Keywords Bone scintigraphy - fluorodeoxyglucose positron emission tomography/computed tomography
- osteoblastic - osteolytic - semiquantitative
Introduction
Bone is one of the most common sites for metastatic spread of tumors. Tumors most
commonly metastasizing to bone are prostate, breast, kidney, lung, and thyroid. In
children, common causes of skeletal metastases include neuroblastoma, Ewing sarcoma,
and osteosarcoma. In men, carcinoma of the prostate accounts for 60% of bone metastases,
while in women, breast cancer accounts for 70% of such metastases. Metastases typically
involve the axial skeleton, which is the region rich in red bone marrow. Bone metastases
could be purely osteoblastic, mixed osteoblastic/osteolytic, or osteolytic. Prostate
cancer metastases are purely osteoblastic, whereas metastases of thyroid and kidney
carcinomas are purely lytic. Mixed osteolytic/osteoblastic lesions occur in carcinomas
of the breast, lung, cervix, ovary, and testis.
Bone metastasis can be identified only when the distortion of the compact bone structure
in the direct X-ray and computed tomography (CT) reach a certain level. Bone scintigraphy
is the easiest and cheapest way to scan the whole body with a higher sensitivity than
the specificity.[1 ],[2 ],[3 ],[4 ],[5 ],[6 ] Whole body bone scan allows scanning the whole skeletal system. Single photon emission
CT (SPECT), particularly SPECT/CT, further increases the sensitivity and specificity
of bone scintigraphy for the detection of bone metastases. F-18 fluorodeoxyglucose
(FDG) positron emission tomography/CT (PET/CT) imaging is commonly used in oncology
for staging of the tumors as well as detecting recurrences and assessing response
to treatments. While bone scintigraphy assesses the osteoblastic activity of bone
metastases, FDG PET scan evaluates the glucose metabolism/glycolysis of the lesions.
Bone scintigraphy has higher sensitivity in detecting osteoblastic bone metastases
than osteolytic ones.[7 ],[8 ] In detecting osteolytic bone metastases, FDG PET/CT has been reported to be superior
to bone scintigraphy.[3 ],[4 ],[7 ] Metastatic bone lesions can be detected in early stage on magnetic resonance imaging
(MRI) with the signal changes in the bone marrow which is hypointense in suppressed
T1 images and hyperintense in T2 images. Diffusion-weighted whole-body MRI was found
to be equivalent to bone scintigraphy and FDG PET/CT in assessing bone metastases
in non-small cell lung cancer.[9 ] There are various studies comparing FDG PET/CT with bone scintigraphy visually for
the detection of bone metastases.[7 ],[10 ],[11 ],[12 ],[13 ],[14 ],[15 ],[16 ],[17 ] However, to the best of our knowledge, there is no study determining and comparing
semiquantitative values of bone lesions on bone scintigraphy and FDG PET in patients
with malignancies.
In this study, we aimed to determine semiquantitative measurement values of osteolytic,
osteoblastic, and mixed-type bone lesions on bone scintigraphy and FDG PET/CT images
in patients with solid organ malignancies.
Materials and Methods
Bone scintigraphy and F-18 FDG PET/CT images of patients with various solid tumors
were selected for further analysis. This retrospective study was approved by the Ethics
Committee of Trakya University Faculty of Medicine.
For bone scintigraphy, the patients were injected 20–25 mCi (740–925 MBq) technetium-99
m methylene diphosphonate, and images were obtained 2–4 h after the injection. Images
included anterior and posterior whole body (10–15 cm/min scan speed), and SPECT (64
images for 20–40 s each), and spot (500–1000 kct) images of the area of interest.
Images were obtained at dual-head gamma camera (Siemens E. CAM, Erlangen, Germany
and Philips BrightView, Milpitas, CA, USA) using low-energy high-resolution collimator
with 120 keV energy settings and 20% window. 1024 × 512 matrix was used for whole
body images and 64 × 64 matrix for SPECT. Images were evaluated visually and semiquantitatively.
For semiquantitative analysis, a region of interest (ROI) was drawn over the bone
lesion and normal bone to obtain maximum lesion to normal bone count ratio (ROImax).
JETStream Workspace version 3.0 was used for this semiquantitative analysis.
