Eur J Nucl Med Mol Imaging (2011) 38:764773

DOI 10.1007/s00259-010-1664-1

REVIEW ARTICLE

SPECT imaging evaluation in movement disorders:

far beyond visual assessment
Kosmas Badiavas & Elisavet Molyvda &
Ioannis Iakovou & Magdalini Tsolaki &
Kyriakos Psarrakos & Nikolaos Karatzas

Received: 27 May 2010 / Accepted: 1 November 2010 / Published online: 2 December 2010
# Springer-Verlag 2010

Abstract Single photon emission computed tomography

(SPECT) imaging with 123I-FP-CIT is of great value in
differentiating patients suffering from Parkinsons disease
(PD) from those suffering from essential tremor (ET).
Moreover, SPECT with 123I-IBZM can differentiate PD
from Parkinsons plus syndromes. Diagnosis is still
mainly based on experienced observers visual assessment
of the resulting images while many quantitative methods
have been developed in order to assist diagnosis since the
early days of neuroimaging. The aim of this work is to
attempt to categorize, briefly present and comment on a
number of semi-quantification methods used in nuclear
medicine neuroimaging. Various arithmetic indices have
been introduced with region of interest (ROI) manual
drawing methods giving their place to automated procedures, while advancing computer technology has allowed
automated image registration, fusion and segmentation to

K. Badiavas (*)
Medical Physics Department, Papageorgiou General Hospital,
Periferiaki odos,
564 03, Thessaloniki, Greece
e-mail: badiavas@auth.gr
E. Molyvda : K. Psarrakos
Medical Physics Department, Aristotle University,
Thessaloniki, Greece
I. Iakovou : N. Karatzas
3rd Nuclear Medicine Department, Aristotle University,
Papageorgiou General Hospital,
Thessaloniki, Greece
M. Tsolaki
3rd Neurology Clinic, Aristotle University,
Papanikolaou General Hospital,
Thessaloniki, Greece

bring quantification closer to the final diagnosis based on

the whole of the patients examinations results, clinical
condition and response to therapy. The search for absolute
quantification has passed through neuroreceptor quantification models, which are invasive methods that involve tracer
kinetic modelling and arterial blood sampling, a practice
that is not commonly used in a clinical environment. On the
other hand, semi-quantification methods relying on computers and dedicated software try to elicit numerical
information out of SPECT images. The application of
semi-quantification methods aims at separating the different
patient categories solving the main problem of finding the
uptake in the structures of interest. The semi-quantification
methods which were studied fall roughly into three
categories, which are described as classic methods, advanced automated methods and pixel-based statistical
analysis methods. All these methods can be further divided
into various subcategories. The plethora of the existing
semi-quantitative methods reinforces the feeling that visual
assessment is still the base of image interpretation and that
the unambiguous numerical results that will allow the
absolute differentiation between the known diseases
have not been standardized yet. Switching to a commonly
agreedideally PC-basedautomated software that may
take raw or mildly processed data (checked for consistency
and maybe corrected for attenuation and/or scatter and
septal penetration) as input, work with basic operators
inference and produce validated numerical results that will
support the diagnosis is in our view the aim towards which
efforts should be directed. After all, semi-quantification can
improve sensitivity, strengthen diagnosis, aid patients
follow-up and assess the response to therapy. Objective
diagnosis, altered diagnosis in marginal cases and a
common approach to multicentre trials are other benefits
and future applications of semi-quantification.

Eur J Nucl Med Mol Imaging (2011) 38:764773

Keywords Parkinson . Neuroimaging SPECT .

123
I-FP-CIT . 123I-IBZM . Semi-quantification ROI
analysis . Statistical parametric mapping

Introduction
Parkinsons disease (PD) is a progressive neurological
disorder with an aetiology which is still mostly unknown.
The associated degeneration of the dopaminergic nigrostriatal neurons results in dopamine depletion in the corpus
striatum [1]. The caudate nucleus and especially the
putamen are the structures mostly associated with this
pathology. Dopamine, the transmitter of the dopaminergic
neurotransmission system, is produced in the dopaminergic
neurons and stored in the synaptic vesicles. In response to
an action potential, dopamine is released from the presynaptic terminal into the synaptic cleft and is taken in the
dopamine receptors of the postsynaptic membrane. The
dopamine transporter (DAT) is the protein that controls
dopamine levels through active reuptake after dopamines
interaction with the postsynaptic receptors. Dopamine
induces postsynaptic potentials and is associated with skills
like movement, focusing and learning, as well as pathologies like parkinsonism and schizophrenia.
The common clinical features that patients suffering
from movement disorders share are referred to as
parkinsonism. The neurodegenerative diseases called
Parkinsons syndromes include PD and Parkinsons
plus syndromes like multiple system atrophy (MSA),
progressive supranuclear palsy (PSP) and corticobasal
degeneration (CBD). Differentiating between these pathologies and essential tremor (ET) is a difficult and
important task as they demand different therapeutic
strategies and have different prognoses. Nuclear medicine
single photon emission computed tomography (SPECT)
plays a major role in differentiating PD patients from ET
patients with 123I-FP-CIT and PD patients from Parkinsons plus syndromes patients with 123I-IBZM, among
other radiopharmaceuticals.
Until now, only a postmortem study (autopsy) could
clearly identify PD due to the presence of Lewy bodies. In
patients, there is no definitive test for this purpose and PD
diagnosis must be based on clinical criteria that range
from the presence of motor symptoms like rest tremor,
bradykinesia, rigidity and loss of postural reflexes to
secondary motor symptoms and non-motor symptoms like
sleep disorders and anosmia. Response to levodopa and
apomorphine challenges are also used for identifying PD.
Rating scales are used to evaluate the motor impairment
and disability in PD patients. The Hoehn and Yahr scale is
commonly used to compare groups of patients and assess
disease progression while the Unified Parkinsons Disease

765

Rating scale (UPDRS) is the preferred scale for assessing

disability and impairment [2].
Since the early days of neuroimaging, many quantitative
methods have been developed in order to assess diagnosis
which is still mainly based on experienced observers visual
assessment of the SPECT images. Various arithmetic
indices have been introduced with region of interest (ROI)
manual drawing methods giving their place to automated
procedures, while advancing computer technology has
allowed automated image registration, fusion and segmentation to bring quantification closer to the final diagnosis
based on the whole of the patients examinations results,
clinical condition and response to therapy.

