that only 25% of the controls showed
a SPECT pattern without any hypoperfusion.
Is bilateral temporoparietal hypoperfusion DAT-sensitive
and -specific?
One problem with the term “bilateral temporoparietal hy-
poperfusion” is that it is not clear exactly which patterns
are being described with this expression. In fact, different
authors using this term assume different patterns and it is
difficult, if not impossible, to decide which definition is
the most appropriate one. We, therefore, used different de-
finitions ranging from extremely narrow to extremely broad
and report the observed frequencies for DAT patients and
controls in Table 4.
Table 4 shows that the extreme definitions are obviously
not appropriate. Although we investigated only 20 con-
trols – and, therefore, the chi
2
-test for significant associa-
tion may seem somewhat hard – this result shows that the
association between bilateral temporoparietal hypoperfu-
sion and DAT is at least not a very close one.
Another implication of the hypothesis of bilateral tem-
poroparietal hypoperfusion in DAT patients is that tempo-
ral and parietal regions of the brain should be affected more
often than frontal and occipital regions. In fact, 88% of
our DAT patients showed at least one temporal or parietal
hypoperfusion, whereas 70% showed frontal and/or occipi-
tal hypoperfusions. A binomial test showed that the fre-
quency of temporal and/or parietal affection was signifi-
cantly higher (p< 0.001) than the 70% observed in frontal
and/or occipital regions. Nevertheless, the observed dif-
ference of only 18% again suggests that at least in later
stages of DAT hypoperfusions are often not restricted to
the temporoparietal regions.
If DAT cannot be characterised easily by a certain pat-
tern of hypoperfusion, it could theoretically be assumed that
the amount of hypoperfusion is perhaps decisive. Since
only 10% of the controls showed hypoperfusions in more
than three regions, this cut-off point was also used in DAT
patients. 48% of the DAT patients had SPECT patterns
with hypoperfusions in more than three regions and this
criterion proved to be significantly associated with diag-
nosis (p< 0.01, chi
2
-test). Fig.1 shows the mean MMSE
for the number of regions with hypoperfusion in DAT pa-
tients.
193
Table 3 Observed frequencies of SPECT patterns in 20 controls
Pattern
a
Frequency Relative frequency (%)
00000000 5 25
00000100 2 10
00010000 1 5
00010100 2 10
00011100 1 5
00100000 1 5
00101000 2 10
00110000 1 5
00111100 1 5
01100000 1 5
01110000 1 5
10100000 1 5
11110011 1 5
a
0 denotes normal perfusion, 1 hypoperfusion; the strings are in
the sequence left frontal lobe – right frontal lobe – left temporal
lobe – right temporal lobe – left parietal lobe – right parietal lobe –
left occipital lobe – right occipital lobe
Table 4 Observed frequencies
of bilateral temporoparietal hy-
poperfusion in 116 DAT pa-
tients and 20 controls for dif-
ferent definitions of bilateral
temporoparietal hypoperfusion
a
l = left, r = right, f = frontal,
t = temporal, p = parietal,
o = occipital
b
0 = normal perfusion, 1 = hy-
poperfusion, X = normal or
hypoperfusion
Definition
a
Classification Patterns
b
Relative frequency (%) p
Chi
2
-
DAT patients Controls test
lt and rt and Extremely 00111100 4 5 0.89
lp and rp and narrow
not (lf or rf or
lo or ro)
lt and rt and Narrow XX1111XX 18 5 0.14
lp and rp
(lt and rt) or Broad XX1X1XXX 48 25 0.053
(lp and rp) XXX1X1XX
lt or rt or lp or Extremely XX1XXXXX 87 75 0.12
rp broad XXX1XXXX
XXXX1XXX
XXXXX1XX
There is a significant linear trend for the MMSE score to
decrease with the number of regions with hypoperfusion
(r= 33, p < 0.01).
To summarise, the data from this study show that the
number of regions with hypoperfusion is more closely as-
sociated with DAT than the existence of any bilateral tem-
poroparietal pattern. The severity of DAT as assessed with
the MMSE is also significantly but not very closely re-
lated to the number of regions with hypoperfusions. Al-
though these data suggest that the importance of bilateral
temporoparietal hypoperfusion in DAT may have been
overemphasised in the past, the importance of pattern
analyses in general should not be renounced prematurely.
The next section describes some more detailed analyses
of the observed patterns.
Does DAT start in the temporal regions?
The hypothesis that the cortical starting point of DAT is the
temporal region is old, and in the meantime some data –
including SPECT findings – have been acquired which
seem to support this hypothesis (Eberling et al. 1992). One
implication of this hypothesis that can be easily tested
with the data from this study is that, if there is only one re-
gion with hypoperfusion, this should be a temporal one.
