| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Nuclear Medicine |
1 From the Division of Nuclear Medicine, Johns Hopkins School of Medicine, 601 N Caroline St, Room 3223A, Baltimore, MD 21287-0817. Received December 8, 2003; revision requested February 12, 2004; final revision received December 9; accepted January 26, 2005. Address correspondence to R.L.W.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: This HIPAA-compliant study had institutional review board approval, and the need for patient informed consent was waived. Three hundred fifty-nine consecutive patients (191 male patients, 168 female patients; mean age, 56.9 years; age range, 8–92 years) underwent PET/CT. PET images were first reviewed by nuclear medicine physicians who had no clinical information regarding the presence or absence of bone metastasis by using a five-point grading system (0, a lesion was definitely negative for metastasis; 1, a lesion was probably negative; 2, a lesion was equivocal; 3, a lesion was probably positive; and 4, a lesion was definitely positive). For lesions assigned a grade of 3 or 4 at PET, CT characteristics such as the presence or absence of morphologic changes or accompanying findings (including bone destruction) were assessed by radiologists on the CT images obtained during the same imaging session.
RESULTS: One hundred seventy-nine lesions in 55 patients were considered to be probable or definite bone metastases at PET. One hundred thirty-three of these lesions in 33 patients were clinically confirmed to be bone metastases at follow-up and/or histopathologic examination. CT revealed osteolytic changes in 41 (31%) and osteoblastic changes in 21 (16%) of the 133 lesions, but no or nonspecific changes were seen at CT in 49 (37%) and 22 lesions (17%), respectively. Of the 179 lesions suspected at PET, 46 ultimately proved to be nonosseous or false-positive for bone metastasis. Of these 46 lesions, 38 were not located in the bone but in adjacent tissues such as the pleura.
CONCLUSION: CT images obtained as part of PET/CT scanning were useful in yielding the precise location of bone lesions and thus helping avoid misdiagnosis of bone metastasis; however, CT revealed morphologic changes in only half of the lesions assigned a grade of 3 or 4 at PET.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Despite the high diagnostic accuracy of FDG PET in many cancers and the advantage of its ability to enable evaluation of the entire body in a single examination, in clinical practice it is often necessary to correlate PET results with findings obtained with other imaging modalities (eg, computed tomography [CT], magnetic resonance [MR] imaging, radiography, bone scintigraphy) to diagnose or confirm that the suspected lesions truly represent metastatic lesions to the bone and not other glycolytic processes. To date, we have only limited data regarding how many lesions with abnormal FDG PET findings in bone also show corresponding morphologic abnormalities, even if they are not suspected of representing definite bone metastasis. Furthermore, we have not clearly defined the number of cases among patients who are suspected of having bone metastases at FDG PET that do not support the results of FDG PET and appear to be false-positive. Relatively recently, combined PET/CT scanners have been developed, and clinical applications have rapidly been initiated (11–13). PET/CT scanners provide detailed anatomic information regarding the sites of the abnormal PET findings. With PET/CT, we can use CT to assess the morphologic features of lesions that are suspected of being bone metastases at PET.
The purpose of our study was to retrospectively evaluate lesion findings at CT performed as part of a PET/CT examination in patients suspected of having metastatic bone lesions—lesions detected with FDG PET as part of the same examination—and to correlate the CT and FDG PET findings.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Patients
The study population comprised 359 consecutive patients (191 male patients, 168 female patients; mean age, 56.9 years; age range, 8–92 years) who underwent combined whole-body PET/CT scanning for evaluation of known or suspected neoplasms between July 2001 and December 2001. Data sets were retrospectively examined. Patients who underwent strictly cardiac or brain PET/CT were not included.
PET/CT Scanning
Whole-body PET scanning was performed by using a combined PET/CT scanner (Discovery LS; GE Medical Systems, Waukesha, Wis). The system permitted the simultaneous acquisition of 35 transaxial PET emission images per field of view with an intersection spacing of 4.25 mm. Transaxial resolution with a Gaussian filter was approximately 4.5 mm full width at half maximum. The field of view and pixel sizes of the reconstructed images were 50 cm and 1.953 mm, respectively, with use of a 128 x 128 matrix size for image fusion. This PET/CT scanner also had four–detector row helical CT scanning capability. The technical parameters used for four–detector row CT imaging were as follows: a section thickness of 5 mm, a pitch of 6:1 (in high-speed mode), a gantry rotation speed of 0.8 second, a table speed of 30 mm per gantry rotation, 140 kVp, and 80 mA. After at least a 4-hour fasting time, patients received an intravenous injection of approximately 15 mCi (555 MBq) of FDG (Gamma Plus, Baltimore, Md). FDG was synthesized with the Hamacher method (14). Approximately 50 minutes later, CT scanning was performed from the meatus of the ear to the midthigh for 35 seconds without breath holding; whole-body emission scanning that took 5 minutes for each bed position was also performed with the same axial coverage. PET images were reconstructed with an ordered-subset expectation maximization iterative reconstruction algorithm that involved two iterations and 28 subsets.