Parkinsonian syndromes and SPECT

Radiopharmaceuticals
Since the late 1980s, D2 postsynaptic receptors have been
visualized with various radiopharmaceuticals allowing the
differentiation between PD and Parkinsons plus syndromes
[36]. In the mid-1990s, studying the dopaminergic neurons
allowed the differentiation between PD and ET through
imaging the DAT [79]. Today in Europe, 123I-FP-CIT with
the commercial name DaTSCAN (GE Healthcare, UK) is the
commonest radiopharmaceutical for DAT SPECT imaging,
while 123I-IBZM (GE Healthcare, UK) is the most common
radiopharmaceutical for imaging the D2 postsynaptic
receptors [10, 11].
123
I-IBZM binds steadily in the striatum 1.53 h after its
i.v. administration and therefore SPECT imaging gives best
results approximately 2 h after injecting the patient.
However, its low specific binding has rendered the use of
semi-quantification methods compulsory. The detailed
procedure as well as other issues concerning the D2
postsynaptic terminal examinations with 123I radiopharmaceuticals can be found in the relevant European Association
of Nuclear Medicine (EANM) guidelines [12, 13].
On the other hand, 123I-FP-CIT, which visualizes the
dopamine transporter, is stable 36 h after its i.v. administration and best imaged 34 h post-injection achieving
high specific binding to the striatum compared to the rest
of the brain parenchyma. The above characteristics make
123
I-FP-CIT ideal for clinical use. As in the previous case,
the detailed procedure and all relevant issues can be found
in the appropriate EANM guidelines [14, 15]. Koch and
coworkers found that combining quantitative results from
both examinations increases the diagnostic accuracy in the
differential diagnosis of parkinsonism in comparison to
specific D2 receptor binding alone [16].
In the USA and other parts of the world, the radiopharmaceuticals 123 I--CIT and 99m Tc-TRODAT, a

766

Eur J Nucl Med Mol Imaging (2011) 38:764773

99m

frontal lobe have been used as reference areas by various

researchers, with the occipital lobe ending up as the
prevalent choice. The final image can be considered as
the sum of the specific uptake in the striatum plus the
nonspecific uptake in the rest of the brain parenchyma. The
most prevalent but ambiguous quantitative index as
considered to be by various researchers and commented
later in the paper, that tries to express the situation is the
specific binding ratio (SBR). It is defined as the ratio of
specific to nonspecific uptake and is related to the density
of the postsynaptic terminals or the dopamine transporters
[24]. Semi-quantification results can be used to strengthen
diagnosis [30], improve sensitivity, aid follow-up and
assess response to therapy [31, 32].

Tc-labelled compound, are also used for the dopamine

transporter examinations [1721]. Booij and coworkers
reviewed SPECT and also positron emission tomography
(PET) imaging of the dopaminergic neurotransmission
system [22].
Structures of interest and visual assessment
The SPECT images of normal volunteers and patients
suffering from ET show similar uptake in the caudate
nucleus and putamen with the striatum presenting as
comma-shaped. However, in parkinsonian syndromes,
the neuronal degeneration is more pronounced in the
putamen than in the caudate nucleus and in severe cases
the affected striatum loses the comma shape and becomes
point-like. Eventually, it loses the ability to selectively
gather radiopharmaceuticals in comparison to the rest of the
brain parenchyma. The reduced uptake is more pronounced
in the contralateral side of the one which suffers from the
heavier movement disorder symptoms [2325].
Visual assessment of the resulting images is the
generally accepted diagnostic method for SPECT examinations of the dopaminergic neurotransmission system [26].
Assessing striatal shape and completeness offers information about left to right asymmetric radiopharmaceutical
uptake which is important because the loss of the
dopaminergic neurons starts ipsilaterally, at the contralateral
side to the symptoms [27]. Visual assessment depends
heavily on the observers experience and naturally shows
inter- and intra-observer variability, dictating the need for
quantification to substantiate the diagnosis [28].

Tracer kinetic models versus semi-quantification

methods
Attempts to quantify uptake have led to the development of
neuroreceptor quantification models using tracer kinetic
modelling [29]. These methods combine the patients blood
samples with the relevant image data, assuming the
existence of compartmental systems. They ascribe the
percent (%) uptake, called binding potential, an index that
is directly related to the density of the presynaptic
transporters or the postsynaptic terminals of the brains
ROIs. As these methods are invasive, demanding multiple
arterial blood sampling, and despite the fact that recent
approaches are less complex, they are not commonly used
in clinical practice.
Alternatively, semi-quantification methods use ratios
formed between radiopharmaceutical-specific uptake in
special ROIs (striatum, caudate nucleus, putamen) and
nonspecific uptake reference areas, the latter being devoid
of nigrostriatal neurons. The occipital lobe, cerebellum and

