Cartilage tumors such as enchondromas, osteochondromas, and chondrosarcomas
are the most common skeletal neoplasms. Histopathologic grades correlate well
with clinical
outcome1. Imaging
studies such as plain radiographs, computed tomography, and magnetic resonance
imaging are essential to assess the anatomic and tissue characteristics of the
lesions. Bone-scan changes reveal osteoblastic activity or increased local
blood flow of the affected regions and indirectly reflect the activity of the
cartilaginous tumors in the bone. It would be ideal if the biologic behavior
of a cartilaginous lesion and its histopathologic grade could be determined in
a noninvasive manner.
Positron emission tomography has been used successfully to detect glucose
uptake by cells with high metabolic activity, such as heart, brain, and tumor
cells2-4.
However, the value and limitations of positron emission tomography for the
diagnosis of cartilaginous tumors have not been fully
investigated5. As
positron emission tomography is an expensive diagnostic modality, it is
important to determine conditions that may cause false-positive or
false-negative results.
In this study, we assessed the value and limitations of positron emission
tomography for the diagnosis of cartilaginous tumors of bone. We hypothesized
that higher-grade chondrosarcomas would have a higher metabolic activity
measured by fluorine-18 fluorodeoxyglucose (18F-FDG) positron
emission tomography in comparison with low-grade chondrosarcomas and benign
cartilage tumors.
Thirty-five biopsy-proven cartilage lesions in twenty-seven patients were
studied with positron emission tomography between 1999 and 2002. Retrospective
review of imaging studies and pathologic specimens was approved by the
institutional review board of our institution. There were eighteen female and
nine male patients. Nine lesions involved the femur; nine, the humerus; six,
the pelvis; three, the scapula; two, the sacrum; and six, other bones. As part
of the diagnostic and preoperative work-up, the lesions were evaluated with
plain radiographs, computerized tomography, magnetic resonance imaging,
technetium-99m bone-scanning, and positron emission tomography. Plain
radiographs and computerized tomography scans demonstrated intraosseous or
extraosseous lobular lesions containing stippled calcific densities. Magnetic
resonance imaging showed decreased signal on T1-weighted images and a lobular
pattern of signal intensity on T2-weighted images. The diagnosis of each tumor
was confirmed by histopathologic examination. The thirty-five lesions included
ten enchondromas, three osteochondromas, twelve grade-I chondrosarcomas, five
grade-II chondrosarcomas, and five grade-III chondrosarcomas.
Benign enchondromas in patients experiencing pain or anxiety were treated
with either curettage and bone-grafting or placement of polymethylmethacrylate
bone cement after the patient gave informed consent. Three symptomatic
osteochondromas were excised. Grade-I chondrosarcomas were associated with
pain (two femoral lesions), bladder and bowel obstruction (two pelvic
lesions), hydronephrosis (one lesion), and lumbosacral nerve compression (one
lesion). All twelve grade-I chondrosarcomas were treated with intralesional or
marginal resection and reconstruction as needed. One patient with a grade-II
chondrosarcoma of the left scapula and another patient with a grade-II
chondrosarcoma of the right side of the pelvis had metastasis to the lung.
Another patient, who originally had had a grade-II chondrosarcoma of the
humerus, presented with an expansile lesion in the scapula three years after
wide resection of the humeral lesion. A scapulectomy was performed, and the
diagnosis was consistent with a dedifferentiated chondrosarcoma. Signs of a
probable impending pathologic fracture developed in a patient with a
dedifferentiated chondrosarcoma in the distal part of the femur, and wide
resection and reconstruction was performed.
Analysis of Conventional Imaging Studies
The imaging characteristics of each lesion were analyzed with reference to
tumor size, location, bone destruction, and invasion into surrounding
structures. The size of the lesion on magnetic resonance images was measured
by calculating the proximal-distal, medial-lateral, and anterior-posterior
diameters. The uptake pattern on the technetium-99m bone scans was recorded as
hot or cold.
Analysis of Histopathologic Specimens
The histopathologic specimens were reviewed by a musculoskeletal
pathologist. The diagnosis of grade-I chondrosarcoma was based on increased
cellularity, the presence of plump chondrocytes with nuclear chromatin,
nuclear hyperchromasia, a high number of binucleated cells, myxoid stromal
changes, and infiltration of bone marrow. A tumor size of >6 cm also
supports the diagnosis of chondrosarcoma. Grade-II chondrosarcoma was
diagnosed on the basis of the presence of greater cellularity; loss of lacunar
differentiation; and the presence of spindle-shaped chondrocytes, greater
nuclear atypia, mitotic figures, multinucleation, and a myxoid stroma. The
diagnosis of grade-III chondrosarcoma was based on greater pleomorphism,
increased cellularity, and a higher rate of mitosis.