Image Analysis
For all patients, PET images only were independently reviewed at a workstation (eINTEGRA; Elgems, Haifa, Israel) by two experienced nuclear medicine physicians (Y.N., with 7 years of experience with PET, and C.C., with 3 years of experience with PET) who had no knowledge of any clinical information, including the primary cancer. PET images were assessed for the presence of bone metastasis by using a five-point grading system in which a score of 0 indicated that a lesion was definitely negative for bone metastasis; a score of 1, that a lesion was probably negative; a score of 2, that a lesion was equivocal; a score of 3, that a lesion was probably positive; and a score of 4, that a lesion was definitely positive.
When we saw intense focal uptake that was apparently within the bone and considered it to definitely represent a metastasis to the bone, a grade of 4 was given, while a grade of 0 was given when we did not see any abnormally increased FDG uptake that suggested bone metastasis. When uptake was seen in or around the bone and the findings were indeterminate for osseous metastasis, we assigned a grade of 2 to the lesion. Grade 1 and grade 3 represent intermediate levels between grades 0 and 2 and grades 2 and 4, respectively. If there were major disagreements about the grading of a lesion—that is, when one reader classified a lesion as grade 3 or 4 and the other did not—the lesion was then reevaluated by both readers together, and a consensus score was determined. Thus, grade 3 or 4 lesions at PET were considered to be suspected bone metastases at PET alone through the consensus of the two readers.
So that we could determine true- and false-positive results at PET, follow-up information for the patients was investigated by one author (D.H.), and the final diagnoses were determined in the following ways: When CT images obtained during PET/CT scanning showed that an area of abnormal uptake corresponded to adjacent tissue rather than bone, the suspected lesion was considered to be negative for bone metastasis. If, at CT performed as a part of the same PET/CT examination, there were definite morphologic findings of metastasis in lesions suspected of being metastatic at PET, such lesions were considered to be true-positive findings of bone metastases. When CT depicted the morphologic changes of metastasis in some of the lesions in a patient who was suspected of having multiple bone metastases at PET, the remaining lesions in the same patient that did not show definite morphologic changes were also considered to be positive.
During follow-up, two patients underwent biopsy, six underwent follow-up PET scanning, nine underwent separate diagnostic CT examinations, two underwent radiography, one underwent MR imaging, and two underwent bone scanning; results of all of these tests were available for correlation. If no progressive findings—including imaging findings—were observed during clinical follow-up, the suspected lesions were considered to be negative. The average clinical follow-up period was 11.4 months (range, 9–14 months). Because granulocyte colony stimulating factor may affect the bone marrow uptake of FDG, it was determined whether a specific indication for the use of granulocyte colony stimulating factor was present for each of the 55 patients who were ultimately suspected of having bone metastasis at PET.
A region of interest was placed by an author (Y.N.) over all suspected lesions with a grade of 3 or 4 by using a 4-pixel circle, and the maximal single pixel value was determined. The standardized uptake value (SUV) was calculated by using the following formula: SUV = ROIRC/(ID/BW), where ROIRC is radioactivity concentration in the region of interest (in becquerels per milliliter), ID is injected dose of FDG (in becquerels), and BW is body weight in grams.
For lesions given a consensus grade of 3 or 4 at PET, evaluation of CT images was performed by two radiologists (Y.N., with 11 years of experience with CT, and D.H., with 6 years of experience) by using CT planes that corresponded to the planes in which the lesion appeared at FDG PET. The radiologists knew that the lesions had been given grades of 3 or 4 at PET, and they used a workstation to display CT scans with bone and soft-tissue windows. Diagnostic certainty in the presence or absence of bone metastasis at CT was expressed by using the same five-point grading system that had been used at PET. When definite morphologic changes within the bone that corresponded to the areas of abnormal uptake were identified, a grade of 4 was given at CT, while a lesion was given a grade of 0 when no abnormal findings indicating bone metastasis were seen. When some morphologic changes were seen in the bone but the finding was indeterminate for bone metastasis, the lesion was scored as grade 2. Grade 1 and grade 3, respectively, were applied to lesions that showed changes that were intermediate between those seen with grade 0 lesions and those seen with grade 2 lesions or between those seen with grade 2 lesions and those seen with grade 4 lesions.
For true-positive lesions that were clearly localized to the bone, CT findings (location [cortex, medulla, or both], morphologic changes [none, nonspecific, osteolytic, osteoblastic], and the presence of an osteosclerotic rim, soft-tissue mass, and/or bone destruction) were recorded. The term nonspecific was used when we determined that the CT findings were not normal but it was difficult to differentiate between neoplastic and degenerative change. If one lesion contained both osteolytic and osteoblastic changes, the predominant anatomic change in the lesion was recorded.