Semi-quantification methods: opening Pandoras box

Classic quantification methods
The measurements of an index that characterize striatal
uptake, like the above-mentioned SBR, requires the
definition of ROIs for the structure investigated and for a
reference area. Many methods for drawing ROIs have been
developed and their influence on the quantitative results has
been studied [33].
Manual positioning of anatomical ROIs
Costa and coworkers, in 1990 while studying D2 receptors
in humans with 123I-IBZM, drew irregular ROIs covering
the striatum and a reference region in order to calculate the
specific binding activity [17]. This simple method is still
used in many centres, expanded by specialists drawing
ROIs covering the relevant substructures, namely caudate
nucleus and putamen (see Fig. 1). Variations of the method
range between using the slice with maximum counts and
adding a number of slices that comprise the whole structure
in the image [31]. An anatomical atlas can be used to aid
drawing in an individualized way or for standard ROI
template creation [23]. However, manually positioning
ROIs is very subjective, can be time-consuming and shows,
as has already been said, intra- and inter-observer variability [28]. The numerical outcome of such a manual
interpretation involves the SBRs for the striatum, caudate
nucleus and putamen, as well as percent differences of the
unilateral structures and popular ratios like the putamen to
caudate ratio. Representative SBR values for normal studies
done in our institution are presented in Table 1 [34]. We use
an ADAC Vertix Plus dual-head camera, low-energy highresolution (LEHR) collimators, 128 projections, 30 s per
step, 128128 matrix, 20% symmetrical energy window
centred around 159 keV, reconstruction using a Butterworth

Eur J Nucl Med Mol Imaging (2011) 38:764773

Fig. 1 Manually positioned
ROIs on the same transverse
slice (a no attenuation correction, b attenuation correction) of
a 123I-FP-CIT SPECT study
(Papageorgiou General
Hospital, Nuclear Medicine
Department)

767

filter, 0.20 cycles/pixel, order 10, and no attenuation

correction (AC).
Drawing individualized ROIs using image fusion
techniques or generalized ROIs using standard ROI
templates
Multimodal image fusion and registration allows combining
structural and functional images to facilitate ROI positioning. Using a brain MRI study of the patient taking the
nuclear medicine examination is a very precise method
although it is difficult to follow in routine clinical practice
as not all patients taking neuroimaging SPECT have a
recent MRI scan that can be used for fusion purposes.
Unfortunately from the radiology point of view but luckily
for this implementation, striatal size and shape as seen in
conventional MRI are similar between patients and healthy
volunteers except in very advanced disease states [35].
Consequently, an MRI study of a single subject or a mean
Table 1 Local normal database for the various quantification methods
(Papageorgiou General Hospital, Nuclear Medicine Department)
Method
Manual ROI positioning
Striatum
Caudate nucleus
Putamen
Two box
Specific binding index
Three box
Total striatal binding potential index
Crescent
Striatum
Putamen
Putamen to caudate ratio

AC

Values SD

No
No
No

2.620.45
3.320.57
1.950.46

Yes

5.770.97

Yes

788.7 (%)

Yes
Yes
Yes

1.680.44
1.370.41
0.800.10

MRI, SPECT or PET set of images of a group of healthy

subjects could be used to form standard ROI templates for
all structures of interest [36, 37]. Additional techniques,
like the use of external markers placed on the patients
head, increase the accuracy of ROI placing and could
optimize the procedure [20, 38].
Avoiding partial volume effect using resolution-independent
methods
Based on the work of Costa and coworkers, a quantification
method that calculates uptake ratios from two large
rectangular ROIs covering striata and one rectangular ROI
that is placed on the occipital lobe has been developed [17].
The approach of using large ROIs, in contrast to the strict
and detailed drawing around the small structures of interest,
is used to eliminate the so-called partial volume effect
(PVE) owing to the poor spatial resolution of the gamma
camera. PVE is the phenomenon due to which counts are
blurred out of the structures physical volume, underestimating and preventing accurate evaluation of uptake,
when very high count small areas are next to very low
count areas, as it happens in striatal borders where the high
uptake striatum gives its place to the nonspecific uptake
brain parenchyma.
Fleming and coworkers proposed an interesting uptake
index, called the specific uptake size index (SUSI), which
describes a new measure of total uptake of an organ and can
either be related in absolute terms to the total activity injected
or to the specific activity of a reference region [39]. This new
index type which depends on the total objects activity is
independent of the resolution of the imaging process and
attributes the activity concentration more successfully than is
conventionally derived from an index of the SBR type.
Tossici-Bolt and coworkers used the SUSI approach for the
accurate determination of the SBR taking into account the
PVE, while extensively commenting on the ambiguity of the

768

Eur J Nucl Med Mol Imaging (2011) 38:764773

specific binding index (SBI) as a universal index [40]. With

their method, operator intervention is limited to the
placement of two large trapezoidal ROIs separately covering
striata, while the reference region is automatically defined as
the whole of the brain parenchyma excluding the striatal
regions. Their retrospective study concerning a group of
normal and abnormal subjects imaged with a dual-head
camera, 128 projections, 128128 matrix, 20% symmetrical
energy window centred around 159 keV, reconstruction
using a Butterworth filter, 0.20 cycles/pixel, order 10, with
first-order Chang attenuation correction and no correction for
scatter and septal penetration, showed 3 and 4% intra- and
inter-operator variability, respectively, while by using 4.5 as
the cutoff value that separated normal and abnormal groups
they achieved a sensitivity, specificity and diagnostic
concordance of 97, 92 and 95%, respectively.
Dobbeleir and coworkers introduced a similar resolutionindependent quantification method and reported 89%
sensitivity and 100% specificity with 94% accuracy in
separating patients with and without abnormal FP-CIT
uptake [41]. Goethals and coworkers applied receiveroperating characteristic (ROC) analysis in the same method
and found 95% sensitivity and 72% specificity for
differentiating between normal and abnormal scans [42].
Semi-automated software packages
A PC-based automated software package offering three
semi-quantification methods to facilitate diagnosis has been
commercially available for some years now (QuantiSPECT,
Mirada Solutions). The package offers the choice of
reconstruction, three-dimensional (3-D) Chang attenuation
correction and three different methods for automated
quantification with the descriptive and commercial names
(a) the two box method based on the quantification
technique of Tossici-Bolt and coworkers mentioned above
Fig. 2 Two large trapezoidal
striatal ROIs and the automatically defined reference region
on the same transverse slice
(a no attenuation correction,
b attenuation correction) of a
123
I-FP-CIT SPECT study
(the two box method,
QuantiSPECT, Mirada
Solutions; Papageorgiou
General Hospital, Nuclear
Medicine Department)