Positron Emission Tomography
Positron emission tomography has been performed for analysis of bone and
soft-tissue tumors at our institution since the early 1990s. The metabolic
uptake of the fluorine-18 fluorodeoxyglucose (18F-FDG) employed in
the positron emission tomography is assessed by obtaining images fifty minutes
after the intravenous administration of 0.14 mCi/kg of 18F-FDG. In
each case, a transmission-corrected scan of the entire body was performed with
use of a positron emission tomography scanner (ECAT EXACT HR or ECAT EXACT;
Siemens/CTI, Knoxville, Tennessee). The positron emission tomography units
were used to obtain the images and to calculate the 18F-FDG uptake
values (Figs. 1-A, 1-B,
1-C,
1-D,
2-A, 2-B,
2-C,
2-D). The resolution of the
ECAT EXACT HR and ECAT EXACT units was 4.6 to 5.4 mm and 4.5 to 6.7 mm,
respectively. Nine pixels of the volumes of interest were placed over the
regions of interest identified by the computerized tomography or magnetic
resonance imaging. Standardized uptake values were calculated according to the
formula:
tissue
concentration(MBq/g)[injected
dose(MBq)/body weight(g)]
As the most aggressive-looking region determines the ultimate pathologic
grade, the highest standardized uptake values were determined after multiple
measurements over the region of interest by a nuclear medicine specialist.
Statistical Analysis
Simple correlation coefficients, the t test or one-way analysis of
variance, and the Mann-Whitney U test or the Kruskal-Wallis test were used to
compare maximal standardized uptake values with the following tumor
characteristics: diagnosis, histopathologic grade, tumor volume, recurrence,
and metastasis6. The
statistical analysis was performed with use of the Statistical Package for the
Social Sciences (SPSS) software (version 10; SPSS, Chicago, Illinois). The
bone scans and the maximal standardized uptake values of the positron emission
tomography were tested for their ability to predict the malignant tumor
(histopathologic grade 0 versus grade I, II, or III and grades 0 and I versus
grades II and III) by calculating sensitivity, specificity, positive and
negative predictive values, accuracy, and 95% confidence
interval6 with use
of EpiCalc 2000 software (version
1.02)7.
Interpretation of Positron Emission Tomography Scan
The results of the positron emission tomography are summarized in
Table I. The average tumor
volume (and standard deviation) was 65.6 ± 53.1 cm3 for the
ten enchondromas, 86.0 ± 36.8 cm3 for the three
osteochondromas, 3743.0 ± 8696.9 cm3 for the twelve grade-I
chondrosarcomas, 2004.0 ± 2062.6 cm3 for the five grade-II
chondrosarcomas, and 295.0 ± 204.3 cm3 for the five
grade-III chondrosarcomas. The maximal standard uptake value in the region of
interest was 1.00 ± 0.63 (range, 0 to 1.82) for the enchondromas, 1.62
± 1.07 (range, 0.91 to 2.86) for the osteochondromas, 0.90 ±
0.91 (range, 0 to 2.85) for the grade-I chondrosarcomas, 3.10 ± 1.11
(range, 2.20 to 5.02) for the grade-II chondrosarcomas, and 10.71 ±
5.71 (range, 6.09 to 20.38) for the grade-III chondrosarcomas (Figs.
1-A, 1-B,
1-C,
1-D,
2-A, 2-B,
2-C,
2-D). There was no significant
difference in the maximal standardized uptake value in the region of interest
between the enchondromas and the osteochondromas (p > 0.05) or between the
enchondromas and the grade-I chondrosarcomas (p > 0.05). When the
enchondromas and osteochondromas were grouped together as benign cartilage
tumors, there was no clear distinction between the benign cartilage tumors and
the grade-I chondrosarcomas with regard to the maximal standardized uptake
value. For example, two very large grade-I chondrosarcomas did not demonstrate
any measurable standardized uptake value. However, the maximal standardized
uptake value significantly correlated with higher tumor grades (correlation
coefficient, r = 0.713; p = 0.0001).
Grade-III chondrosarcomas showed significantly higher standard uptake
values than did lower-grade chondrosarcomas (p = 0.0001). When benign
cartilage tumors and grade-I chondrosarcomas were grouped together as
low-grade lesions, there was a significant difference in maximal standardized
uptake values between the low-grade and high-grade lesions (p = 0.009)
(Table I). Therefore, the
positron emission tomography predicted high-grade chondrosarcomas, especially
when the cutoff for the maximal standard uptake value was higher than
2.33.
The maximal standardized uptake values did not correlate with tumor size or
recurrence (p > 0.05). Although the numbers are limited, tumors with
metastasis showed significantly higher standard uptake values than did
nonmetastatic tumors. The standardized uptake value of each metastatic lesion
differed from that of the primary tumor, suggesting variation in tumor
metabolic activity.
Interpretation of Bone Scans
The bone-scan statistics are summarized in
Table II. The bone scans showed
photopenia in the region of interest of two enchondromas and two large
chondrosarcomas. The rest of the lesions had increased uptake on the bone
scans. With the given number of cases, it was not possible to determine the
tumor types or grades on the basis of the hot or cold bone-scan findings.