For FDG PET/CT study, the patients fasted 6 h before imaging. Blood glucose level
was checked before FDG injection. The patients were given oral contrast1 h before
the study. FDG was injected when the blood glucose level was <150 mg/dl. PET/CT images
were obtained at GE discovery 8 PET/CT camera (GE Medical Systems, Waukesha, USA)
60 min following intravenous injection of 296–555 MBq (8–15 mCi) F-18 FDG. Before
PET image acquisition, a low-dose CT was obtained for attenuation correction and anatomic
localization purposes. PET acquisition was 3 min/bed from top of the head to mid thighs.
PET images were corrected for attenuation on the basis of the CT data and reconstructed
using a standard iterative algorithm and reformatted into transaxial, coronal, and
sagittal views. Maximum intensity projection images were also generated. Both attenuation
corrected and non-corrected PET images as well as PET/CT fusion images were visually
evaluated. Low-dose CT images were also assessed by a radiologist to determine osteoblastic,
osteolytic, and mixed lesions which are consistent or suspicious for bone metastases.
For semiquantitative analysis, maximum standardized uptake value (SUVmax ), and Hounsfield unit (HU) values were measured from the lesions.
Number Cruncher Statistical System 2007 and PASS 2008 Statistical Software (Utah,
USA) program were used for statistical analysis. Mann–Whitney U, Spearman's correlation
coefficient, Pearson's Chi-squared, and Kruskal–Wallis tests were used.
Results
A total of 33 patients were included in this study. Fifteen patients were female and
18 were male with an age range of 37–79 years (mean 60.09 ± 8.77). Patients had various
solid tumors including lung cancer (42.4%), breast cancer (30.4%), prostate cancer
(6.1%), and other (endometrial cancer, pancreatic cancer, malignant melanoma, parathyroid
tumor, renal cell carcinoma, soft-tissue sarcoma, and oral cavity tumor). There were
145 bone lesions (22.8% osteoblastic, 53.1% osteolytic, and 24.1% mixed osteoblastic-osteolytic
metastases) on CT. Distribution of bone metastases included pelvis (24.1%), lower
thoracic spine (17.9%), lumbar spine (16.6%), ribs and sternum (10.3%), lower limbs
(9.7%), upper thoracic spine (7.6%), upper limbs (4.8%), and other regions (9%).
Mean SUVmax of osteolytic bone lesions (7.73 ± 4.35) was higher than mean SUVmax of osteoblastic (6.84 ± 3.03) and mixed (6.88 ± 3.10) lesions, but it was not statistically
significant (0.876) [Table 1 ] and [Table 2 ].
Table 1 Kruskal-Wallis test comparing mean maximum standardized uptake value, region of interest,
and hounsfiled unit values in osteoblastic, osteolytic and mixed lesions
Table 2
Post hoc Mann-Whitney U-test results
Mean ROImax of osteoblastic bone lesions (6.42 ± 4.22) was higher than mean ROImax
of osteolytic lesions (5.33 ± 3.60), but it was not statistically significant (0.077)
[Table 1 ] and [Table 2 ]. Mean ROImax of mixed metastases was 6.32 ± 4.03.
Mean HU SUVmax of osteoblastic bone lesions (344.09 ± 140.62) was higher than mean HU of osteolytic
(233.39 ± 125.29) and mixed (254.86 ± 105.69) lesions and it was statistically significant
(P < 0.01) [Table 1 ] and [Table 2 ].
In osteoblastic metastases, there was no correlation between SUVmax and ROImax, SUVmax and HU, and ROImax and HU values [Table 3 ].
Table 3 Spearman's rho correlation analysis results in between Maximum standardized uptake
value, Maximum region of interest, and hounsfiled unit values
In osteolytic metastases, there was no correlation between SUVmax and ROImax, SUVmax and HU, and ROImax and HU values [Table 3 ].
In mixed metastases, there was no correlation between SUVmax and ROImax, SUVmax and HU, and ROImax and HU values [Table 3 ].
Discussion
Bone scintigraphy and FDG PET/CT imaging play an important role in the management
of patients with malignancies. Bone scintigraphy images, whole body, spot, or SPECT,
are usually assessed visually. Several studies have been published on semiquantitative
analysis of bone scintigraphy. Erdi et al. developed a semiautomated image segmentation
program to determine the total fraction of skeletal involvement with bone metastases.[18 ] Bone scan lesion area, bone scan lesion intensity, and bone scan lesion count were
calculated from identified lesions to determine response to treatment.[19 ] Regional activity concentration of the injected tracer was measured on SPECT images.[20 ] In our study, we obtained maximum uptake ratio of bone lesion to normal bone to
determine the degree of osteoblastic activity of the bone lesions.