(see Fig. 2), (b) the three box method based on the paper
of Costa and coworkers (see Fig. 3) and (c) the crescent
method (see Fig. 4) based on a paper of Lkkegaard and
coworkers [43], as described in detail in the relevant papers
of Morton and coworkers [44, 45].
The numerical outcome of the two box method is the
SBI calculated for each striatum as defined in [45].
Representative SBI values for normal studies done in our
institution, attenuation-corrected using the packages inherent Chang attenuation correction option, are presented in
Table 1 [34]. The quantitative data from the three box
algorithm is the total striatal binding potential index
(TSBPI) for each striatum. Again, representative TSBPI
values for normal studies done in our institution, Chang
attenuation-corrected using the packages inherent Chang
attenuation correction option, are presented in Table 1.
Finally, the crescent method provides uptake values for the
striatum and putamen, as well as the putamen to caudate
ratio. Again, representative values for normal studies done
in our institution, Chang attenuation-corrected using the
packages inherent Chang attenuation correction option, are
presented in Table 1 [34].
The newly launched EXINI DAT is another PC-based
commercially available software for computer-assisted
diagnosis of DAT SPECT scans (EXINI Diagnostics AB,
Lund, Sweden). As input, it uses transverse slices which
have already been reconstructed and/or attenuation corrected in the local departments workstation and transferred
to the PC via the DICOM protocol. It performs fully
automatic quantitative analysis of DAT images and provides the specific uptake values for the striatum, caudate
and putamen. The analysis is done in 3-D (the entire image
volume is used) and the report can be exported to a hospital
information system (HIS). After contacting the company, a
time-limited demo version can be downloaded from the
relevant website for evaluation purposes.

Eur J Nucl Med Mol Imaging (2011) 38:764773

Fig. 3 Three large rectangular
ROIs, one at each striatum and
one defining the reference
region on the same transverse
slice (a no attenuation correction, b attenuation correction)
of a 123I-FP-CIT SPECT study
(the three box method,
QuantiSPECT, Mirada
Solutions; Papageorgiou
General Hospital, Nuclear
Medicine Department)

Advanced automated quantification methods

Multimodal image registration, especially SPECT combined with MRI, is considered to be very effective for
quantifying uptake. Moreover, both modalities are used to
provide generic templates, which are preferable to use than
patient-specific ones in this categorys quantification
methods. However, a specially designed or even automated
image registration algorithm can be very useful in everyday
clinical practice.
Lee and coworkers originally developed an automated
SPECT and MRI image registration algorithm based on
anatomical characteristics using image intensities for
improved precision [46]. In the same paper they review
image registration methods and present a SPECT normalization algorithm for the production of clinically relevant
image contrast. As an extension of their previous work,
they present a computer-aided evaluation system for PD
Fig. 4 Striatal and background
template ROIs on the same
transverse slice (a no attenuation
correction, b attenuation correction) of a 123I-FP-CIT SPECT
study (the crescent method,
QuantiSPECT, Mirada Solutions; Papageorgiou General
Hospital, Nuclear Medicine
Department)

769

detection that uses automated registration and image

labelling [47].
Zubal and coworkers propose an automated quantification procedure named objective striatal analysis, in the
form of MATLAB scripts (The MathWorks Inc., Natick,
MA, USA) computer language [48]. This software
reorients the images according to a template, identifies
the structures of interest (striatum, caudate nucleus and
putamen) and the reference region (occipital lobe),
corrects for attenuation and computes the specific to
nonspecific ratio using the parameter V3 introduced by
Seibyl and coworkers in 1995 [20].
Calvini and coworkers designed BasGan, an algorithm
obviously named after an acronym of the words basal
ganglia, for the automatic segmentation of striatal 123I-FPCIT SPECT [49]. They use a high-definition 3-D striatal
template derived from the Talairach atlas [50]. A blurred
template, derived from a convoluted with a 3-D Gaussian

770

kernel of 10 mm full-width at half-maximum (FWHM)

original template, approximates striatal activity distribution.
The algorithm, which includes PVE correction, makes the
necessary transformations to place the template in the right
initial position while an optimization protocol is responsible
for the final matching. A reference region in the occipital
lobe is used. They report a highly significant correlation
between true binding potential and measured 123I activity in
an anthropomorphic striatal phantom, significantly lower
123
I-FP-CIT uptake in all basal ganglia in a PD group
versus controls and significant correlations with the motor
UPDRS score. The advantages of this algorithm are that
coregistered MRI is not needed, it has a 3-D nature that
uses volumes of interest (VOIs) instead of ROIs and is
freely available on the Web.
Other groups have developed a 3-D automated technique
for the evaluation of 123I-IBZM SPECT that can be added
to the Brain Analysis Software (BRASS, Hermes Medical
Solutions, Stockholm, Sweden) which is a software
package that runs on a Hermes workstation (Nuclear
Diagnostics, Stockholm, Sweden) [51, 52]. BRASS implements a multistep registration of individual patient studies
to a template of healthy controls before the subsequent
application and fine adjustment of a standardized 3-D VOI
map to measure semi-quantitative values. Both the abovementioned template and 3-D VOI map are available with
BRASS, but alternatively each centre has the option to
create its own one. The software calculates mean counts per
voxel values for each VOI and ratios of the entire striatum,
caudate nucleus and putamen to the occipital cortex. Koch
and coworkers proceeded in clinical testing of this software
in a large patient population and found the results to agree
with the conventional time-consuming and observerdependent manual drawing method [53]. Morton and
coworkers compared QuantiSPECT methods, in which both
the two box and the three box methods do the quantification on a summed image that encompasses the whole striata
and the ROIs that the operator has to position are
effectively 3-D VOIs, and BRASS, which performs 3-D
processing referencing a normal image template. They
found that all programs and processing methods could
distinguish between PD and ET, there was a good
correlation between the results from both methods and that
the intra- and inter-operator variability depended on the
amount of operator intervention [45].
Hamilton and coworkers developed an artificial neural
network (ANN) with the ability to discriminate between PD
and ET patients after inserting results of classic quantitative
analysis and manual ROI positioning [54]. The specific
neural network is described as taught by example and not a
priori specified, which seems to be advantageous according
to its developers, in comparison to classic pattern recognition procedures. They studied the role of the putamen to

Eur J Nucl Med Mol Imaging (2011) 38:764773

caudate nucleus uptake ratio as an important marker and

reported a successful separation between 18 PD and ET
patients analysed in a single step with their ANN.