Statistical Analysis
Location-based and primary cancer–based visual grading at CT was assessed by using the 
2 test. The relationship between the SUVs in the bone lesions and visual grading at CT was assessed with the Spearman rank correlation. The significance of the correlations was assessed with the Fisher z test. The SUVs were logarithmically transformed to approximate a normal distribution. The average SUVs in the visual grading groups were compared with the use of a generalized estimating equation model with an exchangeable covariance matrix that was implemented in Proc Genmod (SAS Institute, Cary, NC) and accommodated the dependency of multiple lesions in a patient. In addition, the multiple comparisons were adjusted by using the Bonferroni-Holm method. P values of less than .05 were considered to indicate significant differences.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
True-Positive Lesions
The results of visual grading at CT are given in Table 1. Forty-six lesions (35%) probably or definitely represented bone metastases at CT (ie, were given a CT visual grade of 3 or 4). Table 2 lists true-positive lesions according to location; there were no statistically significant differences in the ratio of positive CT findings among the locations (P = .665, 2 test). In addition, according to the classification of primary cancers, lesions that proved to be bone metastases from lymphoma and digestive cancers (eg, pancreatic or rectal cancer) were consistently diagnosed as negative for metastasis at CT (Table 3) more often than were metastases from other tumor types (P < .01). Table 4 lists the CT findings in the 133 bone metastases that were true-positive findings at PET. CT depicted predominantly osteolytic changes in 41 lesions (31%) and mainly osteoblastic changes in 21 lesions (16%) when bone windows were used, but no or nonspecific changes were seen at CT in 49 (37%) and 22 lesions (17%), respectively. About 10% of the lesions showed soft-tissue masses and/or bone destruction—evidence of bone metastasis—at CT.
|
|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
It is of note that some lesions described as osteolytic or osteoblastic changes in the bone at CT often showed only slight and subtle changes and are not necessarily related to the diagnosis of suspected bone metastasis with confidence. Because a grade of 3 or 4 was not given for these equivocal changes, there was a discrepancy between the number of grade 3 or 4 lesions at CT and the number of osteolytic or osteoblastic lesions. There were no major differences in the prevalence of grade 3 or 4 CT findings according to the anatomic location of the lesions. However, CT more frequently showed normal findings or only minimal morphologic changes, especially in cases of lymphoma or gastrointestinal cancers versus other cancers. Bone metastases from these cancers are predominantly osteolytic and may mainly affect the marrow, whereas those from breast cancer are often osteoblastic.
An osteosclerotic rim of reactive bone can be seen at the margin of slow-growing tumors within the bone. We found such a rim in only 3% of the lesions in this study. The number of lesions with a soft-tissue mass and bone destruction that clearly represented metastatic bone disease was also low, at around 10%. In most lesions in which CT depicted osteolytic or osteoblastic changes, the results were most apparent only after we evaluated the lesions by using the PET findings to guide the CT examination in optimal CT conditions.
It has been reported that metastatic bone disease predominantly involves areas of red marrow such as the skull, the axial skeleton, or the medullary portion of the appendicular skeleton (15), but it has also been suggested that metastases to the bone cortex may be more common than previously expected (16,17). Our data support this latter viewpoint, and involvement of mainly the cortex was reasonably common.
Quantitative and visual analysis revealed that a weak but significant correlation was present between the SUV and the visual grade assigned at CT. In addition, the average SUV in grade 4 lesions (ie, lesions considered definitely positive for metastasis) was significantly higher than the average SUV in lesions with other grades. However, it did not appear possible to differentiate between true- and false-positive lesions with only this quantitative value, although the number of false-positive lesions was limited.
In cases where bone metastasis to the ribs was suspected at PET alone, our results indicated that there were a substantial number of lesions in which the findings did not represent bone metastases but rather represented pleural involvement or metastasis to the lung, chest wall, or even liver. When PET/CT is available, anatomic information obtained at CT may be useful in localizing the lesions and contributing to an accurate diagnoses. Other than rib lesions suspected at PET, there were some lesions in which uptake around the sternum or thoracic vertebrae was misinterpreted as bone metastasis. Therefore, CT evaluation provides useful localization information, especially in thoracic region.
In this series, there was one patient who had five apparently false-positive lesions that showed FDG uptake in bone and were probably due to sarcoidosis. In this patient, no histopathologic confirmation was obtained for the osseous lesions; however, because biopsy of the primary site in the lung revealed sarcoidosis, other foci of intense uptake were also considered to be due to systemic involvement by this noncaseating granulomatous inflammation. Osseous disease has been reported in 1%–13% of patients with sarcoidosis, whose lesions not uncommonly involve the hands and feet (18). A few reports mention FDG PET findings of osseous involvement in sarcoidosis (19,20). In our series, the five FDG-avid sarcoid lesions were seen in the thoracic spine (n = 1), the lumbar spine (n = 2), the rib (n = 1), and the iliac bone (n = 1), and no morphologic changes were seen in corresponding bone regions at CT. An inflammatory process might be responsible for these abnormalities. It is worth noting that FDG can accumulate in many organs other than the lung or skeletal system in sarcoidosis (21).
