AJR:187, September 2006 825

AJR 2006; 187:825829

0361803X/06/1873825
American Roentgen Ray Society
M
E D I C A L

I M A G I N G
A
C E N
T
U
R Y O
F
M
E D I C A L

I M A G I N G
A
C E N
T
U
R Y O
F
Nakahara et al.
CT Thallium-201 SPECT
for Cancer Detection
Head and Neck I magi ng Or i gi nal Res earch
Value of CT Thallium-201
SPECT Fusion Imaging over
SPECTAlone for Detection and
Localization of Nasopharyngeal
and Maxillary Cancers
Tadaki Nakahara
1
Naoyuki Shigematsu
1
Masato Fujii
2
Etsuo Kunieda
1
Takayuki Suzuki
1
Chikako Tanaka
3
Jun Hashimoto
1
Atsushi Kubo
1
Nakahara T, Shigematsu N, Fujii M, et al.
Keywords: CT, head and neck imaging, image fusion,
maxillary cancer, nasopharyngeal cancer, nuclear
imaging, SPECT, thallium-201
DOI:10.2214/AJR.05.0617
Received April 9, 2005; accepted after revision
July 22, 2005.
1
Department of Radiology, Keio University School of
Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo, Japan
160-8582. Address correspondence to T. Nakahara
(n-tadaki0909@k6.dion.ne.jp).
2
Department of Otolaryngology, Keio University School of
Medicine, Tokyo, Japan.
3
Department of Radiology, Tokyo Metropolitan Komagome
Hospital, Tokyo, Japan.
OBJECTIVE. The purpose of this study was to investigate the incremental clinical utility
of CT and high-resolution SPECT fusion imaging.
MATERIALS AND METHODS. Eighteen patients with nasopharyngeal cancer or
cancers around the maxilla were scanned with high-resolution SPECT at the time of initial
diagnosis (18 studies) and during follow-up after chemoradiotherapy (23 studies). SPECT
results were compared with histologic findings or the findings of other imaging tech-
niques. In addition, automatic image registration without fiducial markers was performed
from CT and SPECT data, and the effect of fusion imaging on the localization of abnor-
malities was evaluated.
RESULTS. All of the original 18 untreated lesions showed high uptake. Recurrent tu-
mors had a tendency to show high uptake (seven of nine patients), whereas little or no uptake
generally represented no recurrence (12 of 14 patients) (chi-square test with Yates correction:

2
= 6.80, p < 0.01). In two patients, physiologic uptake in the unilateral prevertebral muscle
was revealed on image fusion. In four of the nine recurrent nasopharyngeal cancers (44%),
SPECT alone could not determine abnormalities in uptake sites, whereas CT/SPECT fusion
imaging clearly localized the sites and was helpful for treatment strategy.
CONCLUSION. High-resolution thallium-201 (
201
Tl) SPECT has a very high detection
rate in patients with nasopharyngeal cancer and cancers around the maxilla. However, the an-
atomic identification or localization of the uptake sites is sometimes difficult without
CT/SPECT fusion imaging. This technique without external markers is practically feasible to
generate clinically valid fusion images.
hallium-201 (
201
Tl) SPECT has
unique features that enable it to
reveal metabolically active tissue
by virtue of its cellular uptake by
malignant cells. This technique has, in fact,
been used for tumor diagnosis. Although it
has been claimed to be of limited utility,
given the poor anatomic localization af-
forded by its functional imaging techniques,
nuclear SPECT is still widely available, to-
gether with recent advances in CT/SPECT
fusion imaging [1].
Fusion imaging of structural and func-
tional data may appear problematic in a clin-
ical setting. With SPECT, which has poor
spatial resolution, tumor uptake is ill de-
fined, and the size of the lesion uptake is
much larger than true lesion size. Therefore,
without high-resolution SPECT data, even
CT/SPECT fusion imaging may not provide
accurate positional identification, especially
in the adjacent small structures of the head
and neck. Improved SPECT is an available
option for evaluating the accuracy of image
fusion and for generating clinically valid fu-
sion images.
Because the scan area includes the supra-
clavicular region when evaluating head and
neck cancers, it is unavoidable that the scan
includes a large field of view and a long ro-
tation radius. As a result, spatial resolution is
remarkably deteriorated. The limitation de-
scribed was overcome by using techniques
adapted from Togawa et al. [2], which are
similar to our methods. Moreover, these
techniques can coregister CT images with-
out fiducial markers in the same fashion as
brain image fusion.
In our study, we retrospectively investi-
gated the incremental clinical utility of CT
high-resolution SPECT fusion imaging over
SPECT alone and the detection rates of
SPECT in patients with nasopharyngeal can-
cer and cancers around the maxilla.
T
D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
a
j
r
o
n
l
i
n
e
.
o
r
g

b
y

1
1
8
.
9
6
.
2
0
9
.
2
0
9

o
n

0
6
/
1
5
/
1
4

f
r
o
m

I
P

a
d
d
r
e
s
s

1
1
8
.
9
6
.
2
0
9
.
2
0
9
.