Voxel-based statistical analysis methods

A somewhat different quantification approach is based on
voxel-based statistical analysis. These methods use automated procedures to manipulate data and apply statistical
models to find the probability that the voxels of interest
have values that lie inside a predefined range or above a
predefined threshold. They usually use predefined templates to register the individual patient data.
Statistical Parametric Mapping and NEUROSTAT
Statistical Parametric Mapping or SPM (Wellcome Department of Cognitive Neurology, University College London,
UK) is a software package that allows voxel-by-voxel
statistical comparison between sets of images [55, 56]. It
was first used for analysing PET and functional MRI
(fMRI) studies. SPM allows the comparison between
groups of patients and a group of healthy subjects. It uses
the Montreal Neurological Institute (MNI) stereotaxic
template as the canvas for all images. After modifications, it can compare a single patient against a database or
another patient [57, 58]. SPM runs on a MATLAB
platform and is available for downloading from the
website of its creators.
Colloby and coworkers applied SPM to 123I-FP-CIT
SPECT in dementia with Lewy bodies (DLB), Alzheimers
disease (AD) and PD [59]. They investigated differences in
striatal binding FP-CIT binding in DLB in comparison with
AD, idiopathic Parkinsons disease (IPD) and control
subjects using SPM. In addition, they investigated the
usefulness of SPM as a decision aid and compared the
findings with visual rating observations. Group differences
were assessed using a two-sample t test and applying a
height threshold of p0.05 corrected, the SPM{t} maps
showed significantly reduced uptake in the striata of DLB
and PD subjects compared to AD subjects and controls.
Using ROC curve analysis they found that SPM demonstrates comparable discriminatory power with visual rating.
NEUROSTAT is an anatomical standardization package
that works on a pixel-by-pixel basis and was developed for
the objective interpretation of functional brain images by
Minoshima and coworkers [60]. It was validated and found
to be as effective as SPM in PET studies of normal brains
with 18F-fluorodeoxyglucose (FDG) [61]. Takada and
coworkers used NEUROSTAT to estimate easily and
objectively the striatal uptake of a PD group and a group
of ET patients after 123I-FP-CIT SPECT by calculating V3

Eur J Nucl Med Mol Imaging (2011) 38:764773

with two methods, VOIClassic, i.e. the software that

automatically calculates the counts in anatomically defined
regions in the brain using stereotactic anatomical standardization included in NEUROSTAT, and from manual ROI
placement in fused MR images [62].
Using scripts on the SPM platform
Buchert and coworkers developed a fully automated expert
system (IBZM tool) to be used for the assessment of
123
I-IBZM SPECT procedures. It can also be used for
123
I-FP-CIT SPECT studies. It was tested on clinically welldocumented patients and managed to differentiate MSA
subjects from controls with 100% sensitivity and 86%
specificity providing improved statistical power compared
with manual VOI analysis [63]. This software, written in
MATLAB, is offered as a tool in the SPM toolbox and uses
predefined VOIs. The tool is freely available on the Web.
Other pixel-based techniques
Acton and coworkers proposed their pixel-based statistical
analysis method which they call logistic discriminant
parametric mapping (LDPM). The operator trained the
discriminant model on a set of patients with unequivocal
diagnosis obtaining logistic discriminant parametric maps
for all subjects. They reported that LDPM showed 100%
sensitivity and 95% specificity in differentiating patients
and age-matched controls [64].
Habraken and coworkers developed a fully 3-D automatic,
observer-independent technique for the quantification of the
neuronal radiotracer binding on a voxel-by-voxel basis, which
they call voxel-based automatic neuronal quantification
(VANQ) [65]. They argue that because 123I-FP-CIT studies
show a high specific binding in striatal areas and consequently show large functional differences between pathologic and
normal studies, linear scaling and nonlinear warping techniques are not optimal for registering such studies. On the other
hand, their technique implements a 3-D registration algorithm
that registers each VOI separately using a 3-D model (VOI
map) which is based on healthy controls as a reference
model. This software automatically calculates the specific to
nonspecific ratio for the patients striatum, an apparent
limitation of the method that uses the whole striatum (not
its substructures), which compares it to a database of healthy
volunteers. They found that this automatic method produces
equally good results to visual assessment.

Concluding remarks
SPECT neuroimaging studies play an important diagnostic
role in movement disorders. Despite the fact that specific