SUVmax is a commonly used parameter on FDG PET/CT studies to assess the metabolic activity
of the lesions which can help to differentiate benign from malignant lesion and determine
the aggressiveness of the tumor. In a study by Cook et al., 81% of lytic bone metastases
showed increased FDG uptake; however, only 40% of sclerotic bone lesions were detected
on FDG PET.[21 ] Abe et al. found that FDG PET was superior to bone scintigraphy in detecting osteolytic
metastases, while bone scintigraphy was superior to FDG PET in detecting osteoblastic
lesions.[12 ] In a study by Hur et al., SUVmax was significantly higher in osteolytic metastasis than in osteoblastic lesions.[16 ] Cook et al. found that the FDG uptake of osteoblastic metastases (mean SUVmax : 0.95) was significantly lower than the FDG uptake of osteolytic metastases (mean
SUVmax : 6.77).[21 ]
As opposed to literature, in our study, we did not find a significant difference in
mean SUVmax and ROImax values of osteoblastic, osteolytic, and mixed lesions and there was also
no correlation between the SUVmax , ROImax, and HU values. Various factors may alter the mean SUVmax and ROImax values. In our cases, some of the lesions seen on CT could be active metastatic
disease and some inactive. For example, some of the sclerotic lesions on CT could
be active osteoblastic metastases and some could be treated old lesions. Increased
uptake on bone scan in a sclerotic lesion could be due to flare phenomenon in a treated
osteoblastic, osteolytic, or mixed metastases.[22 ] Lack of increased activity on bone scan in a sclerotic lesion may be due to treated
very old osteoblastic metastasis. Sclerotic changes on CT in a treated bone metastasis
may last longer than osteoblastic activity on bone scan. Increased uptake on bone
scan due to flare is usually not seen on follow-up bone scan at 6 mos.[23 ] Flare phenomenon in bone metastases has also been reported with FDG PET/CT study.[24 ] Development of fracture in a lytic lesion may cause increased uptake on bone scan.
Some sclerotic, lytic, or mixed lesions on CT may not be metastatic and could be due
to various benign pathologies such as cyst or hemangioma. Measurement of SUVmax is also affected by various factors such as blood glucose level at the time of injection,
duration of the uptake period, body weight, and body composition. SUVmax may be overestimated in sclerotic lesions due to over attenuation correction by CT,
and it may be underestimated in osteolytic lesions due to under attenuation correction
by CT. In small lesions, partial volume averaging may cause erroneous results for
bone scintigraphy and FDG PET scan. For example, uptake of a small osteolytic lesion
on bone scintigraphy may be overestimated and uptake of a small osteoblastic lesion
may be underestimated. ROImax value of bone lesion is affected by underlying normal
bone uptake on planar imaging. SPECT or SPECT/CT may provide more accurate ROImax
values. Lytic lesions, particularly large ones, are not always seen cold on bone scan
as seen in [Figure 1 ] which can further increase mean value of the ROImax. Our study consisted of various
malignancies. Metabolic behavior of the osteoblastic and osteolytic bone lesions may
vary in malignancies.
Figure 1 Fluorodeoxyglucose positron emission tomography/computed tomography maximum intensity
projection image and bone scan in a patient with non-small cell lung cancer. Fluorodeoxyglucose
positron emission tomography demonstrates multiple bone and lymph node metastases
in addition to primary tumor in the left lung. Bone scan demonstrates metastasis in
the left iliac bone and right acetabulum and superior pubic ramus. Computed tomography,
not shown here, demonstrated multiple osteolytic bone metastases. Note that bone scan
shows cold and hot pattern in the left iliac lytic lesion. There is also mildly increased
uptake in right distal clavicle, sternum, and few left anterior ribs
Conclusion
We did not find a significant difference in SUVmax and ROImax values of osteoblastic, osteolytic, and mixed lesions and also lack of
correlation between SUVmax , ROImax, and HU values which could be due to various technical or patient-related
causes or varying metabolism of bone metastases from various malignancies. A study
with a larger number of patients who had untreated and proven bone metastases and
also using SPECT instead planar bone imaging can be valuable to assess the metabolic
and osteoblastic activities of bone metastases.