C
o
p
y
r
i
g
h
t

A
R
R
S
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d

Nakahara et al.
826 AJR:187, September 2006
Materials and Methods
Patients
From April 2002 to September 2004, 41 cases
were studied with high-resolution
201
Tl SPECT in
the detection of primary and recurrent head and
neck cancers above the level of the maxilla. Patients
with no distant metastasis in which localized treat-
ment such as radiation therapy was scheduled were
included in this study. Fifteen men and three
women (mean age, 56 years; range, 2092 years)
with nasopharyngeal cancer (n = 14) or cancers
around the maxilla (n = 4) participated. Eighteen
studies were performed soon after the initial diag-
nosis; 23 studies were done during the follow-up
period. Two patients with nasopharyngeal cancer
were excluded from this study because we could
not obtain DICOM CT data to make fusion images.
All patients underwent CT and
201
Tl SPECT at
the time of initial and follow-up imaging. In our
study, we used CT as a tool to precisely localize ab-
normal uptake on image fusion, rather than for its
diagnostic ability. Therefore, the time interval be-
tween corresponding CT and SPECT scans was not
so restricted (interval within 2 months).
SPECT, CT, and Image Fusion
Thallium-201 SPECT was performed using a
triple-headed rotating gamma camera (GCA-
9300A, Toshiba Medical Systems) equipped with
low-energy, ultrahigh-resolution fan-beam colli-
mators. The energy peak and window level were
set at 71 keV 20%. SPECT scans were started
20 minutes after the IV injection of 148 MBq of
201
Tl chloride. All patients were scanned in the su-
pine position with external head restraints used in
the same fashion as was done in the brain SPECT.
The patients head was introduced into the interior
of a three-headed gamma camera as far as possible
so the nasopharynx was entirely within the scan-
ning field. Although this method can be applied
only to head and neck cancers above the maxilla,
it yields an in-plane spatial resolution of approxi-
Fig. 1Normal and abnormal findings on thallium-201
SPECT images.
A, 20-year-old man in complete remission from T3 N2
M0 nasopharyngeal carcinoma after
chemoradiotherapy. Modified brain
201
Tl SPECT image
shows no pathologic uptake in scan area including
nasopharynx. Physiologic uptake sites in scalp, nasal
cavity, salivary glands, palate, and prevertebral muscle
(arrows) are noted.
B, 60-year-old man with untreated T1 N0 M0
nasopharyngeal carcinoma. CT/SPECT fusion image
shows pathologic uptake in right side of nasopharynx
(arrowhead). Physiologic uptake in nasal cavity,
muscle, parotid gland, and scalp is helpful to confirm
image registration accuracy (CT in gray-scale and
thallium uptake in color).
A B
A B
Fig. 268-year-old man with local recurrence of T4
N2c M0 nasopharyngeal carcinoma 10 months after
chemoradiotherapy.
A, Contrast-enhanced CT scan shows nasopharyngeal
wall thickening on left side. No remarkable change of
structural abnormality is seen when compared with CT
performed 6 months after therapy.
B, CT/SPECT fusion image performed 10 months after
chemoradiotherapy shows moderate uptake in
thickened wall (CT in gray-scale and thallium uptake
in color).
D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
a
j
r
o
n
l
i
n
e
.
o
r
g

b
y

1
1
8
.
9
6
.
2
0
9
.
2
0
9

o
n

0
6
/
1
5
/
1
4

f
r
o
m

I
P

a
d
d
r
e
s
s

1
1
8
.
9
6
.
2
0
9
.
2
0
9
.