Although CT can provide detailed anatomic information as a part of PET/CT, our data suggest that there is a limitation in the characterization of lesions even when the optimal CT window width and level are used. For indeterminate lesions, MR imaging or biopsy may be still necessary for further characterization and confirmation (22). Percutaneous bone biopsy, which has an overall accuracy of about 95% (23), is sometimes performed. This high accuracy is based on results in cases in which radiographic changes were apparent. Because PET/CT is increasingly available, some patients with no obvious morphologic changes potentially could be evaluated with PET/CT-guided biopsy, as well. Further studies are needed to assess the clinical feasibility and effect of PET/CT-guided biopsy, as well as its accuracy.
Because we were able to acquire histopathologic confirmation for only two of our suspected bone lesions, most final diagnoses of lesions were achieved by means of clinical follow-up, including imaging tests. This was a major limitation of our study. Future larger studies in this area should ideally include more biopsy confirmation of lesions, although, owing to ethical considerations, we could not sample for biopsy all sites of lesions that were positive at PET. However, when we detected multiple foci of uptake that suggested bone metastases, it is most probable that they all represented bone metastases. In addition, five intense foci of uptake in a patient with sarcoidosis were categorized without histologic confirmation as false-positive findings caused by sarcoidosis.
In conclusion, CT findings at PET/CT were often useful in evaluating bone metastases suspected at PET and in localizing them precisely, thus helping to avoid misinterpretations. However, among true-positive bone metastases seen at PET, morphologic changes at CT were observed in just half. In addition, because subtle changes at CT were common but did not enable a definite diagnosis of bone metastasis, we were able to obtain an accurate diagnosis of bone metastasis in only one-third of the lesions seen at CT when bone metastasis was suspected at PET. Bone metastases at PET often were not confirmed at CT. The use of other modalities such as MR imaging may be necessary for further evaluation of these lesions.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Abbreviations: FDG = fluorine 18 fluorodeoxyglucose • SUV = standardized uptake value
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, Y.N., R.L.W.; study concepts and design, Y.N., R.L.W.; literature research, Y.N.; clinical studies, C.C., D.H.; data acquisition, Y.N., C.C.; data analysis/interpretation, Y.N., C.C., M.T., D.H., R.L.W.; statistical analysis, Y.N.; manuscript preparation, Y.N., D.H.; manuscript definition of intellectual content, C.C., D.H., R.L.W.; manuscript editing, Y.N., R.L.W.; manuscript revision/review, R.L.W., D.H.; manuscript final version approval, R.L.W.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. S. Andrade, J. J. Blondet, T. Kast, J. Jessurun, and M. A. Maddaus Evaluation of Isolated Rib Lesions With Radionuclide-Guided Biopsy Ann. Thorac. Surg., October 1, 2008; 86(4): 1111 - 1114. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U. Tateishi, C. Gamez, S. Dawood, H. W. D. Yeung, M. Cristofanilli, and H. A. Macapinlac Bone Metastases in Patients with Metastatic Breast Cancer: Morphologic and Metabolic Monitoring of Response to Systemic Therapy with Integrated PET/CT Radiology, April 1, 2008; 247(1): 189 - 196. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Du, I. Cullum, T. M. Illidge, and P. J. Ell Fusion of Metabolic Function and Morphology: Sequential [18F]Fluorodeoxyglucose Positron-Emission Tomography/Computed Tomography Studies Yield New Insights Into the Natural History of Bone Metastases in Breast Cancer J. Clin. Oncol., August 10, 2007; 25(23): 3440 - 3447. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Basu, B. A. Siegel, D. J. McDonald, and B. Nussenbaum Detection of Occult Bone Metastases From Head and Neck Squamous Cell Carcinoma: Impact of Positron Emission Tomography Computed Tomography With Fluorodeoxyglucose F 18 Arch Otolaryngol Head Neck Surg, August 1, 2007; 133(8): 801 - 805. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. V. Taira, R. J. Herfkens, S. S. Gambhir, and A. Quon Detection of Bone Metastases: Assessment of Integrated FDG PET/CT Imaging Radiology, April 1, 2007; 243(1): 204 - 211. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Israel and A. Kuten Early Detection of Cancer Recurrence: 18F-FDG PET/CT Can Make a Difference in Diagnosis and Patient Care J. Nucl. Med., January 1, 2007; 48(1_suppl): 28S - 35S. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