771

radiopharmaceuticals achieve a much higher uptake in the

structures of interest in comparison to the rest of the brain
parenchyma and that the head is a favourable human part to
image, the diagnosis is not trivial in many cases. The visual
assessment of the resulting images done by experienced
observers, although subjective, is still the method of choice
for reaching a diagnosis, as autopsy is the only gold
standard but unfortunately useless for the patient. Searching
for objectivity, many semi-quantification methods have
been introduced providing a variety of carefully chosen
quantitative indexes. The quantitative results range from the
SBR type results, seen in the classic semi-quantification
methods, to the voxel-based statistical analysis results, seen
in the relevant methods. Surprisingly, many semiquantification methods are usually tested against the
subjective visual assessment. In straightforward cases,
almost all semi-quantification methods produce satisfactory
numerical results. However, in patients showing reduced
uptake or special brain anatomy the results are sometimes
inconclusive or controversial. The EANM through its
guidelines encourages the use of semi-quantification methods. The plethora of existing methods reinforces the feeling
that quantification has earned a place in nuclear medicine
neuroimaging. On the other hand, it states that the absolute
semi-quantification method has not been established yet. In
our view, a commonly agreed upon and ideally PC-based,
automated software that may take raw or mildly processed
data as input and produce, with minor operators inference,
validated numerical results that will support the diagnosis is a
good solution. Extensive studies that consider the final
diagnosis as the end point and the visual assessment as one
of the processing techniques under evaluation will better
demonstrate the value of semi-quantitative techniques. Another issue is that the dependence of the semi-quantification
results on the specific gamma camera and associated
equipment forces nuclear medicine centres to create local
(ideally age-dependent) normal databases. In addition, the
results of the ongoing project for a European database of
123
I-FP-CIT SPECT scans of healthy subjects are expected to
be published soon. Their incorporation in various semiquantification packages is then only a matter of time.

Conflicts of interest None.

References
1. Hornykiewicz O. Dopamine (3-hydroxytyramine) and brain
function. Pharmacol Rev 1966;18:92564.
2. Jankovic J. Parkinsons disease: clinical features and diagnosis. J
Neurol Neurosurg Psychiatry 2008;79:36876.
3. Kung HF, Pan S, Kung MP, Billings J, Kasliwal R, Reilley J,
et al. In vitro and in vivo evaluation of [123I]IBZM: a

772
potential CNS D-2 dopamine receptor imaging agent. J Nucl
Med 1989;30:8892.
4. Kung HF, Alavi A, Chang W, Kung MP, Keyes JW, Velchik MG,
et al. In vivo SPECT imaging of CNS D-2 dopamine receptors:
initial studies with iodine-123-IBZM in humans. J Nucl Med
1990;31:5739.
5. Brcke T, Podreka I, Angelberger P, Wenger S, Topitz A, Kfferle
B, et al. Dopamine D2 receptor imaging with SPECT: studies in
different neuropsychiatric disorders. J Cereb Blood Flow Metab
1991;11:2208.
6. Tatsch K, Schwartz J, Oertel WH, Kirsch CM. SPECT imaging of
dopamine D2 receptors with 123I-IBZM: initial experience in
controls and patients with Parkinsons syndrome and Wilsons
disease. Nucl Med Commun 1991;12:699707.
7. Seibyl JP, Wallace E, Smith EO, Stabin M, Baldwin RM, Zoghbi
S, et al. Whole-body biodistribution, radiation absorbed dose and
brain SPECT imaging with iodine-123--CIT in healthy human
subjects. J Nucl Med 1994;35:76470.
8. Booij J, Tissingh G, Boer GJ, Speelman JD, Stoof JC, Janssen
AG, et al. [123I]FP-CIT SPECT shows a pronounced decline of
striatal dopamine transporter labelling in early and advanced
Parkinsons disease. J Neurol Neurosurg Psychiatry 1997;62:133
40.
9. Asenbaum S, Pirker W, Angelberger P, Bencsits G, Pruckmayer
M, Brcke T. [123I]beta-CIT and SPECT in essential tremor and
Parkinsons disease. J Neural Transm 1998;105:121328.
10. Gerasimou G, Tsolaki M, Bostanjopoulou S, Liaros G,
Papanastasiou E, Balaris V, et al. Findings from molecular
imaging with SPET camera and 123I-ioflupane in the differential
diagnosis of Parkinsonism and essential tremor. Hell J Nucl Med
2005;8(2):815.
11. Marshall V, Grosset D. Role of dopamine transporter imaging in
routine clinical practice. Mov Disord 2003;18(12):141523.
12. Tatsch K, Asenbaum S, Bartenstein P, Catafau A, Halldin C,
Pilowsky LS, et al. European Association of Nuclear Medicine
procedure guidelines for brain neurotransmission SPET using
(123)I-labelled dopamine D(2) receptor ligands. Eur J Nucl Med
Mol Imaging 2002;29:BP239.
13. Van Laere K, Varone A, Booij J, Vander Borght T, Nobili F,
Kapucu OL, et al. EANM procedure guidelines for brain
neurotransmission SPECT/PET using dopamine D2 receptor
ligands, version 2. Eur J Nucl Med Mol Imaging 2010;37:434
42. doi:10.1007/s00259-009-1265-z.
14. Tatsch K, Asenbaum S, Bartenstein P, Catafau A, Halldin C,
Pilowsky LS, et al. European Association of Nuclear Medicine
procedure guidelines for brain neurotransmission SPET using
(123)I-labelled dopamine D(2) transporter ligands. Eur J Nucl
Med Mol Imaging 2002;29:BP305.
15. Darcourt J, Booij J, Tatsch K, Varrone A, Vander Borght T,
Kapucu OL, et al. EANM procedure guidelines for brain
neurotransmission SPECT using (123)I-labelled dopamine transporter ligands, version 2. Eur J Nucl Med Mol Imaging
2010;37:44350. doi:10.1007/s00259-009-1267-x.
16. Koch W, Hamann C, Radau PE, Tatsch K. Does combined
imaging of the pre- and postsynaptic dopaminergic system
increase the diagnostic accuracy in the differential diagnosis of
parkinsonism? Eur J Nucl Med Mol Imaging 2007;34:126573.
17. Costa DC, Verhoeff NPLG, Cullum ID, Ell PJ, Syed GM, Barrett
J, et al. In vivo characterization of 3-iodo-6-methoxybenzamide
123I in humans. Eur J Nucl Med 1990;16:8136.
18. Neumeyer JL, Wang S, Millius RA, Baldwin RM, Zia-Ponce Y,
Hoffer PB, et al. [123I]-2 beta-carbomethoxy-3 beta-(4-iodophenyl)tropane: high-affinity SPECT radiotracer of monoamine
reuptake sites in brain. J Med Chem 1991;34:31446.
19. Kuikka JT, Bergstrm KA, Vanninen E, Laulumaa V, Hartikainen P,
Lnsimies E. Initial experience with single-photon emission