C
o
p
y
r
i
g
h
t

A
R
R
S
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d

CT Thallium-201 SPECT for Cancer Detection
AJR:187, September 2006 827
mately 7 mm, which is comparable to the resolu-
tion obtained with PET in the late 1990s. Image
data were obtained for approximately 35 minutes
in 360 rotation with 90 projections in steps of 4
degrees. The matrix size for data acquisition and
image reconstruction was 1.6 mm (128 128).
Transaxial slices 1.6 mm thick were processed us-
ing a Butterworth filter (order, 8; cutoff frequency,
0.18 cycles/pixel) and reconstructed using the or-
dered-subset expectation maximization algo-
rithm. As mentioned previously, full-width at
half-maximum was approximately 7 mm in air at
a distance of 13 cm. Neither scatter nor attenua-
tion correction was performed. Reconstructed im-
ages were displayed at appropriate window levels
to discard background noise.
CT data used in our study were obtained in rou-
tine clinical practice (contrast-enhanced) or to
gather information for determining the radiation
field (unenhanced). CT of the neck was performed
from the orbit to the thoracic inlet with 5-mm col-
limation and a pitch of 1. Contiguous transaxial
CT images 5 mm thick were obtained with a
512 512 matrix. CT data were transferred to a
nuclear imaging workstation in DICOM format.
SPECT and CT data were processed using a med-
ical image processor (GMS-5500A/PI, Toshiba).
SPECT images were automatically resliced using
commercially available software (Automatic Reg-
istration Tool, Toshiba) to correspond to CT im-
ages. Ardekani et al. [3] have described the regis-
tration theory, which rests on the assumption that
uniform regions shown with one technique repre-
sent uniform findings in corresponding regions on
another technique, allowing segmentation of the
CT images into eight ranges of voxel intensity.
Eighty to 130 segments are automatically selected
to perform optimal registration so that the SPECT
voxel intensity variance is minimized. Fiducial
markers are not required for image registration,
and registration error does not exceed 3 mm. Au-
tomatic registration takes less than 5 minutes on
our workstation.
SPECT Image Interpretation
Before interpretation of the
201
Tl SPECT im-
ages, the evaluating nuclear medicine physician
was informed of the patients treatment history (i.e.,
no treatment or follow-up after treatment) but was
not given details of the disease, including the pres-
ence or absence of viable tumor, location, size, or
extent. Tumor uptake was classified as marked,
moderate, slight, or no evidence of tumor uptake. In
addition, the location where abnormal uptake was
most likely to be visualized was recorded before
and after CT/SPECT fusion imaging.
Diagnostic Criteria
Figure 1A shows normal
201
Tl SPECT images at
the superior level of the head and neck. These were
obtained 26 months after chemoradiotherapy in a
patient with stage T3 N2 M0 nasopharyngeal carci-
noma. More than 3 years later, the patient was still
in complete remission. Figure 1B shows SPECT
images with abnormalities in a patient with T1 N0
M0 nasopharyngeal carcinoma in whom abnor-
mally increased uptake is shown in the right na-
sopharyngeal wall. As shown in Figure 1A, it is
common to find nonpathologic uptake in the scalp,
nasal cavity, salivary glands, and ocular muscles,
whereas bone uptake is barely visible [4]. Also,
prevertebral muscles and masseters are frequently
visualized. It is not rare to see uptake in the palate
at the level of the maxilla.
Concerning initial diagnosis, all suspicious le-
sions were confirmed to be malignant by biopsy.
Tumor location was determined by fiberscopic ex-
amination. In the case of recurrent tumors, histo-
logic results obtained within 4 weeks before or after
SPECT were used for diagnosis. If histopathology
was not obtained or if biopsy specimens were insuf-
ficient for definitive diagnosis, the diagnosis and
location of recurrences were established by follow-
up imaging if the scans showed disease progression
within 12 months. If follow-up imaging showed le-
sion regression after a minimum 6-month interval
or lesion stability for a minimum of 1 year, or it
failed to show any evidence of masses, the patients
were considered to have a benign condition.
Statistical Analysis
Patients were classified into two groups accord-
ing to SPECT findings (because of the small num-
ber of patients): those with SPECT showing
marked or moderate uptake and those with SPECT
showing little or no uptake. The groups pathologic
A B C
Fig. 364-year-old woman with local recurrence of T4 N1 M0 nasopharyngeal carcinoma 6 months after chemoradiotherapy.
A, Thallium-201 SPECT scan reveals moderate uptake near right side of nasopharynx (arrow).
B, CT/SPECT fusion image shows uptake to be mainly located in clivus (CT in gray-scale and thallium uptake in color).
C, SPECT scan after stereotactic radiosurgery to clival lesion shows remarkably reduced uptake in recurrent lesion.
D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
a
j
r
o
n
l
i
n
e
.
o
r
g

b
y

1
1
8
.
9
6
.
2
0
9
.
2
0
9

o
n

0
6
/
1
5
/
1
4

f
r
o
m

I
P

a
d
d
r
e
s
s

1
1
8
.
9
6
.
2
0
9
.
2
0
9
.