Fourteen of our DAT patients showed only one region
with hypoperfusion; the left temporal lobe was affected
eight times and the right temporal lobe four times, whereas
the left frontal and left parietal lobes were each affected
only once. The probability of the observed accumulation in
the temporal regions is p< 0.001 (binomial test with the
assumption that hypoperfusions in all regions were equally
probable = 1/8). This result strongly supports the hypoth
e-
sis that the temporal region is usually the first one af-
fected. Nevertheless, it must be mentioned that in the four
controls who showed hypoperfusion in only one region,
the temporal region was affected in two cases (and the
right parietal region in two cases). Although the low N of
controls necessitates care, this observation might suggest
that the temporal regions may be especially vulnera
ble in
general and not only for DAT. Although the left tem
poral
lobe was affected twice as often as the right temporal lobe
in this sample, this difference is not statistically signifi-
cant with the low number of patterns showing only one re-
gion with hypoperfusion.
Relationships between hypoperfusions in different areas
Table 5 shows phi coefficients for the associations between
each pair of brain regions when 2*2 tables of the respec-
tive regions are statistically analysed. In this case, phi co-
efficients are comparable to correlations and can have val-
ues between – 1 and 1.
Especially the parieto-occipital region shows high and
significant associations. A closer look at this region shows
that if the left occipital lobe is affected, then in 95% of
cases in this study the left parietal region is also affected.
The respective relative conditional frequency for the right
occipital region and the right parietal region is 89%. A
194
Fig.1 Means and standard deviations of the Mini Mental State
Examination (MMSE, Folstein et al. 1975) for the number of re-
gions with hypoperfusion in SPECT
Table 5 Association (phi co-
efficients) between hypoperfu-
sions in different brain regions
of N = 116 DAT patients
*P< 0.05, chi
2
-test
**P< 0.01, chi
2
-test
***P< 0.001, chi
2
-test
fr = frontal, te = temporal, pa =
parietal, oc = occipital/l = left,
r = right
fr/l fr/r te/l te/r pa/l pa/r oc/l
fr/r 0.41***
te/l 0.12 –0.18*
te/r –0.10 0.23* 0.14
pa/l 0.14 –0.01 0.21* 0.07
pa/r –0.05 0.17 0.09 0.26** 0.50***
oc/l –0.06 –0.25** 0.09 –0.15 0.63*** 0.26**
oc/r –0.13 –0.08 0.02 0.11 0.38*** 0.53*** 0.65***
plausible explanation for these associations is that after
affection of the parietal regions, the ipsilateral occipital
regions are affected next. All directly neighboured regions
– ipsilateral and contralateral – with exception of the left
frontal and temporal region and the left and right tempo-
ral region show significant associations. although these
associations are clearly lower than in the parieto-occipital
regions.
An interesting question is whether the additional in-
volvement of the occipital region is accompanied by fur-
ther functional decline. We, therefore, compared the MMSE
scores of patients who showed only temporoparietal af-
fection with those who also showed occipital affection. In
fact, the first group had a significantly higher mean MMSE
score (22.1) than the second group (18.3, p< 0.05, t-test).
A similar analysis comparing the frontal with the tem-
poroparietal region showed no significant difference. Thus,
hypoperfusion in the occipital region seems to occur pre-
dominantly after affection of the parietal region and is ac-
companied by further cognitive decline.
Do SPECT data contribute to the accuracy of diagnosis?
Thirteen of our controls and 22 of our DAT patients were
tested with the MMSE and the more difficult delayed word
recall test (DWR15). We performed discriminating analy-
ses with MMSE and DWR15 as predictors. These vari-
ables discriminated significantly (p< 0.05) between diag-
noses and classified 95% of DAT patients and 92% of con-
trols correctly. Inclusion of the number of brain areas
showing hypoperfusion as an additional predictor did not
alter this result at all. Five of the controls and 54 of the
DAT patients were tested with the MMSE and the easier
delayed word recall test (DWR8). MMSE and DWR8 dis-
criminated again significantly (p< 0.05) between diag-
noses and classified 87% of DAT patients and 100% of
controls correctly. Again, additional inclusion of the num-
ber of brain areas showing hypoperfusion did not improve
classification. Thus, using only the MMSE and a delayed
word recall test, good discrimination of diagnoses could
be reached and there was no case in which this discrimi-
nation could be improved by taking the best discriminat-
ing SPECT variable into account. Although the N of con-
trols is low, and therefore caution is necessary, in these data,
SPECT findings are clearly outperformed by variables of
cognitive function in their ability to differentiate between
controls and DAT patients.