Eur J Nucl Med Mol Imaging (2011) 38:764773

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

tomography using iodine-123-labelled 2-carbomethoxy-3-(4iodophenyl)tropane in human brain. Eur J Nucl Med 1993;20:7836.
Seibyl JP, Marek KL, Quinlan D, Sheff K, Zoghbi S, Zea-Ponce
Y, et al. Decreased single-photon emission computed tomographic
[123I]-CIT striatal uptake correlates with symptom severity in
Parkinsons disease. Ann Neurol 1995;38:58998.
Kung HF, Kim HJ, Kung MP, Meegalla SK, Plssl K, Lee HK.
Imaging of dopamine transporters in humans with technetium99m TRODAT-1. Eur J Nucl Med 1996;23:152730.
Booij J, Tissingh G, Winogrodzka A, van Royen EA. Imaging of
the dopaminergic neurotransmission system using single-photon
emission tomography and positron emission tomography in
patients with parkinsonism. Eur J Nucl Med 1999;26:17182.
Innis RB, Seibyl JP, Scanley BE, Laruelle M, Abi-Dargham A,
Wallace E, et al. Single photon emission computed tomographic
imaging demonstrates loss of striatal dopamine transporters in
Parkinson disease. Proc Natl Acad Sci U S A 1993;90:119659.
Booij J, Tissingh G, Winogrodzka A, Boer GJ, Stoof JC, Wolter
EC, et al. Practical benefit of [123I]FP-CIT SPET in the
demonstration of the dopaminergic deficit in Parkinsons disease.
Eur J Nucl Med 1997;24:6871.
Staffen W, Mair A, Unterrainer J, Trinka E, Ladurner G.
Measuring the progression of idiopathic Parkinsons disease with
[123I] beta-CIT SPECT. J Neural Transm 2000;107:54352.
Benamer TS, Patterson J, Grosset DG, Booij J, de Bruin K, van
Royen E, et al. Accurate differentiation of parkinsonism and
essential tremor using visual assessment of [123I]-FP-CIT SPECT
imaging: the [123I]-FP-CIT study group. Mov Disord 2000;15:503
10.
Asenbaum S, Brcke T, Pirker W, Podreka I, Angelberger P,
Wenger S, et al. Imaging of dopamine transporters with iodine123--CIT and SPECT in Parkinsons disease. J Nucl Med
1997;38:16.
Verhoeff NP, Kapucu O, Sokole-Busemann E, van Royen EA,
Janssen AG. Estimation of dopamine D2 receptor binding
potential in the striatum with iodine-123-IBZM SPECT: technical
and interobserver variability. J Nucl Med 1993;34:207684.
Ichise M, Meyer JH, Yonekura Y. An introduction to PET and
SPECT neuroreceptor quantification models. J Nucl Med
2001;42:75563.
Habraken JB, Booij J, Slomka P, Sokole EB, van Royen EA.
Quantification and visualization of defects of the functional
dopaminergic system using an automatic algorithm. J Nucl Med
1999;40:10917.
Booij J, Speelman JD, Horstink MWIM, Wolters EC. The clinical
benefit of imaging striatal dopamine transporters with [123I]FPCIT SPET in differentiating patients with presynaptic parkinsonism from those with other forms of parkinsonism. Eur J Nucl Med
2001;28:26672.
Seibyl JP, Marek K, Sheff K, Baldwin RM, Zoghbi S, Zea-Ponce
Y, et al. Test/retest reproducibility of iodine-123-betaCIT SPECT
brain measurement of dopamine transporters in Parkinsons
patients. J Nucl Med 1997;38:14539.
Baulieu JL, Ribeiro MJ, Levilon-Prunier C, Tranquart F, Chartier
JR, Guilloteau D, et al. Effects of the method of drawing regions
of interest on the differential diagnosis of extrapyramidal
syndromes using 123I-iodolisuride SPET. Nucl Med Commun
1999;20:7784.
Badiavas K, Iakovou I, Molyvda E, et al. Creation of a
customized normal database for 123I-FP-CIT SPECT studies in
a major nuclear medicine department in Northern Greece. Eur J
Nucl Med Mol Imaging 2010;37 Suppl 2:S399.
Seppi K, Schocke MFH. An update on conventional and advanced
magnetic resonance imaging techniques in the differential diagnosis of neurodegenerative parkinsonism. Curr Opin Neurol
2005;18:3705.