C
o
p
y
r
i
g
h
t

A
R
R
S
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d

Nakahara et al.
828 AJR:187, September 2006
findings were compared using the chi-square test
with Yates correction.
Results
Detection and Localization of
Untreated Primary Tumors
All 18 original, untreated lesions showed
marked or moderate uptake, resulting in a de-
tection rate of 100%. Most of the lesions were
large (mean, 42 21 mm; range, 1380 mm),
although SPECT also detected two T1 na-
sopharyngeal cancers [5]. SPECT images in
one of the two T1 tumors are shown in
Figure 1B. Abnormal uptake sites in all na-
sopharyngeal cancers could be accurately lo-
calized on SPECT images (14 of 14 patients),
whereas the locations of 50% of cancers
around the maxilla (two of four patients)
could not be determined on SPECT alone.
Detection and Localization of
Recurrent Primary Tumors
Fifteen patients were studied with
201
Tl
SPECT during the follow-up period. Because
some patients underwent follow-up SPECT
several times, a total of 23 SPECT studies
were evaluated. Diagnostic accuracy was
based on pathologic results or clinical follow-
up in 19 of the 23 SPECT studies; the remain-
ing four studies showed abnormal uptake in
bone where biopsy could not be performed.
There were nine nasopharyngeal cancer re-
currences and 14 complete remissions at the
time of SPECT examination. Recurrent tumors
had a tendency to show marked or moderate
uptake (seven of nine patients), whereas little
or no uptake generally represented no recur-
rence (12 of 14 patients) (chi-square test with
Yates correction:
2
= 6.80, p < 0.01). In two
patients, marked or moderate uptake was visu-
alized unilaterally in the nasopharynx. In these
patients, CT/SPECT fusion imaging con-
firmed uptake in the normal prevertebral mus-
cle. In the nine recurrent nasopharyngeal can-
cers, five (56%) showed abnormal uptake in
the swollen nasopharyngeal wall. Figure 2
shows a patient in whom nasopharyngeal wall
thickening barely changed 10 months after
chemoradiotherapy; however, a histologically
confirmed local recurrence was identified on
SPECT. In this patient, chemotherapy using
TS-1 was initiated after the SPECT study.
SPECT alone was unable to localize the abnor-
mal uptake site in four other recurrent tumors
(44%), whereas CT/SPECT fusion imaging lo-
calized three recurrences, mainly in the clivus.
Figure 3 shows a patient who underwent ster-
eotactic radiation therapy for a recurrent lesion
in the clivus. In another patient, fusion imaging
was useful in localizing a recurrent lesion in an
unexpected site (Fig. 4).
Discussion
Thallium-201 can be used to assess many
malignant tumors including aerodigestive
cancers. We have previously reported that
201
Tl uptake may be associated with the vi-
ability of esophageal cancer, and that its
chemoradiotherapeutic effect can be evalu-
ated with
201
Tl SPECT [6]. Thalium-201
SPECT has been reported to have a very
high accuracy in the detection of head and
neck cancers [79], which is consistent with
our results. Therefore,
201
Tl SPECT may be
useful in addition to established imaging
techniques such as CT or MRI.
As mentioned in our previous study, in
which we applied SPECT to esophageal can-
cer, the various limitations lowered the effec-
tiveness of SPECT in tumor diagnosis, in-
cluding relatively poor spatial resolution,
noise, and strong physiologic uptake near the
tumor [6]. In contrast, the method used in the
present study is advantageous for the follow-
ing reasons: the short distance between the
patients head and the collimator can lead to
high resolution and sensitivity, SPECT with a
fan-beam collimator provides higher resolu-
tion than that using a parallel-hole collimator
without significant loss of sensitivity, and
there are no physiologic uptake sites other
than prevertebral muscle near the nasophar-
ynx or maxilla.
Various reports describing CT/SPECT fu-
sion imaging can be found in the literature
[1012]. In these reports, external markers
were used for images registration; this differs
from our method. In the case of CT/SPECT,
patients are equipped with external markers for
A B C
Fig. 454-year-old man with cervical spine recurrence (or metastasis) of T4 N0 M0 nasopharyngeal carcinoma 11 months after chemoradiotherapy.
A, Thallium-201 SPECT scan shows moderate uptake near left posterior side of nasopharynx (arrow). Uptake site cannot be determined on SPECT scan alone.
B, CT/SPECT fusion image shows that uptake is located in left lateral mass of first cervical spine (CT in gray-scale and thallium uptake in color).
C, T1-weighted MR image after contrast enhancement confirms recurrence (or metastasis) in same location (arrow) as shown in B.
D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
a
j
r
o
n
l
i
n
e
.
o
r
g