Limitations of Positron Emission Tomography
The positron emission tomography provided maximal standardized uptake
values, representing the glucose metabolic profile of the lesion, in the
regions of interest that had been identified on plain radiographs, bone scans,
computed tomography scans, and magnetic resonance images. However, several
large chondrosarcomas had low standardized uptake values, even though they
caused considerable local morbidity, including one case of bowel obstruction
(Figs. 3-A,
3-B,
3-C, 3-D) and another case of
obstruction of the inferior vena cava, because of their large size. When the
tumor was located in the pelvis, adjacent structures such as the intestine,
bladder, and muscles could not be clearly distinguished from the peripheral
borders of the large expansile lesion because of the limited resolution of the
positron emission tomography (4.5 to 6.7 mm).
Four patients were found to have lung metastasis on the computerized
tomography scan, and two patients showed lung metastasis on the positron
emission tomography scan. The positron emission tomography had the advantage
of demonstrating metastatic lesions both in bone and in soft tissues,
including the lung. In two patients, very small nodules, measuring 6 mm, that
were confirmed by lung biopsy were not visible on positron emission
tomography.
Another patient had Paget disease in the sclerotic phase involving the
pelvis and the left tibia. A bone scan revealed markedly increased uptake, but
the positron emission tomography did not detect the changes in the tibia. In
one patient with a recurrent enchondroma in the distal part of the femur, the
positron emission tomography showed an increased standard uptake value in the
left groin and pelvic region. One year later, repeat positron emission
tomography showed no abnormal uptake in the groin and pelvic region. These
findings were consistent with lymphadenopathy that resolved
spontaneously2,8.
The odds ratio, sensitivity, specificity, and positive and negative
predictive values with reference to the standardized uptake values and the
grades of the lesions are summarized in
Table II. For these
calculations, increased uptake on a bone scan was interpreted as an indication
of malignancy and a standardized uptake value of >2.3 was interpreted as
indicating a high-grade chondrosarcoma.
In previous studies, positron emission tomography has shown variable
results in distinguishing between benign and malignant bone
tumors9-12.
Although maximal standardized uptake values higher than 2 or even 3 suggest a
malignant lesion, there have been cases of giant cell tumors or fibrous
dysplasia having standard uptake values as high as
11.213. Moreover,
Schulte et al.10
found that histiocytic or giant-cell-containing lesions showed an average
standardized uptake value of 3.52 ± 1.42.
We selected cartilage tumors for our study because they consist of
chondrocytes with variable differentiation rather than a mixture of cells of
different cellular origins. A goal of this study was to determine whether
positron emission tomography could resolve the diagnostic dilemma encountered
in distinguishing low-grade chondrosarcomas from enchondromas.
Aoki et al. initially reviewed the results of positron emission tomography
of six benign cartilage tumors and seven
chondrosarcomas12
and found the average standardized uptake value for the chondrosarcomas to be
2.23 ± 0.74. Our findings suggest that positron emission tomography may
underestimate the clinical behavior of a grade-I chondrosarcoma as exemplified
by the pelvic lesions in our series (Figs.
3-A,
3-B,
3-C, 3-D). Our study
demonstrated limitations of positron emission tomography for evaluating
cartilaginous lesions. First, the resolution of the study is approximately 4.5
to 6.7 mm, and the critical margins of extraosseous lesions cannot be linearly
defined, especially when the surrounding structures have normal physiologic
uptake. To circumvent this limitation, simultaneous computerized tomography
and positron emission tomography may be performed in the clinical
setting14. This
allows simultaneous assessment of anatomical information obtained from the
computerized tomography and functional data obtained from the positron
emission tomography. Anatomical landmarks provided by computerized tomography
greatly facilitate the assignment of biological abnormalities of anatomical
structures, thereby improving the specificity of the
test15.
Another limitation of positron emission tomography is that certain tumors
can be misrepresented depending on their biology and vascularity.
Cartilaginous tissue has a tendency to be hypovascular or avascular. Moreover,
in our series, the larger tumors did not necessarily have higher standardized
uptake values than the smaller tumors did. An explanation is that each tumor
has different mechanisms of cell survival and proliferation under nutrition
deprivation16. In
addition, some tumors may not have sufficient metabolic activity to be
detectable by positron emission tomography. For example, positron emission
tomography scans of bronchoalveolar carcinomas, carcinoid tumors, prostatic
carcinomas, and low-grade non-Hodgkin lymphoma can show negative
findings17-21.
The region of interest should be determined on the basis of the clinical
history and other imaging modalities. It is important to note that FDG uptake
is not specific for
tumor15,22.
Any area of increased metabolic activity can appear on positron emission
tomography images. False-positive findings are frequently related to
inflammatory or infectious processes such as pneumonia, septic arthritis, and
diverticulitis17,23.
Appropriate selection and timing of the scan in defined clinical situations
along with the knowledge of potential pitfalls will reduce interpretation
errors. Positron emission tomography may not be the appropriate initial
screening tool for cartilaginous tumors of bone. However, despite its
limitations18, it
may be helpful for distinguishing grades-II and III chondrosarcomas from
benign chondroid lesions and grade-I chondrosarcoma. ?