Eur J Nucl Med Mol Imaging (2011) 38:764773

36. Koole M, Laere KV, de Walle RV, Vandenberghe S, Bouwens L,
Lernahieu I, et al. MRI guided segmentation and quantification of
SPECT images of the basal ganglia: a phantom study. Comput
Med Imaging Graph 2001;25:16572.
37. Rojas GM, Raff U, Quintana JC, Huete I, Hutchinson M. Image
fusion in neuroradiology: three clinical examples including MRI of
Parkinson disease. Comput Med Imaging Graph 2007;31:1727.
38. van der Wee NJ, van Veen JF, Stevens H, van Vliet IM, van Rijk
PP, Westenberg HG. Increased serotonin and dopamine transporter
binding in psychotropic medication-nave patients with generalized
social anxiety disorder shown by 123I--(4-iodophenyl)-tropane
SPECT. J Nucl Med 2008;49:75763.
39. Fleming JS, Bolt L, Stratford JS, Kemp PM. The specific uptake
size index for quantifying radiopharmaceutical uptake. Phys Med
Biol 2004;49:N22734.
40. Tossici-Bolt L, Hoffmann SMA, Kemp PM, Mehta RL, Fleming
JS. Quantification of [123I]FP-CIT SPECT brain images: an
accurate technique for measurement of the specific binding ratio.
Eur J Nucl Med Mol Imaging 2006;33:14919.
41. Dobbeleir AA, Hambye AE, Vervaet AM, Ham HR. Quantification of iodine-123-FP-CIT SPECT with a resolution-independent
method. World J Nucl Med 2005;4(4):25261.
42. Goethals I, Dobbeleir A, Ham H, Santens P, DAsseler Y. Validation
of a resolution-independent method for the quantification of 123IFP-CIT SPECT scans. Nucl Med Commun 2007;28:7714.
43. Lkkegaard A, Werdelin LM, Friberg L. Clinical impact of
diagnostic SPET investigations with a dopamine re-uptake ligand.
Eur J Nucl Med Mol Imaging 2002;29:16239.
44. Morton RJ, Guy MJ, Marshall CA, Clarke EA, Hinton PJ.
Variation of DaTSCAN quantification between different gamma
camera types. Nucl Med Commun 2005;26:11317.
45. Morton RJ, Guy MT, Clauss R, Hinton PJ, Marshall CA, Clarke
EA. Comparison of different methods of DatSCAN quantification.
Nucl Med Commun 2005;26:113946.
46. Lee JD, Huang CH, Weng YH, Lin KJ, Chen CT. An automatic
MRI/SPECT registration algorithm using image intensity and
anatomical feature as matching characters: application on the
evaluation of Parkinsons disease. Nucl Med Biol 2007;34:44757.
47. Lee JD, Chen CW, Huang CH. Computer-aided evaluation system
for Parkinsons disease using image registration and labeling.
Proceedings of the 29th Annual International Conference of the
IEEE EMBS. 2007: 8447.
48. Zubal IG, Early M, Yuan O, Jennings D, Marek K, Seibyl JP.
Optimized, automated striatal uptake analysis applied to SPECT brain
scans of Parkinsons disease patients. J Nucl Med 2007;48:85764.
49. Calvini P, Rodriguez G, Inguglia F, Mignone A, Guerra UP,
Nobili F. The basal ganglia matching tools package for striatal
uptake semi-quantification: description and validation. Eur J Nucl
Med Mol Imaging 2007;34:124053.
50. Talairach J, Tournoux P. Co-planar stereotaxic atlas of the human
brain. New York: Thieme; 1988.
51. Radau PE, Linke R, Slomka PJ, Tatsch K. Optimization of
automated quantification of 123I-IBZM uptake in the striatum
applied to parkinsonism. J Nucl Med 2000;41:2207.

773
52. Ppperl G, Radau P, Linke R, Hahn K, Tatsch K. Diagnostic
performance of a 3-D automated quantification method of
dopamine D2 receptor SPECT studies in the differential diagnosis
of parkinsonism. Nucl Med Commun 2005;26:3943.
53. Koch W, Radau PE, Hamann C, Tatsch K. Clinical testing of an
optimized software solution for an automated, observerindependent evaluation of dopamine transporter SPECT studies.
J Nucl Med 2005;46:110918.
54. Hamilton D, List A, Butler T, Hogg S, Cawley M. Discrimination
between parkinsonian syndrome and essential tremor using
artificial neural network classification of quantified DaTSCAN
data. Nucl Med Commun 2006;27:93944.
55. Friston KJ, Ashburner J, Frith CD, Poline JB, Heather JD,
Frackowiak RS. Spatial registration and normalization of images.
Hum Brain Mapp 1995;3:16589.
56. Friston KJ, Holmes AP, Worsley KJ, Poline JP, Frith CD,
Frackowiak RS. Statistical parametric maps in functional imaging:
a general linear approach. Hum Brain Mapp 1994;2:189210.
57. Scherfler C, Seppi K, Donnemiller E, Goebel G, Brenneis C,
Virgolini I, et al. Voxel-wise analysis of [123I]-CIT SPECT
differentiates the Parkinson variant of multiple system atrophy
from idiopathic Parkinsons disease. Brain 2005;128:160512.
58. Seppi K, Scherfler C, Donnemiller E, Virgolini I, Schocke MF,
Goebel G, et al. Topography of dopamine transporter availability
in progressive supranuclear palsy: a voxelwise [123I]-CIT
SPECT analysis. Arch Neurol 2006;63:115460.
59. Colloby SJ, OBrien JT, Fenwick JD, Firbank MJ, Burn DJ,
McKeith IG, et al. The application of statistical parametric
mapping to 123I-FP-CIT SPECT in dementia with Lewy bodies,
Alzheimers disease and Parkinsons disease. Neuroimage
2004;23(3):95666.
60. Minoshima S, Koeppe RA, Frey KA, Kuhl DE. Anatomic
standardization: linear scaling and nonlinear warping of functional
brain images. J Nucl Med 1994;35:152837.
61. Hosaka K, Ishii K, Sakamoto S, Sadato N, Fukuda H, Kato T, et
al. Validation of anatomical standardization of FDG PET images
of normal brain: comparison of SPM and NEUROSTAT. Eur J
Nucl Med Mol Imaging 2005;32:927.
62. Takada S, Yoshimura M, Shindo H, Saito K, Koizumi K, Utsumi
H, et al. New semiquantitative assessment of 123I-FP-CIT by an
anatomical standardization method. Ann Nucl Med 2006;20
(7):47784.
63. Buchert R, Berding G, Wilke F, Martin B, von Borczyskowski D,
Mester J, et al. IBZM tool: a fully automated expert system for the
evaluation of IBZM SPECT studies. Eur J Nucl Med Mol Imaging
2006;33:107383.
64. Acton PD, Mozley PD, Kung HF. Logistic discriminant
parametric mapping: a novel method for the pixel-based
differential diagnosis of Parkinsons disease. Eur J Nucl Med
1999;26:141323.
65. Habraken JB, Booji J, Slomka P, Sokole EB, van Royen EA.
Quantification and visualization of defects of the functional
dopaminergic system using an automated algorithm. J Nucl Med
1999;40:10917.