b
y

1
1
8
.
9
6
.
2
0
9
.
2
0
9

o
n

0
6
/
1
5
/
1
4

f
r
o
m

I
P

a
d
d
r
e
s
s

1
1
8
.
9
6
.
2
0
9
.
2
0
9
.

C
o
p
y
r
i
g
h
t

A
R
R
S
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d

CT Thallium-201 SPECT for Cancer Detection
AJR:187, September 2006 829
CT, and they receive radiation exposure only
for image fusion. CT is the first choice and is
necessary for evaluating head and neck can-
cers. However, in clinical practice it seems
somewhat difficult to perform routine CT ex-
aminations and successive SPECT studies in
keeping with external markers. Although
coregistration inaccuracy is a potential prob-
lem with CT/SPECT fusion images without fi-
ducial markers, algorithms permitting image
fusion in the absence of markers have been in-
tensively developed for brain imaging [3,
1315]. Organs around the base of the skull are
rigid; thus, image fusion accuracy in this re-
gion is comparable to that of a phantom study
(error < 3 mm). In addition, high spatial reso-
lution and certain landmarks that show physi-
ologic uptake can enhance its accuracy.
In the present study, we did not focus on di-
rect comparison of SPECT with other imag-
ing techniques. Our method is insufficient for
the evaluation of head and neck cancers be-
cause of the limited scan area. As shown in
Figure 1B, in most cases with untreated can-
cers, CT alone can detect abnormalities that
are suspect for malignancy. In contrast, it is
sometimes difficult to detect recurrent tumors
because of structural abnormalities persisting
long after therapy (Figs. 2 and 3). Further-
more, some patients with untreated advanced
cancers with bone destruction extending to
the clivus had CT after chemoradiotherapy
that showed somewhat improved nasopharyn-
geal wall thickening but persistent clival de-
struction. According to clinical records of di-
agnostic CT and our retrospective reviews, it
was difficult to differentiate residual viable
tumors from soft tissues with no viable cells
in that morphologically changed area. Tha-
lium-201 SPECT can detect viable tumors
even after therapy [6]; thus, CT/SPECT fu-
sion images may have significant clinical im-
pact. Indeed, treatment was changed in the
patients described in Figures 24.
Fluorine-18-FDG PET is useful for evalu-
ating head and neck cancers. In our study, it
took approximately 35 minutes to produce
high-resolution SPECT images, which is
longer than the scanning time required to ob-
tain high-resolution whole-body tomographic
images using recent PET devices. In this re-
gard, PET seems methodologically superior
to
201
Tl SPECT. One advantage of
201
Tl
SPECT over
18
F-FDG PET may be that brain
uptake is negligible in SPECT scans, whereas
it is very high in
18
F-FDG PET scans. This
makes evaluation around the clivus easier
with
201
Tl SPECT than with
18
F-FDG PET
[9]. Indeed, recurrence in the clivus is not
rare. In PET/CT, the high uptake of
18
F-FDG
in the brain has no effect on the detection of
tumor in the clivus or any other regions in the
head and neck. However, false-positive
18
F-
FDG PET results have been reported in pa-
tients with nasopharyngeal cancer after radia-
tion therapy [16], which may not be able to be
resolved even in PET/CT. We believe that the
appropriate selection of imaging techniques
can enhance the utility of high-resolution
201
Tl SPECT combined with CT in patients
with nasopharyngeal cancer or cancers
around the maxilla.
In summary,
201
Tl accumulated well in na-
sopharyngeal cancer and cancers around the
maxilla, resulting in a very high sensitivity on
high-resolution SPECT. The
201
Tl uptake was
significantly higher in recurrent tumors than
in benign conditions. However, the anatomic