Discussion
SPECT patterns observed in 20 controls and 116 DAT pa-
tients were evaluated. Although the temporal and parietal
region showed hypoperfusion more frequently in DAT pa-
tients than the frontal and occipital region, bilateral tem-
poroparietal hypoperfusion seems less characteristic for
DAT than previously suggested. In these data, none of four
different bilateral temporoparietal patterns was significantly
associated with diagnosis. However, the most appropriate
definition of bilateral temporoparietal hypoperfusion (bi-
lateral temporal and/or parietal hypoperfusion with or with-
out additional defects) led to patterns that only just missed
significant association with diagnosis (p= 0.053). More-
over, the sample of controls was not large and is conspic-
uous in that the control patients – although finally diag-
nosed as “normal” – visited our memory clinic. Therefore,
we feel sure that with a larger and/or less conspicuous sam-
ple of controls, a significant association between diagno-
sis and presence of bilateral temporoparietal hypoperfu-
sion could be demonstrated. Nevertheless, a significant
association does not mean that there is a close relationship,
and in keeping with recent work (Holman et al. 1992,
Horn et al. 1995), these data show that broadly defined bi-
lateral temporoparietal hypoperfusion is found in only ap-
proximately 48% of DAT patients and also approximately
25% of controls with cognitive problems. In looking for
an alternative SPECT indicator of DAT, the amount of hy-
poperfusions as indicated by the number of brain regions
showing hypoperfusion was evaluated. This indicator –
more than three regions with hypoperfusion – was signif-
icantly associated with diagnosis, and was more DAT-spe-
cific in this sample in that only 10% of the controls
showed this feature. Moreover, in DAT patients the num-
ber of brain regions with hypoperfusion correlated signif-
icantly with the severity of dementia as measured by the
MMSE. Nevertheless, further research is needed to decide
whether this purely quantitative consideration of SPECT
data, which completely ignores the location of hypoperfu-
sion, is really an alternative or supplementation to other
approaches.
In keeping with earlier publications (e.g. Eberling et al.
1992), the data from this study suggest that the cortical
decline in DAT usually begins in the temporal regions: the
temporal lobes were affected in 12 of the 14 cases which
had hypoperfusion in only one region. If the occipital lobe
was affected, then the ipsilateral parietal lobe was also af-
fected in approximately 90% of cases. Hypoperfusion in
the occipital lobe yielded significantly lower MMSE scores
than in the temporoparietal region. Thus, in DAT final
cortical involvement seems to extend usually from the
parietal to the occipital lobe, accompanied by further cog-
nitive impairment. Conversely, involvement of the frontal
lobe did not significantly influence the MMSE score com-
pared to the level that was observed with only tem-
poroparietal affection.
DAT patients and controls could be well discriminated
by the use of the MMSE and two delayed verbal recall tests
(correct classification in about 90% of the cases). Addi-
tional inclusion of SPECT findings did not improve clas-
sification in any single case. Therefore, SPECT findings
seem to be less useful than is sometimes suggested for the
differentiation between controls and DAT patients. To the
best of our knowledge, the performance of SPECT data in
the differentiation of diagnoses has not yet been tested in
comparison with other diagnostic tools.
To summarise, our results are not surprising if earlier
and recent literature is evaluated critically. The diagnosis
195
196
of DAT remains primarily a clinical one, and the relation-
ship between DAT and SPECT patterns is complex. Nev-
ertheless, SPECT patterns contribute importantly to the de-
scription of changes in cerebral perfusion that are caused
by DAT and other demential diseases. In this respect,
HMPAO-SPECT can contribute valuable and clinically rel-
evant information to other imaging approaches (Hampel
et al. 1997) in the field of the aetiopathology and diagno-
sis of dementia.