identification or localization of the uptake site
was sometimes unclear without CT/SPECT
image fusion. Fusion imaging without exter-
nal markers was practically feasible to gener-
ate clinically valid images and was useful not
only for excluding physiologic uptake, as
shown in prevertebral muscles, but also for lo-
calizing recurrent tumor.
References
1. Keidar Z, Israel O, Krausz Y. SPECT/CT in tumor
imaging: technical aspects and clinical applica-
tions. Semin Nucl Med 2003; 3:205218
2. Togawa T, Yui N, Kinoshita F, Shimada F,
Omura K, Takemiya S. Visualization of nasopha-
ryngeal carcinoma with Tl-201 chloride and a three-
head rotating gamma camera SPECT system. Ann
Nucl Med 1993; 7:105113
3. Ardekani BA, Braun M, Hutton BF, Kanno I,
Iida H. A fully automatic multimodality image reg-
istration algorithm. J Comput Assist Tomogr 1995;
19:615623
4. Schomacker K, Schicha H. Use of myocardial im-
aging agents for tumour diagnosis: a success story?
Eur J Nucl Med 2000; 27:18451863
5. Nakahara T. Tumor diagnosis in nuclear medicine:
clinical usefulness and application to radiation ther-
apy [in Japanese]. (letter) Jpn J Clin Radiol 2005;
50:212214
6. Nakahara T, Togawa T, Nagata M, et al. Compari-
son of barium swallow, CT and thallium-201
SPECT in evaluating responses of patients with
esophageal squamous cell carcinoma to preopera-
tive chemoradiotherapy. Ann Nucl Med 2003;
17:583591
7. Gregor RT, Valdes-Olmos R, Koops W, et al. Pre-
liminary experience with thallous chloride Tl 201-
labeled single-photon emission computed tomogra-
phy scanning in head and neck cancer. Arch Oto-
laryngol Head Neck Surg 1996; 122:509514
8. van Veen SA, Balm AJ, Valdes-Olmos RA, et al.
Occult primary tumors of the head and neck: accu-
racy of thallium 201 single-photon emission com-
puted tomography and computed tomography
and/or magnetic resonance imaging. Arch Oto-
laryngol Head Neck Surg 2001; 127:406411
9. Mukherji SK, Gapany M, Phillips D, et al. Thal-
lium-201 single-photon emission CT versus CT for
the detection of recurrent squamous cell carcinoma
of the head and neck. AJNR 1999; 20:12151220
10. Scott AM, Macapinlac H, Zhang J, et al. Image reg-
istration of SPECT and CT images using an external
fiduciary band and three-dimensional surface fitting
in metastatic thyroid cancer. J Nucl Med 1995;
36:100103
11. Chajari M, Lacroix J, Peny AM, et al. Gallium-67
scintigraphy in lymphoma: is there a benefit of im-
age fusion with computed tomography? Eur J Nucl
Med Mol Imaging 2002; 29:380387
12. Yamamoto Y, Nishiyama Y, Monden T, Matsumura
Y, Satoh K, Ohkawa M. Clinical usefulness of fu-
sion of
131
SPECT and CT images in patients with
differentiated thyroid carcinoma. J Nucl Med 2003;
44:19051910
13. Thurfjell L, Lau YH, Andersson JLR, Hutton BF.
Improved efficiency for MRI-SPECT registration
based on mutual information. Eur J Nucl Med 2000;
27:847856
14. Friston KJ, Ashburner J, Frith CD, Poline J-B,
Heather JD, Frackowiak RSJ. Spatial registration
and normalisation of images. Human Brain Map-
ping 1995; 2:165189
15. Woods RP, Mazziotta JC, Cherry SR. MRI-PET
registration with automated algorithm. J Comput
Assist Tomogr 1993; 17:536546
16. Liu S-H, Chang JT, NG S-H, Chan S-C, Yen T-C.
False-positive fluorine-18 fluorodeoxy-D-glucose
positron emission tomography finding caused by
osteoradionecrosis in a nasopharyngeal carcinoma
patient. Br J Radiol 2004; 77:257260 D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
a
j
r
o
n
l
i
n
e
.
o
r
g

b
y

1
1
8
.
9
6
.
2
0
9
.
2
0
9

o
n

0
6
/
1
5
/
1
4

f
r
o
m

I
P

a
d
d
r
e
s
s

1
1
8
.
9
6
.
2
0
9
.
2
0
9
.

C
o
p
y
r
i
g
h
t

A
R
R
S
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
;

a
l
l

r
i
g
h
t
s

r
e
s
e
r
v
e
d