References
American Psychiatric Association (1987) Diagnostic and statistical
manual of mental disorders (DSM-III-R) 3rd edn, Washington
Burns A, Philpot M, Costa D, Ell P, Levy R (1989) The investiga-
tion of Alzheimer’s disease with single photon emission to-
mography. J Neurol Neurosurg Psychiatry 52:248–253
Eberling JL, Jagust WJ, Reed BR, Baker MG (1992) Reduced tem-
poral lobe blood flow in Alzheimer’s disease. Neurobiol Aging
13:483–491
Folstein MF, Folstein SE, McHugh PR (1975) “Mini Mental
State”. A practical method for grading the cognitive state of pa-
tients for the clinician. J Psychiat Res 12:189–198
Frackowiak RSJ, Pozzili C, Legg NJ, et al. (1981) Regional cerebral
oxygen supply and utilization in dementia: a clinical and phys-
iological study with oxygen-15 and positron tomography. Brain
104:753–778
Friedland RP, Prusiner S, Jagust W, Budinger TF (1984) Bitempo-
ral hypometabolism in Creutzfeld-Jakob disease: positron emis-
sion tomography with
18
F-2-fluorodeoxyglucose. J Comput As-
sist Tomogr 8:978–981
Friedland RP, Brun A, Budinger TF (1985) Pathological and
positron
emission tomographic correlations in Alzheimer’s dis-
ease. Lancet 1:228
Geaney D, Soper N, Shepstone B, Cowen P (1990) The effect of
central cholinergic stimulation on regional cerebral blood flow
in Alzheimer’s disease studied by single photon emission to-
mography. Lancet 335:1484–1487
Gemmel H, Sharp O, Besson J (1987) Differential diagnosis in de-
mentia using the cerebral blood flow agent 99m TC HMPAO:
a Spect study. J Comp Assis Tomog 11:398–402
Grünwald F, Horn R, Rieker O, Klemm E, Menzel C, Möller H-J,
Biersack H-J (1993) HMPAO-SPECT bei Demenz vom Alz-
heimer-Typ und Major Depression mit mnestischen Störungen
[HMPAO-SPECT in dementia of the Alzheimer type and Ma-
jor Depression with memory impairment]. Nuklearmedizin 32:
128–133
Habert M, Spampinato U, Mas J, et al. (1991) A comparative tech-
netium 99m hexamethylpropylene amine oxime SPECT study
in different types of dementia. Europ J Nuclear Med 18:3–11
Hachinski VC, Iliff LD, Chiak E, et al. (1975) Cerebral blood flow
in dementia. Arch Neurol 32:632–637
Hampel H, Padberg F, Kötter HU, Teipel SJ, Ehrhardt T, Hegerl
U, Stübner S, Möller H-J (1997) Diagnostik und Therapie der
Demenz vom Alzheimer-Typ. Münch med Wschr 139:35–41
Holman BL, Johnson KA, Gerada B, Carvalho PA, Satlin A (1992)
The scintigraphic appearance of Alzheimer’s disease: a prospec-
tive study using Technetium-99m-HMPAO SPECT. The Jour-
nal of Nuclear Medicine 33:181–185
Horn R, Grünwald F, Rieker O, Klemm E, Menzel C, Möller H-J,
Biersack H-J (1995) HMPAO-SPECT in memory disorders. In
Grünwald F, Kasper S, Biersack H-J, Möller H-J (eds) Brain
SPECT imaging in psychiatry. de Gruyter, Berlin, pp 97–112
Kuhl DE, Metter EJ, Riege WH (1984) Pattern of local cerebral
utilization determined in Parkinson’s disease by the [
18
F]fluo-
rodeoxyglucose method. Ann Neurol 15:419–424
Kuwabara Y, Ichiya Y, Otsuka M, et al. (1990) Differential diag-
nosis of bilateral parietal abnormalities in I-123 IMP SPECT
imaging. Clin Nucl Med 15:893–899
McGeer PL, Kamo H, Harrop R, et al. (1986) Comparison of PET,
MRI, and CT with pathology in a proven case of Alzheimer’s
disease. Neurology 36:1569–1574
Montaldi D, Brooks D, McColl J, Wyper D, Patterson J, Barron E,
McCulloch J (1990) Measurements of regional cerebral blood
flow and cognitive performance in Alzheimer’s disease. J Neu-
rol Neurosurg Psychiatry 53:33–38
Poremba M (1993) Demenzen [Dementias]. In: Brandt T, Dichgans
J, Diener HC (eds) Therapie und Verlauf neurologischer Er-
krankungen [Therapy and course of neurologic diseases]. Kohl
-
hammer, Stuttgart, pp 890–911
Salmon E, Frack G (1989) Positron emission tomographic study in
Alzheimer’s disease and Pick’s disease. Arch Gerontol Geriatr
Suppl 1:241–247
Salmon E, Sadzot B, Maquet P, Degueldre C, Lemaire C, Rigo P,
Comar D, Franck G (1993) Differential diagnosis of Alzhei-
mer’s disease with PET. The Journal of Nuclear Medicine 35:
391–398
Volles E (1991) Sogenannte degenerative Krankheiten. In: Fröscher
W (ed) Neurologie. de Gruyter, Berlin, pp 347–368
Welsh KA, Butters N, Hughs JP, Mohs RC, Heyman A (1992) De-
tection and staging of dementia in Alzheimer’s disease. Arch
Neurol 49:448–452
Wolfe N, Reed BR, Eberling JL, Jagust WJ (1995) Temporal lobe
perfusion on single photon emission computed tomography pre-
dicts the rate of cognitive decline in Alzheimer’s disease. Arch
Neurol 52:257–262
